Exploration of maintenance technology of aviation new energy power battery under machine learning

The existing prediction models on the market have uncertainties and errors, which cannot accurately predict the remaining life of batteries. These problems lead to excessive or delayed maintenance of batteries, thereby affecting their performance and safety. Therefore, machine learning-based maintenance technology for aviation new energy power batteries has become a new research direction. This study used machine learning algorithms to construct a battery status classification and prediction model by collecting a large amount of aviation new energy power battery data. Through experiments, the model significantly improved its accuracy in battery status recognition tasks, with a maximum of 96%. It can accurately identify tasks such as battery health status, lifespan prediction, and fault detection. The model had made progress in balancing accuracy and recall, with a high F1 value. This meant that the model could accurately identify the abnormal state of the battery while minimizing misjudgment of normal batteries as much as possible. These research results provided important support for battery maintenance in the aviation field, with the potential to improve battery availability, extend service life, and ensure flight safety while reducing maintenance costs.


Introduction
The increasing demand for new energy power batteries in the aviation industry has made the maintenance of aviation new energy power batteries a key research direction.Effective battery maintenance technology can ensure the reliability and safety of aircraft while extending the service life of batteries.However, traditional battery maintenance methods have certain limitations in accuracy and efficiency, making it difficult to meet the requirements of high precision, instantaneity, and intelligence in the aviation industry.Therefore, machine learning-based maintenance technology for aviation new energy power batteries has become a current research hotspot.
Many scholars and researchers use machine learning algorithms such as deep learning, support vector machines, and decision trees, to perform state recognition, life prediction, and fault detection on aviation new energy power batteries.Cheng and Zhu researched maintenance and upkeep strategies for new energy vehicle power batteries to help promote the establishment of a reasonable use and maintenance system for new energy vehicles and promote the sustainable development of the new energy vehicle industry [1].Wang introduced the current status of battery usage, proposed key points for battery maintenance, and believed that the installation of UPS (Uninterruptible Power Supply) power supply system in the weak current system of the rail transit system requires proper maintenance work [2].Chen et al. emphasized the importance of the initial charging of helicopter cadmium nickel aviation alkaline batteries and emphasized the precautions to be taken during the charging and discharging process to improve the battery's efficiency and ensure helicopter air safety [3].These research results demonstrated IOP Publishing doi:10.1088/1742-6596/2764/1/012033 2 the enormous potential of machine learning in the maintenance of aviation new energy power batteries, which can improve the accuracy and predictive ability of battery status.However, there are still some challenges in current research, such as data preprocessing, feature extraction, and model optimization, which require further research and improvement.
The research significance of this article is to propose a comprehensive, efficient, and accurate research method for the maintenance of aviation new energy power batteries through the use of machine learning algorithms.This article explored data preprocessing techniques suitable for aviation environments and extracted effective features of battery status through reasonable feature extraction methods.At the same time, this article constructed a machine-learning model and trained and optimized it to achieve accurate classification and prediction of battery status.The research method of this article aimed to improve the reliability and practicality of maintenance technology of aviation new energy power batteries and provide scientific basis and technical support for battery maintenance decisionmaking in the aviation industry.

Overview
The maintenance technology of aviation new energy power batteries is generally used for monitoring, maintaining, and managing batteries in the aviation field.As one of the important energy systems of aircraft, the performance and reliability of this type of battery are self-evident for the importance of aviation safety and efficiency.Effective battery maintenance technology can ensure the normal operation of batteries, predict and solve potential faults, and improve battery life and performance [4][5].By monitoring the working status of the battery, combined with sensors and monitoring systems, the performance and health status of the battery are monitored in real-time.The battery state prediction model can predict the future performance and remaining life of batteries.By combining battery usage and historical data, the lifespan and performance degradation trend of the battery can be predicted.Based on the predicted results, optimization strategies can be developed, such as charging control, temperature management, and recycling strategies, to extend battery life and improve performance [6][7].After efficiently analyzing the abnormal behavior and current and voltage fluctuations and identifying the types and causes of battery faults, corresponding maintenance and replacement measures are taken.

Function
The development and application of this battery maintenance technology have a significant impact on the aviation industry.It can improve the energy efficiency and performance of aircraft, further promoting the transformation of the aviation industry towards sustainable development.This technology can improve the energy density and cycling stability of batteries, thereby increasing the aircraft's endurance and flight safety [8][9].It can also reduce the operational costs of aircraft.By accurately monitoring and predicting the health status of batteries, timely detection of battery failures and abnormalities can avoid maintenance and replacement costs caused by battery failures and reduce aircraft downtime on the ground, thereby improving operational efficiency and reliability.This technology also helps to improve flight safety.Batteries are a key energy storage device for aircraft, and their reliability and performance are directly related to flight safety [10][11].By adopting technologies such as machine learning and data analysis, the working status, temperature, voltage, and other parameters of the battery can be monitored in real time, and potential faults and problems can be warned in advance to ensure the normal operation and safety of the battery.
In summary, the application of aviation's new energy-power battery maintenance technology can improve aircraft energy efficiency, reduce operating costs, and improve flight safety.This can promote significant breakthroughs in sustainable development and innovative technologies in the aviation industry, making positive contributions to the future development of the aviation transportation industry [12][13].

Shortcomings in current technology
In practical situations, identifying battery failure modes and abnormal behavior is not an easy task.The working state of batteries is influenced by various factors such as temperature changes, charging and discharging cycles, and load changes.The complexity of these factors leads to low accuracy in fault diagnosis, leading to misjudgment or omission.The current life prediction model has uncertainties and errors, which cannot accurately predict the remaining life of batteries.These issues lead to excessive or delayed maintenance of the battery, thereby affecting its performance and safety [14][15].During the flight, battery monitoring data needs to be collected and transmitted to ground systems in a timely and reliable manner for analysis and processing.However, the complexity and highly dynamic operating environment of aircraft may lead to difficulties in data collection and transmission, limiting accurate monitoring and analysis of battery status.The maintenance technology of aviation new energy power batteries often requires advanced sensors, monitoring systems, and data analysis tools.These devices and technologies typically have high costs and complexity, increasing the difficulty and cost of implementing maintenance techniques.This may limit the popularization and application range of technology [16][17].
The existence of these problems greatly hinders the development of the aerospace technology field.This article further studies and develops machine learning technology to improve the accuracy of fault diagnosis and life prediction, and solve the problems and shortcomings of data collection and transmission.

Battery monitoring model under machine learning
For the collection of battery monitoring data, real-time collection is carried out through sensors or monitoring devices, and it is associated with the fault status of the battery.The collected data needs to be preprocessed to ensure the quality and accuracy of the data.The missing value processing and outlier detection preprocessing operations are performed on the data, and the formula is as follows: Where x is the original data; n x is the normalized data;  and  are the mean and standard deviation of the original data, respectively.Afterwards, we perform data smoothing operations on the data.Data smoothing is the process of smoothing data to remove random noise and outliers, reduce data fluctuations, and improve readability.
Where t S represents the smoothed value; Xt represents the value of the original data; a is the smoothing coefficient.For the battery monitoring model, current fluctuation is selected as the feature for selection.Based on the monitoring data and preprocessed features, a decision tree machine learning algorithm is selected to construct a fault diagnosis model [18][19].Figure 1 shows the simple structure of the battery monitoring model of the decision tree algorithm.

Battery health assessment
Analytic Hierarchy Process is used to determine the weights of indicators.The evaluation indicators for battery health are divided into capacity decay, internal resistance measurement, cycle life, temperature, and environment.A paired comparison matrix is created to compare the relative importance of each indicator to other indicators.In an N × N matrix, N is the number of indicators.By using a scale of 1-9, each pair of indicators is compared, where 1 represents the same importance and 9 represents the absolute importance difference, as shown in Table 1.Next, the average values of each column are calculated to standardize the matrix, as shown in Table 2. Then the weight of each row is calculated, which is the average divided by their total, as shown in Table 3.The indicators are sorted based on their weights, and the indicator with the highest weight is considered the most important [20].

Optimization of battery life
(1) High temperature is the main factor that shortens the lifespan of batteries.Controlling the working temperature of the battery and avoiding excessive temperature and extreme temperature exposure can slow down the aging rate of the battery and extend its service life.This can be achieved through the use of heat dissipation systems, temperature sensors, and temperature management algorithms.
(2) A reasonable charging strategy can have a significant impact on battery life, and overcharging or discharging can hurt battery life.By optimizing charging management strategies, limiting charging voltage and current, and adopting appropriate charging algorithms such as constant current and constant voltage charging, battery loss can be reduced and battery life can be extended.
(3) The number of cycles of a battery also has a significant impact on its lifespan.It is necessary to minimize the frequency or amplitude of charging and discharging cycles as much as possible and adopt a strategy of alternating deep and shallow cycles, which can slow down the capacity decay rate of the battery and extend its lifespan.
(4) Reasonably managing the load of the battery and avoiding high or low loads can reduce the stress of the battery and extend its service life.This can be achieved through methods such as load controllers and power management algorithms.
(5) The working environment of batteries also has an impact on their lifespan.Avoiding adverse environmental conditions such as extreme temperature, humidity, and vibration and providing appropriate ventilation and cooling measures can improve the lifespan and performance stability of batteries.

Testing methods
The purpose of conducting experiments is to predict the lifespan of aviation new energy power batteries through machine learning algorithms and to evaluate the accuracy and reliability of the algorithms.To better compare the advantages and disadvantages of the method and technology proposed in this article, the experimental group was selected as the method and technology proposed in this article, while the traditional method and technology were used as controls to conduct experimental research on the two methods.After obtaining the available dataset of aviation power batteries, including charging and discharging cycle data, temperature information, and capacity degradation indicators, the decision tree model was used as a machine learning algorithm for life prediction.A computer device was prepared for experiments and algorithm development and training, with Windows as the operating system.The Python programming language and the PyTorch machine learning library were installed for algorithm development, data processing, and model training tasks.In testing, accuracy and F1 value were selected for performance indicators, and multiple experiments were conducted on the same indicator to avoid experimental errors.

Results and discussion
The experiment was conducted according to the above steps.During the experiment, the experimental process was strictly followed and the results were recorded and statistically analyzed.Matlab was used to statistically analyze the data and create charts.The final results of the experiment are as follows: Figure 2 shows the accuracy comparison results, and Figure 3 shows the F1 value comparison results.In the experiment in Figure 2, the accuracy of the model was tested.In terms of test results, the maximum accuracy of the proposed method model was 96%, and the minimum accuracy was 84%; the maximum accuracy of traditional method models was 88%, and the minimum accuracy was 78%.High accuracy indicates that the model has strong classification or prediction ability for input data.For battery maintenance technology research, high accuracy means that the model can accurately identify the state or problem of the battery, thereby accurately predicting the battery life, identifying the health status of the battery, or detecting battery faults.This can improve the efficiency and reliability of aviation's new energy-power battery maintenance.In the aviation industry, the performance and health status of batteries are crucial for flight safety.Machine learning models have high accuracy, which can help engineers or maintenance personnel more accurately determine the status of batteries and take corresponding maintenance measures promptly, thereby improving the usability and extending the service life of batteries.In the experiment in Figure 3, the F1 value of the proposed method model ranged from 83% to 94%, and the F1 value of the traditional method model ranged from 79% to 87%.The F1 value of the method model in this article was higher than that of the traditional model.For the research of battery maintenance technology, it is important to accurately identify the normal state of the battery.The higher the value is, the more accurately the model can determine the normal state of the battery and minimize the possibility of mistakenly identifying normal batteries as abnormal.It combines accuracy and recall, providing a comprehensive evaluation while balancing these two indicators.A high F1 value means that the model can achieve a good balance in battery state recognition tasks while possessing high accuracy and comprehensiveness.

Conclusions
Through testing, this article successfully applied machine learning technology to improve the maintenance technology of aviation's new energy power batteries.Decision time algorithms were used to classify and predict battery status, and the accuracy and F1 value of the model were evaluated.Research has shown that this battery maintenance technology can significantly improve the accuracy of battery status recognition while maintaining high accuracy and recall rates.By collecting a large amount of battery data and conducting training, the model can accurately identify tasks such as battery health status, lifespan prediction, and fault detection, which provides reliable support for battery maintenance and improves the comprehensive recognition ability of battery status.This means that the model can accurately identify the abnormal state of the battery while minimizing erroneous judgments of normal batteries, which improves the accuracy and efficiency of maintenance decisions.These research results provide important support for battery maintenance in the aviation field, with the potential to improve battery availability, extend service life, and ensure flight safety while reducing maintenance costs.

Figure 1 .
Figure 1.Structure display of electrical monitoring model.

Figure 2 .
Figure 2. Comparison of accuracy data.In the experiment in Figure2, the accuracy of the model was tested.In terms of test results, the maximum accuracy of the proposed method model was 96%, and the minimum accuracy was 84%; the maximum accuracy of traditional method models was 88%, and the minimum accuracy was 78%.High accuracy indicates that the model has strong classification or prediction ability for input data.For battery maintenance technology research, high accuracy means that the model can accurately identify the state or problem of the battery, thereby accurately predicting the battery life, identifying the health status of the battery, or detecting battery faults.This can improve the efficiency and reliability of aviation's new energy-power battery maintenance.In the aviation industry, the performance and health status of batteries are crucial for flight safety.Machine learning models have high accuracy, which can help engineers or maintenance personnel more accurately determine the status of batteries and take corresponding maintenance measures promptly, thereby improving the usability and extending the service life of batteries.

Table 2 .
Calculation of average values.

Table 3 .
Calculation of weights