Advanced Noncontact Sensors for Blade Tip-Timing Systems

The conference paper addresses noncontact sensors for measuring vibrations of machine blades using the Blade Tip Timing method. An overview of sensors based on the electromagnetic principle suitable for the harsh environment of low-pressure stages of steam turbines is presented. Special attention is further given to magnetoresistive sensors and possibilities for increasing their sensitivity. Variants of the sensors with controlled, adaptive, and intelligent operating principles are discussed. The properties and advantages of each version of the solution are demonstrated.

and shaped into a rectangular signal.This signal is processed in a precise counter, where the passage time of each blade of the machine is determined.Based on the detected time differences, individually parameters of their vibrations are determined for each blade.The system is synchronized with an OPR (One per revolution) sensor, which detects the passage of a phase mark (PM) on the shaft.The signal processing and evaluation chain in the BTT system is shown in Fig. 2. In addition to the Tip-Timing time differences, the system may include a parallel branch that evaluates the amplitudes of the pulses.This method is referred to as "amplitude-impulse methods" here.Supplementary evaluations in this branch can help obtain information about certain important static characteristics of the blades, such as tip clearance, blade elongation, and unusual behavior in case of blade damage.

Electromagnetic non-contact sensors for BTT
The most commonly used principles of sensors for BTT systems are optical, capacitive, and electromagnetic sensors.BTT sensors based on different electromagnetic principles offer several advantages when used in steam turbines.They provide high accuracy and precision in measuring blade vibrations, allowing for reliable monitoring of turbine health and performance.Electromagnetic BTT sensors are resistant to harsh operating conditions, such as high temperatures, steam, and vibration, making them suitable for steam turbine applications.Furthermore, some of these sensors are immune to electromagnetic interference, ensuring accurate and reliable signal detection.Additionally, they offer real-time monitoring capabilities, enabling prompt detection of any abnormalities or faults in the turbine blades.Moreover, the installation of electromagnetic BTT sensors is relatively straightforward and does not require significant modifications to the turbine structure.These sensors can be easily integrated into existing turbine control and monitoring systems.Finally, electromagnetic BTT sensors provide continuous monitoring, allowing for proactive maintenance and preventing potential catastrophic failures.
On the other hand, capacitive sensors rely on changes in electrical capacitance, which can be affected by the presence of steam and condensation, leading to inaccurate measurements.Steam turbines are also subject to high temperatures, which can impact the performance and reliability of capacitive sensors.Similarly, optical sensors depend on the transmission and reflection of light, and the presence of steam and dirt can interfere with the optical signals, causing inaccuracies.Additionally, the high temperatures and harsh conditions within a steam turbine can affect the performance and durability of optical sensors, making them unsuitable for this application.Therefore, electromagnetic BTT sensors are preferred due to their resilience, accuracy, and suitability for the challenging environment of steam turbines.
At the Institute of Thermomechanics, we have been conducting research on electromagnetic noncontact sensors for several years.We have developed several types of sensors and compared their properties on a test model wheel in the laboratory, as well as during measurements on low-pressure stages of steam turbines.The following overview briefly lists the advantages and disadvantages of each sensor type: • Eddy-current -positives: effective for measuring vibrations in non-ferromagnetic materials, high sensitivity, disturbance-resistive, speed independent; negatives: close proximity of the sensor is required, lower durability.• Hall-effect -positives: the ability to measure static displacements; negatives: narrow frequency band, low sensitivity, high values of offset, drift, and noise.• Induction -positives: passive, simple, temperature and environmental resistance; negatives: larger active area, lower frequency band, speed dependance.• Magnetoresistive -positives: symmetrical, low-noise, high sensitive, direction-sensitive, temperature and humidity resistance, frequency band 0 to 5 MHz (with a magnetic circuit 300 kHz), static calibration, measurement of static quantities: blade elongation, blade tip axial deflections, ejection of the blade roots from the disk, axial shift of the disk, stator temperature; negatives: smaller signal amplitude, sensitivity to external magnetic fields.Through our research, we have found that electromagnetic sensors offer significant advantages for BTT systems, such as their ability to withstand challenging environments and provide accurate measurements.These sensors have proven to be reliable and suitable for steam turbine applications.The magnetoresistive type of electromagnetic sensor has emerged as the most advantageous.Its notable features include a wide frequency bandwidth and the ability to measure static quantities and perform static calibration.

Magnetoresitive (MR) noncontact sensor
The active element of the MR sensor is formed by 4 magnetoresistors arranged into a Wheatstone bridge (see Fig. 3).The symmetrical arrangement brings high sensitivity, suppression of the temperature dependence and a distinguished common mode rejection.The anisotropy of the MR sensors is the basis for their directional sensitivity.The original frequency range of the MR chip is 0 to 5 MHz which is reduced to 0 to 300 kHz by the magnetic circuit.The output signal of the MR sensor is independent to the velocity of the blade passage and it can be tested, calibrated and mainly used statically.
Experiments were conducted to increase the sensitivity of the magnetoresistive sensor.Magnetic materials with higher magnetic induction values, particularly Samarium and Neodymium materials, were tested to generate the working magnetic field.As evident from Fig. 5 and 6, sensors with these magnets exhibited higher signal amplitudes.However, their disadvantage is that the strong magnetic field can affect the magnetoresistive bridge of the sensor to such an extent that its function is disrupted, or the magnetoresistors become magnetized to a degree that the sensor is destroyed.For this reason, the use of modern magnetic materials was abandoned, and the sensors are now constructed with a Ferrite magnet.

Controlled and adaptable sensor
The requirement for increased sensitivity of the magnetoresistive sensor cannot be addressed by increasing the supply current of the bridge due to sensor overheating.A possible solution is to operate the MR sensor in controlled mode.The principle is that the sensor is operated in two regimes, passive and active, depending on the actual position of the machine part being measured.When a moving machine component, such as a turbine blade, approaches the controlled sensor to sense the position of the rotating machine parts, the sensor is brought into an active state.In the active state, the measuring bridge of the magnetoresistive sensor is fed by a direct current Imax from a controlled direct current source, which active state lasts for a period of t1.Then the sensor goes into a passive state that lasts for t2.During this time, the measuring bridge of the magnetoresistive sensor is supplied with a direct current Imin.Preferably, the minimum power loss at the sensor bridge occurs when Imin = 0. Of the total time t = t1 + t2, the sensor of the magnetoresistive sensor is powered only for the time t1.
If the control of the supply current of the bridge is derived from the sensor output signal, we are talking about an adaptable sensor.The sensor control electronics are equipped with a sample-and-hold unit SH MAX and a comparator Co1 MAX that continuously compares the amplified (AMP) sensor output signal to the maximum value.When the signal drops by the selected voltage difference ΔV, the power supply control circuit I REG of the measuring bridge is activated, which connects the high current Imax.After the output signal passes the zero voltage level, the second comparator Co2 REF is activated and flips the R-S flip-flop.Then the bridge is powered by the current Imin.A block diagram of the adaptable sensor according to this description can be seen from Fig. 7. Fig. 8 shows the corresponding waveforms of the signals.At room temperature, the sensitivity of the sensor can be increased by approximately 20 times compared to a conventional circuit, at 200 ° C approximately 6 times.

Intelligent sensor
When utilizing non-contact electromagnetic sensors with an internal magnetic field source in the presence of magnetized blades, a potential problem arises.The magnetic fields produced by the sensors can interact with the magnetized blades, resulting in a reduction of the effective magnetic field.Consequently, the amplitudes of the output pulses from the sensors may not reach sufficient levels.This interference between the sensor's magnetic field and the blade's magnetization can hinder the precise detection of blade passage and lead to signal degradation.The diminished pulse amplitudes can impede accurate measurements and compromise the overall performance of the sensor system.To address this issue, careful demagnetization of the blades becomes essential, ensuring that they are free from residual magnetization.Proper demagnetization procedures should be carried out with care to eliminate any interference between the magnetized blades and the sensor's magnetic field, enabling reliable and accurate measurements.Demagnetizing blades is achieved using either an alternating or pulsed current source that generates the corresponding magnetic field.Demagnetization of the blades is not a straightforward process.If, for instance, one end of a blade is demagnetized, the other end of the same blade may simultaneously become magnetized.Furthermore, there are interconnections between the blades, so demagnetizing one blade can increase the residual magnetization in the neighboring blades.The complexity arises from the need to carefully control the demagnetization process to ensure that each blade is effectively demagnetized without inadvertently magnetizing adjacent blades.The presence of these interconnections and the potential for residual magnetization pose challenges in achieving complete demagnetization of the blades.Close attention and precision are required during the demagnetization process to minimize these effects and ensure successful demagnetization of the individual blades.
Imperfectly demagnetized blades exhibit distinct amplitudes of blade pulses.We have addressed this issue by developing a specialized intelligent sensor.The magnetoresistive sensor is enhanced with intelligent electronics.These electronics measure the amplitudes and offsets of individual blade pulses within a few rotations.Subsequently, the controlling microprocessor of the unit calculates the gain correction and amplifier offset value to ensure that the resulting pulses have sufficient and approximately equal amplitudes.This guarantees that no pulse will be missed or appear as a double pulse during comparison.The operational principle of this unit is depicted in Fig. 9, while the implementation of the microprocessor-controlled electronics can be seen in Fig. 10.Fig. 11 illustrates the result of signal processing by the intelligent electronics of the magnetoresistive sensor.The green waveform represents the original signal, which is solely amplified.In the latter part of the waveform, there is a region with very low pulse amplitudes that could be misidentified.The intelligent electronics have modified the waveform -depicted in red -such that all signals are correctly converted to a rectangular waveform after processing in the comparator -in blue.

Conclusion
A new magnetoresistive sensor has been developed for Blade Tip Timing systems.The sensor's sensitivity was subsequently increased through special hardware modifications and operational modes with controlled and adaptive operation.The intelligent sensor adjusts the gain and offset individually for each blade.This ensures precise and accurate blade passage detection for every blade.The ability to customize the gain and offset settings enhances the performance and reliability of the sensor system.The developed sensors were successfully tested under laboratory and operational conditions at lowpressure stages of steam turbines.

Fig. 2 .
Fig. 2. The signal processing and evaluation chain in the BTT system.
By passing a ferromagnetic blade tip under the sensor, the equilibrium of the bridge is distorted and a voltage pulse is generated at the output of the bridge.The magnitude of this output voltage pulse can be expressed as   =       =       sin  =  0     sin  sin 2 (1) where cS is the total sensitivity of the sensor in [m/A]; VS is the supply diagonal voltage [V]; Ht [A/m] is the tangential component (in the direction y) of the total magnetic field intensity HB [A/m] during the passage of a blade along the active element of the sensor; γ is the angle of the sensed tangential component of the magnetic field intensity Ht to the axis of the bridge in the tangential plane; cS0 is the sensitivity of the sensor for γ = 45 [deg].

Fig. 5 .
Fig. 5. Dependence of the signal amplitude on the radial distance between the sensor (with SmCo magnet) and the measured object.

Fig. 6 .
Fig. 6.Signal amplitude in dependence on the used type of magnet.