Table of contents

How have advances in technology changed astronomy over the past 100 years? In this very brief review I will give examples of just how closely discovery has followed technical innovation, allowing observations that would previously have seemed impossible. Obviously great progress has come from spaceflight allowing observations from outside the Earth's atmosphere. But two other factors have also been crucial—the development of efficient detectors of electromagnetic radiation and the application of computers to both instrument control and data analysis. Driving down experimental errors in pursuit of ever-more accurate measurements has been important too. I will particularly highlight advances in cosmology, the nature of galactic nuclei and the discovery of exoplanets.

Chemical species tomography (CST) has been deployed in a wide range of applications in the last two decades for multi-dimensional measurement of gaseous flow fields. CST offers unique capability for spatiotemporally resolved imaging of multiple thermochemical parameters. It is fundamentally robust, highly sensitive, and adaptable to industrial processes and large-scale combustion systems. The instrumentation methods used to implement CST measurements are critical in determining the physical and chemical variables that may be imaged by CST. In this perspective, we have three main objectives: (a) discuss recent advances in CST instruments from the viewpoint of optics and electronics; (b) highlight the on-going challenges for systems to address the ever-increasing requirements on temporal and spatial resolutions; and (c) consider potential developments for next-generation CST instrumentation.

Chemical energy carriers are crucial for addressing challenges that arise from time lag, large distances, and temporal fluctuations in renewable energy production, which lead to unbalanced energy production and demand. The thermochemical utilization of chemical energy carriers such as solid fuels must be decarbonized to achieve a climate-neutral circular economy as soon as possible, while remaining important for reliable electricity generation and stable economics. To accomplish this, extensive fundamental research is required to understand the underlying chemical and physical processes that can potentially be realized at an industrial scale. This paper reviews optical diagnostics used for particle-level combustion studies for clean power generation applications. The focus is on particle-resolved optical experiments for oxy-fuel coal combustion, biomass combustion, and utilization of iron in regenerative oxidation–reduction schemes. Previous studies are categorized by fuel and reactor types, investigated parameters, and experimental methodology. Phenomenological aspects of the multi-stage combustion process at the particle level are outlined using examples of bituminous coal and iron particle burning in hot gas. A selection of experimental studies is highlighted, with a particular methodological focus on measuring quantities at the particle level. These representative examples address relevant parameters, including particle number density, particle size and shape, surface temperature, ignition and combustion time, gas flame structure, gas temperature and species, nanoparticle formation, gas velocity, and particle dynamics. Finally, issues and problems that require further effort to improve diagnostics for solid fuel combustion studies are discussed.

High-quality metal (carbide)–carbon eutectic materials based on high-temperature fixed points (HTFPs) are widely used in radiometry and thermometry as reference standards. HTFPs on the base of iron–carbon (Fe–C) binary eutectic alloys, with a nominal melting temperature of about 1154 °C (just above the copper freezing point of 1084.62 °C), are one of the promising candidates among the eutectic materials. To establish new HTFPs as reference metrological tools for high-temperature thermometry, their performance should be thoroughly investigated regarding reproducibility and stability. In this work, two large-area (8 mm aperture, 107 mm cavity/thermowell length) Fe–C fixed-point cells were constructed and studied in detail using a radiation thermometer and two different thermocouples (TCs). Three different furnaces were used to explore the thermal behaviors of the cells at various furnace gradients and furnace offsets. The melting temperature at the inflections point of the melting curves of the cells studied across extensive measurement campaigns demonstrated good performance with repeatability of less than 9 mK (assessed from four successive runs) and reproducibility—less than 100 mK (at different furnaces and furnace offsets). The melting temperature agreement between both cells in the same experimental conditions was better than 30 mK. In addition, the equivalence of the developed large-area cells and a small-area radiometric cell (3 mm cavity aperture, and 35 mm cavity length) were comparatively examined in the same experimental conditions. The coherence of the obtained results for the melting temperature of large-area Fe–C cells indicates the feasibility of using large-volume cells for precise calibration of both radiation thermometers and TCs.

Building a free-head gaze tracking model with high accuracy, simple equipment, and not limited to wearing glasses is a challenge. In this paper, a monocular free-head gaze-tracking method based on machine learning are proposed. Two lightweight, high-precision and real-time gaze-tracking models are constructed, which can measure the 2D gaze point and 3D gaze direction respectively. In addition, we combined our gaze-tracking technology with electric sickbed to create an eye-gaze control based electric sickbed system that allows the patient to control the sickbed with their eyes. The experimental results show that the measurement errors of the two models on the MPIIGaze dataset are 4.84 cm and 4.8^∘ respectively. After commissioning, user feedback has shown that this eye-gaze controlled electric sickbed system can enhance the lives of patients.

Accurate thickness measurement of thin coatings (typically 50–500 μm) on carbon fibre-reinforced polymer composites is a major challenge in the manufacturing and maintenance processes of modern aircraft. Different from the conventional material-dependent technique for prediction, a machine learning-enabled strategy with an artificial neural network configuration is used with no requirement of prior knowledge of the type of coating or substrate under test. In the test, an open microwave cavity resonator sensor is directly placed on a coated composite, and any variation of the coating material, coating thickness and conductivity of the composite alters the resonance frequency. Principal component analysis is employed in the signal pre-processing for the dimensionality reduction of the raw measurement data. In terms of the root-mean-square error, the maximum value for the calibration approach is approximately 15 μm and that for the machine learning-based approach is 12 μm. The sensor system developed enables real-time on-site assessment of coated composite structures and thus offers a new approach for non-destructive evaluation 4.0 with improved efficiency, accuracy and automation.

The quality of the measured signature is influenced not only by the instrument's precision but also by the selected measurement configuration. In optical scatterometry, the purpose of measurement configuration optimization (MCO) is to select an optimal or suboptimal combination of measurement conditions, such as the angles of incidence, azimuth, polarization and wavelength, to achieve higher measurement accuracy. This analysis not only requires an effective optimization strategy but is also time-consuming. In this work, we propose a general MCO method that incorporates error propagation theory and condition-number-based error estimation technique, by which the MCO problem can be formulated as an optimization problem for the condition number of the coefficient matrix in the linear estimation of parameter deviations. The method is demonstrated on a multi-wavelength Mueller matrix scatterometry measuring a Si grating. With the help of the neural-network-based surrogate model, the feasibility of the method is verified by making a comparison with Latin hypercube sampling. Fitting results of the measured and calculated Mueller matrix spectra obtained at the selected optimal measurement configuration show a good agreement. The proposed method is promising to provide an alternate solution to globally evaluate the MCO problem in optical scatterometry and other measurement scenarios.

Bearing intelligent fault diagnosis has been researched comprehensively in recent years. However, the scarcity of labeled training samples and various working conditions seriously hinder the widespread application of deep learning based fault diagnosis methods. To address this problem, we propose a transfer multiscale adaptive convolutional neural network (TMACNN), which significantly enhances the performance of deep learning based methods on few-shot and cross-domain bearing fault diagnosis in terms of network architecture and transfer strategy. On the one hand, we design a novel multiscale adaptive convolutional neural network (MACNN) framework that effectively improves the feature extraction and generalization abilities for bearing fault diagnosis by introducing mega-scale convolutions and continuous stacked multiscale convolutions. On the other hand, we propose an innovative transfer strategy for the MACNN that freezes the six stacked multiscale convolutional feature extraction units and fine-tunes the mega-scale convolution unit and the classifier, which are more suitable for few-shot transfer learning. In experiments on the CWRU dataset and Paderborn dataset, our proposed TMACNN outperforms various advanced baseline models for few-shot and cross-domain bearing fault diagnosis.

Data-driven deep learning methods have been widely used in the fault diagnosis of rolling bearings, while general network structures are complex with numerous parameters and computationally intensive calculations, leading to limited real-time performance and delayed fault detection. To address these challenges, this paper presents a novel hybrid framework, termed FKP-SGECNN, for efficient and accurate bearing fault identification. The proposed framework combines the strengths of kernel principal component analysis (KPCA), Fisher criterion, spatial group-wise enhance network (SGENet), and convolutional neural network. In the proposed framework, FKP incorporates Fisher criterion to optimize the kernel functions in KPCA, effectively reducing information redundancy in the input data. Furthermore, SGENet is integrated to streamline the network structure and enhance the model's generalization capability, while maintaining high diagnostic accuracy. The performance of the hybrid framework implies a great potential, which was evaluated by several case studies using multi-class data of bearing faults.

It is observed that previous research studies focusing on few-shot feature data for fault diagnosis or anomaly detection have a limitation, that is, feature extraction methods to solve few-shot feature data problems will also have scenarios where they may not always be applicable. In this paper, a motor anomaly detection model with generalization performance is proposed to meet the anomaly detection needs in the above scenarios. The model consists of a reinforcement unit and a diagnosis unit. Firstly, the reinforcement unit extracts the adjacent features with different timestamps through ensemble learning. Secondly, the temporal convolutional network (TCN) model is nested to increase the receptive field of the reinforcement unit. Additionally, a residual network is introduced to improve the generalization performance. Finally, features obtained from the reinforcement unit are used for final anomaly detection through neural networks in the diagnosis unit. Experimental results indicate that the proposed model achieve an anomaly detection accuracy of 97.96% in factory motor dataset, while the model has the superior generalization ability.

Rolling bearings play a vital role in ensuring the safe operation of rotating machinery. However, in many application scenarios, the collected data has a low signal-to-noise ratio and the samples with faults are rare, which affects the generalization capability of the model, making it impossible to achieve accurate diagnosis. To solve this problem, the selection of time-frequency (TF) maps was considered in this paper through reinforcement learning. The TF maps are built by four classical TF characterization methods such as short-time Fourier transform and synchro squeezing transform. And the match-reinforcement learning time frequency selection (MRLTFS) fault diagnosis model is proposed to extract the fault-related features. Experiments show that the proposed MRLTFS method is superior to existing methods in robustness, generalization and feature selection capability.

Simultaneous localization and mapping (SLAM) is one of the prerequisite technologies for intelligent mobile robots to accomplish various tasks in unknown environments. In recent years, many excellent SLAM systems have emerged, but most of them have a basic assumption that the environment is static, which results in their poor performance in dynamic environments. To solve this problem, this paper presents SCE-SLAM: a novel real-time semantic RGB-D SLAM system that is built on the RGB-D mode of ORB-SLAM3. SCE-SLAM tightly combines semantic and geometric information. Considering the real-time requirements, the semantic module provides semantic prior knowledge for the geometric module using the latest and fastest object detection network YOLOv7. Then, a new geometric constraint method is proposed to filter dynamic feature points. This method takes full advantage of depth images and semantic information to recover three-dimensional (3D) feature points and the initial camera pose. A 3D coordinate error is used as a threshold, and SCE-SLAM removes dynamic points using the K-means clustering algorithm. In this way, SCE-SLAM effectively reduces the impact of dynamic points. Furthermore, we validate SCE-SLAM with challenging dynamic sequences of the TUM dataset. The results demonstrate that SCE-SLAM significantly improves the localization accuracy and system robustness in all kinds of dynamic environments.

Approximate 5 min are needed to obtain diffraction efficiency of a grating in a wavelength range with 151 sampling points by the currently widely used measurement technique. The time-consuming nature restricts its usage in many dynamic measurements. To this end, we developed a fast measurement technique, and the corresponding measurement is completed in a high speed within 1 s. Meanwhile, dynamic measurement of the efficiency of a grating from 600 nm to 900 nm during a heating procedure is successfully carried out. Our technique could find applications in other circumstances where high-speed measurement is needed.

Coal dust leakage occurs most often during transportation to a power plant. Owing to the transportation method, the transported high-pressure coal dust can damage weak points in the pipeline wall and leak into the air, leading to serious safety hazards. To address this, this study proposes a coal dust parameter estimation model that combines domain-adaptive segmentation with extreme corrosion and a particle mass prediction regression model to extract the key characteristic signals of leaking coal dust and evaluate the production environment safety status. First, the connected domain is applied to segment the overlapping particles and extract two-dimensional image information. Subsequently, a regression model was constructed to predict the particle mass, which was mapped with the coal dust thickness model, density, and projected area and applied to environmental dust concentration characterization. The experiment samples included 3000 coal dust images captured from production links in power plants. A statistical analysis showed that the proposed model improved the accuracy and reliability of coal dust detection.

The ultrasonic pulse-echo method plays a significant role in ultrasonic thickness measurement. However, when the measured thickness is relatively thin, it is universal for the reflected ultrasonic echoes to overlap in the time domain which constrains the achievable resolution using the conventional pulse-echo method. A deep learning network for multi-echo overlapping ultrasonic signals separation under the constraint of an ultrasonic transducer with specific center frequency and bandwidth was proposed to enhance the axial resolution. With this model, the adaptive separation of overlapping signals with an unknown number of echoes is achieved by extracting detailed features of the overlapping signals. An overlapping signal training dataset based on truncated Nakagami functions is constructed to train the network. Simulation experiments under different degrees of overlap and noise levels were conducted for the performance analysis. Simulation results manifested the performance advantage of proposed separation method over some traditional separation methods. Further, ultrasonic thickness measurement experiments were performed on thin-walled aluminum alloy plates with various thicknesses, demonstrating feasibility and practicability of the network over traditional separation methods.

The impact of contact resistance (R_c) on 1/f noise measurements was studied to demonstrate improved accuracy with a novel three-point probe (3pp) method, introduced here, versus the typical two-point probe (2pp) arrangement. It was shown for n⁺-Si that using indium to lower R_c decreased the noise spectral density (NSD) over 100× for 2pp. In the 3pp configuration, NSD was reduced by another ∼5×, independent of indium use (i.e. spectra overlapped), suggesting that 3pp avoids the impact of R_c on NSD. More heavily doped n⁺⁺-Si also showed improvements with 3pp and 2pp/indium versus 2pp/bare. Expectedly, 3pp provided less of a benefit relative to 2pp/indium as R_c was already small due to highly degenerate doping. Measurement drift also improved with 3pp. These results have implications on 1/f measurement accuracy for the broadly used 2pp arrangement. The 3pp method provides a better noise floor for NSD testing because it is not convoluted with signal from R_c, does not require metallization and has improved simplicity and versatility yet performs equal to four-point probe methods.

Aiming at the problem of inaccurate measurement results caused by factors such as gas leakage during the calibration of the standard meter gas flow standard device of the same type and small diameter, this paper proposes a similarity assessment method of single flow point energy consumption mapping based on Fréchet distance (FD). First, through the statistical analysis of the historical energy consumption data of gas flow standard device, a discrete standard energy consumption mapping is formed. Next, the energy consumption mapping to be tested is standardized and the similarity of the energy consumption mapping is determined by the binary matrix accelerated search of the FD between the energy consumption mapping to be tested and the standard energy consumption mapping. Finally, the similarity threshold is calculated by combining the standard deviation of the FD and the reliability of the gas flowmeter calibration process is evaluated accordingly. The test shows that the method in this paper can effectively determine the reliability of the gas flowmeters calibration results.

The current deep learning based machinery fault diagnosis models still face challenges in effectively capturing rich multi-scale feature information and dynamically compensating training loss when dealing with imbalanced dataset. This paper presents a novel approach for machinery fault diagnosis using multi-scale feature focused network and adaptive cost-sensitive loss. Firstly, a multi-scale feature focused network is constructed with improved multi-scale CNN and point-wise attention mechanism module, in which the former can synthetically fuse the features at different scales to expand the coverage of the equivalent receptive field, and the latter can further refine fine-grained features and filter out irrelevant feature interference. Then, an adaptive cost-sensitive loss function is designed to adjust the cost matrix in the training process, dynamically assigning more loss weights for small samples that are difficult to distinguish. The experimental results of planetary gearbox fault diagnosis demonstrate that the proposed approach exhibits superior diagnostic performance compared to other existing methods.

Atomic force microscopy (AFM) cantilevers are commonly made from two material layers: a reflective coating and structural substrate. Although effective, this can result in thermally induced cantilever deflection due to ambient and local temperature changes. While this has been previously documented, key aspects of this common phenomenon have been overlooked. This work explores the impact of thermally induced cantilever deflection when in- and out-of-contact, including the topographic scan artefacts produced. Scanning thermal microscopy probes were employed to provide direct cantilever temperature measurement from Peltier and microheater sources, whilst permitting cantilever deflection to be simultaneously monitored. Optical lever-based measurements of thermal deflection in the AFM were found to vary by up to 250% depending on the reflected laser spot location on the cantilever. This highlights AFM's inherent inability to correctly measure and account for thermal induced cantilever deflection in its feedback system. This is particularly problematic when scanning a tip in-contact with the surface, when probe behaviour is closer mechanically to that of a bridge than a cantilever regarding thermal bending. In this case, measurements of cantilever deflection and inferred surface topography contained significant artefacts and varied from negative to positive for different optical lever laser locations on the cantilevers. These topographic errors were measured to be up to 600 nm for a small temperature change of 2 K. However, all cantilevers measured showed a point of consistent, complete thermal deflection insensitivity 55% to 60% along their lengths. Positioning the reflected laser at this location, AFM scans exhibited improvements of up-to 97% in thermal topographic artefacts relative to other laser positions.

In engineering practice, several factors affect the various types of information during the collection process. For example, information data measurement errors are caused by equipment performance and the working environment. During the transmission of detection information, signal distortion caused by energy loss and signal interference causes unpredictable detection errors in collected data. Through the study of fractional calculus theory, it was found that it is suitable for studying nonlinear, noncausal, and nonstationary signals, and has the dual functions of improving detection information and enhancing signal strength. Therefore, under the influence of many factors, we applied the fractional difference algorithm to the field of information-data processing. A multisensor detection data fusion algorithm based on the fractional partial differential equation was adopted to establish online detection data. A multi-sensor detection data fusion algorithm based on a fractional partial differential equation was established, which effectively fuses the information data detection errors caused by various influencing factors and significantly improves the detection accuracy of information data. The effectiveness of this method was experimentally demonstrated by its application.

For the degraded performance of the fault diagnosis model caused by massive normal samples and scarce fault samples under unbalanced conditions, a new fault diagnosis method based on a hybrid sampling algorithm and energy entropy, namely HSEEFD is proposed in this paper. In the proposed method, Empirical Modal Decomposition is employed to decompose the vibration signals into Intrinsic Mode Functions (IMFs), and the energy entropy feature of each IMF component is extracted to construct a feature vector matrix. Then, a new hybrid sampling algorithm using Tomek's Links algorithm, Euclidean distance, K-means algorithm, and synthetic minority over-sampling technique (SMOTE), namely TSHSA is designed to balance the extracted features. Tomek's Links algorithm is used to identify and remove the confusable majority class samples at the boundary. Euclidean distance is applied to find the suspected noise points in minority class samples and remove them. The k-means algorithm is employed to cluster the minority class samples and SMOTE is used to deal with each cluster according to the density of the clusters to synthesize new features. Finally, the support vector machine is applied to classify faults and realize fault diagnosis. The experiment results on the actual imbalanced data show that the proposed HSEEFD method can effectively improve the accuracy (AUC) of the fault diagnosis under unbalanced conditions by increasing the AUC value by more than 2.1%, and the AUC and G-mean by more than 0.7%, 2.1%, respectively.

The miniature fiber optic spectrometer is smaller, cheaper and has a wide range of applications. However, the measurement error is larger. In order to solve this problem, the adaptive iterative optimization method for spectral calibration is proposed. In this study, a trinity neural network model is built based on spectral wavelength segmentation to improve the calibration degree. Based on the 'pseudo-label', a self-optimization method for spectral calibration is proposed to reduce the amount of data required. This study optimizes the measurement accuracy without changing the structure of the spectrometer. And the self-optimization of calibration model in practical application is realized. After experiment, the calibration degree of the calibration model can reach 75.72%. After a self-optimization, it can be increased to 87.45%. The calibration time of 401 spectral values (380 nm–780 nm) is less than 0.01 s. The results show that the operator can use this method to calibrate spectral data without having optical knowledge. This method has low cost, high calibration speed, good reliability and application value.

This article presents a fiber interferometric sensor (FIS) for measuring the velocity amplitude of an oscillatory vibrating object, with a focus on velocity mode measurement in applications using the Kibble balance principle. The sensor uses the range-resolved interferometry method to measure the displacement of the moving object and employs a multi-harmonic sine-fit algorithm to estimate the displacement amplitude and frequency, thereby determining the velocity amplitude. This article provides a comprehensive explanation of the experimental setup and the measurement techniques employed, as well as a detailed analysis of the uncertainty budget, with the performance validation of the FIS benchmarked against a commercial interferometer within a Kibble balance setup. The velocity amplitude of a coil of the Kibble balance, oscillating with an approx. amplitude of 20 μm and a frequency of 0.25 Hz, was measured using the sensor and found to be 31.282 31 μm s⁻¹ with a relative deviation of −1.9 ppm compared to a commercial interferometer. The high performance of the FIS, especially with regard to non-linearity errors, and the small size of the measuring head enable universality of integration into a wide variety of measurement systems, also including the use as general-purpose vibration and displacement sensor.

High dynamic range surfaces are prone to overexposure and underexposure in the camera imaging process, which poses a challenging issue for three-dimensional (3D) reconstruction. Capturing multiple images of high dynamic range surfaces at different exposure times is widely used for 3D reconstruction, and the exposure sequence can significantly affect reconstruction performance. However, most existing methods select the exposure sequence based on the overall performance of the image or groups of image pixels, without considering individual pixels in the image. To circumvent the problem, this paper presents the adaptive optimal exposure sequence selection based on time cost function for the 3D reconstruction of high dynamic range surfaces. Different from existing methods that use the overall performance, the acceptable exposure time range for each pixel in the image is determined independently. Furthermore, the time cost function is established to evaluate the priority of each exposure time and used to select the optimal exposure sequence iteratively. Experimental results verify both the reconstruction performance and suitability of the proposed method.

The matrix inversion used to derive the 3D point cloud from the phase is a key factor that limits the reconstruction speed improvement in 3D imaging technique based on phase-shift fringe analysis. To avoid computing matrix inversion, we generate two offline loadable matrices determined solely by the calibration parameters of the triangulation model. Compared with the original method that requires matrix inversion, online computation was accelerated by about 15 times without incurring additional memory overhead. Experimental results demonstrated that the proposed method significantly improves the speed of 3D shape measurement and ensures high accuracy.

To reveal the discrete characteristics of the instantaneous frequency of the electromagnetic radiation (EMR) waveform induced by coal and rock fracture, the uniaxial compression experiments for coal and rock samples were carried out, and the EMR signals with full waveform were acquainted and stored. The empirical wavelet transform is used to filter and de-noise the EMR waveform, and then the short-time Fourier transform is used to analyze the time-frequency characteristics of the waveform. The discrete characteristics of the instantaneous frequency with a larger amplitude and the relationship between the centroid frequencies and peak-to-peak values (Vpps) of the EMR waveforms are statistically analyzed. The results show that the centroid frequency of 0–100 kHz is negatively correlated with the Vpp, and the relationship between them shows a logarithm function relation. The instantaneous frequency of the EMR waveform of coal and rock fracture has significant discrete characteristics. In detail, for the rock sample, the instantaneous frequencies with relatively large amplitude are mainly 4.5 kHz, 19.5 kHz, 22.0 kHz, and 27.5 kHz; for the coal sample, the instantaneous frequencies are mainly 1.0 kHz, 4.5 kHz, 9.0 kHz, and 74.0 kHz. This discrete characteristic is determined by the natural properties and fracture characteristics of the sample. Compared with the homogeneous rock samples, the internal cracks of the coal samples are well developed and show strong anisotropy, resulting in the discrete characteristics of the instantaneous frequency being relatively weaker. The findings have certain guiding significance for optimizing the design of the EMR monitoring frequency band and improving the pertinence and accuracy of the monitoring and early warning for coal and rock dynamic disasters.

This article proposes a novel approach to address the issues of low accuracy in fault diagnosis and the difficulty in installing sensors on rolling bearings in mechanical and electrical equipment systems. To accomplish fault diagnosis of rolling bearings, a network structure algorithm based on convolutional neural network (CNN) and support vector machine (SVM) is presented, which incorporates the electric motor current signal. Firstly, the collected electric motor current signal is subjected to a wavelet filter with a soft-hard threshold to eliminate the noise. Secondly, the processed data is fed as input to a one-dimensional CNN to perform feature extraction and dimensionality reduction. Finally, the dimensionality-reduced features are processed by a SVM to diagnose rolling bearing faults. The research results indicate that the proposed method significantly improves the accuracy of rolling bearing fault diagnosis compared to other approaches, with an accuracy of up to 99.01%. This study introduces an innovative approach that can be applied to the field of rolling bearing fault diagnosis, offering valuable insights for research and application in this domain.

Deep learning methods have been widely investigated in machinery fault diagnosis owing to their powerful feature learning capability. However, high accuracy is hard to achieve due to the limited fault information in a single domain when the data volume is small. In this paper, an optimized Hilbert curve (OHC) method is developed, which can generate a novel domain to highlight the fault impulses of vibration signals. To fully mine the fault information, a bidirectional-channel convolutional neural network with an attention mechanism is further proposed, in which two channels are constructed and a transmission channel selection is conducted by a novel improved convolutional block attention module. First, the OHC images and the time-frequency representations are obtained by OHC and wavelet transform respectively. Second, the two types of representations are fed into the channels respectively for feature learning. Finally, the learned features are allocated to different attention mechanism for feature fusion and classification. The proposed method is evaluated via the datasets of rolling bearings and planetary gearboxes, and results show that it outperforms the comparison methods.

The construction of a high-temperature gas calibration source is of great significance since it can provide an effective high-temperature experimental environment for, e.g. verifying high-temperature measurement techniques and studying high-temperature combustion mechanisms. Here, we try to obtain a high-temperature gas field on a multi-jet burner by controlling the gas supplies to it. We use OH planar laser-induced fluorescence (OH-PLIF) to observe the compositional uniformity of the field and coherent anti-Stokes Raman scattering (CARS) to investigate the temperature uniformity of the field. We find from OH-PLIF images that the distribution of OH between the adjacent jets becomes more uniform with the increasing flow rate of CH₄, and the flow rate of the co-flow N₂ around jets also affects the uniformity of OH distribution. The measured temperature distribution by CARS is consistent with the OH distribution. At the jet outlet location, the temperature distribution had a periodic variation and gradually became more uniform with the height increased from the jet outlet. We find that the flow rate of CH₄ and co-flow N₂ and the radiative heat transfer rate play an important role in temperature distribution for the multi-jet burner. Also, the results show that a wide range of temperatures can be constructed by regulating the recipe of the gas supplies, and the highest temperature achieved in this work is 2457 K.

The CYGNO experiment aims to study rare events related to the search for low-mass dark matter and solar neutrino events. One of the main components of background comes from cosmic rays that generate long tracks in the detector's images. The interaction of such particles with the gas releases a variable energy profile along its trajectory to form tracks with multiple cores that can be easily reconstructed erroneously by being split into more than one cluster. Thus, this work offers a newly adapted version of the well-known density-based spatial clustering of applications with noise (DBSCAN) algorithm, called iDDBSCAN, which exploits the directional characteristics of the clusters found by the DBSCAN to improve its clustering efficiency when dealing with multi-core tracks. This paper provides a detailed explanation of this algorithm, covering its parameter validation and evaluating its influence when integrated into the experiment's event selection routine. To generate background events, data acquisition was performed with the detector installed in an overground laboratory, leaving it exposed to natural radiation. To produce signals in the energy range of interest for the experiment, a ⁵⁵Fe radioactive source was used. The achieved results showed that the iDDBSCAN algorithm is capable of improving the background rejection of the experiment, through a more accurate reconstruction of the tracks produced by natural radiation such as cosmic rays, without deteriorating its signal detection efficiency and energy estimation.

The meshing performance is a significant indicator for evaluating the machining quality of spiral bevel gear. The flank contact accuracy (FCA) and flank geometric accuracy (FGA) are important aspects of meshing performance, and both are indispensable. In the actual production process, due to the lack of professional meshing performance analysis (MPA) software, the flank geometric structure cannot be reflected based on the FCA, and the actual meshing state cannot be reflected by the FGA, which cannot meet the manufacturing requirements of high-quality gear. Therefore, a MPA method for deviation surface of spiral bevel gear is proposed based on the gear measuring center using a one-dimensional probe. By constructing the deviation surface, the flank meshing model is established, and the meshing parameters are determined. Aiming at the complexity of solving the meshing model, a method for decreasing the dimension of the meshing equations is proposed by establishing the mathematical relationship between the rotation angle and the normal vector, which improves the solution accuracy of the meshing point. The intersection points of the adjacent transmission error curves are taken as the meshing in and out points, and then the actual position of the contact pattern is accurately determined. The instantaneous contact ellipse is obtained by using the method of a cylindrical surface intercepting the deviation surfaces, which not only simplifies the solution algorithm, but also obtains the contact pattern more in line with the actual situation. Finally, the transmission error and contact pattern are digitally characterized. The experimental results show that the digital and actual patterns have good consistency. Flank geometry and gear meshing are organically linked based on the proposed method, and the actual meshing state can be predicted according to the flank topology deviation, which provides important theoretical support for the improvement of the machining quality and meshing performance of the spiral bevel gear.

In recent years, deep learning has made significant strides and found extensive applications in tool wear prediction. However, most methods based on deep learning rely on large-scale datasets for training and do not consider the method under variable working conditions. To address these limitations, this paper proposes a novel pre-trained fine-tuning model based on the graph-labeling graph neural network (GGPT) specifically designed to train on small-scale datasets. The GGPT model comprises two essential components: a feature extractor and a graph attention module (GAT). The feature extractor uses a residual connection constructed by maximum pooling and average pooling to improve the reliability of local feature learning. The GAT module enhances the global feature learning ability of the model by aggregating local features extracted by the feature extractor. In addition, Improved Pseudo-task Regularization is introduced to enable the model to converge better on small datasets. Experimental results substantiate the superiority of the GGPT model over other deep learning-based methods, particularly when confronted with variable working conditions for tool wear prediction.

The Young's modulus and fracture strength of single and bilayer graphene (BLGr) grown by chemical vapour deposition (CVD) were determined using atomic force microscopy-based membrane deflection experiments. The uncertainty resulting from instrument calibration and the errors due to the experimental conditions like tip wear, loading position, and sample preparation were investigated to estimate the accuracy of the method. The theoretical estimation of the uncertainty on the Young's modulus linked to the calibration is around 16%. Finite element simulations were performed to determine the effects of membrane shape and loading position on the extraction of the Young's modulus. Off-centre loading results in the overestimation of the Young's modulus while deviation from the circular shape leads to an underestimation of the stiffness. The simulated results were compared with experiments. With all these sources of errors taken into account, the Young's modulus and fracture strength of CVD-grown single layer graphene are found equal to 0.88 ± 0.14 TPa and 134 ± 16 GPa, respectively. For CVD BLGr, the mean values of the Young's modulus and fracture strength are equal to 0.70 ± 0.11 TPa and 95 ± 11 GPa, respectively.

In view of the future space gravitational wave detection mission, a closed-loop monofilament torsional pendulum micro-thrust measurement system has been developed according to the requirements of the drag-free control of the satellite platform for the thrust of the micro-thruster. In this paper, we present the design of the micro-thrust measurement system, the electromagnetic force generation device, and the angular displacement differential measurement. By calibrating the electromagnetic force generation device, the functional relationship between micro-thrust and torsional angular displacement of the torsion scale is obtained, and then the micro-thrust measurement is carried out. The analysis considers the impact of the experimental device's structure and environment on the micro-thrust measurement system. The range for measuring thrust is 1–246 μN, with a resolution of 0.1 μN and a relative uncertainty of 1.174%. Ensure that the thrust resolution and accuracy of the micro-thruster are met according to the measurement requirements.

Being able to characterize impulsive-like signals and extract their transitory features is difficult due to the presence of noise and irrelevant signal components in real signals. To address these problems, a brand-new time-frequency (TF) analysis technique called the transient-extracting wavelet transform is developed. This method is put forth by first investigating which TF coefficients can represent the fundamental TF properties of impulsive signals, and then designing an extraction operator to get the most related TF coefficients while simultaneously removing the unrelated ones. The signal reconstruction of this method is also analyzed. Additionally, a transient feature extraction approach is suggested for pinpointing the impulse's occurrence timing, which is essential for correctly identifying the fault type. The analysis shows that the suggested method is more able to analyze impulsive-like data and is an effective bearing defect detector.

The fault diagnosis method of bearing-rotor system based on infrared thermography can reflect the global fault information of the equipment, which is an advanced non-contact monitoring measure. Current diagnosis methods focus on the analysis of single-scale input images, and the network only accepts fixed-size inputs. The model is not capable of fusing multi-scaled fault features while processing single-scaled images, and compression for the size constraint will cause geometric distortion. This paper proposed a novel fault diagnosis model ReSPP (Residual Network with Least Spatial Pyramid Pooling (LSPP)) based on the improved LSPP. LSPP solves the problem of fault feature distortion caused by fixed-size constraint of the network. By replacing the single-scaled training with the proposed multi-scaled training method applied to fault diagnosis, the weight parameters of ReSPP pool the deep fault features of the bearing-rotor system at multiple scales, retaining the critical fault semantic information. The proposed model solves the fault feature loss in the process of feature extraction by improving the subsampling residual block. Experimental results show that ReSPP with multi-scaled training method (ReSPP-MSTM) classifies the fault conditions of the bearing-rotor system with an average diagnostic accuracy of 99.18%.

Aiming to extract the weak composite fault characteristics of a rolling bearing under harsh operation conditions, a novel composite fault diagnosis method for bearings based on adaptive circulant singular spectrum analysis (ACiSSA) is proposed. The proposed method is able to adaptively obtain the eigenvalue of a non-stationary vibration signal in any dimension, and effectively reassemble the same frequency components and improve the signal-to-noise ratio (SNR). Specifically, circulant singular spectrum analysis is utilized to decompose the raw signal, and the optimal parameters, i.e. the embedding dimension and threshold value of cumulative contribution, are selected to maximum kurtosis through the grey wolf optimization method. The signal is reconstructed with high SNR according to the effective singular spectrum components. Envelope demodulation analysis is then implemented to extract the characteristic defect frequency in the reconstructed signal. Finally, feature extraction performance is quantitatively evaluated, and experimental results show that the proposed ACiSSA method is able to extract more sensitive features under more noisy conditions compared with other common methods, with higher computational efficiency.

Instrumented indentation is a technique used for measuring the mechanical response of materials to deformation. The accuracy of these measurements relies on force and displacement values recorded during the indentation process. This paper presents a miniature fused silica parallelogram flexure to serve as a reference object for calibrating the applied indentation force based on Hooke's law, F = kz, where k is the stiffness of the reference object and z is the displacement of the flexure under force F. This study discusses two independent experimental approaches to determine the stiffness of the flexure. In both cases, the stiffness was calculated from the measured flexure displacement while known forces are applied. In the first method, the displacement is measured using a built-in fiber-based interferometer while reference masses of 100 mg and 200 mg are placed on the flexure. A stiffness of $k = (80.3\pm 0.79)\;\mathrm{kN}\,\mathrm{m}^{-1}$ is determined with this method. The second method uses a commercial indenter to apply a cyclic force of 100 mN magnitude on the flexure while recording the displacement of the indenter probe. Experiments using a 5 μm and a 50 μm conospherical tip estimate the stiffness to be $k = (80.74\pm 0.79) \; \mathrm{kN}\,\mathrm{m}^{-1}$, and $(80.55\pm 0.33) \; \mathrm{kN}\,\mathrm{m}^{-1}$ respectively.

Disclaimer: Certain commercial equipment, instruments are identified in this paper to foster understanding. Such identification does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the materials or equipment identified are necessarily the best available for the purpose.

Magnetoacoustic emission (MAE) holds great promise for evaluating the mechanical properties of ferromagnetic materials. To refine the problems of the current theoretical and numerical models of MAE, a theoretical MAE model that considers the microscopic dependence of the hysteresis properties is proposed in this paper. The microstructure (dislocation density and grain size) and the correlation of MAE jumps are considered and incorporated into the model. Then, the influences of magnetization parameters and microstructure parameters on the envelope of the MAE signal are analyzed by the proposed theoretical model. The proposed theoretical model is then fully evaluated by simulations and experiments. The MAE experiments are conducted on ferromagnetic specimens with different hardnesses, and the MAE signals with different hardnesses are simulated by inverting the basic parameters of the MAE model with the genetic algorithm. Further, the crucial hysteresis parameters of the specimens are calculated using the results of microscopic measurements and the calculated parameters agree well with inversion results from experimental signals. The results demonstrate that the proposed theoretical model is valid for the MAE signal simulation. The trends of different hardnesses can be predicted by the MAE simulation signals. Moreover, the model can be used for theoretical analysis of the microscopic dependence of the MAE signal.

With advances in navigation technology, unmanned aerial vehicles (UAVs) have become widely available to all sections of society. Given the potential hazards of UAVs, as seen from their use in the Russia–Ukraine war and security incidents at London and Baghdad airports, counter-UAV technology is receiving unprecedented attention. This paper describes a method for taking control of a UAV indirectly by navigation spoofing and luring it to a designated area from which it can be neutralized. Our contributions are threefold: first, we analyze the limitations of the traditional state estimation and control model for UAV navigation based on the integration of the Global Navigation Satellite System and Inertial Navigation System. Second, we propose a particle hypothesis and planning model that is insensitive to the particular UAV navigation system, and decompose the UAV navigation spoofing process into two steps: identification and planning. Finally, to overcome the poor heading angle prediction accuracy of unary polynomial regression, we propose a support vector regression algorithm that improves the prediction accuracy from 1.5° to 1.01° under navigation spoofing. Experimental results using the proposed navigation spoofing method prove that machine learning offers significant advantages in modeless identification.

Determining the number of sources under low signal-to-noise ratio (SNR) and signal interference with the same frequency and modulation presents a significant challenge. To address this challenge, we propose a novel method for detecting the number of signal sources from single-channel that leverages signal reconstruction and deep learning. The method employs subspace projection based on the Hankel matrix to reconstruct the measured single-channel signals, effectively suppressing noise. Furthermore, we incorporate the correlation of information and the integrity of feature in the signal, by fusing the in-phase component, quadrature component, and frequency spectrum feature of the reconstructed complex signal into a one-dimensional feature suitable for convolutional neural network (CNN). To address the source number detection task, we design a one-dimensional CNN based on convolutional block attention module, transforming it into a classification problem. Finally, experimental measurements demonstrate the effectiveness of our proposed method, with an detection accuracy of 94% even at an SNR of −10 dB.

In special measurement situations, a vision measurement system needs to measure objects through a viewing observation window made of a planar transparent medium. The refraction at the interface between a medium and air leads to large measurement errors in the pin-hole imaging model. Most existing solutions require complex processes of calculating refracted light equations and/or matching stereo images. Here, we propose a refractive three-dimensional (3D) reconstruction method for fringe projection with a planar medium. We derive a coordinate conversion relationship from a pseudo-object point caused by light refraction to a real-object point based on flat refractive geometry. Then, we integrate the relationship into regular fringe projection framework for unbiased 3D reconstruction. Two experiments, including 3D shape measurements of a step and a regular sphere are performed. The results verify the effectiveness and accuracy of the proposed refractive 3D reconstruction method.

The double probe method is a plasma in situ diagnostic technology. Compared with Langmuir single probe, it has less influence on the background plasma and can obtain relatively accurate results. However, it can only collect some high-energy electrons in the plasma, and cannot directly measure the electron density (N_e). In this paper, a double probe N_e diagnosis approach based on Bidirectional Long Short-Term Memory (BLSTM) is proposed. After the training is completed, the accurate prediction of N_e can be realized by using the double probe data, which solves the problem that the double probe cannot directly measure N_e. In the plasma simulation environment of the laboratory, the plasma source is controlled to generate plasma with different densities, the current–voltage (I–V) characteristic data of the double probe at the same position are used as features, and the N_e calculated by the triple probe is used as the label to train the BLSTM model. The mean square error is used as the loss function, the root mean square error (RMSE) and the prediction accuracy (Acc) are used as the evaluation indicators. The BLSTM network is evaluated according to the evaluation indicators and the hyperparameters are adjusted. After about 100 iterations, the RMSE of the BLSTM network to N_e can be reduced to about 0.03. The final network is evaluated on a separate test set. The results show that in the range of 2 × 10¹³m⁻³–3 × 10¹⁴ m⁻³, the model can predict N_e more than 95% accurately. This approach extends the application of the double probe method and is of great significance for improving the accuracy of plasma diagnostic methods. If it is applied to ionospheric plasma diagnosis, it can reduce the amount of data collected by the probe and improve the spatial resolution of ionospheric detection.

Due to the influence of various sudden and abnormal factors during the tracking environment, equipment working conditions, and operation process, there are inevitable outliers in the tracking measurement process of aircraft such as carrier rockets, artificial satellites, and missiles. Therefore, the prerequisite for ensuring the reliability of processing results is to timely and accurately detect and correct outliers. In this paper, we proposed the sliding window-based variable degree B-spline function method (SWVD B-spline), which can handle isolated outliers and spotted outliers. SWVD B-spline uses variable degree B-spline function to model observed data in sliding windows, which can detect and correct outliers point by point along with window sliding. Then, we propose an initial window data selection method to remove outliers in initial windows to ensure the processing effect. In addition, because there are often inflection points in external ballistic velocity measurement data, differential evolution is used to optimize variable degree B-spline in windows that include inflection points to improve processing accuracy. The experimental results verify that SWVD B-spline can handle various outliers rapidly and efficiently.

Measurement data plays an important role in the control system, but the data collected by sensors often has measurement noise, which makes the states of the system cannot be accurately revealed. Unscented Kalman filter (UKF) is a highly accurate and robust filtering algorithm, but its limitation is the requirement of prior knowledge of the exact dynamic mathematical model, which is a critical issue to be addressed in practice. In this work, a data-driven dynamic data reconciliation scheme called nonlinear auto regressive Elman neural network (ENN) with exogenous inputs combined with UKF (NARX-ENN-UKF) is proposed, where nonlinear auto regressive ENN with exogenous inputs is used for dynamic data-driven modeling, and then UKF is applied for dynamic data reconciliation of the measurements based on the trained model. The scheme is applied to a DC/AC inverter experimental system and a self-developed sliding electrical contact experimental system to verify the effectiveness of NARX-ENN-UKF.

To address the problem of the difference in temperature between the exposed and amorce-filled state of the bridge wire in the fiber optic thermometry, this paper proceeds from a mathematical model of the temperature rise of the hot bridge wire electro-explosive device (EED), the relationship between the critical ignition temperature and the current of hot bridge wire EED is discussed. Combining thermal ignition theory, the technique of measurement and calibration the critical ignition temperature of hot bridge wire EED is proposed, and the technique is validated by injection versus radiation. The results show that the critical ignition temperature of the tested hot bridge wire EED obtained by testing and calibration is 337.53 °C. The 50% ignition field strength of the hot bridge wire EED under radiation conditions obtained by the up-and-down method test is 1271 V m⁻¹, while the equivalent ignition field strength determined from the critical ignition temperature under the same radiation conditions is 1345 V m⁻¹, the error between the two is only 0.49 dB, which demonstrates the high accuracy of the proposed measurement and calibration technique.

Domain adaptation (DA) based intelligent fault diagnosis (IFD) methods have attracted great attention in recent years. The key motivation of DA methods is to extract the domain-invariant features. In most cases, the one-dimensional convolutional neural networks are adopted as the feature extractor, in which the kernels are usually single and fixed. However, the monitoring data for IFD usually involve various scale information, and the feature representation extracted by the above models may be incomprehensive. Moreover, the target domain data is only used to narrow the distribution discrepancy in an unsupervised way, which may lead to ignorance of class information of the target domain. To address these issues, in this paper, a two-stage multi-scale domain adversarial fault diagnosis method is proposed. A multi-scale feature extractor with different kernel sizes is designed to acquire more discriminative domain-invariant features. Meanwhile, pseudo label learning is adopted to provide the transfer learning process with the pseudo labels of the target domain, which are generated by a pre-trained network in the first stage and then are optimized through iterations in the second stage. The maximum mean discrepancy is also adopted to enhance the ability of the model for marginal distribution alignment, which can make the model more robust. Thirty-eight transfer tasks from two different datasets were conducted to evaluate the effectiveness of the proposed method. The experimental results demonstrated that the proposed method achieved higher average diagnosis accuracy compared with several popular methods. The superiority of our proposed method was further explained by visualization of learned features.

A multistage singular spectrum analysis based method is presented to extract the useful component from the residue stage by stage. In every stage of the decomposition, the sum of the signal-dominated SSA components is treated as the denoised signal. The signal-to-noise ratios of the denoised signals first increase and then decrease. A measurement called spectrum overlap factor (SOF) is proposed to estimate the optimal stage which achieves the highest SNR. First, the factor is calculated to measure the spectrum overlap degree between the residue and the denoised signal in every stage. Then, the curve of the SOF with respect to the number of stages is analyzed. Further, the minimum of the SOFs, which indicates the less spectrum overlap, allows for the estimation of the optimal stage. The proposed strategy avoids inappropriate parameter selection effectively since the estimation of the optimal stage is automatic. Besides, simulation results show that the proposed method has satisfactory denoising performance in different test scenarios.

Sphericity is an essential geometric parameter for determining the processing manufacturing quality of spherical components. This study presents a sphericity measurement method based on telecentric imaging mechanism with image distortion correction and eccentricity error compensation. A sphericity measurement system is designed, and the projection images of the cross-section on the equatorial plane of the measured sphere are acquired using a line scan camera with bilateral telecentric lenses. An analytical camera model for this sphericity measurement system is established by considering major sources of image distortion. The subpixel edge points of the projection images are extracted and converted to cross-sectional profiles on the equatorial plane according to the camera model. An eccentricity error compensation strategy for contour matching of the obtained cross-sectional profile on the equatorial plane is also developed to obtain the three-dimensional contour point coordinates of the sphere. Sphericity is then calculated according to the least square criterion. The effectiveness of the proposed method is verified by measuring two steel balls with diameter in 20 mm and 30 mm. The sphericity measurement error is reduced by 0.3383 mm and 0.4653 mm respectively after image distortion correction and is reduced by 0.2268 mm and 0.0795 mm respectively after eccentricity error compensation. The proposed method can be extended to form error measurements for opaque components with rotary structure on the outer contour.

Fault diagnosis of axle box bearings is an important technology to improve the service safety and economy of high-speed trains. However, it is difficult to obtain sufficient fault samples in actual train operation, which limits the application of deep learning methods in the field of high-speed train fault diagnosis. Therefore, in this paper, a simulation and test data fusion-driven fault diagnosis method for axle box bearings with few samples is proposed. In this method, a bearing dynamics model is constructed to obtain simulation data for fault bearings, and the accuracy of the constructed dynamics model is verified by experimental data. Data fusion methods are proposed which use massive simulated fault samples and only normal test samples to construct new samples for the pre-training of deep learning diagnostic models. The parameter transfer strategy is used to fine-tune the diagnostic model with a small amount of test fault samples. A real train experiment is carried out for validating the proposed methods which use the real fault bearings that fail during operation. The results show that the proposed method via the fusion of simulation and test data has higher generalization ability and diagnostic accuracy in small sample situations.

Federal learning based on parameter sharing under the assumption that the data obey independent identical distribution (IID has already achieved good results in areas such as fault diagnosis. Data collected by the decentralized devices often do not obey IID. However, when faced with the scenario of client data obeying Non-IID distribution, its diagnostic accuracy is usually weak. Based on this, we did an investigation on the mechanism causing this phenomenon and found that it was attributed to the weight shift of the network. Therefore, based on the stimulus response principle, we investigated the network similarity of federal clients under different data distributions and explain the reasons for the weight shift. Firstly, it was pointed out that there are differences in the regions where the network is activated when performing different classification tasks. Then, similarity metric federal learning (FedSiM) was proposed based on the principle that there are differences between the activated regions. Finally, experiments were designed on the Case Western Reserve University bearing failure dataset for different degrees of IID cases. The results show that FedSiM can improve the diagnostic accuracy by 15.8 percentage points in the case of Non-IID, and a few shared FedSiM methods to further improve the accuracy were also given.

Due to rolling bearings usually operate under fluctuating working conditions in practical engineering, the raw vibration signals generated by bearing faults have nonlinear and non-stationary characteristics. Additionally, there is a lot of noise interference in the collected bearing vibration signal, which indicates that it is difficult to extract bearing fault information and obtain a satisfactory diagnosis accuracy via using traditional method. Deep learning provides a shining road to address this issue. Nevertheless, traditional deep network model has the shortcomings of poor generalization performance and weak robustness in the feature learning. To improve fault recognition accuracy and obtain a favorable anti-noise robustness, this paper proposes a novel bearing fault diagnosis approach based on attention mechanism-guided residual convolutional variational autoencoder (AM-RCVAE). Firstly, the improved residual module is constructed to overcome the convergence difficulty problem caused by network degradation and promote the model generalization performance by replacing the batch normalization (BN) layer in the traditional residual module with the adaptive BN layer. Subsequently, by incorporating the convolutional block attention module and the improved residual module into convolutional variational autoencoder, a deep network model termed as AM-RCVAE is presented to automatically learn fault features from the original data and perform fault diagnosis tasks. The effectiveness of the proposed approach is verified via two experimental cases. Moreover, the recognition accuracy and diagnostic performance of the proposed approach have been certain improved compared with several representative methods.

Currently, rolling bearings operate in harsh environments, resulting in acquired signals with a low signal-to-noise ratio. In light of this, this paper proposes an improved variational modal decomposition(VMD) combined with refine composite multi-scale fuzzy entropy (RCMFE) and linear support vector machine (LSVM) for fault diagnosis. Firstly, the sailfish optimization (SFO) algorithm is employed to optimize the important parameter combinations in the VMD algorithm, using the envelope entropy as its objective function. The analysis includes both simulated and real measured signals with varying signal-to-noise ratios. The results demonstrate that, compared to traditional manual parameter setting and empirical modal decomposition methods, this approach effectively addresses the parameter setting issue of VMD in the signal decomposition process. Additionally, it successfully eliminates noise to extract the fault characteristic signal hidden within the original signal. Secondly, the RCMFE algorithm is introduced to overcome the problem of commonly used dimensioned and dimensionless indicators being influenced by load and speed when used as characteristic indicators. By analyzing the influence of load and speed on the RCMFE value, the results demonstrate its strong stability as a feature indicator, unaffected by these factors. For the intelligent classification of failure type and damage degree, LSVM is chosen as the classification method. Analysis results indicate that the distribution characteristics of RCMFE values align better with LSVM compared to the common radial basis function support vector machine, resulting in a significant improvement in diagnosis accuracy.

Hot ring rolling (HRR) is an advanced incremental metal-forming technology to manufacture high-performance rings. Nowadays, the mechanical or laser-sensors measurement methods can only obtain one value of the dimensions on the ring circumference per time unit, the measurement data is single and the information is limited. This paper presents a real-time vision measurement method for multi geometric information of the rolled ring such as ring's diameter, center position, circularity and growth speed in HRR process. Firstly, the deep learning model is constructed to intelligently and quickly identify rolled ring targets from complex image backgrounds, then the adaptive linear gray scale transformation algorithm is proposed to adjust the gray scale contrast of the image according to changes in ring temperature, the interference of oxide scales and flying chips are eliminated by morphology operation. Finally, the least square method is used to fit the contour points, and further the rolled ring's geometric information and corresponding variation trend are calculated. Several experiments are conducted on a vertical HRR mill, the geometric information of the rolled ring during the HRR is measured in real-time under the interference of harsh working conditions. For the ring with outer diameter of 350 mm, the measurement error of ring's outer and inner diameter is less than 0.8 mm, the average processing time per image takes about 80 ms. The research provides a basis for accurate measurement of HRR process.

The remaining useful life (RUL) of bearings in space inertia actuators is crucial for performance maintenance requirements. But it is quite difficult to accurately predict the RUL of space bearings due to the significant intermittency and nonstationary properties caused by cage friction faults commonly occurring during the operation of the actuator. This paper proposes a data-driven method for RUL prediction of space bearings by incorporating the gated recurrent unit network with a novel data pre-screening approach. In the proposed method, a stacked autoencoder and clustering approach are introduced into the data pre-processing method, and a health index called Overrun-Distance is constructed for lifetime assessment. To verify the proposed method, a series of vibration tests on flywheels equipped with space bearings are conducted and used for RUL evaluation. The results show that the proposed RUL prediction method is applicable to space bearings for RUL prediction with high accuracy and effectiveness.

Bearings are key components of rotating machinery, and their fault diagnosis is essential for machinery operation. Bearing vibration signals belong to time series data, but traditional convolutional neural networks (CNNs) or recurrent neural networks cannot fully extract the fault features from these signals. To address the insufficient feature extraction and poor noise resistance, this paper proposes a fault diagnosis model based on continuous wavelet transform (CWT), CNN with channel attention, bidirectional long short-term memory network (BiLSTM) and residual module. Firstly, a parallel dual-path feature extraction mechanism is constructed which takes time-domain signals and time–frequency images transformed via CWT as the input respectively. Then BiLSTM extracts the time features of the signal as one path, and the CNN with efficient channel attention extracts the spatial features as the other path. This parallel neural network contributes to better feature extraction. Then, the residual module is applied to extract the global features to further improve the feature extraction ability and noise immunity. The experimental results demonstrate that the proposed model on the Case Western Reserve University dataset has better diagnostic accuracy under different working conditions and different signal-to-noise ratios than other methods. In addition, the model shows good generalization performance on Jiangnan University dataset.

For different transportation agencies that monitor vehicle overloads, develop policies to mitigate the impact of vehicles on infrastructure, and provide the necessary data for road maintenance, they all rely on precise, detailed and real-time vehicle data. Currently, real-time collection of vehicle data (type, axle load, geometry, etc) is typically performed through weigh-in-motion (WIM) stations. In particular, the bridge WIM (BWIM) technology, which uses instrumented bridges as weighing platforms, has proven to be the most widely used inspection method. For most of the BWIM algorithms, the position of the vehicle's axle (i.e. vehicle wheelbase) needs to be measured before calculating the axle load, and the identification of the axle load is very sensitive to the accuracy of the vehicle wheelbase. In addition, the vehicle's wheelbase is also important data when counting stochastic traffic flow and classifying passing vehicles. When performing these statistics, the amount of data is often very large, and the statistics can take years or even decades to complete. Traditional manual inspection and recording approaches are clearly not up to the task. Therefore, to achieve automatic measurement of the on-road vehicles' wheelbase, a framework based on computer vision and view geometry is developed. First, images of on-road vehicles are captured. From the images, the vehicle and wheel regions can be accurately detected based on the You Only Look Once version 5 (YOLOv5) architecture. Then, the residual unified network model is improved and an accurate semantic segmentation of the wheel within the bounding box is performed. Finally, a view geometry-based algorithm is developed for identifying vehicle wheelbase. The accuracy of the proposed method is verified by comparing the identified results with the true wheelbases of both two-axle vehicles and multi-axis vehicles. To further validate the effectiveness and robustness of the framework, the effects of important factors, such as camera position, vehicle angle, and camera resolution, are investigated through parametric studies. To illustrate its superiority, the developed vehicle wheelbase measurement algorithm is compared with two other advanced vehicle geometry parameter identification algorithms and the results show that the developed algorithm outperforms the other two methods in terms of the degree of automation and accuracy.

As the significant ancillary equipment of coal-fired power plants, coal mills are the key to ensuring the steady operation of boilers. In this study, a fault diagnosis model was proposed on the basis of a dynamic model of a coal mill and deep belief network (DBN). First, a dynamic coal mill model that considered the joint influence of drying, ventilation and grinding forces was established. Parameters in the model were identified by designing a two-phase optimization method based on the genetic algorithm. Then, this model was used for simulating the common faults of coal mills under a variety of operating conditions and obtaining extensive data. On this basis, the DBN fault diagnosis model was established and the combination of parameters was optimized by use of an orthogonal experiment. Finally, the validity of the model was verified by using the actual operation data of the coal mill. Compared with the dynamic models built in previous studies, that constructed in this paper can significantly improve the capability to simulate and analyze the coal mill. The convergence rate of the designed two-phase optimization method was improved. The experimental results show that the proposed method of coal mill fault diagnosis based on the dynamic model and DBN has an accuracy of 95%, which proves that this method has excellent application potential.

In the study of the high-speed dynamic balance of flexible rotors, rotor unbalance positioning is a challenging topic. Particularly for slender rotors, the axial position of the unbalance has an important influence on the high-speed dynamic balance. The unbalance at different axial positions is not the same or even opposite in different rotor mode vibration behaviors. If the unbalance position of a rotor can be identified, the actual unbalance of the rotor can be reduced from the root. This balance method has the same effect in each vibration mode of the rotor; hence, low-speed dynamic balance can be realized to replace high-speed dynamic balance, considerably saving on costs. Deep learning based on few labeled samples can achieve good results for the identification of unbalanced positions; however, there are infinite potential positions of unbalance in the actual rotor. It is difficult to collect sufficient labeled samples to train a reliable intelligent diagnostic model. Fortunately, a large number of rotor vibration datasets labeled with different unbalance positions are available using the rotor dynamic model, and the unbalance position data calculated using the dynamic model contain diagnostic knowledge related to the rotor unbalance position data measured in the rig. Hence, inspired by transfer learning, this study proposed a transfer learning method using dynamic model simulation and experiment data for flexible rotor unbalance fault location. Cross-domain deep transfer recognition of rotor unbalance position was realized.

In this study, the indoor localization problem for human movement is built as a sparse representation, and a learnable sparse coding algorithm is proposed by introducing a nonconvex penalty function as the sparse constraints. The good properties of the penalty are able to help reduce the iteration cost to find a sparse solution indicating the target locations. The sparse code can achieve its optimal value from the trained neural network model with parametrization. It is instead of a piece of algorithmic processing in the conventional methods requiring a large iteration. A real-world experiment using a Bluetooth low energy wireless sensor network validates the proposed method and shows the improvement of the error reduction at the first five iteration points on average compared to iterative shrinkage and thresholding algorithm (ISTA), and achieves competitive performance in the localization tasks for learned-ISTA-based algorithms.

An artificial intelligence (AI) technical system is the foundation of the internet of things. While in the technical system, the material identification technology is the key point. Material identification technology can be applied to AI robots and smart skin. However, most of the traditional material identification technologies have limitations such as high cost and complicated operation, while the low-cost ones have the problem of low accuracy. To accurately, efficiently, and simply identify objects is a bottleneck that researchers urgently need to overcome. Based on the thermoelectric effect, a material identification theory is proposed and a prototype is fabricated in this paper. A series of experiments are carried out to test the performance of the prototype. According to the results, the prototype can precisely identify metal materials and nonmetal materials less than 3 s. There are many merits of the prototype, such as high identification accuracy, low cost, and low maintenance difficulty. Besides, considering complex surface of materials may decrease the identification accuracy, this paper proposes a kind of flexible thermal conductive material to improve identification accuracy. It is verified that the identification accuracy is improved with the flexible thermal conductive material.

It is still a main challenge to achieve accurate and reliable vehicle positioning in Global Position System (GPS)-denied environments using low-cost sensors. Although existing methods have obtained a certain performance improvement, there is still room for further enhancement in positioning accuracy due to the lack of effective observation information during GPS outages. To address this challenge, this paper proposes an effective fusion positioning methodology based on enhanced observation information. Firstly, a data denoising algorithm based on convolutional denoising autoencoder is developed to effectively filter the noises in inertial sensors, so as to provide more clean data for subsequent modeling. Then, a differential fusion strategy is designed to selectively fuse multiple observation information from monocular camera, low-cost GPS and wheel speed sensor with extended Kalman filter, which can further improve the positioning accuracy. Finally, a long short-term memory-based error prediction model is constructed to learn the mapping relationship between the result of data denoising and differential fusion. In the case of GPS outages, the error model provides accurate corrections to compensate for position errors of inertial navigation system. The performance of the proposed methodology was evaluated on real-world data collected in complex urban environments. Experimental results indicate that the proposed positioning methodology can obtain significant accuracy enhancement.

Intelligent perception of a scraper conveyor straightness and attitude monitoring of mechanical supporting equipment in the stope have practical and theoretical values for mining. This study proposed an optical fiber curvature sensor and a scraper conveyor's curve reconstruction method. The optical fiber curvature sensor comprises the fiber grating strain sensing optical cables, the flexible substrate, and the packaging material. The coordinate positions of each monitoring point are obtained through the strain–curvature conversion relationship and the slope recurrence algorithm, and then the reconstruction curve is obtained by fitting. The finite element simulation verifies the feasibility of the curve reconstruction method used for the deformation monitoring via optical fiber curvature sensors. The reconstruction error analysis results show that the root mean square error of reconstructions for two kinds of 2D plane bending and 3D space bending are 2.98%, 1.89%, and 3.13%, respectively. Their mean absolute errors are 8.9, 3.56, and 9.82 mm, respectively, verifying the feasibility and high accuracy of the proposed curve reconstruction equation. The research results provide a theoretical basis for the shape perception and straightening control of scraper conveyors in the intelligent working surface.

Complexity measures typically represented by entropy are capable of detecting and characterizing underlying dynamic changes in a system, and they have been considerably studied for machine condition monitoring and fault diagnosis. Various entropies have been developed based on Shannon entropy to meet actual demands. Nevertheless, currently existing research works about complexity measures mainly focus on experimental studies, and their theoretical studies are still ongoing and not fully explored. In previous studies, it was theoretically and experimentally proved that two complexity measures including correlation dimension and approximate entropy have a 'bilateral reduction' effect. Since sample entropy and fuzzy entropy are two more advanced complexity measures that were developed based on the concept of correlation dimension and approximate entropy, this paper continues conducting theoretical and experimental investigations on sample entropy and fuzzy entropy and exploring their theoretical properties to enrich the domain of complexity measure analysis and its applications to machine condition monitoring. Specifically, this paper theoretically proves and verifies that sample entropy and fuzzy entropy still have a similar 'bilateral reduction' effect with correlation dimension and approximate entropy, and they are indeed complexity measures. The relationships between sample entropy, fuzzy entropy, and their key parameters during their calculation are numerically and experimentally studied. Bearing and gear run-to-failure datasets are used to investigate the effectiveness of sample entropy and fuzzy entropy for bearing and gear condition monitoring, and experimental results of sample entropy and fuzzy entropy are well-matched with the theoretical 'bilateral reduction' effect of sample entropy and fuzzy entropy. Overall, this paper will provide a guideline for correct uses of sample entropy and fuzzy entropy for engineering applications, especially for machine condition monitoring.

In recent years, intelligent condition monitoring and diagnosis based on deep learning have made great progress. However, traditional diagnostic methods mostly perform vibration analysis based on accelerometer signals, ignoring the influence of sensors on the mass load of the measured object. On the other hand, conventional transfer learning (TL) methods are mostly based on global distribution alignment to achieve intelligent diagnosis under variable working conditions. In this paper, a deep global subdomain adaptation network (DGSAN) is proposed to solve the intelligent diagnosis problem under variable working conditions based on vibration image and TL. First, visual measurement is introduced in vibration extraction. Based on the phase vibration extraction method, the vibration feature information is obtained from the visual vibration image to construct the vibration dataset. Then, the proposed DGSAN establishes a multi-layer domain adaptive network to minimize the difference in feature distribution and realize fine-grained feature distribution alignment of fault data under variable working conditions. Comparative experiments are carried out on the vibration image datasets of rotor-bearing systems, and the results show that the proposed method achieves high-precision transfer intelligent diagnosis.

Integrity monitoring with a Kalman filter (KF) has recently attracted significant attention. In this paper, a computationally efficient architecture of a KF-based receiver autonomous integrity monitoring (RAIM) algorithm is discussed for aviation applications to ensure reliable operations of Global Navigation Satellite Systems (GNSS). It is built on the Schmidt KF navigation processor to model time-correlated measurement errors. Reasons for important design choices of the algorithm are clarified. Different strategies are adopted to efficiently include the contributions of past KF measurements in fault detection as well as protection level (PL) calculations. Module-wise most significant numerical complexity is also analyzed in detail. The algorithm performance is studied with simulated Global Positioning System (GPS) and Navigation with Indian Constellation (NavIC) signals for a number of scenarios. They comprise different configurations related to the number of satellites, geometry, total duration, and aircraft dynamics. Fault detection performance of presented KF RAIM is shown to be superior to another innovation-based test with a moving time window. It is demonstrated that KF RAIM running on a single-core virtual machine can complete processing within a small fraction of each time interval. The performance is also analyzed by restricting CPU usage. The processing time of GPS-NavIC KF RAIM at every interval is shown to be consistently less than that of standalone GPS in all scenarios. Therefore, dual constellations not only result in lower PLs, but also require shorter execution times. An explanation for faster execution times with dual GNSS is provided using the numerical complexity of different modules.

Depth estimation is one of the key technologies in some fields such as autonomous driving and robot navigation. However, the traditional method of using a single sensor is inevitably limited by the sensor's performance. Therefore, a precise and robust method for fusing LiDAR and stereo cameras is proposed. This method fully combines the advantages of the LiDAR and stereo cameras, which can retain the advantages of the high precision of the LiDAR and the high resolution of images respectively. Compared with the traditional stereo matching method, the texture of the object and lighting conditions have less influence on the algorithm. Firstly, the depth of the LiDAR data is converted to the disparity of the stereo camera. Because the density of the LiDAR data is relatively sparse on the y-axis, the converted disparity map is up-sampled using the interpolation method. Secondly, in order to make full use of the precise disparity map, the disparity map and stereo-matching are fused to propagate the accurate disparity. Finally, the disparity map is converted to the depth map. Moreover, the converted disparity map can also increase the speed of the algorithm. We evaluate the proposed pipeline on the KITTI benchmark. The experiment demonstrates that our algorithm has higher accuracy than several classic methods.

Remaining useful life (RUL) prediction is the main approach to guarantee the reliability of lithium-ion batteries. This paper proposes an interpretable hybrid method to predict the RUL distribution with changeable form. The method integrates prior knowledge from the statistical model and regular patterns learned from monitoring data based on the data-driven model. The predicted compound distribution provides more information compared to point estimation and distribution with fixed form. The general hybrid framework contains a component learner, a fusion model with a stacking strategy, and a prognostic distribution algorithm with adaptive sampling weights. The stacking fusion model is implemented by a one-dimensional convolution neural network. The sampling weights are estimated by optimal estimation. The statistical model describes the individual capacity degradation path based on the Wiener process. The data-driven model learns the degradation process from historical data based on convolutional gated recurrent neural network (CNN-GRU) and Monte Carlo dropout simulation. The comparative experiments between the proposed method and existing methods were carried out. The experiment results show that the proposed hybrid method performs well.

Rolling bearings are essential parts of rotating equipment. Due to their unique operating environment, bearings are vulnerable to failure. Graph neural network (GNN) provides an effective way of mining relationships between data samples. However, various existing GNN models suffer from issues like poor graph-structured data quality and high computational consumption. Moreover, the available fault samples are typically insufficient in real practice. Therefore, an improved graph convolutional network (GCN) is proposed for bearing fault diagnosis with limited labeled data. This method consists of two steps: graph structure data acquisition and improved graph convolution network building. Defining edge failure thresholds simplifies the generated weighted graph-structured data, thereby enhancing data quality and reducing training computation costs. Improvements to standard GCNs can effectively aggregate data features of different receptive field sizes without noticeably raising the computational complexity of the model. Experiments with limited labeled data are conducted on two public datasets and an actual experimental platform dataset to verify the superiority of the proposed method. In addition, experiments on imbalanced datasets also fully demonstrate the robustness of the proposed method.

Feature engineering is a difficult task, and for real signal data, it is difficult to find a certain feature that can easily distinguish all classes. Multiple features can provide more information, which means the fusion of multi-feature learning strategies has potential significant advantages. Based on this premise, this paper proposes a multi-class framework based on the multi-featured decision to distinguish all the different classes, and takes automatic dependent surveillance-broadcast (ADS-B) signal data as an example, first extracts the phase features and wavelet decomposition features of the signal data, then selects the features with high discrimination between classes, then proposes a one-dimensional residual neural network based on 16 convolutional layers to learn the unique features of different features and classes separately, and finally proposes a novel multi-featured decision method based on voting method and a priori probability. Results show that the proposed one-dimensional residual neural network has better performance metrics on the test set compared to some machine learning-based and neural network-based algorithms, with classification accuracies of 86.1%, 84.6% and 83.6% on wavelet decomposition features, raw features and phase features, respectively, on ADS-B preamble signals. The proposed feature decision framework based on the voting method and a priori probability has a recall, precision and F1 value of 80.24%, 89.89% and 84.79% on ADS-B preamble signals, respectively.

Intelligent fault diagnosis of rotating equipment is increasingly reliant on algorithms that are driven by big data. By contrast, signal processing was once widely utilized for fault diagnosis in machinery as a classical tool for signal analysis due to its capability to investigate the fault-related mechanism and almost no demand on the number of data samples. This investigation was motivated by the notion that signal processing and data-driven algorithms are combined to exploit their respective characteristics and strengths. Furthermore, in engineering practice, numerous complex factors such as time-variable operating conditions of equipment, non-stationary properties of signals, and extremely limited samples available for model training, can make it difficult to learn discriminative features from input data, thereby diminishing the diagnostic accuracy. In this paper, a novel framework of vibration amplitude normalization (VAN) enhanced fault diagnosis is proposed. Firstly, after dissects deeply the effects of the time-varying speed conditions on vibration signal and its characteristics, VAN technique is proposed for non-stationary signal processing to obtain the approximate stationary signal, so as to facilitate the subsequent state characteristics mining from the vibration signal. Then, two VAN enhanced fault diagnosis methods—i.e. signal amplitude normalization integrated with shallow learning by cascade and VAN integrated with deep learning by embedding—are developed to capture discriminative features from approximate stationary signal for fault diagnosis under conditions of variable speed and extremely limited samples. Finally, the feasibility and effectiveness of the proposed methods are verified using actual vibration datasets measured on test rig and in-site wind turbines. The number of samples required to achieve the same diagnostic accuracy is reduced by an average of 60%, demonstrating the superiority.

Recently, the fault diagnosis of rotating machinery based on deep learning has achieved increasingly widespread applications. However, it is often difficult to achieve the expected results by relying on a single sensor due to the limited information obtained by the single sensor and the susceptibility to the influence of the additive noise. To address the above problems, this paper proposes a multi-sensor fusion fault diagnosis method for rotating machinery based on improved fuzzy support fusion and self-normalized spatio-temporal network to enhance feature learning while achieving multi-sensor data fusion. This method includes a data pre-processing module, a fusion module and a fault recognition module. In the first module, a complete ensemble empirical mode decomposition with adaptive noise algorithm is introduced to decompose and reconstruct the multi-source sensor signals, thereby reducing the impact of environmental noise on data quality. In the fusion module, a data fusion algorithm based on improved fuzzy support is designed to achieve the data-level fusion of multi-source sensors. By introducing the self-normalized properties into the convolutional structure with bi-directional gated recurrent unit, a self-normalized spatio-temporal network is designed in the fault recognition module to perform the fault diagnosis of rotating machinery. The experimental results show that the proposed method can achieve high quality data-level fusion and outperforms the state-of-the-art fault diagnosis methods in terms of fault classification.

Pre-seismic anomaly detection plays a crucial role in reducing economic losses and casualties caused by earthquakes. This paper proposes a novel four-step approach for pre-seismic anomaly detection. In the first step, a series of pre-seismic features are extracted by analyzing the earthquake catalog and geomagnetic signals. In the second step, the multi-view learning strategy is employed to obtain fusion features. In the third step, multiple seismic stations in one seismic zone are treated as a seismic station network, and a pre-seismic anomaly detection model is constructed based on the station network. In the final step, four evaluation indicators are introduced to comprehensively evaluate the detection results. Verification results show that the proposed method is effective and achieves better performance than other existing methods.

The inspection of corrosion under insulation (CUI) has been identified as a significant challenge in the petroleum and chemical process industries. As some of the most effective strategies, pulsed eddy current (PEC) techniques have proved effective for the measurement of the CUI of pipelines. In this paper, we propose a circular-arc array (CAA) to improve the measurement efficiency for the PEC inspection of thermally insulated pipelines. Based on the PEC system model for inspecting the CUI of pipelines, the magnetic field distribution of the CAA with multiple excitors was investigated. It is shown that the coverage of induced magnetic field gets much larger than that of the single excitor to realize high-efficiency measurements. Moreover, a sparsely distributed receiver array is designed to further improve the signal-to-noise ratio by eliminating the waviness effect due to multiple excitors. Finally, experiments were conducted, and the results demonstrated the effectiveness of the proposed method for the inspection of thermally insulated pipelines.

Global Navigation Satellite System Interferometric Reflectometry (GNSS-IR) ground height retrieval technology is based on global navigation satellite system (GNSS) signal reflection, which can achieve efficient and high-precision ground retrieval. However, errors cannot be avoided. And whether there is a linear relationship between the height of the station and the error is unknown. This research uses Hi-Target geodetic GNSS receivers, smart phone devices (Honor 60) and smart tablet devices (Huawei MatePad Pro) to collect a total of 5 d data from DOY65 to DOY69 in 2023, with the station heights of 0.8 m, 1.0 m, 1.2 m, 1.4 m and 1.6 m, respectively. The experimental results show that each satellite can effectively establish a linear relationship between the inversion error and the station height, which can be used in the error compensation research of different station heights under the limitation that the height of reflector is between 0.8 m and 1.6 m. Simultaneously, the error is related to the influence of comprehensive factors such as reflector type, satellite number, and data-receiving equipment. Secondly, two clustering methods, k-means and k-media, are introduced to cluster a and b in the linear relationship y = ax + b of each satellite, and it is proved that the linear relationship between inversion error and station height is obviously related to ground reflection surface (plastic track and concrete ground). Finally, it is verified that the height measurement accuracy of low-cost smart electronic equipment (Root Mean Square Error (RMSE): 0.047 m and 0.042 m) is worse than that of GNSS (RMSE: 0.010 m), but it still has good measurement performance. All in all, this study provides an essential technical reference for the error compensation of different station heights and for the application of GNSS-IR with low-cost smart electronic equipment. Due to its low-cost advantage, it has great potential in developing other surface parameter inversion of GNSS-IR technology.

In recent years, unmanned autonomous driving technology has attracted increasing attention from people, and become a research hotspot. Currently, the integration of strapdown inertial navigation system (SINS) and global navigation satellite systems (GNSSs) is the most common and effective navigation and positioning scheme for unmanned ground vehicles (UGVs) and unmanned aerial vehicles. However, this integrated system is unable to maintain a reliable positioning solution in challenging environments due to the inherent weakness of GNSS signals and the accumulation of SINS positioning errors over time. To address this issue, this paper proposes an integrated scheme based on an asynchronous Kalman filter for SINS, GNSS and two-dimensional (2D) laser Doppler velocimeter (LDV). In the proposed scheme, the SINS and 2D-LDV are tightly coupled to improve the robustness of the integrated system, and the error parameters between the 2D-LDV and the SINS are calibrated in real time during the validity of the GNSS signal. In addition, the designed asynchronous Kalman filter method evaluates the validity of the GNSS and 2D-LDV measurements in real time based on Mahalanobis distance of innovation vector and statistical property principle. Two groups of long-distance, high-mobility vehicle experiments conducted in challenging environments verify the validity of the proposed scheme. The experimental results show that the proposed SINS/GNSS/2D-LDV integrated navigation scheme has good environmental adaptability and reliability, and can maintain high horizontal and vertical positioning accuracy despite frequent GNSS signal failures, which can meet the needs of UGVs.

The increase in the number of channels for extracting bearing fault features can to some extent enhance diagnostic performance. Therefore, this article proposes a SENet (squeeze and excitation network)—TSCNN (two flow convolutional neural network) model with high accuracy and generalization characteristics for fault diagnosis of rolling bearings. Firstly, use convolutional pooling layers to construct a basic diagnostic model framework. Secondly, due to the unsatisfactory performance of feature extraction solely on one-dimensional frequency domain signals or two-dimensional time-frequency signals, there may be misjudgments. Therefore, a dual stream convolutional model is integrated to process both one-dimensional and two-dimensional data. Fast Fourier transform is used to process one-dimensional frequency domain data, and continuous wavelet transform is used to process two-dimensional time-frequency maps. Once again, integrating the SENet module into the dual stream diagnostic model, the addition of attention mechanism can enable the model to better understand key features of input data. Finally, the data obtained from the processing of two channels is fused and classified in the Softmax layer. This article uses the rolling bearing fault standard data from Case Western Reserve University and the American Society for Mechanical Fault Prevention Technology, and verifies through multiple controlled experiments that the model established in this article has high accuracy and good generalization characteristics.

In practical industrial applications, rolling bearing generally operates under variable conditions and its vibration signal significantly fluctuates in amplitude and frequency. This increases the feature distribution differences of the bearing fault samples and makes the health status identification of the bearing more difficult. To this end, a new intelligent fault diagnosis method for bearing under time-varying speed conditions is proposed based on time-characteristic order (TCO) spectrum and multi-scale domain adaptation network (MSDAN). Firstly, by using the good noise robustness and high time–frequency aggregation properties of the synchrosqueezed wave packet transform (SSWPT), the TCO spectrum method based on SSWPT is proposed to eliminate the impact of speed fluctuation, reducing the distribution shift of bearing data under time-varying speeds. Secondly, an MSDAN model based on global-local feature fusion is established to extract the domain-invariant features closely related to the bearing fault state from the TCO spectrum. Finally, the local maximum distribution discrepancy is introduced to capture the discriminative fine-grained features. The feasibility of the proposed method is verified in various transfer tasks on two different bearing datasets with time-varying speeds. Compared with some state-of-art methods, the proposed method can eliminate the sample distribution differences under time-varying speeds, significantly improving its accuracy and generalization performance in cross-domain fault diagnosis of rolling bearing.

The strap-down inertial navigation system (SINS)/celestial navigation system (CNS) integrated navigation system has been extensively employed in the aerospace field owing to its unique advantages. The internal parameter errors of the star sensor constitute one of the primary factors that affect the navigation accuracy. In this paper, an error suppression method based on diametrical stargazing has been designed to suppress the navigation errors caused by the internal parameter errors of the star sensor. First, a SINS/CNS integrated navigation model based on Kalman filter has been established. Then the influence of the principal point error, image-plane tilt error, and distortions of the optical system on the measurement accuracy of the star sensor have been analyzed. Further, the process of introducing the errors into the measurement equation has been obtained. Finally, for multiple errors, two stargazing methods have been designed to suppress the impact of these errors on the navigation accuracy, whereas the constant errors of the inertial device can be modulated. The simulation results demonstrate that the proposed method can improve the positioning accuracy of the system, and verify the effectiveness of the algorithm.

Traditional LiDAR simultaneous localization and mapping (SLAM) methods rely on geometric features such as lines and planes to estimate pose. However, in unstructured environments where geometric features are sparse or absent, point cloud registration may fail, resulting in decreased mapping and localization accuracy of the LiDAR SLAM system. To overcome this challenge, we propose a comprehensive LiDAR SLAM framework that leverages both geometric and intensity information, specifically tailored for unstructured environments. Firstly, we adaptively extract intensity features and construct intensity constraints based on degradation detection, and then propose a multi-resolution intensity map construction method. The experimental results show that our method achieves a 55% accuracy improvement over the pure geometric LiDAR SLAM system and exhibits superior anti-interference capability in urban corner scenarios. Compared with Intensity-SLAM, the advanced intensity-assisted LiDAR SLAM, our method achieves higher accuracy and efficiency.

To address the shortcomings of existing bearing remaining useful life (RUL) prediction process such as low accuracy and reliance on expert experience for parameter estimation, this paper proposes a bearing RUL prediction method combining relevance vector (RV) machine (RVM) and hybrid degradation model. The bearing degradation characteristics are extracted from the acquired vibration acceleration signals, the time-varying 3σ criterion is then used to determine the bearing first predicting time, and the sequence from initial failure time point to the inspection time is regressed by differential kernel parameter RVM to obtain the different sparse RVs. A mixed degenerate model combined single exponential, weighted double exponential, and polynomial is used to fit the sparse RVs to obtain the fitted curve clusters. The similarity based on bidirectional Hausdorff distance is used to select the best degradation curve, and to extrapolate the best degradation curve to the failure threshold. The experimental results indicate that the proposed method overcomes the widespread drawbacks of monotonicity and trend bias in model-based methods, and has better prediction efficiency than the conventional exponential models.

Among thermoelectric materials with thermoelectric effect, lead telluride (PbTe) is widely used because of its high performance and chemical stability in the medium temperature region. In this study, PbTe was creatively employed to develop a transient thin films heat flux sensor (THFS) using magnetron sputtering technique based on the transverse Seebeck effect. After static and dynamic calibration, the rise time of the THFS is 35 μs and the sensitivity is up to 7.9 μV(kW·m⁻²)⁻¹, it can be measured in transient high heat flux testing environments without the need for a signal amplifier. In the experiment of measuring transient heat flux with explosion driven shock tube, the THFS has the advantages of high resolution and high dynamic response, which provides scientific basis for the study of explosion thermal damage effect and has important significance.

Rotating machinery is one of the most common components in the industry. Therefore, timely and accurate fault diagnosis of rotating machinery is essential for the regular operation of equipment. At present, some achievements have been made in rotating machinery fault diagnosis based on a large number of marked fault data. However, most of the machines are in a normal state in actuality. Especially, the machines run under different loads, so it is costly to collect a large number of labeled fault data under different load distributions. To solve rotating machinery fault diagnosis in different load conditions with limited samples, a domain adaptive group convolutional neural network is proposed. Firstly, the least squares generative adversarial networks were used to expand the limited target sample data. By changing the objective function, the two defects of the low quality of the vibration signal generated by the traditional generative adversarial networks and the unstable training process are optimized. Secondly, the raw vibration signals in the source domain are pre-trained by the group convolutional neural network, and the group training network effectively reduces network parameters. Finally, the source domain signals and target domain signals were trained in domain adversarial networks to diagnose different distributed data in target domains. The proposed method is validated by collecting the raw vibration signals of gears under different loads and different health states, and the effectiveness of the proposed method is proved. Experimental validation shows that the method proposed in this paper achieves an average accuracy improvement of more than 12% compared to other existing methods.

This work proposes a signal preprocessing framework that cancels out impulse noise in the non-stationary gearbox vibration signals of a ship unloader, which reduce the interference of non-cyclic impulses on the damage detection. The algorithm involves three main steps: (i) Preliminary localization of impulse noise, including suppressing components related to normal machine operation, segmenting the signal, and localizing abnormal sub-segment signal based on higher-order statistical criteria. (ii) Detection of impulse boundaries. The influence of random impulse noise on the signal envelope spectrum is studied, and an impulse indicator is constructed to accurately detect the boundaries of impulse noise in abnormal sub-segment signals. (iii) Cancellation of impulse noise. Replace the identified random impulse with Gaussian white noise. Laboratory and industrial data confirm the excellent performance of the proposed algorithm. It can accurately identify and cancels out the non-cyclic impulse noise in the non-stationary gearbox vibration signal, which facilitates the subsequent damage detection.

In this paper, an inspection device to detect the inner wall defects of oil tubes is proposed, mainly consisting of a micro laser distance sensor and a stepper motor. When the inspection device moves along the axial direction inside the tube, the stepper motor drives the micro laser distance sensor to rotate and scan the inner wall of the tube so that the coordinate data of the 3D point cloud can be obtained. In order to obtain the pose transformation matrix for the laser distance sensor coordinate system with respect to the device flange coordinate system, a particle swarm optimization algorithm is introduced. Finally, experiments are carried out on two real oil tubes and point cloud data are obtained which provide 3D information about the inner walls of the tubes. The measuring error of tube eccentric wear is within 0.2 mm, which verifies the effectiveness of this novel method.

Vibration amplitude and frequency are the two most important indicators that characterize the health status of high-speed rotating blades, but the signal obtained by blade tip timing (BTT) technology, one of the best rotating blade vibration monitoring methods, is seriously nonuniform and under-sampled, which makes these two indicators difficult to identify. In view of this problem, the paper proposes a parameter identification method for the nonuniform and under-sampled BTT signal based on extended Discrete Fourier transform and compressed sensing (CS), with the Fourier integral transformation as the goal. It realizes the frequency analysis of nonuniform under-sampled signals by constructing and optimizing the transformation basis function instead of the exponential basis in the traditional FFT transformation in the extended frequency range, and then constructs a CS model through the obtained blade vibration frequency. The complete waveform of the blade vibration is restored by using a small number of under-sampled signals, thus obtaining the blade vibration amplitude and vibration frequency. On the one hand, the method proposed in this paper breaks through the limitation of Nyquist's sampling theorem, and the number of analytical spectral lines is no longer limited to the number of sampling points, which improves the frequency resolution. On the other hand, only a small number of measurement signals can be reconstructed to achieve a complete vibration signal. The feasibility and reliability of the proposed method are verified by mathematical modeling, simulation analysis, and experimental testing. The results indicate that when the number of sensors is greater than or equal to four, the time domain and frequency domain signals of blade vibration can be accurately analyzed based on the proposed method, the vibration amplitude error is less than 0.01 mm, the frequency error is less than 0.1 Hz, and it has good anti-interference performance.

Sensor linearization is an important aspect for enhancing the efficiencies of measuring systems. Conventional methods use additional circuits and/or software models to achieve linearization. The direct interface technique obviates the requirement for intermediate electronic circuits, including linearization circuits, between sensors and embedded systems. Sensor linearization strategies need to be explored to obtain maximized performances from directly interfaced sensing systems. In this work, the performances of both hardware- and software-based strategies for linearization of directly interfaced thermistor sensors are evaluated. The experimental results show that the hardware-based linearization approach (1P_Shunt, 2P_Shunt) can yield a maximum linearized output range, i.e. from 0 °C to ∼70 °C with <1% full scale span (FSS) nonlinearity error (NLE). The sensitivities in both cases are found to be comparable. In 1P_Shunt, the linearized range is found to be almost independent of the β-values, ranging from 3012 K to 3924 K. A high-speed timer, via minimization of quantization errors, allows a significant reduction in measurement time while maintaining the linearized range. With the artificial neural network based linearization approach, a linearized range up to 100 °C and beyond can be achieved. A shallow network with optimum architecture (1-5-1), with Bayesian regularization and log-sigmoid as an activation function, is found to be sufficient to yield <1% FSS NLE.

Obstacles that intrude into the rail area can lead to serious rail accidents, so obstacle detection technology is an essential guarantee for the safe operation of fully automatic trains. To meet the high-performance requirements of onboard obstacle detection, an efficient feature-aware convolutional neural network (EFA-Net) is proposed in this paper. The multi-scale aware feature pyramid network (MA-FPN) is designed as feature fusion network to extract multi-scale context information. In the detection head, the dynamic awareness block is used to refine the features. A joint representation branch and the generalized focal loss function are introduced to optimize the training effect. The experiments are based on the dataset of real-world rail transit environment. The results show that EFA-Net can achieve a detection accuracy of 90.4% mAP at a detection speed of 20.4 frames per second, and the lightweight design significantly reduces the computational complexity of the proposed model. Compared with other classical detectors, EFA-Net has the best comprehensive performance.

This paper proposes a novel composite measurement method for defect detection, which integrates PMFP (permanent magnetic field perturbation) testing and lift-off measurement. The PMFP method is used to detect defects, while the permanent magnet is utilized to construct a magnetic bridge for lift-off measurement. In PMFP testing, the variation of lift-off can significantly affect the signal. To mitigate this effect, the proposed method compensates for the defect signal using the measured lift-off information, which effectively reduces the impact of lift-off variation during the testing. Therefore, this paper constructs a magnetic dipole model based on the PMFP method and analyzes the numerical relationship between the defect signal and lift-off. Based on this, a composite lift-off measurement method for PMFP with a signal compensation function is proposed. This multifunctional sensing method has a compact structure, which significantly reduces the impact of lift-off on the defect signal. The experimental results demonstrate that the confidence level can reach 93% after compensating for the amplitude of the defect signal. The proposed method exhibits excellent performance in practical applications.

Simultaneous localization and mapping (SLAM) plays a key role in 3D environment modeling and mobile robot environment perception. However, the traditional discrete-time Laser-inertial SLAM methods are not robust due to the imbalanced registration steps between a single LiDAR frame and the global map. This paper proposes a tightly coupled laser-inertial pose estimation and map building method that uses B-spline curves to represent continuous-time trajectory and achieve high robustness of the registration steps. To ensure efficiency, the proposed method separates the SLAM task into an odometer module and a mapping module. The odometer module performs a coarse pose estimation, while the mapping module performs a fine one and builds a global map with 3D LiDAR points. B-spline curves are utilized to integrate both IMU measurement constraints and LiDAR point constraints in the proposed mapping module, which can enhance the association of consecutive LiDAR frames in the optimization step. Besides, the explicit expression of the Jacobi matrix derivation for B-spline-based laser residuals is also introduced to furtherly improve the computation efficiency. Both indoor and outdoor experiments are conducted on a self-collected dataset and a public dataset. Experimental results show that the proposed method can achieve superior performance than the baseline method LIO-mapping.

Applying methods such as deep learning improves the efficiency of bearing fault diagnosis and reduces trains' operation and maintenance costs. However, in practical applications, the deficiency of historical data and the imbalance of data types often limit the effectiveness of the diagnosis. The variability between operating conditions also restricts the availability of transfer learning including domain adaptation. To address this challenge, a digital twin (DT) framework is established to fill the data for train fault diagnosis. A train bearing dynamics model is optimized using virtual-reality mapping in the DT framework with measured health data as a baseline to generate data closer to reality. Finally, the fault diagnosis uses a hybrid dataset that mixes measured and simulated data as a source domain for transfer learning. The Case Western Reserve University dataset is used as an example, and the accuracy reaches up to 99.40%, which verifies the method's effectiveness.

Relaxation oscillator-based circuits are widely used for interfacing various resistive and capacitive sensors. The electrical equivalent of most resistive and capacitive sensors is represented using a parallel combination of resistor and capacitor. The relaxation oscillator-based circuits are not suitable for parallel R–C sensors. In this paper, we propose a modified circuit for parallel R–C sensors. The proposed relaxation oscillator-based circuit is based on a dual-slope and charge transfer technique to measure the resistance and capacitance of parallel R–C sensors separately. In addition, the paper provides a detailed analysis and design considerations for the oscillator design by taking into account the various sources of non-idealities. A method to reduce the error by using single-cycle averaging is also introduced. To verify the analyzed design criteria, the circuit is tested with multiple operational amplifiers with different non-idealities. Experimental results verify the performance of the proposed circuit. The circuit is tested for a range from 10 pF to 42 pF and 100 kΩ to 1 MΩ for parallel R–C sensors with an error of less than 1.5%. The circuit is tested with a fabricated water-level sensor. The result confirms the efficacy of the proposed circuit.

In recent years, deep learning has been increasingly applied to fault diagnosis and has attracted significant attention and research interest. Deep reinforcement learning (RL), with its capabilities in feature extraction and interactive learning, is highly suitable for fault diagnosis problems because it can acquire knowledge solely via system feedback. Despite its advantages, this method also has limitations, such as low training efficiency and unstable performance. Therefore, this study presents a novel diagnostic approach based on system feedback for rolling bearing fault diagnosis. This approach builds upon the original deep Q-network (DQN) approach, which incorporates an interactive dual network structure and experience replay optimisation for RL intelligence. This method introduces two major improvements. First, a dual network cyclic update scheme is implemented, assigning each dual network specific responsibilities to ensure training stability. Second, a novel experience playback system is introduced, which improves the efficiency of experience utilisation while circumventing the risk of overfitting. Compared with the original DQN method, the proposed approach and its two enhancement strategies provide significant advances in training efficiency, stability and diagnostic accuracy. Our experimental results indicate that this novel methodology has the potential to make valuable contributions in the area of rotating machinery fault diagnosis.

In this work, a low-cost double split complementary split ring resonator (CSRR) microwave sensor has been designed and fabricated on an FR4 substrate for permittivity characterization of liquids. This modified CSRR structure is proposed to make use of the benefits of the dielectric resonator technique compared to other types of microwave methods. The resonator structure comprises two slits in each of the rings normal to each other while maintaining one slit in each ring along the feed line for maximum excitation of the resonator. The resonator with substrate dimensions of $20\;{\text{mm}}\; \times \;30\;{\text{mm}}$ acts like an inductance-capacitance-resistance (LCR) circuit and operates at $2.501\;{\text{GHz}}$ of the industrial scientific and medical band. Eleven different liquid samples (each with $150\,\mu l$ of volume) covering a wide permittivity range of $1 - 111$ have been used to measure the transmission coefficient $\left( {{S_{21}}} \right)$ using vector network analyzer. In comparison with many recently reported works, the sensitivity of the sensor is found to be higher with average and normalized sensitivity of $35.07\;{\text{MHz/unit}}$ permittivity and $1.4\%$ respectively. Fit equations are developed for both real and imaginary parts of permittivity, and using the fit equations the permittivity of three unknown samples are determined with less than $3.7\%$ error. The sensor adopts a unique orientation of the slits to enhance the E-field intensity for better interaction of the samples with the sensor. A compact form factor, low production cost, future integration possibilities, and high sensitivity are the distinct highlights of the sensor. The sensor promises to be a potential candidate for situations that demands low-cost and highly accurate measurements and may be a good alternative to commercial sensors for dielectric characterization, concentration analysis, and impurity measurement of liquids.

Effective health indicator (HI) construction can help equipment managers detect the abnormal state of rotating machinery quickly. However, although the current deep learning-based HI construction methods have good life prediction value, most of them lose the ability to detect device anomalies and little work has been done on model interpretability. Therefore, an interpretable HI construction method based on semi-supervised autoencoder (AE) latent space variance maximization (SSALSVM) was proposed to monitor the health status of bearings. In order to fully excavate degradation features inside the device and make the model focus on the encoding process, a deep convolutional neural network (DCNN) is used as the encoding layer, while only a layer of fully-connected layer is used as the decoding layer. In addition, to enable the latent space to capture the device early degradation point (EDP) successfully, an auxiliary layer is added to the output of the encoder layer. Simultaneously, for improving the sensitivity of the indicator to capture equipment abnormal state and highlight the difference between equipment health state and degradation state, the constraint of variance maximization is added into the latent space. The model optimizing process was presented by observing the projected variance of the test set in latent space of each epoch model. The validity of the proposed HI was verified by comparison experiments on two datasets.

Data-driven methods have been recognized as an efficient tool for multivariate statistical process control. Contribution plots are also well known as a popular tool in principal component analysis, which is used for isolating sensor faults without the need for any prior information. However, studies carried out in the literature have unified contribution plots in three general approaches. Furthermore, they demonstrated that correct diagnosis based on contribution plots is not guaranteed for either single or multiple sensor faults. Therefore, to deal with this issue, the present paper highlights a new contribution formula called relative variation of contribution. Simulation results show that the proposed method of contribution can successfully perform the fault isolation task, in comparison with partial decomposition contribution and its relative version (rPDC) based on their fault isolation rate.

The velocity information estimated by the global navigation satellite system (GNSS) receiver is an important element for the dynamic alignment of the inertial navigation system, and it is of great significance to analyze it deeply and meticulously. A variety of GNSS velocity measurement models show different characteristics in a changeable environment, and this status quo is bound to break the monotonous situation in which the Doppler model is widely used. In this regard, this paper applies different GNSS velocity measurement models to strap-down interial navigation system (SINS) dynamic alignment and continuous observation. In addition, aiming at the shortcomings of the traditional time-differenced carrier phase (TDCP) algorithm, an optimization method is deduced from the formula level, and two effective constraint algorithms are given. Then, according to the vehicle test results, comprehensively compare the integrated navigation performance of various speed measurement models, and analyze the improvement effect of the proposed TDCP algorithm. This paper provides a summary for the comprehensive study of GNSS velocity measurement model and the application of optimized carrier phase to integrated navigation, which has certain practical value.

This article studies the human motion tracking problem using the disturbance rejection adaptive filter with an inertial sensor. Due to the complexity of human motion, conventional inertial methods easily suffer from motion accelerations (MAs) and ferromagnetic disturbances (FDs). Here, a disturbance rejection adaptive filter is presented to segregate the exterior disturbances from the inertial sensor observations before attitude estimation. This method estimates the MA and FD by Kalman filters. According to disturbance intensity, Sage–Husa adaptive strategies based on fuzzy rules are designed to adjust noise covariance. Moreover, gravity and geomagnetic field estimation are applied as multiplicative extended Kalman filter observations to solve external disturbances problems. Finally, the effectiveness and superiority of the proposed method are verified by an example of human motion tracking.

In practical applications of fault diagnosis, several factors, including fluctuations in load, changes in equipment condition, and environmental noise effects, could cause a classifier that's been trained on the source domain to be ill-suited for matching data from the target domain. Unsupervised domain adaptation techniques have been developed to tackle this issue, but they typically demand access to fully labeled source domains, ignoring concerns of privacy regarding source domain data. Therefore, we consider a new research scene for source-free unsupervised domain adaptation (SFUDA), which exclusively relies on a source model trained on source domain sample without requiring access to fully labeled source domain data. This paper introduces a SFUDA approach that utilizes knowledge distillation (KD), which involves two stages: (1) generalizing the source model by applying domain augmentation techniques and LS methods that enhance the model's potential to enhance its generalization capability; (2) adapting the target model using a KD framework to achieve knowledge migration; and in addition, mutual information structure regularization is added to consider the internal data structure, thus enhancing the model's adaptability. To evaluate the efficacy of our approach, we perform experiments on two datasets—the Case Western Reserve University dataset and the Paderborn University dataset, comprising 24 transfer tasks. Our experiments demonstrate the effectiveness of the domain augmentation technique, mutual information regularization, and the proposed method.

This paper presents a broadband ultrasonic rangefinder, with merits of a small blind area and high accuracy, based on a multi-frequency piezoelectric micromachined ultrasonic transducer (MF-PMUT) array. The coupling bandwidth between six types of resonant cells with adjacent resonant frequencies was studied and optimized, and these cells were then integrated into an array to broaden the bandwidth of the device. The array device showed a wide −6 dB fractional bandwidth of 108% in silicon oil. A wide bandwidth can contribute to obtaining the maximum steady-state with less excitation and reducing its residual ring-down, resulting in a small blind area. Pulse-echo ranging experiments demonstrated that the blind area was effectively reduced to 5 mm due to the multi-frequency array. For short distance measurements (<250 mm), the error was well controlled within ±0.3 mm and the 3σ accuracy was 0.445 mm. Compared with earlier reported PMUTs designed for bandwidth enhancement, the proposed broadband MF-PMUT array is much simpler in design and fabrication, revealing it as a promising candidate for short-distance high-accuracy measurements.

Fiber-optic distributed acoustic sensing (DAS) systems based on phase-sensitive optical time-domain reflection technology have been widely used for perimeter security and oil and gas pipeline safety monitoring. To address the problem of low recognition accuracy of high-sampling-rate long-sequence signal data (length greater than or equal to 1000 points) collected by the DAS system, we propose a CDIL-CBAM-BiLSTM network model based on feature fusion. The model uses a modified circular dilated convolutional neural network to extract detailed temporal structure information from each signal node, and combines it with bidirectional long short-term memory network using feature fusion to dig deeper into the data. Meanwhile, a convolutional block attention module was introduced to improve the model performance. The experimental results based on 5040 training samples and 2160 test samples show that the proposed model can achieve an average recognition accuracy of more than 99$\%$ for six real disturbance events under perimeter security scenarios, and the recognition time was less than 2 ms. In addition, our method achieved the highest recognition accuracy compared with other methods used in the experiments and can be extended to other areas, such as pipeline safety monitoring and industrial inspection measurements.

This study proposes a new method for the precision measurement of the 5-DOF error motion of rotating machines. The developed seven-probe error separation system uses two measurement units, consisting of an upper four-probe unit and a lower three-probe unit, for the separation of error signals. From the seven probe output signals, the profile signals of the measured surface and the 5-DOF error motion signals, including random components, can be precisely separated. An analysis was carried out to examine the effects of the intersection term by the 5-DOF error motion signals, which is an additional measurement error cause. Furthermore, an effective calculation process capable of calculating a precise surface profile and error motion signal by compensating for the effect was proposed. Numerical validation tests and measurement experiments were performed. This showed it was possible to make accurate measurements by separating the 5-DOF error motions and surface profile signals and using a compensation calculation process. In addition, the effect of the stability error of the probe, which can be a significant factor in increasing measurement uncertainty, on the measurement results was observed and analyzed. Based on the geometric relationship of the measurement system, a process for estimating the expected effect of the stability error of the probe on the measurement result was presented. By applying this process, it is possible to determine the scale of probe stability error to secure the measurement reliability required when selecting a probe used in the system.

Each chemical process industry system possesses unique process knowledge, which serves as a representation of the system's state. As graph-theory based methods are capable of embedding process knowledge, they have become increasingly crucial in the field of process industry diagnosis. The fault representation ability of the diagnosis model is directly associated with the quality of the graph. Unfortunately, simple fully connected graphs fail to strengthen the internal connections within the same process but weaken the interactive connections between different processes. Moreover, each node in the graph is considered equally important, making it impossible to prioritize crucial system monitoring indicators. To address the above shortcomings, this paper presents a spatial weighted graph (SWG)-driven fault diagnosis method of complex process industry considering technological process flow. Initially, the physical space sensor layout of the technological process flow is mapped into the spatial graph structure, where each sensor is regarded as a node and these nodes are connected by the k nearest neighbor algorithm. Subsequently, according to the mechanism knowledge, the sensors in the process are divided into different importance categories and weight coefficients are assigned to their nodes. The similarities between these weighted nodes are calculated, and the resulting edge information are used to construct the SWGs. Finally, the SWGs are input to a graph convolutional network, facilitating fault representation learning for fault diagnosis of complex process industry. Validation experiments are conducted using public industrial datasets, and the results demonstrate that the proposed method can effectively integrate the process knowledge to improve the fault diagnosis accuracy of the model.

This work aims to realize multiple mechanical fault diagnosis for transformers in the incubation period. A double antenna radio frequency identification (RFID) sensor is used to obtain vibration signals generated by a transformer. The measured raw vibration signals have prominent nonlinear characteristics; besides, they are mixed with vast noises, such as electromagnetic interference, measurement disturbance and so on, making it challenging to identify distinguishable features from the measured data. The long short-term memory (LSTM), which exhibits satisfactory performance in dealing with large-scale nonlinear time-series signals, is adopted to extract features from high-dimensional raw signals. Because the parameters of the LSTM cell and support vector machine significantly determine the accuracy of diagnosis, these parameters are determined using the chaotic quantum particle swarm optimization algorithm. Moreover, the sparse periodical attention (PSA) mechanism is used to enhance the LSTM model's performance by focusing on global feature learning and determining the input data length. The experimental results verify that the exploited RFID sensor realizes reliable data wireless transmission within 17.5 m. Moreover, the PSA-optimized LSTM approach achieves satisfactory trade-off between diagnosis accuracy and computation complexity while dealing with multiple mechanical fault diagnosis of a transformer in early stages.

Active Pixel sensors play a crucial role in enabling successful low-light scientific experiments due to their inherent advantages and capabilities. Such devices not only offer high spatial resolution but also feature individual pixels with integrated amplifiers, allowing for direct signal amplification at the pixel level. This results in reduced readout noise and improved signal-to-noise ratio (SNR), which are particularly vital when dealing with limited photon counts in low-light environments. This holds particularly true for scientific CMOS (sCMOS) sensors, acknowledged as an advanced evolution of Active Pixel sensors. However, despite their advantages, such sensors can still exhibit limitations such as higher cost and presence of noise artifacts that should be closely investigated. In particular, CYGNO project fits in a global effort aimed at direct detection of Dark Matter particles. CYGNO collaboration intends to build a detector based on a Time Projection Chamber making use of Gas Electron Multipliers for the amplification of ionization electrons. The GEM multiplication process produces photons that can be readout by a high-resolution sCMOS sensor. Such detection system is being designed to have enough sensitivity to detect low-energy particles and to measure released energy with enough granularity so to reconstruct direction and energy profile along their trajectories. The image sensor has an important role in the detector performance, having a direct impact on the SNR of the experiment. This work proposes a study on the performance of three different sCMOS sensors with respect to their sensitivity to low-energy particles and their intrinsic noise, which are of the utmost importance for various scientific experiments.

The vibration signal of rolling bearing with variable operating conditions contains complex interference components, which will cause low fault diagnosis accuracy, especially in strong noise case. To solve this problem, we proposed a noise reduction method of rolling bearing with variable operating based on empirical wavelet transform and adaptive time-frequency peak filtering (EWT-ATFPF). Firstly, empirical wavelet transform is used to obtain different frequency intrinsic mode functions (IMFs). Secondly, a modified adaptive window length formula for time-frequency peak filtering (TFPF) is constructed by combining the sampling ratio index and a fault sensitivity indicator that calculated by kurtosis and correlation coefficients of IMFs, which can better characterize the impact components. Thirdly, to balance noise reduction effect and the fidelity of IMFs, we proposed an improved TFPF method by adaptively adjusting its windows length. The adaptive method could be carried out using the proposed fault sensitivity indicator and window length formula, and the denoising IMFs could be obtained by ATFPF. Finally, the denoising vibration signal is reconstructed by using the denoising IMFs. The performance of fault diagnosis of the proposed method is verified by using simulated signal and bearing fault test data. The results show that the proposed EWT-ATFPF method could effectively achieve noise reduction under variable operating conditions.

The complex electromechanical coupling effect of an industrial robot's electromechanical system significantly influences the dynamic characteristics and stability of the joint servo transmission system. This study proposes a kind of electromechanical coupling dynamic model for the industrial robot's joint servo transmission system, incorporating multiple parameters, and investigates the vibration response characteristics of the system. Firstly, an electromechanical coupling analysis of the industrial robot's joint servo transmission system is conducted, simplifying it as a 'motor-reducer-load' system and establishing the corresponding electromechanical coupling vibration mathematical model. Secondly, a dynamic model of the robot's joint servo transmission system is developed to analyze its speed characteristics. Finally, the model is used to analyze the electromechanical coupling vibration process of the joint servo transmission system with the introduction of mechanical and electrical parameters and other factors to obtain the vibration characteristics of the industrial robot joint servo transmission system under electromechanical coupling. The experimental results show that this model can better analyze the electromechanical coupling vibration characteristics of the joint servo transmission system, which is significant for improving the performance and fault traceability of the industrial robot joint transmission.

The existing federated structure protects data privacy with only a certain level of confidentiality, and it is difficult to resist the reconstruction of other clients' data by malicious participants inside the federated and the illegal manipulation by external attackers or interceptors on the shared information. Besides, the average fusion algorithm used in the cloud center is difficult to eliminate the negative impact of outliers on model updates, and it cannot handle and fuse the time delay or even packet loss that occurs in the information obtained from each local client promptly. Therefore, to make the federated learning (FL) mechanism with stronger privacy protection ability and security, while effectively avoiding the negative impact of outliers on the aggregation of model parameters. We innovatively establish multi-Level FL based on cloud-edge-client collaboration and outlier-tolerance for fault diagnosis. At first, we build a multi-level FL network framework based on the cloud-edge-client collaborative approach for restricted sharing of network parameters level by level without data communication. Then, the edge-side performs Euclidean metrics on the restricted shared model parameters uploaded to the primary edge by each client, and uses them to identify outliers to evaluate and weight them for outlier-tolerance; Then, an outlier-tolerance mechanism is designed based on a centralized Kalman filtering algorithm that is to adjust the modeling error weights adaptively; Lastly, the cloud center performs asynchronous aggregation on the model parameters uploaded asynchronously by the highest-level edge based on a sequential Kalman filtering algorithm and transmitted the optimal model parameters back along the original path. Finally, the effectiveness of the proposed method is verified on the collected dataset.

Defect detection of lithium batteries is a crucial step in lithium battery production. However, traditional detection methods mainly rely on the human eyes to observe the bottom defects of lithium battery products, which have low detection accuracy and slow detection speed. To solve this practical problem, an improved YOLOv5s model is proposed in this paper. Firstly, a new layer of the network output layer is added to improve the detection effect of small defects. Secondly, to extract important information in the feature maps, the convolutional block attention module attention mechanism is added to the YOLOv5s model. Finally, a new position loss function is used to improve the accuracy of the position prediction of the model. The experimental results indicate that the improved YOLOv5s model can accurately and quickly detect three types of defects on the bottom surface of lithium batteries. Specifically, the loss and mean average precision (mAP) of the improved YOLOv5s model are 0.033 94 and 67.5% respectively. Compared with the traditional YOLOv5s model, the loss of the improved YOLOv5s model is reduced by 31%. As well as, the mAP of the improved YOLOv5s model is increased by 4.3% on the lithium battery defect dataset. Compared with the YOLOv3, YOLOv3-spp, retinanet and YOLOv4, the mAP of the improved YOLOv5s model increased by 5.4%, 0.7%, 11.9% and 3.7% respectively. Compared with other improved YOLOv5 algorithms used in various fields, the mAP of the proposed model on the lithium battery dataset is the highest. The detection speed of the improved YOLOv5s model reaches 111 frames per second, which can meet the real-time detection requirements. The improved YOLOv5s model has board application prospects in the industrial production of lithium batteries.

In order to address the problem that one- dimensional convolutional neural networks is difficult to extract the local correlation information and mine multi-scale information of rolling bearing fault signals under variable working conditions, a novel fault diagnosis method for rolling bearings based on Markov transition field (MTF) and multi-scale Runge–Kutta residual attention network (MRKRA-Net) is proposed in this paper. Firstly, the original signal is encoded into a two-dimensional image using the MTF method. Then, a multi-scale network is constructed using pre-activation Runge–Kutta residual blocks to extract multi-level features. Secondly, a feature-guided attention mechanism is designed and embedded into the network model to enhance its generalization ability. Finally, the MRKRA-Net model is validated on two different bearing datasets, and the results show that compared with other popular intelligent fault diagnosis methods, MRKRA-Net has higher fault diagnosis accuracy and stronger robustness under both given and variable working conditions.

Accurate characterization of the thermoelectric module (TEM) is a prerequisite for its application. Among the physical parameters of the TEM, the heat flow is crucial but it is the most difficult parameter to measure. This paper summarizes the previously reported heat flow measurement methods and proposes a heat flow meter (HFM) for TEM characterization with a sandwich structure composed of copper-glass-copper. The key factors affecting the precision of the HFM are analyzed with the assistance of simulation to obtain its design principles and fabrication procedures. A self-designed calibration device is used to calibrate the HFM to understand the relationship between heat flow and temperature. The HFM can characterize the TEM's heat flow with its normalized root mean square error of no more than 2%. The accurate characterization of the TEM shows great potential to contribute to the development of TEM technologies.

Water stress is a significant environmental factor that hampers plant productivity and leads to various physiological and biological changes in plants. These include modifications in stomatal conductance and distribution, alteration of leaf water potential & turgor loss, altered chlorophyll content, and reduced cell expansion and growth. Additionally, water stress induces changes in the emission of volatile organic compounds across different parts of the plants. This study presents the development of an electronic nose (E-nose) system integrated with a deep neural network (DNN) to detect the presence and levels of water stress induced in Khasi Mandarin Orange plants. The proposed approach offers an alternative to conventional analytical methods that demand expensive and complex laboratory facilities. The investigation employs the leaf relative water content (RWC) estimation, a conventional technique, to evaluate water stress induction in the leaves of 20 plants collected over a span of 9 days after stopping irrigation. Supervised pattern recognition algorithms are trained using the results of RWC measurement, categorising leaves into non-stressed or one of four stress levels based on their water content. The dataset used for training and optimising the DNN model consists of 27 940 samples. The performance of the DNN model is compared to traditional machine learning methods, including linear and radial basis function support vector machines, k-nearest neighbours, decision tree, and random forest. From the results, it is seen that the optimised DNN model achieves the highest accuracy of 97.59% in comparison to other methods. Furthermore, the model is validated on an unseen dataset, exhibiting an accuracy of 97.32%. The proposed model holds the potential to enhance agricultural practices by enabling the detection and classification of water stress in crops, thereby aiding in water management improvements and increased productivity.

Despite achieving considerable success, the fault diagnosis methods will still be disturbed by noisy labels, this causes the model's degradation and reduced diagnostic precision. Focused on solving the above issues, a robust intelligent fault diagnosis approach for rotating machinery under noisy labels is proposed. Firstly, we maintain two deep neural networks (DNNs) and alternatively execute parameters updating and models optimization by referring to the Co-teaching strategy, which can maximize filtering different error types and implement pre-training of DNNs. Secondly, adopting a two-component Gaussian mixture model (GMM) to fit training dataset's cross-entropy (CE) loss and realize the clean and noisy labels division according to the threshold. Then, a data augmentation method called Mixup operation is employed in semi-supervised learning (SSL) to increase noise robustness and avoid error accumulation, subsequently, performing fine-tuning and correction for clean and noisy samples. Challenging experiments on a transmission gearbox dataset under different noisy labels levels show that the proposed method has robustness to noise and significantly surpasses other approaches, which provides an important reference value for accurate fault diagnosis for rotating equipment parts with noisy labels.

Remaining useful life (RUL) prediction plays an important role in prognostics and health management (PHM) and can significantly enhance equipment reliability and safety in various engineering applications. Accurate RUL prediction enables proactive maintenance planning, helping prevent potential hazards and economic losses caused by equipment failures. Recently, while deep learning-based methods have swept the RUL prediction field, most existing methods still have difficulties in simultaneously extracting multiscale global and local degradation feature information from raw multi-sensor monitoring data. To address these issues, we propose a novel multiscale global and local self-attention-based network (MGLSN) for RUL prediction. MGLSN consists of global and local feature extraction subnetworks based on self-attention, which work in parallel to simultaneously extract the global and local degradation features of equipment and can adaptively focus on more important parts. While the global network captures long-term dependencies between time steps, the local network focuses on modeling local temporal dynamics. The design of parallel feature extraction can avoid the mutual influence of information from global and local aspects. Moreover, MGLSN adopts a multiscale feature extraction design (multiscale self-attention and convolution) to capture the global and local degradation patterns at different scales, which can be combined to better reflect the degradation trend. Experiments on the widely used Commercial Modular Aero-Propulsion System Simulation (CMAPSS), New CMAPSS (N-CMAPSS), and International Conference on Prognostics and Health Management 2008 challenge datasets provided by NASA show that MGLSN significantly outperforms state-of-the-art RUL prediction methods and has great application prospects in the field of PHM.

Tactile object recognition is very important in robot sorting and handling system. The existing researches on tactile recognition mainly focus on rigid grippers, which cannot be directly applied to soft grippers with better safety and applicability. In this paper, a tactile sensor for soft gripper is proposed. The tactile sensor consists of a carbon composite Velostat and a flexible printed circuit designed for soft gripper. The sensor is equipped with signal scanning, amplifying and processing systems. A number of performance tests were carried out on the designed tactile sensor to prove that it has good performance in sensing sensitivity, loading stability, repeatability. In addition, tactile sensors are applied to tactile classification. The two fingers of the flexible gripper were loaded with a sensor and nine different objects were contacted. The tactile information of the objects was obtained through the stable gripper, and the tactile data was trained and classified by the support vector machine. Finally, the tactile recognition accuracy of the double-exponent data was 91.1%.

This paper proposes a design method for a fiber Bragg grating (FBG) displacement sensor with a hyperbolic flexible hinge structure. A compact FBG displacement sensor with strong micro displacement measurement capability was fabricated. The final assembly testing showed that the sensor achieved good linearity with a linearity of 0.53%F·S at a range of 50 mm. The sensitivity was measured as 24.45 pm mm⁻¹, hysteresis error as 1.1%FS, and repeatability error as 0.659%. The testing results demonstrated that the fabricated FBG displacement sensor exhibited strong practicality and good stability, making it suitable for long-term displacement monitoring of slope safety.

The global navigation satellite system (GNSS) and a micro-electro-mechanical system inertial measurement unit (MEMS IMU) can be integrated to form a robust navigation system, providing continuous and accurate location information on dynamic vehicles. However, the performance of a GNSS/MEMS IMU integrated navigation system is easily affected by the accuracy of measurement noise matrix, which is significantly reduced in the urban environment. The adaptive Kalman filter can improve the measurement noise matrix reliability by constructing robust equivalent weight factors; however, GNSS observations are easily misjudged as exception sources that can deteriorate the filter stability. To address these problems, a novel integrated algorithm with a displacement vector constraint is proposed based on the kinematic characteristics of vehicles, which are considered as judgment factors. In this case, the measurement noise can be adjusted accordingly only when the GNSS observations are judged as abnormal. In this study, two types of vehicle navigation experiments were conducted: a low-speed wheeled robot in a campus environment and a dynamic vehicle in an actual urban environment. The experimental results showed that the three-dimensional navigation accuracy was improved by 47% and 55% compared with the conventional algorithm under the two different environments, respectively.

The predictive maintenance approach in the fault diagnosis of rotating machines is becoming important in industries in order to minimize operational cost and to improve the reliability of machines. Recent studies have focused on developing an effective prediction model to detect machinery faults at various speeds using machine learning techniques. In this research study, machine-learning-based models are developed to detect multi-component faults such as gear faults, bearing faults, and unbalanced shafts, by training the four speed data sets using decision trees, support vector machines and artificial neural networks. The machine learning models are improvised by fusing the vibration signals (X, Y and Z) and sound signals with the feature selection algorithm, minimal redundancy-maximum relevance (mRMR). The research work is extended by training a model with any three speeds and testing the trained model with the remaining speed. The state-of-the-art algorithms used to train all four speeds yield poor performance in the prediction of faults at testing speeds. A linear and quadratic discriminant analysis (QDA) is chosen based on its multivariate discrimination capability for better fault prediction at testing speed. The proposed combination of QDA with mRMR selection of fused vibration and sound feature set perform well in the detection of multiple faults at variable speeds.

Microfabricated piezoresistive accelerometers with purely axially deformed piezoresistive beams have demonstrated high performance. However, the conventional design of such accelerometers requires complex theoretical calculations and specific dimensional conditions to achieve purely axial deformation, which inevitably increases the difficulty and cost of the design and manufacturing. We propose an innovative structure that can simply realize pure axial deformation of piezoresistive beams by eliminating the transverse displacement at both ends without tedious calculations. An accelerometer based on the structure was fabricated; both static and dynamic performances were tested. The results showed that the accelerometer had high sensitivity (2.44 mV g⁻¹ with 5 V bias, without circuit amplification), low cross-axis sensitivity (1.56% and 0.49 %, respectively), and high natural frequency (11.4 kHz), with a measurement range of 0–100 g. This design method provides an easy approach for designing high-performance piezoresistive sensors.

Considering that existing robust filtering algorithms rely on the selection of initial values of degree of freedom (DOF) parameters in outlier interference environments and cannot effectively cope with unknown non-stationary heavy-tailed measurement noises (HMN), a Gaussian–Pearson type VII (PTV) adaptive mixture distribution-based outlier-robust Kalman filter (GPTVMAKF) is proposed. In order to determine whether the current measurement is a normal value or an outlier, a judgment factor subject to the Beta-Bernoulli distribution is introduced. PTV distribution is used to model HMN caused by outliers, and two Gamma distributions are used to model the two different DOF parameters, which can make the PTV distribution have the adaptive adjustment ability. By introducing the inverse Wishart distribution as the prior distribution of the measurement noise covariance, which is adaptively estimated to cope with the unknown time-varying measurement noises. The state and parameters are jointly estimated by variational Bayesian. Finally, the simulation experiments verify that the proposed GPTVMAKF can obtain more accurate state estimation than existing filters in the environments with varying degrees of HMN and unknown non-stationary HMN.

Accurate determination of the optical properties of ultra-thin dielectric films is an essential and challenging task in optical fiber sensor systems. However, nanoscale thickness identification of these films may be laborious due to insufficient and protracted classical curve matching algorithms. Therefore, this experimental study presents an application of a radial basis function neural network in phase diffraction-based optical characterization systems to determine the thickness of nanoscale polymer films. The non-stationary measurement data with environmental and detector noise were subjected to a detailed analysis. The outcomes of this investigation are benchmarked against the linear discriminant analysis method and further verified by means of scanning electron microscopy. The results show that the neural network has reached a remarkable accuracy of 98% and 82.5%, respectively, in tests with simulation and experimental data. In this way, rapid and precise thickness estimation may be realized within the tolerance range of 25 nm, offering a significant improvement over conventional measurement techniques.

Fringe projection technology is usually used for three-dimensional (3D) measurement, but there will be large errors when the measurement is overexposed or the surface is too dark. Polarizing filters can be used to eliminate the highlights of the image, but they may also cause the image to be too dark and affect the measurement accuracy. Otherwise, to ensure measurement accuracy, the complexity of the operation of the polarization system will be increased. In order to ensure the measurement accuracy and improve the measurement efficiency, a 3D measurement method of quantified exposure time is proposed in this paper, and the intensity response function of the light intensity captured by the camera and the exposure time is established, and more meaningful variable value intervals can be selected by the method of this paper. It can avoid the complex polarization bidirectional reflectance distribution function model, and the required exposure time can be calculated quantitatively within the already screened intervals. Then it is combined with the image fusion algorithm to generate the optimal fringe pattern. Experimental results demonstrate that this method significantly eliminates the effects of highlights in the image. The fuzzy transition area between the black and white fringes is effectively reduced, and the edge information of the fringes is correctly restored.

There are numerous redundant points in the point cloud model of ring forgings obtained by 3D laser scanner. How to remove the redundant points while keeping the model characteristics unchanged is a critical issue. This paper proposes a point cloud simplification algorithm based on the joint entropy evaluation theory. Firstly, the K-D tree is used to search for the K-neighbors of the sampled points. Secondly, a surface is fitted to the spatial neighborhood of the sampled points using the least squares method. The curvature operator of the sampled points is derived on the fitted surface using Riemannian geometry theory. After that, an energy operator is defined by using the normal vectors and distances of the sampled points and their neighborhood points. The joint entropy values of all points in the model are determined based on the probability distributions of these two operators in the local neighborhood. Finally, the data points are sorted by entropy value. Data points with high entropy values are put into the data set U1. Data points with low entropy values are clustered through the K-Means algorithm of swarm optimization. The redundant points outside the cluster centers are removed, and the cluster centers are put into the data set U2. The final simplification results are obtained by integrating data sets U1 and U2. The experimental results show that the point cloud simplification algorithm proposed in this paper is effective and feasible.

This paper describes the steps taken to improve the measurement speed of a combined low-coherence and confocal refractive index measurement system. The instrument measures the refractive index and thickness of transparent plates using a fibre-based low-coherence interferometer with a line-scan spectrometer. The spectrometer allows on-line dispersion measurement which is necessary to derive the sample thickness t as well as both the phase and group refractive indices n_p, n_g. The measurements were performed on a sample with six surfaces consisting of three glass windows mounted in a lens tube. Experimental results show that a measurement time of 4.4 s for the multi-layered object, which has a total thickness of approximately 10.5 mm, can be achieved whilst maintaining an accuracy of better than 0.1% for n_p, n_g, and t. This represents an approximately hundredfold improvement over previously published measurements.

As a non-contact, couplant-free and nondestructive technique, the laser ultrasonic technology has great potential for detecting surface defects. In this paper, we propose a surface defect identification method based on broadband laser-generated Rayleigh wave with wavenumber filtering in an area scanning mode. Firstly, a particular lowpass filter is constructed in frequency-wavenumber domain to extract the scattered wave generated by the defects, and then a surface defect image can be reconstructed based on the broadband scattered wave wavenumbers. Secondly, a threshold denoising method is employed to enhance the signal-to-noise ratio and improve the image sharpness. Thirdly, the experimental validation is carried out, in which a laser ultrasonic detection system is used to detect the surface defects of aluminum alloy specimens and identify their different parameters. The experimental results verify that the proposed method can identify the location, size and orientation of surface defects effectively, meanwhile, its imaging effect shows significantly superior to that of the conventional reflected wave energy and standing wave energy methods. Furthermore, the correlation between maximum local scattered wave energy and the depth of defects is found, which can characterize the severity of the surface defects.

Point cloud registration is a fundamental problem for 3D laser scanning technology, which is extensively applied in geographic entity modelling such as 3D reconstruction of urban roads and buildings. Registration accuracy is one of the main focuses for these applications. However, noisy points, limited overlap, varying data sources, and differing measuring accuracy may cause changes between point cloud sets and reduce registration accuracy. This paper introduces an automatic plane registration method for urban roads and building scenes, which does not need manual on-site deployment. Firstly, plane primitives are extracted using voxel-based filtering region growth. Next, corresponding planes for the extracted primitives are identified by leveraging saliency features and constructing adjacency matrices. Finally, plane registration is achieved using a weighted plane coordinate conversion model. Through real-world scene experimentation, an overall accuracy of 10 cm and a segmental registration accuracy of 5–6 cm is achieved with our method, outperforming both feature point-based and global point cloud registration approaches in terms of efficiency and accuracy.

This work presents an experimental setup for acquiring synchronized wall and flow field measurements in a turbulent boundary layer (TBL). Instantaneous measurements of the convective heat transfer distribution at the wall and of the flow field are carried out simultaneously, using synchronized wall-parallel planar particle image velocimetry and infrared thermography. A low-thermal-inertia heated-thin-foil sensor is embedded in the wall beneath the TBL to measure the wall temperature maps with enough temporal resolution. The unsteady energy balance of the heated foil can be solved to restore the instantaneous value of the convective heat transfer coefficient on the wall. A detailed description of the sensor design is included. Furthermore, owing to the relatively low signal-to-noise ratio of instantaneous temperature fluctuation measurements, a recipe for data processing is proposed. A study to characterize the uncertainty of the experimental setup design is also performed. The overall agreement in the correlation between wall heat transfer and velocity fields with the literature supports the validity of the proposed approach. This solution is potentially interesting for flow control purposes, where sensing is performed at the wall.

Hydraulic turbines play an important role in stabilizing electrical grids due to their fast response times. Comparative to other hydraulic turbines, a jet deflector of a Pelton turbine enables rapid response for load reduction. However, jet deflectors are subjected to high forces, torques and the splashing water impinges on the runner and elements within the casing. Thus, the design of the jet deflector is important. The analytical calculations are complex due to the splashing and backscattering of the deviated jet. Though the prediction of forces and deflector torque based on the impulse–momentum equation is feasible, but results are not sufficiently accurate, and thus measurements are required. Subject of this study is a test rig equipped with the physical model of a Pelton turbine injector with deflector control and a high-quality torque transducer for precise measurement. Calibration of instruments used, and uncertainty analysis of the test rig has been carried out. The overall uncertainty of the torque resulted in ±1.58%. The experimental results are compared with the analytical results to verify the primary data. It is observed that as the deflector initiates its action, the torque undergoes an increasing trend. At a deflection angle of 12°, the torque reaches its maximum value, and subsequently, upon achieving complete deflection, the torque is observed to attain its minimum magnitude. The empirical relationship of the deflector torque uncertainty has been developed using the regression analysis of the evaluated uncertainty data.

High fidelity near-wall velocity measurements in wall bounded fluid flows continue to pose a challenge and the resulting limitations on available experimental data cloud our understanding of the near-wall velocity behavior in turbulent boundary layers. One of the challenges is the spatial averaging and limited spatial resolution inherent to cross-correlation-based particle image velocimetry (PIV) methods. To circumvent this difficulty, we implement an explicit no-slip boundary condition in a wavelet-based optical flow velocimetry (wOFV) method. It is found that the no-slip boundary condition on the velocity field can be implemented in wOFV by transforming the constraint to the wavelet domain through a series of algebraic linear transformations, which are formulated in terms of the known wavelet filter matrices, and then satisfying the resulting constraint on the wavelet coefficients using constrained optimization for the optical flow functional minimization. The developed method is then used to study the classical problem of a turbulent channel flow using synthetic data from a direct numerical simulation (DNS) and experimental particle image data from a zero pressure gradient, high Reynolds number turbulent boundary layer. The results obtained by successfully implementing the no-slip boundary condition are compared to velocity measurements from wOFV without the no-slip condition and to a commercial PIV code, using the velocity from the DNS as ground truth. It is found that wOFV with the no-slip condition successfully resolves the near-wall profile with enhanced accuracy compared to the other velocimetry methods, as well as other derived quantities such as wall shear and turbulent intensity, without sacrificing accuracy away from the wall, leading to state of the art measurements in the $y^+\lt 1$ region of the turbulent boundary layer when applied to experimental particle images.

The aim of this study is to develop a contour extraction method for the investigation of the temporal and spatial characteristics of cavitation evolution in high-speed inducers. Cavitation experiments were conducted, and corresponding snapshots were captured using a high-speed camera subsequently. The raw images were processed while using threshold value filtering techniques to extract only relevant bubble information. Proper orthogonal decomposition was employed to explore the evolving characteristics in a quantitative manner. Results show that the cavitation evolution process is an unsteady process when viewed from the side. In high net positive suction head (NPSH) environments, multiple small-scale spatial modals contribute significantly to the effective proportion of the cavitation structure. The attached sheet cavitation exists in the blade edge which is the main embodiment of cavitation area evolvement. Under low NPSH conditions, the primary flow feature can be reflected by a low-mode flow pattern, associated with the inducer's rotation effect. The evolving regulation of cavitation area in developed stage is closely submitted to the trigonometric function. And the exhibited spatial coefficient exposes the hidden reflux vortex feature. The proposed image separation techniques enable the quantification of morphological features, which can provide technical details for the digital design of inducers.

Simultaneous chemiluminescence (CL) imaging of multiple chemical species in unsteady and/or turbulent combustion environments requires multiple high-speed camera/intensifier systems which can be cost prohibitive. This study presents a novel optical configuration for simultaneous, spatially resolved CL imaging of two species using a single high-speed camera. The superiority of the technique over existing approaches is demonstrated via a series of experiments in a methane fueled McKenna burner operated at steady and unsteady conditions. Results demonstrate capabilities of the present system to provide spatially resolved measurements of hydroxyl (OH*) and methylidyne (CH*) species without parallax or path length errors in different flame configurations.

This article reports the use of a new measurement technique based on micro thermography for determining the thermal contact resistances (TCRs) between filled polymers and solids. The thermal conductivity of polymers can be significantly increased by using thermally conductive fillers. For numerous applications, not only is a high intrinsic thermal conductivity required but also a good thermal transfer between the filled polymer and an adjacent solid surface. The physical principles of thermal transport when considering this type of contact have not yet been investigated in detail, and only a few experimental results are available. The most common measurement techniques determine a macroscopic resistance and project it onto the contacting surface. However, the heterogeneous microstructure of a filled polymer causes the TCR to be a volumetric phenomenon in the overall boundary region. The utilized IR camera system takes thermal images with a spatial resolution of less than $15\,\,\unicode{x03BC} {\text{m}}$ per pixel. The new method resolves the TCRs spatially and gives new insights into the microscale effects on the particle level. In addition to the common zero-gap extrapolation for the extraction of TCRs, we propose another evaluation method that considers all microscale effects of the boundary layers and evaluates TCR as a volumetric phenomenon. For the first systematic study, samples consisting of two aluminum substrates and a filled epoxy polymer were prepared and investigated. We studied the effects of filler size, filler material, filler volume fraction, and surface structure, focusing on monomodally filled polymers with filler amounts between $30$ and $60{\text{v}}\%$. The obtained results and the uncertainties of the new method are discussed within this paper.

Fabrics play a pivotal role in human life and production, and surface defects can directly affect the quality and value of fabrics. Many methods for fabric defect detection have been proposed, but tiny defects are still difficult to be detected effectively, and the accuracy of defect localization and classification is low. To address these issues, a modified YOLOX network called YOLOX-CATD is proposed, which was supplemented with a coordinate attention module (CAM) and tiny defect detection layer (TDDL) for fast and efficient detection of fabric defects, especially tiny defects. Firstly, the anchor-free network is used as the detection framework to avoid the influence of hyperparameters of the setting anchor. Secondly, a CAM is proposed to enhance the representation of the object of interest in the input feature map and suppress the background regions. Finally, a TDDL is added to introduce high-resolution features to improve the localization accuracy of tiny defects. The experimental results on the Aliyun Tianchi Fabric dataset and NEU-DET demonstrate the superiority and generalization of the modified model. The mean average precision (mAP) of YOLOX-CATD on the fabric defect dataset is improved by 5.67% compared to the original YOLOX, and the detection speed can reach 35–36 frames per second (FPS). This proves that YOLOX-CATD can obtain excellent fabric defect detection performance and meet the urgent need for real-time detection in industrial applications.

This study investigates two unavoidable noise factors: electronic noise (EN) and radiation scattering, associated with detectors and their electronics. This study proposes a novel methodology to estimate electronic and scattering noise separately. It utilizes mathematical tools, namely, Kanpur theorem-1 (KT-1), standard deviation, similarity dice coefficient parameters, and experimental computerized tomography technique. Four types of gamma detectors: CsI (Tl), ${\text{LaB}}{{\text{r}}_3}\left( {{\text{Ce}}} \right)$, NaI (Tl) and HPGe are used with their respective electronics. A detector with integrated circuit electronics imparts significantly less (∼33% less) EN in data than detectors with distributed electronics. KT-1 signature is proposed as a scattering error estimate. The developed empirical expression shows that scattering noise depends strongly on the mass attenuation coefficients of detector crystal material and weakly on their active area. The difference between predicted and estimated relative scattering is 14.6%. The methodology presented in this study will assist the related industry in selecting the appropriate detector of optimal diameter, thickness, material composition and hardware as per requirement.

Capacitively coupled electrical resistance tomography (CCERT) is an innovative technique for electrical resistance tomography based on capacitively coupled contactless conductivity detection (C⁴D). Despite its potential, there are only a few studies on image reconstruction algorithms for CCERT. To address this, a CCERT measurement system is developed, and the compressed sensing (CS) theory is applied to the inverse problem imaging of CCERT to improve the image reconstruction quality and speed. Firstly, a mathematical CCERT image reconstruction model is constructed under CS theory and three algorithms under CS theory are employed to solve the convex optimization problem of the reconstruction model. Then a thresholding operation is applied to obtain the post-processed image and compared it with the classical linear back projection and Landweber algorithms. The simulation results demonstrate that the Barzilai–Borwein gradient projection for sparse reconstruction (GPSR-BB) algorithm yields more satisfactory imaging results than the other four algorithms. Additionally, three sensitivity matrix optimization methods for GPSR-BB algorithm are compared and find that the new method of screening the rows of the sensitivity matrix to zero is more effective than the other two common optimization methods in terms of reconstructed image quality. Finally, the static experiments of void fraction measurement are conducted using the developed CCERT system. The results indicated that the absolute error of void fraction measurement by GPSR-BB algorithm was less than 6.59% in the range of void fraction from 0.90% to 66.29%.

As intelligent power grid construction advances, substation inspection becomes crucial, particularly in identifying meter readings. Existing meter reading methods are mainly based on the relationship between pointer and scale. However, these methods commonly suffer from the issue of over-relying on prior reading information, limiting readings to known meters. Hereby in this paper, we propose a method reaching more accurate and robust performance for meter reading by utilizing the unnoticed scale value. We determine the meter pointer direction with the aid of Hough transform and the pointer distribution. To detect meters from the scene as well as obtain scale values and pointers from the meter dial, we build an object detection network, named Lite-FCOS, whose backbone adopts a fast global context network that is lightweight and is of powerful feature extraction capabilities. For training these, meter dial detection dataset and dial reading information dataset are constructed. Lite-FCOS achieves 94.4 mAP50 and 96.7 mAP50 on the above two datasets with only 4.2 M parameters and 56.2 FPS. The entire pointer meter reading recognition process only spends 52 ms on an RTX 3080Ti with a successful meter reading rate (SMR rate) of 89.6%, which indicates that the proposed method achieves promising accuracy and speed.

As plate-like structures are widely used in railway vehicle systems, structural health monitoring and nondestructive testing of those structures are important. The Lamb wave, as an ultrasonic guided wave propagating in plate-like structures, is an effective tool for fault diagnosis. The Capon method is widely used in Lamb wave based damage imaging due to its superior performance in suppressing background noise compared with the conventional delay-and-sum method. As the conventional Capon method is sensitive to modeling errors like the inaccuracy of the look-direction, diagonal loading is usually adopted to regularize the inverse of the covariance matrix. However, the degree of diagonal loading is related to the accuracy of the given look-direction and is hard to determine accurately. Aimed at improving the imaging quality of the conventional Capon method, an adaptive Capon method with adaptive diagonal loading is proposed. In the proposed method, the dispersion and the amplitude caused by wave diffusion are first compensated for each imaging point using the virtual time reversal technique. The correlations among the compensated waveforms after windowing are used to determine the degree of diagonal loading. In the process of regularizing the inverse of the covariance matrix, a large diagonal loading factor will be applied when the correlations among the compensated waveforms are large. As correlations are closely related to the signal phase, both the amplitude and the phase of scattering signals are utilized in the proposed method. The numerical and experimental validations on an aluminum plate are carried out to verify the effectiveness of the proposed method. The results show that compared with the conventional Capon method, the proposed method can obtain images with higher imaging quality, fewer artifacts, and lower noise.

In the electromagnetic tomography system, the soft-magnetic shield is employed to avoid external interference at low frequency. However, the influence of the soft-magnetic shield on the measurement lacks theoretical exploration, and the design of the soft-magnetic shield is arbitrary. To solve this problem, the effect of different thicknesses and locations of the soft-magnetic shield on the measurement is compared by numerical simulation. It is observed that the soft-magnetic shield not only enhances the boundary measurement data and sensitivity but also improves the quality and stability of the reconstructed images. The smaller the distance between the soft-magnetic shield and the pipe, or the bigger the thickness of the soft-magnetic shield, the better the improvement. The physical mechanism of the above results is explicated by the equivalent magnetic circuit analysis method. Based on the above investigation, the optimal structure of the soft-magnetic shield is designed, and it is verified by experiments.

The cable-free seismograph is an instrument for capturing seismic wave signals, which is popularly employed in underground resource exploration tasks in complex field environments due to its high flexibility. However, the efficiency of exploration operations is low because the arrangement and recovery of instruments are done manually. Therefore, autonomously arranging and recovering cable-free seismographs will hold great significance. Nevertheless, the instrument surface will produce highlight under intense illumination in field environments, which can have a negative impact on the accuracy of segmentation and contour extraction of the instrument. It will significantly reduce the visual positioning accuracy of the instrument and ultimately fail in recovery tasks. To address this problem, we analyze the characteristics of highlight on cable-free seismographs in field environments and develop a method to remove the highlight by weighting the color features of the non-highlight pixels in the neighborhoods of highlight pixels. We realize the detection of highlight pixels based on the classical modified specular-free model and propose an image terrain map model to select the weighted neighborhoods of highlight pixels to ensure that no background pixels are involved in the process of highlight removal. The experimental results show that our method meets the requirements for contour extraction and object segmentation in the autonomous instrument recovery tasks. Furthermore, it outperforms currently existing highlight removal methods in the task of highlight removal for cable-free seismographs in field environments.

Aluminum surface defect detection plays a crucial role in the manufacturing industry. Due to the complexity of aluminum surface defects, the existing defect detection methods have false and missed detection problems. To address the characteristics of aluminum surface defects and the problems of existing methods, we propose a weight-guided feature fusion and non-local balance model to improve the detection effect. Firstly, we design the feature extraction network cross-stage partial ConvNeXt, which achieves adequate feature extraction while reducing the model's size. In addition, we propose a weight-guided feature fusion and non-local balanced feature pyramid (WBFPN). Specifically, we design a weight-guided feature fusion module to replace the simple feature fusion method so that the WBFPN can suppress interference information when fusing feature maps at different scales. The non-local balancing module captures the long-range dependencies of image features and effectively balances small target defects' detail and semantic information. Finally, the confidence loss was redefined to effectively solve the problem of poor detection effect caused by the imbalance of positive and negative samples. Experimental results show that the average accuracy of the proposed model reaches 91.9%, and the detection speed is high, which meets the requirement of real-time defect detection.

We demonstrate a broadband mid-infrared frequency comb spectrometer utilizing virtually imaged phased array (VIPA) with the potential of quantitative, fast thermometry over large dynamic range. A modified Boltzmann plot method for attaining quantitative spectral interpretation in the dispersive VIPA system is proposed and realized, despite the broadband spectral overlapping and spatially variant instrumental broadening. The developed method allows accurate interpretation of the comb spectra even without reaching single-comb-tooth resolution, and can be readily extended to other measurements with overlapped spectra. The method is validated with measurements of the v₃ band of methane, and a single-shot temperature accuracy of within 10 K is experimentally demonstrated over the range of 300–600 K at a rate of 200 Hz. The present method takes full benefit of the unique advantages of the high-resolution, broadband VIPA-based frequency comb spectroscopy and has the potential to offer fast and precise thermometry. As far as we know, this is the first study to use a mid-infrared VIPA for temperature measurement with frequency comb spectroscopy.

A mid-infrared interference-free laser absorption technique for simultaneously measuring rotational temperature, vibrational temperature, and CO concentration was developed for application to shock-tube studies on thermochemical non-equilibrium CO over 1000–3000 K. Three transition lines in the fundamental vibrational band of CO (P(0, 21), near 4.87 μm, P(1, 21), near 4.93 μm, and P(0, 37), near 5.05 μm) were selected. The P(0, 21)/P(1, 21) line pair was used for vibrational temperature measurements whereas the P(0, 21)/P(0, 37) line pair was used for rotational temperature measurements. Spectroscopic parameters for developing the technique were measured: line strengths and collisional broadening data in Ar were obtained at 1040–2940 K. Validation experiments for the thermometry system were performed in shock-heated thermal-equilibrium CO/Ar mixtures at 1050–3010 K and 1.1–2.8 bar. The time-dependent rotational and vibrational temperatures were measured during the vibrational relaxation processes of CO. The technique showed high sensitivity in detecting the rotational and vibrational temperatures. The measured rotational temperature agreed well with the temperature calculated using the measured pressure and isentropic relationship. The measured vibrational temperature showed good agreement with the predictions using the Landau and Teller theory and Millikan and White relationship. The time-dependent CO concentration during the oxidation processes of n-heptane over a wide temperature range (1350–2750 K) was measured considering n-heptane as one of the alternative fuels for the scramjet. The interference-free laser absorption strategy showed good flexibility in detecting the CO concentration at ultra-high temperatures. The measured results showed overall good agreement with the predictions from two detailed mechanisms and one skeletal mechanism. The reactivity of n-heptane was found to be insensitive to the temperature increase at ultra-high temperatures (>2100 K).

In the field of prostate cancer based on surface-enhanced Raman scattering (SERS), the repeatability of detection is crucial for its practical application in clinical medicine. We propose to use a new SERS substrate based on a polymer planar waveguide in combination with microfluidic technology and utilized in the label-free detection of prostate-specific antigen (PSA) in serum. The experimental results show that the detection limit of the substrate for PSA sample solution is as low as 10 pg ml⁻¹. Additionally, the optimal transmission length for PSA is determined to be 9 mm. Further examinations encompass repetitive testing of serum samples spiked with PSA at a concentration of 4 ng ml⁻¹, revealing a relative standard deviation (RSD) of 5.01%. This outcome is compared with single-point detection, manifesting a notable enhancement of 9.81% in terms of RSD, signifying a marked improvement in repeatability. Notably, our method not only capitalizes on the long-range averaging effect to enhance the repeatability of Raman signals, but also circumvents sample damage resulting from direct laser irradiation.

This paper aims to investigate the dynamic characteristics of an aeroengine rotor subjected to aerodynamic excitation. Firstly, the dynamic model of the turbo-shaft engine rotor system is established by means of finite element (FE) method, taking into account the blade-tip clearance-induced aerodynamic force and oil-film force induced by squeeze film damper. Next, the critical speeds and mode shapes of the rotor are achieved by calculation, simulation and experiment to illustrate the validity of the established model. Then, using bifurcation diagram, rotating orbit and time-domain waveform, the dynamic responses of the turbo-shaft engine high-pressure (HP) rotor with aerodynamic excitation are analyzed, and the effects of blade-tip clearance, inlet and outlet angles on the dynamic responses of the rotor are discussed, respectively. The numerical results reveal that the system maintains a stable-periodic motion at low-speed region. In contrast, at high-speed region, there is a noticeable appearance of instability. Besides, the magnitude of the response decreases with the growth of the blade-tip clearance. The growth of the inlet angle will result in an increase of the magnitude of the response, and the magnitude of the response reduces with the growth of the outlet angle. The aerodynamic force at resonance is greater than that at non-resonance. The research results can assist in understanding the dynamic characteristics of the rotor subjected to aerodynamic excitation, and provide theoretical guidance for the structural design of an aeroengine rotor system.

The procedure of feature matching is one of the most important components in binocular or multi-ocular stereoscopic photogrammetry. In this paper, a feature-point matching algorithm based on the technique of particle tracking velocimetry is proposed for the scenario of measuring complex surface morphology by dense-point three-dimensional reconstruction. The core idea is to mix the epipolar-line constraint of line-of-sight (LOS) with the measure of a global similarity pairing and estimate the depth of each feature point in an iterative way. Experimental test is conducted to verify the algorithm performance by measuring the surface topology of a wave-like model. The result demonstrates that the feature-point matching algorithm is superior to traditional LOS method in terms of accuracy and robustness. Moreover, replacing the first module of coarse matching in the proposed algorithm by LOS will save the computational cost significantly without sacrificing the measurement accuracy.

Drug abuse is currently a growing concern, with the abuse of narcotic drugs, represented by ketamine, being particularly serious. Chronic use and overdose of such drugs can lead to hallucinations and serious health risks. Therefore, it is essential to establish a fast and high-precision detection method for strict control of drug abuse. In this study, we propose a fast detection process for small instruments, and the use of data augmentation and transfer learning techniques makes it possible to build a high-precision detection model based on small data sets. We also propose an EfficientNet-based Ketamine quantitative detection network. We maintain its lightweight architecture while introducing large kernel convolution to obtain a stronger feature representation without increasing the number of parameters. Our proposed detection process is time-saving and effective, reducing the time required to acquire large data sets, and our proposed model predicts all test data perfectly, solving the problem of bias of other networks in predicting low concentration samples, and breaking the limit of optical biosensor detection accuracy from 1 ng ml⁻¹ to 0.1 ng ml⁻¹.

This paper presents the development and experimental validation of an Internet of Things (IoT) thermography system for in-situ and real-time monitoring of wall thermal transmittance. The solution proposed has been derived from the upgrade of the Comfort Eye sensor, which is an infrared-based sensor adopted for non-intrusive indoor environmental quality monitoring in occupied buildings. In this work, the system has been used to detect potential building envelope inefficiencies and track building performance trends in a continuous way. The methodology is based on the ISO 9869-2 standard but it has been applied to an entire wall and during its normal functioning without the need of operators. The data management has been performed with a dedicated IoT architecture that allows the synchronised collection of quantities required for transmittance calculation, i.e. indoor and outdoor air temperatures together with the thermographic maps of the wall. The measurement technique has been validated in a real building through the comparison with the results obtained using a heat flux meter (HFM). An uncertainty analysis with Monte Carlo simulation has also been performed to evaluate the overall uncertainty of the method. The values obtained are coherent with those measured with the HFM and the infrared system has proved to be able to provide thermal transmittance measurements with an expanded uncertainty of ±0.038 W m⁻²K⁻¹ with coverage factor k = 2. The innovative methodology described can be used for U-value estimation without the need for extra measuring tools.

The serious pollution of particulate matter (PM) causes many adverse effects on human health and industrial production. High-precision, real-time monitor of atmospheric particulate concentration is becoming increasingly important. In this study, a method to find the optimal detection angle to measure the mass concentration of PM is proposed. To find an optimal angle, a motorized rotary system is used to measure the light scattering signal of particles from 0° to 180°. The new method obtains parameters such as mass concentration factor, deviation, mass concentration coefficient and deviation coefficient, which can be measured more accurately. Through theoretical calculation and experimental analysis, the optimal detection angle is 45°. At the optimal angle, the correlation between the scattered light intensity and the mass concentration of spherical SiO₂ particles reaches 0.9836. Then, the mass concentration of PM is inverted and calculated at the optimal angle. The results showed that the correlation coefficient R² of different particle sizes is larger than 0.952, and the measurement results were highly correlated.

The atmospheric boundary layer (ABL) height (ABLH) is an essential index for climate and weather studies in the troposphere. Obtaining abundant and precise observations is crucial for modeling the top of the ABL. The radio occultation can provide observation data of Earth's atmosphere with high vertical resolution and comprehensive spatial coverage, benefiting from its measurement mode. Thanks to the improved quality of Constellation Observing System for Meteorology, Ionosphere, and Climate (COSMIC)-2 lower atmosphere observations, its substantial amount and higher-quality atmospheric profile data are well suited for ABLH detection. Therefore, 1.4 million COSMIC-2 refractivity profiles from October 2019 to September 2020 were employed to determine the ABLH, and its retrieval performance is assessed using the official COSMIC-2 product (zbalmax) and radiosonde data. The retrieval results show that most of the effectively detected ABLH occultation events occur over the ocean. Comparing COSMIC-2 ABLH and zbalmax, consistent accuracy is found with a standard deviation of 0.22 km. Similar biases are observed each month. And the inversion accuracy in the Northern Hemisphere is higher than in the Southern Hemisphere. The ABLH derived from radiosonde data and COSMIC-2 has a high correlation coefficient of no less than 0.75. The better accuracy happened in the DJF (i.e. from December to February) for both comparisons with two datasets. Additionally, the distribution and seasonal variations were revealed by modeling verified COSMIC-2 ABLH. The fifth generation European Centre for Medium-Range Weather Forecasts (ECMWF) atmospheric reanalysis of the global climate (ERA5) surface temperature and self-calibrated Palmer Drought Severity Index are used to analyze the impact of temperature and drought on ABLH. The results indicate that higher surface temperatures lead to a smaller ABLH, whereas drought at the surface leads to a larger ABLH. Therefore, the higher precision ABLH can be derived from COSMIC-2 after verification, and we can reveal its seasonal variation and influencing factors on this basis.

Precise point positioning (PPP) technique is practical in estimating the zenith tropospheric delay (ZTD). To comprehensively study the factors affecting its accuracy in the polar region, different elevation cutoff angles for each single system and multi-system combination, different tropospheric mapping functions, tropospheric gradients and also seasonal variations are assessed. Meanwhile, European Centre for Medium-Range Weather Forecasts (ECMWF) products are adopted to validate the accuracy of PPP-derived ZTD. The results show that the root mean square (RMS) value of the estimated ZTD and ECMWF products is around 4 mm when the elevation cutoff angle is set to 10 degrees or less. From the view of system combination, it improves the ZTD accuracy in comparison with each individual system. The dual-system combination has 21.8%, 19.0% and 19.1% improvements in its ZTD accuracy compared with the Global navigation satellite system (GLONASS), Beidou navigation satellite system (BDS) and Galileo satellite navigation (GAL) systems. The four-system combination can reach 7.7% further improvement in ZTD estimation than that of the dual-system. For mapping functions, Vienna mapping function 3 (VMF3) function has the best accuracy among the four mapping functions of Niell mapping function (NMF), global mapping function, VMF1 and VMF3, while NMF does not perform as well as others. Considering the tropospheric gradients could get higher accuracy in ZTD estimation than ignoring them. The tropospheric delay estimated in winter is more accurate than that in summer and specifically the RMS value in winter is about 0.9 mm lower than that in summer. Overall, from investigation the estimation of ZTD in the polar region can be optimally handled with a multi-system and VMF3 mapping function at an elevation cutoff angle of 7°, considering the tropospheric gradients.

Porous media flows are common in both natural and anthropogenic systems. Mapping these flows in a laboratory setting is challenging and often requires non-intrusive measurement techniques, such as particle image velocimetry (PIV) coupled with refractive index matching (RIM). RIM-coupled PIV allows the mapping of velocity fields around transparent solids by analyzing the movement of neutrally buoyant micron-sized seeding particles. The use of this technique in a porous medium can be problematic because seeding particles adhere to grains, which causes the grain bed to lose transparency and can obstruct pore flows. Another non-intrusive optical technique, planar laser-induced fluorescence (PLIF), can be paired with RIM and does not have this limitation because fluorescent dye is used instead of particles, but it has been chiefly used for qualitative flow visualization. Here, we propose a quantitative PLIF-based methodology to map both porous media flow fields and porous media architecture. Velocity fields are obtained by tracking the advection-dominated movement of the fluorescent dye plume front within a porous medium. We also propose an automatic tracking algorithm that quantifies 2D velocity components as the plume moves through space in both an Eulerian and a Lagrangian framework. We apply this algorithm to three data sets: a synthetic data set and two laboratory experiments. Performance of this algorithm is reported by the mean (bias error, B) and standard deviation (random error, SD) of the residuals between its results and the reference data. For the synthetic data, the algorithm produces maximum errors of B & SD = 32% & 23% in the Eulerian framework, respectively, and B & SD = −0.04% & 3.9% in the Lagrangian framework. The small-scale laboratory experimental data requires the Eulerian framework and produce errors of B & SD = −0.5% & 33%. The Lagrangian framework is used on the large-scale laboratory experimental data and produces errors of B & SD = 5% & 44%. Mapping the porous media architecture shows negligible error for reconstructing calibration grains of known dimensions.

Transfer condition monitoring across wind turbines still exists two issues. (1) The model of single turbine trained is directly used for other ones in the wind farm to result in low monitoring accuracy because of individual features of each turbine; (2) if the model is trained separately for each turbine, it is not only time-consuming and inefficient, but also is difficult to transfer across turbines. For solving the two issues, a transfer condition monitoring method across wind turbines using feature alignment and parameter fine-tuning is proposed. First, the representative normal turbine in a wind farm is screened by calculating the correlation between the monitoring parameters of wind turbines; second, the input features are constructed by using the massive normal supervisory control and data acquisition data of the representative turbine. The intelligent condition monitoring model of wind turbines is designed by using convolutional neural networks for spatial features and long-term and short-term memory networks for time features. Finally, feature alignment is performed using the normal data of the representative turbine and a large amount of historical normal data from other turbines to reduce data distribution differences. The aligned data is inputted into pre-training monitoring models for parameter fine-tuning. The personalized features of each turbine are incorporated to achieve high-precision transfer condition monitoring across turbines. The proposed method is verified by the real-world turbine data from a wind farm cooperated with us. The results show that the proposed transfer condition monitoring method using feature alignment and parameter fine-tuning can accurately monitor and identify the health states of wind turbines.

Direct interfacing technique (DIT) eradicates additional circuit requirements for sensor-embedded system interface and digitization of analog signals. This technique provides advantages in designing an efficient, portable, and low-cost sensor system. Pulsed eddy current testing (PECT) systems are used for thickness and defect measurements of conductive materials. Circuitous sensor-interfacing methods and tedious data interpretation processes make PECT systems inapt for miniaturization and portable applications. In this work, DIT is used in conjunction with PECT for thickness estimation of conductive material. Change in the de-energizing time of a single coil probe with sample thickness, with respect to air, is used as a signal. The curve fitting method yields a maximum relative error of ≈2% in the thickness estimation. Effects of temperature and liftoff on system accuracy are also investigated. A liftoff compensation method using a 3-signal data group is proposed. It is shown that for thicknesses in the range of 0.508 mm–3.175 mm and liftoffs up to 3.000 mm (step size: 0.500 mm), the proposed scheme produces a maximum relative error of 5.2%. The DIT and PECT combination can be applied for different structural eddy current testing in the future.

Maintenance is a critical aspect of complex products through entire life cycle, often requiring coordination of production planning and available resources, while previous studies appear to have rarely addressed. With this in mind, this paper presents a prescriptive maintenance framework based on digital twins (DTs) for reducing operational risk and maintenance costs of complex equipment clusters. Virtual entities are firstly constructed for each single asset in multiple dimensions, which use real-time or historical sensing data collected from the physical entities to predict the corresponding remaining useful life (RUL). Then such RUL information is incorporated into a stochastic programming model with chance constraints to enable dynamic decision making. In particular, a risk-based optimization model is formulated to take full account of the physical distances between facilities and production gaps. Further, a dual-sense pyramidal transformer model is proposed to sense important details of data in both time and space while capturing temporal dependencies at different scales. Compared to existing data-driven approaches, the proposed DT-based alternative achieves dynamic real-time interaction between physical and virtual units driven by both models and data, while virtual verification based on high-fidelity models ensures high reliability of maintenance decisions, which has also been validated in an aero-engine maintenance case study.

Wheel profile wear, or bogie suspension failure will result in hunting instability of the vehicle system which will lead to dramatic changes in the lateral motion of the wheelset. The occurrence of hunting instability will worsen the ride performance of the vehicle or even cause the danger of wheelset derailment. Thus, the in-time and effective detection of wheelset lateral motion is very important for the vehicle hunting stability evaluation. By installing a certain amount of displacement sensors on the rail side along the track and detecting the distance between the wheel and rail when the wheelset passing by, the lateral displacements of the wheelset in measuring points can be obtained. The detection method of wheelset lateral motion by the wayside is studied, which includes the measuring principle, structure design, sensor installation, data acquisition and data analysis, and the application of the method to on-site detection for typical instability cases are demonstrated. At the same time, innovate on the existing on-board hunting instability monitoring method, proposed an evaluation method of hunting instability for wayside lateral movement monitoring system. The results show that: installing 20 eddy current sensors on both sides of the straight track to measure the lateral displacement of the wheelset can meet the requirement of detecting the vehicle hunting instability. At the same time, analyzing the measurement data and fitting the maximum value of the signal collected by each sensor can obtain the lateral trajectory of the wheelset. The wheelset lateral displacement can be used to monitor the occurrence of vehicle hunting instability if periodic motion is detected. The amplitude of the motion can determine the hunting instability level, and the frequency can be used to describe the wheel wear level and calculate the wheel/rail contact equivalent conicity. This method is able to provide a reliable data support for the vehicle operation safety. The proposed monitoring method and monitoring index are also of reference value to the development of other wayside hunting instability monitoring systems.

Sparse filtering (SF) has received considerable attentions in the machinery fault diagnosis thanks to its ability to extract the fault-related features using their sparsity. However, the existing SF methods have dilemmas with the empirical selection of model parameters, the loss of fault-related information caused by a screening way for the target mode, and the singularity of results induced by some large-amplitude random impulses (LARIs). Hence, a manifold learning-assisted SF method is proposed for machinery fault-related feature enhancement in this study. First, an improved intrinsic component filtering (ICF) is presented for extracting the multiple modes with feature enhancement, where the parameters of ICF are adaptively determined by using the optimization object to avoid the empirical selection of parameters. Second, the manifold learning is introduced to compress the enhanced multiple modes to overcome the loss of fault-related information; thus the intrinsic manifolds are obtained for disclosing the buried fault-related features and suppressing the band-in noise. Third, an adaptively weighting strategy for these intrinsic manifolds is constructed to obtain a final representative mode for conducting the machinery fault diagnosis. Meanwhile, the LARIs coupled with intrinsic manifolds are calibrated according to their statistical information to resolve the singularity of the representative features. Simulation and experiments show that the proposed method is more effective in extracting fault-related features than some existing methods.

To enhance the signal extraction performance at weak noise intensity in virtual anechoic chamber, this study investigates the correlation between the generation conditions of a stochastic resonance (SR) system and its output performance. To achieve optimal resonance effects, an adaptive system called symmetric piecewise bistable SR (SPBSR) is proposed. This system improves its structure by modifying the potential function to facilitate the occurrence of SR. Meanwhile, it combines a regional multi-role strategy and particle swarm optimization algorithm to determine the optimal structural parameters. The adaptive optimization process utilizes update rules that balance global and local optimal solutions, thereby mitigating the tendency to quickly converge to a local optimum. Experimental results demonstrate that the proposed system exhibits excellent performance within a noise intensity range of 0–10 dB, with a correlation coefficient of over 0.75, which can effectively suppress noise interference. In practical signal processing, the system excels at accurately extracting signal characteristics, resulting in improved similarity and smoothness of the detected signal. Both simulation and experimental results validate this algorithm's strong practical relevance.

Electromagnetic testing (ET) is a method in the nondestructive evaluation (NDE) methodology for detecting and evaluating cracks in engineering structures by measuring the distribution of the electromagnetic field. With the development of advanced deep learning (DL) techniques, DL is getting attention in crack detection problems through the measuring signals of NDE systems. However, obtaining the necessary big data for such experiments is a time-consuming and expensive task. Moreover, the quality of the crack detection results heavily depends on the quality and spatial resolution of the input image data. In this research, we solve the above-mentioned problems by proposing a DL-based method on the magnetic image for both the image super-resolution and image denoise. We first build a magnetic image simulation framework for the ET system, and thus, it is possible to build a large dataset for training the DL model. Second, we build a DL model for enhancing the quality of the magnetic image in both the spatial resolution and the denoise signal. This approach helps to reduce the cost of experiments (i.e. reducing the number of sensors for each experiment), sensor fabrication process, and noise removal of the ET system. We evaluate the proposed approach on several cracks with different sizes and shapes of aluminum specimens. The evaluation metrics, such as the structural similarity index measure (SSIM) and root mean square error show good performance of the proposed approach and compared to traditional methods.

The emissivity of a material changes with temperature. The knowledge of emissivity plays an important role in the estimation of radiation heat load in cryogenic systems. As the emissivity values of different materials at cryogenic temperatures are scarcely available in the literature, room-temperature emissivity values have been extensively used to estimate the radiation heat load in cryogenic systems. This may lead to a significant deviation between the predicted and actual radiation properties at cryogenic temperatures. Therefore, in the present work, an apparatus is developed based on a calorimetric technique for measuring the emissivity of an opaque material around 82 K. The novelties of the apparatus are its compact size, ease of sample handling, shorter time required to reach thermal equilibrium and, most importantly, capacity to measure the emissivity of a sample of high roughness. To understand the effectiveness of low and high emissivity-coated heat radiators for the system, a theoretical and an experimental approach has been followed. It is found that the high-emissivity heat radiator leads to a significant reduction in the time required to reach thermal equilibrium compared to a low-emissivity heat radiator. To verify and validate this setup, the emissivities of Aeroglaze Z306 (high emissivity) and Cu (low emissivity) are measured and compared with the values reported in the literature. Finally, this work has been extended to measure emissivity at cryogenic temperature for the first time for PU1, SG121FD and indigenously developed novel materials, such as black paints, adhesive and activated charcoals of different granule sizes.

Fault diagnosis of asynchronous motors has become a pressing need in the metallurgical industry. Due to the complex structure of asynchronous motors, fault types and fault characteristics are diverse, with strong nonlinear relationships between them, which leads to the difficulty of fault diagnosis. To efficiently and accurately diagnose various motor faults, we propose a fault diagnosis method based on an optimal deep bidirectional long short-term memory neural network. First, the three-phase current, multidimensional vibrational signal, and acoustic signal of the asynchronous motor are collected and construct diverse and robust data sample set to enhance the generalization ability of the model. Next, a modified 3D logistics-sine complex chaotic map (3D LSCCM) is constructed to improve the global and local search capabilities of the pigeon swarm optimization algorithm (PIO). Then, we construct a deep bidirectional long short-term memory network (Bid-LSTM) with attention mechanism to mine high-value fault characteristic information. Meanwhile, the optimal hyper-parameters of the deep ABid-LSTM are explored using the modified PIO to improve the training performance of the model. Finally, the fault data samples of asynchronous motor are induced to train and test the proposed framework. By fusing diverse data samples, the proposed method outperforms conventional deep Bid-LSTM and achieves fault diagnosis accuracy of 99.13%. It provides a novel diagnostic strategy for motor fault diagnosis.

Low-power electromagnetic-acoustic transducer (EMAT) is crucially important for safety-critical equipment in industry, especially for potential explosives and inflammable petrochemical equipment and facilities. When the excitation power is very low, the corresponding echoes are overwhelmed in noise and related measurement would be inaccurate. To solve this problem, this paper presents a new echo reconstruction method based on a deep stacked denoising autoencoder (DSDAE) for nondestructive evaluation. First, the uses of reference signals and new data structure are to improve the training efficiency. A hybrid method based on variational mode decomposition and wavelet transform is used to obtain clean reference signals as inputs of the deep network. Then, the modified network structure and loss function aim to improve the ability of feature extraction and reconstruct clean echoes from low-power EMAT signals. To validate the effectiveness of the proposed method, the experiments of self-excitation and receiving A-scan inspections of stepped specimens with different thicknesses are conducted at some excitation voltages, as low as 25 V. The results indicate that the proposed DSDAE shows better and more stable denoising performance than some popular processing methods for different specimens and excitation voltages. It greatly improves the signal-to-noise ratio of the reconstructed signal to 20 dB. When applying to thickness measurement of specimens, its relative error is lower than 0.3%, which provides a practical and accurate tool for low-power EMAT testing.

In global navigation satellite system (GNSS) data processing, precise point positioning (PPP) with ambiguity resolution (PPP-AR) is a versatile technique that aims to achieve centimetre-level accuracy by resolving integer ambiguities in carrier phase observations. However, the inherent errors and biases in the satellite signals can degrade the performance of PPP-AR solutions. To mitigate such errors, this research proposed to argument PPP-AR using third-generation BeiDou Navigation Satellite System (BDS-3) multi-frequency observations and the observable-specific signal biases (OSBs) generated at the Centre National D'Etudes Spatiales (CNES). To test the proposed technique, both BDS-3 and Galileo observations from the multi-GNSS experiment network were used, in consideration that the latter also transmits multi-frequency signals. Before demonstrating the impact of CNES bias products on PPP-AR, the quality of BDS-3 and Galileo signals was assessed. The results indicated that the modernised frequencies had the best signal strength. The mean standard deviations for the estimated OSB for different receivers were close to each other in both constellations. Besides, the positioning results in different processing schemes unveiled a comparable positioning accuracy, and slightly better in the quad-PPP strategy using the Galileo constellation in both static and kinematic modes. Galileo also attained better ambiguity fixing rates and convergence time than BDS-3. Finally, there were slight differences in the magnitude of the estimated phase residuals for distinct frequency signals between BDS-3 and Galileo, including the interoperable and compatible signals.

Rotating machinery is widely utilized as mechanical equipment in the industrial field. However, due to the complex working conditions, the existing fault diagnosis methods have failed to address good results in practical applications. To improve the fault diagnosis performance of rotating machinery in a noisy environment, a new multi-scale convolution neural network (MSCN) based on a self-calibrating attention module is proposed. First, this thesis constructs a multi-scale convolution layer with a wide convolution kernel to form an efficient sampling structure at the filter level, which can filter out incoherent noise from the signals and extract rich features. Second, a multi-scale self-calibrating attention module is implemented with two identical self-calibrating convolutional networks to continuously focus on significant embeddings and adaptively combine information from different spatial dimensions. Third, the multi-dimensional characteristics are integrated by the feature cascade layer, and then the fault modes are identified though the classifier layer under noise. Finally, based on the Case Western Reserve University datasets and Paderborn University bearing datasets, the experimental results show that our proposed MSCN can significantly enhance the fault identification ability to rotate machinery in a noisy environment.

Rolling bearings are critical and easily damaged components of mechanical equipment. In practical engineering applications, the collected signals typically contain a large amount of noise, which makes fault diagnosis difficult. Based on this, this paper proposes an adaptive time–frequency demodulation method for rolling bearing diagnosis. The proposed method first obtains the complex envelope of the vibration signal in the time–frequency domain using the S transform (ST), and the time–frequency coefficient of ST can be used as the complex envelope, which is proved in detail in this paper. Subsequently, the complex envelope of the optimal slice frequency is obtained by frequency slicing to significantly weaken the interference of irrelevant noise and highlight the fault characteristics. An indicator is proposed to adaptively select an optimal slice frequency component that contains the most fault information. Finally, the slice envelope spectrum of the optimal slice frequency is obtained using Fourier transform for fault diagnosis. The feasibility of the proposed method is verified using the simulated signal. The application results of the bearing inner and outer ring fault experimental signals indicate that the proposed method is more accurate and effective for bearing fault diagnosis. Comparisons with other commonly used methods also verified the superiority of the proposed method.

The aim of this work has been to derive and statistically evaluate the accuracy of second-order and third-order polynomials to determine vapour pressure deficit (VPD). These polynomials take air temperature and relative humidity measurements to determine VPD without the use of an exponential function, as proposed by F W Murray in 1967. Replacing the exponential function with a 2nd or 3rd order polynomial may be beneficial in ultra-low power microcontroller-based measurement applications where; code size, memory usage and power requirements are critical design drivers. However, oversimplification may impact precision. This work presents alternative 2nd order and 3rd order equations that have been derived from a Murray equation dataset where VPD isothermal datasets were plotted against relative humidity. These linear relationships allow y = mx + c analysis where, (i) 'c' can be set to zero with a offset in the relative humidity data, and, (ii) 'm' can be derived from a 2nd of 3rd order polynomial where 'm' = f(T) and is derived using Excel-based fitting of the gradients from the isothermal datasets. The resulting 'm' = f(T) 2nd and 3rd order polynomials presented R² values of 0.998 04 and 0.999 98 respectively. A Bland-Altman statistical assessment was performed, where the Murray equation (reference) dataset is plotted against the difference between the reference and polynomial datasets using the same air temperature and relative humidity inputs. The difference datasets presented 2-sigma (95% confidence interval) variances for the 2nd and 3rd order polynomials as <±0.1 kPa and <±0.01 kPa respectively. The 2nd and 3rd order polynomials also resulted in a bias values of <0.0037 kPa and <0.0013 kPa respectively. These results suggest that a 3rd order polynomial equation could be used to determine VPD in ultra-low-power microcontroller measurement applications, with minimal impact on VPD measurement precision.