Compensation method for LVDT sensor measurement accuracy based on key component position adjustment

Abstract

The application discloses a kind of LVDT sensor measurement precision compensation method based on key component position adjustment, the position of lever fulcrum is optimized by finite element analysis, the mechanical model containing temperature variable is constructed, the influence of the thermal expansion coefficient of sensor material at different working temperatures on the mechanical properties of lever is considered, to ensure that the lever is uniformly stressed within the full range, and maintain the optimal stress state within the full temperature range, reduce the core movement deviation and nonlinear error caused by lever structure and temperature change, improve the measurement accuracy of sensor. Iterative learning algorithm is used to compensate the measurement data, based on multiple measurement data, the residual is corrected in reverse, the compensation parameters are corrected in reverse, and the momentum term is introduced to accelerate the algorithm convergence, and the momentum coefficient is adaptively adjusted according to the dynamic response characteristics of the sensor. This dynamic adjustment compensation method can better adapt to the measurement error change under different working conditions, greatly improves the compensation effect.

Description

Technical Field

[0001] This invention belongs to the field of sensor accuracy compensation technology, specifically relating to a method for compensating the measurement accuracy of an LVDT sensor based on the adjustment of the position of key components. Background Technology

[0002] An LVDT (Linear Variable Differential Transformer) sensor is a high-precision sensor widely used in various precision measurement and position feedback systems. Its main working principle is based on electromagnetic induction, measuring the linear displacement change of an object. When the measured object undergoes a displacement change, the LVDT sensor can reflect this change in real time. An LVDT sensor consists of three main parts: a primary coil and two secondary coils, where the two secondary coils are symmetrical to each other. The primary coil generates the electromagnetic field, while the two secondary coils receive changes in the magnetic field generated by the primary coil. By measuring the voltage difference between the two secondary coils, the sensor can accurately detect the magnitude and direction of the displacement.

[0003] LVDT sensors also come in several other design forms, including pneumatic, integrated, and lever-type. Each structure offers different advantages and characteristics depending on the application requirements and measurement environment. The lever-type LVDT sensor cleverly utilizes the lever principle, enabling a large torque to be generated even with a small force applied to the input end. This torque amplification effect causes the lever to move, which in turn displaces the iron core on the other side. As the iron core moves, the mutual inductance between the coils of the LVDT sensor changes, ultimately leading to fluctuations in the output current, thus accurately measuring the displacement change. Therefore, the lever-type LVDT sensor has very high measurement accuracy, meeting the application requirements of precision instruments and equipment with extremely high displacement measurement demands.

[0004] However, existing lever-type LVDT sensor structures suffer from a significant problem: nonlinear errors due to their structural characteristics. This issue stems from the working principle of the lever structure itself. Although this principle can generate a large torque with a small input force, thus driving the iron core to move, this movement is not entirely linear but involves a certain degree of arc rotation. The iron core in the lever structure does not move smoothly along a straight line but undergoes a certain degree of rotation. This rotational motion causes a deviation between the iron core's trajectory and the ideal linear displacement. Due to this rotational nature, the magnetic field experienced by the iron core during its movement is not uniform, leading to a nonlinear relationship between the sensor's output signal and the actual displacement. This nonlinear error directly affects the sensor's measurement accuracy, causing errors in displacement measurement. This nonlinear error not only reduces the sensor's measurement accuracy but also makes it difficult for the sensor to meet the requirements of applications requiring high-precision feedback.

[0005] In terms of data processing, existing technologies that rely solely on fixed compensation coefficients cannot adapt to individual sensor differences and complex measurement environment variations, and therefore cannot effectively compensate for measurement errors. For example, the measurement error of sensors fluctuates significantly under different operating temperatures, and with increasing usage time, measurement accuracy gradually decreases due to wear of mechanical components and changes in electrical characteristics, failing to meet the requirements of high-precision displacement measurement. Furthermore, existing technologies lack effective mechanisms for handling the repair and replacement of critical sensor components. When critical components such as levers or coils are repaired or replaced, it is impossible to accurately assess changes in the overall sensor performance and adjust the compensation algorithm, further reducing sensor reliability and measurement accuracy. Summary of the Invention

[0006] To address the aforementioned problems in the prior art, this invention provides a method for compensating the measurement accuracy of an LVDT sensor based on the adjustment of the position of key components. The technical problem to be solved by this invention is achieved through the following technical solution: This invention provides a method for compensating the measurement accuracy of an LVDT sensor based on the adjustment of the position of key components, comprising: Based on the current operating temperature, the lever fulcrum in the lever-type LVDT sensor is dynamically fixed at the position after finite element analysis and optimization, and the pre-stored standard sensor model is loaded into the data processing unit in the lever-type LVDT sensor. Measurement data is obtained by sensing changes in the physical quantities of the target in real time using a lever-type LVDT sensor; The system monitors measurement data in real time. When the measurement data exceeds a preset error threshold, it performs multiple rounds of iterative compensation processing using an iterative learning-based compensation algorithm. The optimal compensation result is output when the iteration termination condition is met. In each round of iterative compensation processing, the compensation parameters are updated based on the compensation result of the previous round by calculating the reverse correction of the residual. Momentum adaptive adjustment is introduced in the iterative compensation processing to accelerate convergence. The compensation parameters for the first round of iterative compensation processing are given according to a standard sensor model. The termination condition includes: the residual of the current round of iterative compensation processing converges to a preset minimum value or the number of iterative compensation processes reaches the maximum number of iterations. After replacing or repairing key components in a lever-type LVDT sensor, the key parameters of the standard sensor model are corrected according to the technical specifications of the key components, so as to fine-tune the key parameters in the compensation algorithm based on iterative learning. When the usage time of the lever-type LVDT sensor reaches a preset time interval or the number of uses reaches a preset threshold, a standard value of the full range is applied to the lever-type LVDT sensor using a metrological standard source to update the standard sensor model. When the optimal compensation result shows abnormal fluctuations of a preset number of faults and is not significantly related to environmental factors, a fault tree analysis method is used to comprehensively test each hardware component of the lever-type LVDT sensor to ensure that the lever-type LVDT sensor works normally and stably.

[0007] In one embodiment of the present invention, the process of confirming the positions after finite element analysis and optimization corresponding to different operating temperatures includes: By constructing a mechanical model that includes temperature variables, the force variation law of the fulcrum position as temperature changes is described. At different operating temperatures, finite element analysis was used to obtain the force distribution of the lever and the optimal position of the fulcrum under different operating temperature conditions, so as to generate the temperature-fulcrum position mapping relationship, thereby confirming the position after finite element analysis and optimization corresponding to different operating temperatures.

[0008] In one embodiment of the present invention, the thermal stress formula in the mechanical model that includes temperature variables is as follows: ; in, This represents the thermal stress generated under temperature changes. This represents the elastic modulus of the lever material. This indicates the change in lever length caused by temperature changes. Indicates the original length of the lever. Indicates the coefficient of thermal expansion of a material. It represents the amount of temperature change.

[0009] In one embodiment of the present invention, the process of each round of iterative compensation processing includes: Calculate the residual for the current round based on the measurement data, the compensation parameters to be adjusted, and the ideal values ​​obtained based on the standard sensor model. Based on the residuals from the previous round, the sum of squares of the residuals is minimized using an optimization method to adjust and update the compensation parameters to be adjusted for the current round.

[0010] In one embodiment of the present invention, when the optimization method is gradient descent, the update formula for the compensation parameters is as follows: ; in, Indicates the first The compensation parameters corresponding to each round, Indicates the first The compensation parameters corresponding to each round, Indicates the learning rate. Indicates the first The gradient of the objective function corresponding to the compensation parameters of each round.

[0011] In one embodiment of the present invention, momentum adaptive adjustment is introduced in the iterative compensation process to accelerate convergence, including: In the iterative compensation process, the momentum coefficient is adaptively adjusted according to the dynamic response characteristics of the sensor. The momentum coefficient is increased during the rapid change phase of the sensor response to accelerate the iteration speed, and decreased when the sensor reaches steady-state measurement to avoid overshoot.

[0012] In one embodiment of the present invention, the momentum coefficient is expressed as follows: ; in, Indicates the first The momentum coefficient of the round. This represents the initial momentum coefficient. Indicates the first The change in sensor response in each round, This represents the maximum value of the sensor's response change. This represents the empirical constant obtained from the fitting.

[0013] In one embodiment of the present invention, the parameters of the standard sensor model include mechanical parameters and electrical parameters; The expressions for the mechanical parameters used to correct the mechanical parameters are as follows: ; in, This indicates the force that the modified lever can withstand. This represents the corrected elastic modulus. This represents the corrected moment of inertia. express, Indicates lever displacement; The expression used to correct electrical parameters is as follows: ; in, This represents the corrected output signal of the sensor. This indicates the inductance of the coil in the corrected sensor. express The current flowing through the coil in the sensor after time correction. This indicates the resistance of the coil in the corrected sensor. express The current flowing through the coil in the sensor after time correction.

[0014] In one embodiment of the present invention, a comprehensive inspection of all hardware components of the lever-type LVDT sensor is performed using a fault tree analysis method, including: Identify fault events based on abnormal fluctuations; Based on different failure modes and the working principles of the components of the lever-type LVDT sensor system, the failure event is decomposed into multiple sub-events. The root cause of the failure is determined by evaluating the probability of each failure mode one by one.

[0015] In one embodiment of the present invention, the data processing unit uses a triple redundant acquisition channel to acquire sensor electrical signals. The triple redundant acquisition channel includes a main channel, a slave channel, and a backup channel. The three channels work independently and compare the acquired data in real time. When the difference between the acquired data corresponding to the main channel and the slave channel exceeds a preset fault tolerance range, the system automatically switches to the backup channel for data acquisition, marks the fault, starts the channel self-test program, and promptly locates and repairs the fault.

[0016] The beneficial effects of this invention are: The solution provided in this invention optimizes the lever fulcrum position through finite element analysis, constructs a mechanical model incorporating temperature variables, and considers the influence of the material's thermal expansion coefficient on the lever's mechanical properties at different operating temperatures. This ensures uniform force on the lever across the entire measurement range and maintains optimal force conditions across the entire temperature range, reducing core movement deviations and nonlinear errors caused by lever structure and temperature variations, fundamentally improving the sensor's measurement accuracy. An iterative learning algorithm is used to compensate for the measurement data. Based on multiple rounds of measurement data, the compensation parameters are corrected in reverse according to the residuals, and a momentum term is introduced to accelerate algorithm convergence. The momentum coefficient is adaptively adjusted based on the sensor's dynamic response characteristics. This dynamically adjusted compensation method better adapts to changes in measurement errors under different operating conditions, significantly improving the compensation effect compared to the fixed compensation coefficients of existing technologies. Attached Figure Description

[0017] Figure 1 This is a schematic diagram illustrating the steps of a method for compensating the measurement accuracy of an LVDT sensor based on the adjustment of the position of key components, provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the iterative compensation process in an LVDT sensor measurement accuracy compensation method based on key component position adjustment provided in an embodiment of the present invention. Figure 3 This is a flowchart illustrating the error analysis process in an LVDT sensor measurement accuracy compensation method based on key component position adjustment, provided in an embodiment of the present invention. Detailed Implementation

[0018] The present invention will be further described in detail below with reference to specific embodiments, but the implementation of the present invention is not limited thereto.

[0019] Most existing technologies use fixed compensation coefficients to handle measurement errors, which cannot be dynamically adjusted according to the individual characteristics of the sensor and the complex and ever-changing measurement environment. This makes it difficult to effectively cope with error changes under different operating conditions. For example, in environments with frequent mechanical vibration or strong electromagnetic interference, the measurement data fluctuates greatly, and fixed compensation cannot accurately correct it, resulting in persistent measurement deviations and affecting the accuracy of equipment control and monitoring.

[0020] When critical components of a sensor are repaired or replaced, existing technologies cannot reassess the sensor's performance and adapt the compensation algorithm based on the new component's specifications. For example, after replacing a lever or coil, the measurement errors caused by changes in electrical and mechanical properties cannot be compensated for because the model parameters have not been corrected. This reduces the long-term reliability and stability of the sensor, and increases equipment maintenance costs and downtime risks.

[0021] To address the issues of poor error compensation and lack of personalized processing mechanisms, this invention provides a method for compensating the measurement accuracy of an LVDT sensor based on the adjustment of key component positions, such as... Figure 1 As shown, it may include: S1, dynamically fix the lever fulcrum in the lever-type LVDT sensor to the position after finite element analysis and optimization according to the current operating temperature, and load the pre-stored standard sensor model into the data processing unit in the lever-type LVDT sensor.

[0022] The process of confirming the positions corresponding to different operating temperatures after finite element analysis and optimization may include: By constructing a mechanical model that includes temperature variables, the force variation law of the fulcrum position as temperature changes is described. At different operating temperatures, finite element analysis was used to obtain the force distribution of the lever and the optimal position of the fulcrum under different operating temperature conditions, so as to generate the temperature-fulcrum position mapping relationship, thereby confirming the position after finite element analysis and optimization corresponding to different operating temperatures.

[0023] In the process of optimizing the lever fulcrum position through finite element analysis, this invention considers the influence of the material's thermal expansion coefficient on the lever's mechanical performance at different operating temperatures. A mechanical model incorporating temperature variables is constructed to ensure that the fulcrum position maintains the optimal force state across the entire temperature range. Furthermore, the optimized fulcrum coordinates at different temperatures are recorded and stored in the system configuration file for real-time recall.

[0024] Specifically, when optimizing the lever fulcrum position, the influence of temperature on the mechanical properties of the LVDT sensor needs to be fully considered, especially the thermal expansion effect of the material, which directly relates to the force state of the lever at different operating temperatures. To ensure the uniformity of force on the lever across the entire temperature range and the measurement accuracy of the sensor, it is first necessary to construct a mechanical model that includes temperature variables to describe the force variation law of the fulcrum position with temperature changes.

[0025] In the mechanical model, considering the influence of the material's coefficient of thermal expansion on the lever deformation, it can be described by the corresponding thermal stress formula. The thermal stress formula in the mechanical model that includes temperature variables is as follows: ; in, This represents the thermal stress generated under temperature changes. This represents the elastic modulus of the lever material. This indicates the change in lever length caused by temperature changes. Indicates the original length of the lever. Indicates the coefficient of thermal expansion of a material. It represents the amount of temperature change.

[0026] This thermal stress formula describes the effect of temperature changes on the lever length, thus affecting the lever's stress state. In the model, thermal stress directly affects the lever's deformation and the optimal position of the fulcrum.

[0027] Based on this mechanical model, the force distribution of the lever and the optimal position of the fulcrum under different operating temperatures can be obtained through numerical simulation analysis. This process is performed using finite element analysis (FEA) to generate a set of temperature-fulcrum position mapping relationships. Whenever the operating temperature changes, the system can automatically call the optimal fulcrum position coordinates stored in the system configuration file based on the real-time temperature and previously optimized fulcrum position data, thereby ensuring that the force on the lever remains optimal at all temperatures and avoiding measurement errors caused by thermal effects. This can significantly improve the measurement accuracy of the lever-type LVDT sensor under temperature variations. The key purpose of this model and optimization algorithm is to enable the sensor to operate stably even in harsh temperature environments by incorporating the consideration of temperature effects, thereby reducing temperature-induced errors.

[0028] S2 obtains measurement data by sensing changes in the physical quantities of the target in real time through a lever-type LVDT sensor.

[0029] Specifically, the measurement process begins by using a lever-type LVDT sensor to detect changes in the target physical quantity in real time and convert them into electrical signals. The data processing unit simultaneously acquires these electrical signals and, based on a preset error threshold, determines in real time whether the measurement data exceeds the allowable accuracy range. If it does not exceed the range, the corresponding measurement data is output; otherwise, the subsequent error compensation process is triggered. The preset error threshold can be set according to the sensor's nominal accuracy and actual application requirements.

[0030] S3 monitors measurement data in real time. When the measurement data exceeds the preset error threshold, it performs multiple rounds of iterative compensation processing on the measurement data through a compensation algorithm based on iterative learning. When the conditions for ending the iteration are met, it outputs the optimal compensation result.

[0031] In each round of iterative compensation processing, the compensation parameters are updated based on the compensation results of the previous round by calculating the reverse correction of the residuals. Momentum adaptive adjustment is introduced in the iterative compensation processing to accelerate convergence. The compensation parameters of the first round of iterative compensation processing are given according to the standard sensor model. The conditions for the end of the iteration include: the residual of the current round of iterative compensation processing converges to the preset minimum value or the number of iterative compensation processing reaches the maximum number of iterations.

[0032] Specifically, once the error compensation process is initiated, an iterative learning-based compensation algorithm is used to perform multiple rounds of iterative compensation processing on the measurement data. This algorithm is based on the measurement data from each round, and in each iteration, the compensation parameters are corrected in reverse based on the residual from the previous round, gradually approaching the optimal compensation effect. In essence, the core idea of ​​this compensation algorithm is to continuously correct the compensation parameters through multiple rounds of measurement data and residual feedback, gradually approaching the optimal compensation effect. Each iteration is based on the compensation result of the previous round, updating the compensation parameters by calculating the inverse correction of the residual. The initial values ​​of the compensation parameters are given according to the standard sensor model, ensuring a relatively accurate starting point in the early stages of the iteration.

[0033] The process of each round of iterative compensation processing, such as Figure 2 As shown, it may include: Calculate the residual for the current round based on the measurement data, the compensation parameters to be adjusted, and the ideal values ​​obtained based on the standard sensor model. Based on the residuals from the previous round, the sum of squares of the residuals is minimized using an optimization method to adjust and update the compensation parameters to be adjusted for the current round.

[0034] Specifically, in each round of iterative compensation processing, it is assumed that the actual measurement value of the sensor is... The ideal value calculated based on the standard sensor model is ,in, This indicates the measured voltage value corresponding to the actual measured value. This indicates the compensation parameters to be adjusted.

[0035] The expression for the calculated residual is as follows: ; The residual reflects the deviation between the actual measured value and the ideal value. Therefore, embodiments of the present invention adjust the compensation parameter based on this residual. The compensation parameter is corrected by minimizing the sum of squares of the residuals, and the following objective function can be used: ; Wherein, objective function This represents the total error function for the entire dataset corresponding to the measured data. This indicates the number of measurement data points. To obtain the optimal compensation parameters, it is necessary to minimize them using optimization methods. The specific optimization method can be gradient descent or other optimization algorithms.

[0036] When the optimization method is gradient descent, the update formula for the compensation parameters is as follows: ; in, Indicates the first The compensation parameters corresponding to each round, Indicates the first The compensation parameters corresponding to each round, Indicates the learning rate. Indicates the first The gradient of the objective function corresponding to the compensation parameters of each round.

[0037] Understandably, the iterative compensation process will continue until the preset conditions for ending the iteration are met. There are two main conditions for ending the iteration: first, when the residual of the iteration compensation process converges to the preset minimum value, it indicates that the compensation has reached the predetermined accuracy; second, when the number of iterations reaches the maximum number of iterations, it indicates that the compensation process has been fully performed and the accuracy cannot be further improved through more iterations.

[0038] Introducing adaptive momentum adjustment into iterative compensation processing to accelerate convergence can include: In the iterative compensation process, the momentum coefficient is adaptively adjusted according to the dynamic response characteristics of the sensor. The momentum coefficient is increased during the rapid change phase of the sensor response to accelerate the iteration speed, and decreased when the sensor reaches steady-state measurement to avoid overshoot.

[0039] Understandably, the embodiments of the present invention introduce momentum adaptive adjustment in the iterative compensation process. The momentum coefficient is adaptively adjusted according to the dynamic response characteristics of the sensor. The momentum coefficient is increased during the rapid change phase of the sensor response to accelerate the iteration speed, and the momentum coefficient is decreased when the steady-state measurement is reached to avoid overshoot. The adjustment rule of the momentum coefficient is obtained by fitting a large amount of simulation and experimental data to obtain an empirical formula and storing it in the algorithm module.

[0040] Specifically, the momentum coefficient can generally be viewed as the influence of the error correction from the previous iteration on the current iteration result, thereby accelerating the convergence speed. The momentum coefficient is adaptively adjusted based on the sensor's dynamic response characteristics. During periods of rapid sensor response change, increasing the momentum coefficient allows the algorithm to make larger corrections to the error in each iteration, thus accelerating convergence. However, upon reaching steady-state measurement, to avoid overshoot due to excessive adjustment, the momentum coefficient needs to be decreased, allowing the algorithm to converge more smoothly. The adaptive adjustment rule for the momentum coefficient can be obtained through empirical formulas derived from fitting a large amount of simulation and experimental data. The expression for the momentum coefficient is as follows: ; in, Indicates the first The momentum coefficient of the round. This represents the initial momentum coefficient. Indicates the first The change in sensor response in each round, This represents the maximum value of the sensor's response change. This represents the empirical constant obtained from the fitting, used to adjust the growth rate of the momentum coefficient as a function of the response.

[0041] Using the above formula, the momentum coefficient can be adjusted in real time according to the sensor's current dynamic response. When the response is fast, A larger momentum coefficient accelerates the iteration process; however, upon reaching a steady state... As the momentum coefficient decreases, it avoids overshoot and unnecessary oscillations. This adaptive momentum coefficient adjustment mechanism, combined with the actual dynamic characteristics of the sensor, can effectively improve the convergence speed and accuracy of the iterative learning-based compensation algorithm. By continuously adjusting the momentum coefficient, the algorithm can achieve rapid convergence during periods of rapid change, while maintaining meticulous adjustments as it approaches stability, thus ensuring that the final error is minimized while avoiding overcompensation.

[0042] It is understood that the embodiments of the present invention employ multi-round measurement data and a compensation algorithm based on iterative learning. By introducing a momentum term to adaptively adjust the momentum coefficient, the compensation parameters can be dynamically corrected, effectively compensating for measurement errors. Compared with the existing technology of fixing the compensation coefficient, this greatly improves the compensation effect and accuracy, playing a core role in improving measurement accuracy.

[0043] S4. After replacing or repairing key components in the lever-type LVDT sensor, the key parameters of the standard sensor model are corrected according to the technical specifications of the key components, so as to fine-tune the key parameters in the compensation algorithm based on iterative learning.

[0044] This invention identifies the serial number of a single lever-type LVDT sensor connected to the system. Based on the serial number, it retrieves the sensor's historical maintenance records and special process parameters from the database. Combining this personalized information, it fine-tunes key parameters such as step size and convergence criteria in the iterative learning algorithm, achieving precise compensation for individual sensors. The database is updated in real-time with the sensor production and after-sales maintenance systems. For repair and replacement information in the sensor's historical maintenance records, if critical sensitive components such as levers or coils are involved, the system automatically retrieves the detailed technical specifications and calibration data of the corresponding components, reassesses the overall sensor performance, and corrects relevant electrical and mechanical parameters in the standard sensor model to adapt to the characteristics of the repaired sensor, ensuring the continued effectiveness of the compensation algorithm.

[0045] Specifically, when a critical component is replaced or repaired, the electrical and mechanical parameters in the standard sensor model must first be updated using the technical specifications of the repaired component. If the elastic modulus of the lever changes, or the resistance of the coil differs due to replacement, the impact of these changes on the overall sensor performance needs to be recalculated, thereby correcting the parameters in the sensor model.

[0046] The mechanical response of the lever can be adjusted using the following formula to correct the mechanical parameters: ; in, This indicates the force that the modified lever can withstand. This represents the corrected elastic modulus. This represents the corrected moment of inertia. express, This represents the lever displacement. Using this formula, the new mechanical response is calculated in the replaced or repaired lever components, thereby correcting the sensor's mechanical performance model.

[0047] When correcting electrical parameters, especially when critical components of the sensor (such as coils) are replaced or repaired, it is necessary to recalculate the impact of electrical characteristics on the sensor output. The main electrical parameters to consider include resistance R and inductance L, which directly affect the output signal of the LVDT sensor.

[0048] Assuming the sensor output signal Based on the electrical characteristics of the coil, such as resistance and inductance, the correction of these electrical characteristics can be expressed by the following formula: ; in, This represents the corrected output signal of the sensor. This indicates the inductance of the coil in the corrected sensor. express The current flowing through the coil in the sensor after time correction. This indicates the resistance of the coil in the corrected sensor. express The current flowing through the coil in the sensor after time correction. In this formula, This is the response to the updated current. To maintain the accuracy of the output signal, it is adjusted according to the new electrical parameters. and The compensation parameters are then recalculated to ensure that the measurement results are not affected by changes in electrical characteristics.

[0049] The purpose of the above correction process is to ensure that even after repairing or replacing key components of the sensor, the compensation algorithm can still accurately compensate for measurement errors and maintain the high precision performance of the LVDT sensor.

[0050] This invention retrieves historical maintenance records and special process parameters based on the sensor serial number, fine-tunes key parameters of the iterative learning algorithm, and re-evaluates and corrects sensor model parameters after the repair and replacement of key components, ensuring the continued effectiveness of the compensation algorithm. This solves the problem that existing technologies cannot cope with individual differences in sensors and performance changes after repair, and is crucial for maintaining the high precision performance of sensors.

[0051] S5. When the usage time of the lever-type LVDT sensor reaches the preset time interval or the number of uses reaches the preset threshold, the standard value of the full range is applied to the lever-type LVDT sensor using a metrological standard source to update the standard sensor model. When the optimal compensation result shows abnormal fluctuations of the preset number of faults and is not significantly related to environmental factors, the fault tree analysis method is used to comprehensively test each hardware component of the lever-type LVDT sensor to ensure that the lever-type LVDT sensor works normally and stably.

[0052] The preset time interval and number of uses can be set based on sensor reliability test data and field operation experience.

[0053] Specifically, during step S5, in addition to updating the standard sensor model parameters, information such as environmental parameters and stability indicators of the calibration signal source are recorded simultaneously during the calibration process as a basis for subsequent data analysis. When abnormal fluctuations of the optimal compensation result appear for a preset number of faults and are not significantly related to environmental factors, a deep fault diagnosis process is initiated to conduct a comprehensive inspection of the sensor hardware and peripheral circuits. The fault diagnosis process is constructed based on the fault tree analysis method and covers common hardware fault modes and troubleshooting order.

[0054] When the deep fault diagnosis process is initiated, the system relies on Fault Tree Analysis (FTA) to conduct fault investigation, comprehensively checking the sensor hardware and peripheral circuits. The core idea of ​​FTA is to start from the final fault event and trace upwards step by step to determine the cause of the fault and possible fault modes. This method, by constructing a fault tree, ensures that various hardware faults and circuit problems can be systematically and methodically investigated.

[0055] A comprehensive inspection of all hardware components of the lever-type LVDT sensor is performed using fault tree analysis, which may include: Identify fault events based on abnormal fluctuations; Based on different failure modes and the working principles of the components of the lever-type LVDT sensor system, failure events are classified. The failure is broken down into multiple sub-events, and the root cause of the failure is determined by evaluating the probability of each failure mode one by one.

[0056] Understandably, the basic workflow of fault tree analysis is as follows: Identifying fault events based on abnormal fluctuations Then, based on different failure modes The working principles of the system components, and fault events Break it down into multiple sub-events until the root cause can be found. Let's define a hardware failure as... The probability of a basic event occurring in the fault tree can then be expressed by the following formula: ; in, Indicates a fault event The probability of occurrence express The probability of occurrence This represents the number of all possible sub-events. By evaluating the probability of each failure mode one by one, it is possible to identify the faulty component or link in the system that is most likely to cause abnormal fluctuations.

[0057] The in-depth fault diagnosis process includes the following key steps: First, a comprehensive inspection of the sensor hardware is conducted, including the sensor's resistance, the state of the induction coil, and the mechanical condition of the lever. Second, the peripheral circuits connected to the sensor are inspected, including components such as signal amplifiers, data acquisition units, and power supplies. Finally, fault tree analysis is used to deduce possible fault modes, gradually eliminating non-fault factors, and ultimately pinpointing the possible hardware or circuit problem.

[0058] The purpose of this in-depth fault diagnosis process is to ensure the stable operation of the sensor's hardware and circuitry during long-term operation and to promptly identify potential fault risks. This embodiment of the invention automatically initiates a comprehensive calibration procedure according to predetermined rules, updates the standard sensor model, and records environmental information for fault diagnosis. Compared to existing technologies that lack an effective calibration update mechanism, this approach ensures long-term stable and high-precision sensor measurements and allows for the timely detection and resolution of potential problems.

[0059] The data processing unit uses a triple-redundant acquisition channel to acquire sensor electrical signals. The triple-redundant acquisition channel includes a main channel, a slave channel, and a backup channel. A flowchart illustrating the error analysis process in the compensation method is shown below. Figure 3 As shown, the three channels work independently and compare the collected data in real time. When the difference between the collected data of the main channel and the corresponding data of the slave channel exceeds the preset fault tolerance range, the system automatically switches to the backup channel for data collection, marks the fault, starts the channel self-check program, and promptly locates and repairs the fault.

[0060] In this embodiment of the invention, while marking the faults in both the master and slave channels, a channel self-test program is initiated. By sending specific test signals to the sensors, the program checks key components such as the channel signal conditioning circuit and the analog-to-digital conversion module, thereby promptly locating and repairing the faults.

[0061] Specifically, during the data acquisition process, the data output of the master and slave channels can be represented as signals. and , representing the output signals of the main channel and the slave channel, respectively. To detect the signal difference between the two channels, the system calculates the difference between them. : ; when Exceeding the preset fault tolerance range When the system detects a potential failure in the master-slave dual-channel configuration, it automatically switches to the backup channel and records the timestamp and type of the master-slave channel failure.

[0062] After switching to the backup channel, the system initiates a channel self-test program. This program checks the functionality of key components such as the master / slave channel signal conditioning circuits and analog-to-digital conversion modules by sending specific test signals to the sensors. Assume the system sends the following test signals... The system requires that the response of the test signal should meet the expected standard, and it will compare the actual measured signal with the test signal. Compared with standard response signal The difference in the test signal is used to locate the fault. The error of the test signal can be expressed by the following formula: ; If error Greater than the preset threshold If this is the case, it indicates that there is a problem with some key components in the channel, and the system will further diagnose whether components such as the analog-to-digital converter and signal conditioning circuit are faulty.

[0063] The LVDT sensor measurement accuracy compensation method provided in this invention optimizes the lever fulcrum position through finite element analysis, constructs a mechanical model incorporating temperature variables, and considers the influence of the material's thermal expansion coefficient on the lever's mechanical properties at different operating temperatures. This ensures uniform force on the lever across the entire measurement range and maintains optimal force conditions across the entire temperature range, reducing core movement deviations and nonlinear errors caused by lever structure and temperature variations, fundamentally improving the sensor's measurement accuracy. An iterative learning algorithm is used to compensate the measurement data. Based on multiple rounds of measurement data, the compensation parameters are corrected in reverse according to the residuals, and a momentum term is introduced to accelerate algorithm convergence. The momentum coefficient is adaptively adjusted based on the sensor's dynamic response characteristics. This dynamically adjusted compensation method better adapts to measurement error changes under different operating conditions, significantly improving the compensation effect compared to the fixed compensation coefficients of existing technologies.

[0064] It should be noted that, in the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0065] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A method for compensating the measurement accuracy of an LVDT sensor based on the adjustment of the position of key components, characterized in that, include: Based on the current operating temperature, the lever fulcrum in the lever-type LVDT sensor is dynamically fixed at the position after finite element analysis and optimization, and the pre-stored standard sensor model is loaded into the data processing unit in the lever-type LVDT sensor. Measurement data is obtained by sensing changes in the physical quantities of the target in real time using a lever-type LVDT sensor; The system monitors measurement data in real time. When the measurement data exceeds the preset error threshold, it performs multiple rounds of iterative compensation processing on the measurement data through a compensation algorithm based on iterative learning. When the conditions for ending the iteration are met, the system outputs the optimal compensation result. In each round of iterative compensation processing, the compensation parameters are updated based on the compensation results of the previous round by calculating the reverse correction of the residuals. In the iterative compensation process, an adaptive momentum adjustment is introduced to accelerate convergence; The compensation parameters for the first round of iterative compensation processing are given according to the standard sensor model; the conditions for the end of the iteration include: the residual of the current round of iterative compensation processing converges to a preset minimum value or the number of iterative compensation processing reaches the maximum number of iterations; After replacing or repairing key components in a lever-type LVDT sensor, the key parameters of the standard sensor model are corrected according to the technical specifications of the key components, so as to fine-tune the key parameters in the compensation algorithm based on iterative learning. When the usage time of the lever-type LVDT sensor reaches a preset time interval or the number of uses reaches a preset threshold, a standard value of the full range is applied to the lever-type LVDT sensor using a metrological standard source to update the standard sensor model. When the optimal compensation result shows abnormal fluctuations of a preset number of faults and is not significantly related to environmental factors, a fault tree analysis method is used to comprehensively test each hardware component of the lever-type LVDT sensor to ensure that the lever-type LVDT sensor works normally and stably.

2. The method for compensating the measurement accuracy of an LVDT sensor based on the adjustment of the position of key components according to claim 1, characterized in that, The process of confirming the positions corresponding to different operating temperatures after finite element analysis and optimization includes: By constructing a mechanical model that includes temperature variables, the force variation law of the fulcrum position as temperature changes is described. At different operating temperatures, finite element analysis was used to obtain the force distribution of the lever and the optimal position of the fulcrum under different operating temperature conditions, so as to generate the temperature-fulcrum position mapping relationship, thereby confirming the position after finite element analysis and optimization corresponding to different operating temperatures.

3. The method for compensating the measurement accuracy of an LVDT sensor based on the adjustment of the position of key components according to claim 2, characterized in that, The thermal stress formula in the mechanical model that includes temperature variables is as follows: ； in, This represents the thermal stress generated under temperature changes. This represents the elastic modulus of the lever material. This indicates the change in lever length caused by temperature changes. Indicates the original length of the lever. Indicates the coefficient of thermal expansion of a material. It represents the amount of temperature change.

4. The method for compensating the measurement accuracy of an LVDT sensor based on the adjustment of the position of key components according to claim 1, characterized in that, The process of each round of iterative compensation includes: Calculate the residual for the current round based on the measurement data, the compensation parameters to be adjusted, and the ideal values ​​obtained based on the standard sensor model. Based on the residuals from the previous round, the sum of squares of the residuals is minimized using an optimization method to adjust and update the compensation parameters to be adjusted for the current round.

5. The method for compensating the measurement accuracy of an LVDT sensor based on the adjustment of the position of key components according to claim 4, characterized in that, When the optimization method is gradient descent, the update formula for the compensation parameters is as follows: ； in, Indicates the first The compensation parameters corresponding to each round, Indicates the first The compensation parameters corresponding to each round, Indicates the learning rate. Indicates the first The gradient of the objective function corresponding to the compensation parameters of each round.

6. The method for compensating the measurement accuracy of an LVDT sensor based on the adjustment of the position of key components according to claim 1, characterized in that, The method of introducing adaptive momentum adjustment in the iterative compensation process to accelerate convergence includes: In the iterative compensation process, the momentum coefficient is adaptively adjusted according to the dynamic response characteristics of the sensor. The momentum coefficient is increased during the rapid change phase of the sensor response to accelerate the iteration speed, and decreased when the sensor reaches steady-state measurement to avoid overshoot.

7. The method for compensating the measurement accuracy of an LVDT sensor based on the adjustment of the position of key components according to claim 6, characterized in that, The expression for the momentum coefficient is as follows: ； in, Indicates the first The momentum coefficient of each round, This represents the initial momentum coefficient. Indicates the first The change in sensor response in each round, This represents the maximum value of the sensor's response change. This represents the empirical constant obtained from the fitting.

8. The method for compensating the measurement accuracy of an LVDT sensor based on the adjustment of the position of key components according to claim 1, characterized in that, The parameters of the standard sensor model include mechanical parameters and electrical parameters; The expressions for the mechanical parameters used to correct the mechanical parameters are as follows: ； in, This indicates the force that the modified lever can withstand. This represents the corrected elastic modulus. This represents the corrected moment of inertia. express, Indicates lever displacement; The expression used to correct electrical parameters is as follows: ； in, This represents the corrected output signal of the sensor. This indicates the inductance of the coil in the corrected sensor. express The current flowing through the coil in the sensor after time correction. This indicates the resistance of the coil in the corrected sensor. express The current flowing through the coil in the sensor after time correction.

9. The method for compensating the measurement accuracy of an LVDT sensor based on the adjustment of the position of key components according to claim 1, characterized in that, The method of fault tree analysis is used to comprehensively test all hardware components of the lever-type LVDT sensor, including: Identify fault events based on abnormal fluctuations; Based on different failure modes and the working principles of the components of the lever-type LVDT sensor system, the failure event is decomposed into multiple sub-events. The root cause of the failure is determined by evaluating the probability of each failure mode one by one.

10. The method for compensating the measurement accuracy of an LVDT sensor based on the adjustment of the position of key components according to claim 9, characterized in that, The data processing unit uses a triple redundant acquisition channel to acquire sensor electrical signals. The triple redundant acquisition channel includes a main channel, a slave channel, and a backup channel. The three channels work independently and compare the acquired data in real time. When the difference between the acquired data of the main channel and the slave channel exceeds the preset fault tolerance range, the system automatically switches to the backup channel for data acquisition, marks the fault, starts the channel self-test program, and promptly locates and repairs the fault.

Images