System and method for measuring opening and closing states of electromagnetic valve

Abstract

The invention discloses a system and method for measuring the opening and closing state of an electromagnetic valve, and relates to the technical field of electromagnetic valves.The method comprises the steps that after an electromagnetic valve control signal is triggered, dynamic inductance L at each moment t is obtained through real-time backstepping by means of a voltage and current sampling module integrated in a driving loop and an inductance analysis module in a local microcontroller MCU; gaussian filtering processing and derivative calculation are carried out on the dynamic inductance L in the edge calculation module, first-order derivative and second-order derivative characteristics are extracted, and an inductance mutation factor index Cimp is obtained through calculation. According to the method, clamping stagnation, rebound and discontinuous abnormal states occurring in the opening and closing process of the electromagnetic valve are comprehensively reflected, a sudden change threshold value Cthr is set and compared, and a set of dynamic opening and closing behavior diagnosis mechanism with high resolution and high adaptability is constructed. According to the method, high-frequency dynamic response extraction and abnormal trend quantification can be completed within 100 ms after the control signal is triggered, and the method has higher early warning capability and non-intrusive compatibility.

Description

Technical Field

[0001] The present invention relates to the technical field of solenoid valves, and in particular to a system and method for measuring the opening and closing states of solenoid valves. Background Art

[0002] With the increasing intelligence of rail transit equipment and the increasing automation of train control systems, solenoid valves, as key actuators for controlling the opening and closing of pneumatic circuits, require real-time monitoring and health diagnosis of their operating status, becoming a key component in ensuring vehicle safety and operational reliability. In particular, in rail train air compressor systems, solenoid valves frequently respond to commands such as braking, suspension, and door control. The accuracy and speed of their opening and closing operations directly impact vehicle handling performance. However, traditional solenoid valve opening and closing states are difficult to identify in a non-invasive, high-resolution manner. Therefore, a method for measuring the opening and closing states of solenoid valves that can online acquire the actual motion trajectory of the valve core and structural disturbance signals is urgently needed.

[0003] Currently, solenoid valves are used in rail transit train braking systems to control the release or retention of compressed air in air brake systems. If the solenoid valve exhibits issues such as "spool sticking," "incomplete opening and closing," or "hysteretic response," it can seriously affect the braking system's response efficiency and even cause safety accidents. Currently, most solenoid valve status measurement methods rely solely on indirect judgments based on the control signal's logic state, current waveform characteristics, or port pressure feedback. These methods suffer from low resolution, delayed response, and an inability to characterize intermediate processes. For example, while coil current fluctuations can partially reflect drive changes, they cannot reveal microstructural issues such as internal valve spool sticking, rebound, or discontinuous opening and closing. Furthermore, pressure response often lags behind structural movement and is significantly affected by load disturbances, making it difficult to provide a reliable indicator for evaluating opening and closing behavior. This often leads to overlooking or misdiagnosing potential hazards during the initial stages of system operation or during periods of minor anomalies, reducing the accuracy and timeliness of fault warnings.

[0004] The above defects are mainly due to the potential nonlinear mechanical behavior in the opening and closing process of the solenoid valve, including friction resistance between the valve core and the guide, deviation of the return spring force, and stroke abnormalities caused by insufficient driving magnetic field. Once phenomena such as valve core jamming, rebounding, or adhesion without release occur, although the opening and closing signals have been output in the control logic, the actual valve body has not completed the expected action, which will lead to air path switching failure, abnormal system response, and even malfunction or multiple restarts. Especially in train braking or critical action scenarios, if such anomalies are not detected and suppressed in time, it may cause serious failure of train control system stability and potential safety accidents. Summary of the Invention

[0005] In view of the deficiencies in the prior art, the present invention provides a system and method for measuring the opening and closing status of a solenoid valve, which solves the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: comprising the following steps: S1. After the solenoid valve control signal is triggered, the dynamic inductance L is collected, and the product of the second-order derivative and the first-order derivative of the dynamic inductance L is calculated to obtain the inductance mutation factor index Cimp; S2. Compare the inductance mutation factor index Cimp with the preset mutation threshold Cthr. Based on the comparison result, trigger the disturbance analysis mechanism to collect the instantaneous acoustic signal S during the solenoid valve operation. Perform Fourier transform on the instantaneous acoustic signal S, extract the target frequency band set Bres, and calculate the spectrum disturbance deviation value Fvar. S3. Input the inductance mutation factor index Cimp and the spectrum disturbance deviation value Fvar into the comprehensive opening and closing state scoring model to calculate the comprehensive diagnostic score Qdiag. Compare the comprehensive diagnostic score Qdiag with the preset diagnostic level range to determine the opening and closing state of the solenoid valve, issue an alarm based on the comparison result, and limit the execution of subsequent control instructions.

[0007] Preferably, said S1 includes S11; S11. By integrating a sampling circuit module into the solenoid valve control drive circuit of the rail transit train air compressor system and setting a collection strategy, the voltage signal and current signal at both ends of the solenoid valve coil are collected in real time; The voltage and current signals are transmitted to an inductance analysis module that is synchronously input into a local microcontroller MCU. The inductance analysis module uses the voltage and current signals combined with the known PWM duty cycle to reversely calculate the dynamic inductance L at each time t. The acquisition strategy is set to have a sampling frequency of no less than 5kHz, and a sampling time window covering the interval from 0ms to 100ms after the control signal is triggered; The circuit module is arranged on a PWM control signal path in a solenoid valve control drive circuit.

[0008] Preferably, said S1 further includes S12; S12. The dynamic inductance L at each moment t is integrated into an inductance sequence in chronological order, and the inductance sequence is uploaded to the edge computing module embedded in the control host for analysis and processing. The edge computing module performs a filtering preprocessing on all the dynamic inductances L in the inductance sequence through a set denoising processor. The filtering preprocessing smoothes all the dynamic inductances L in the inductance sequence by using a Gaussian convolution kernel to weaken the derivative deviation caused by high-frequency electromagnetic interference and mechanical jitter. The smoothed inductance sequence is then passed to the derivative calculation module for feature extraction to obtain the first-order derivative and second-order derivative of the dynamic inductance L.

[0009] Preferably, said S1 and S13; S13. Based on the first-order derivative and second-order derivative of the dynamic inductance L, the mutation response factor is calculated for any time t within the entire sampling time window, and then the maximum value of the mutation response factor at all times t is extracted to obtain the inductance mutation factor index Cimp, which is used to measure the maximum nonlinear disturbance intensity reflected by the solenoid valve during the opening and closing process.

[0010] Preferably, said S2 includes S21; S21. Select N groups of opening and closing inductance behavior samples under the same temperature, pressure, and driving conditions, where the N groups of opening and closing inductance behavior samples are the inductance mutation factor indicators Cimp of multiple groups of solenoid valves in a historical healthy state, and then set the 95th percentile of the inductance mutation factor indicators Cimp of the multiple groups of solenoid valves as the mutation threshold Cthr; The inductance mutation factor index Cimp obtained in real time is compared with the mutation threshold Cthr to determine the inductance disturbance during the opening and closing process of the solenoid valve; The specific comparison contents are as follows; If the inductance mutation factor index Cimp in the current opening and closing cycle exceeds the mutation threshold Cthr, it is judged that the current opening and closing state is abnormal, and the subsequent disturbance analysis mechanism is automatically triggered, and the time of the abnormal event is recorded. If the inductance mutation factor index Cimp in the current opening and closing cycle is less than or equal to Cthr, the current opening and closing state is judged to be normal, and it is automatically recorded as a healthy opening and closing cycle, and the disturbance analysis mechanism is skipped, and the next control cycle is continued.

[0011] Preferably, the S2 further includes S22; S22. After the disturbance analysis mechanism is triggered, an acoustic signal S is collected by a MEMS micro-microphone module disposed outside the solenoid valve housing. The microphone module collects sound pressure at a sampling frequency of 20 kHz within the first 50 ms after the solenoid valve drive signal is triggered, obtaining the acoustic signal S at each time instant t. After the acquisition is completed, the local control module performs short-time Fourier transform analysis on the acoustic signal S at each time t, and combines the preset target frequency range judgment rules to screen out abnormal frequency bands from the complete spectrum to form a target frequency band set Bres. The target frequency range includes two sub-intervals: 800Hz–1.2kHz and 2kHz–3.5kHz, which are used to capture the acoustic spectrum disturbance characteristics caused by sticking friction and micro-impact behavior.

[0012] Preferably, said S2 further includes S23; S23, extracting the amplitude-frequency response of the acoustic signal S at each time t in the frequency domain to obtain the current frame sound spectrum intensity value of each target frequency point; Call the benchmark spectrum curve that matches the current ambient temperature and opening and closing conditions from the historical health sample set built into the device to obtain the standard average spectrum line under the target frequency band set Bres; Compare the relative deviation between the current sound spectrum intensity value and the corresponding reference spectrum value at each frequency point, and sum the squares of the deviations to obtain the spectrum disturbance deviation value Fvar, which is used to quantitatively reflect the intensity and concentration of acoustic anomalies during the current opening and closing process. The specific calculation form is: ; Where k represents the frequency point index number, fk represents the kth frequency point, S(fk) represents the amplitude-frequency response of the current acoustic signal S at the kth frequency point fk, represents the amplitude-frequency response of the healthy sample acoustic signal S at the kth frequency point fk; The formula for the spectrum disturbance deviation value Fvar is derived from the frequency domain statistical anomaly detection model. It is a relative deviation measurement method for comparing the "current frame spectrum" with the "reference spectrum" in the classic signal processing field. It is widely used in fault diagnosis, voiceprint analysis, and modal recognition. This formula is based on the idea of ​​normalized spectrum deviation and has been improved in the following aspects: The key target frequency band set Bres is introduced to selectively weight the abnormally sensitive frequency range; the relative deviation square sum is used instead of the absolute difference to strengthen the central judgment of the deviation; the amplitude-frequency response of the healthy sample acoustic signal S at the kth frequency point fk is used Normalization is performed to make the results more consistent across different devices or backgrounds.

[0013] Preferably, the S3 further includes S31; S31. The inductance mutation factor index Cimp and the spectrum disturbance deviation value Fvar are used as input feature vectors, mapped to the standard scoring space respectively through a nonlinear normalization function, and input into a preset comprehensive opening and closing state scoring model. The comprehensive opening and closing state scoring model establishes a boundary interval in the scoring space through normal state samples and abnormal state samples in the training data, and outputs the comprehensive diagnostic score Qdiag as the health index of the current opening and closing state of the solenoid valve.

[0014] Preferably, the S3 further includes S32; S32. Based on the healthy samples and the early fault samples, statistical threshold values ​​are calculated to preset a diagnostic level range. The diagnostic level range includes a first diagnostic threshold F1, a second diagnostic threshold F2, and a third diagnostic threshold F3. The comprehensive diagnostic score Qdiag is compared with the diagnostic level range to determine the opening and closing status of the solenoid valve. An alarm is issued based on the comparison result, and the execution of subsequent control instructions is restricted. The specific comparison content is as follows: When the comprehensive diagnostic score Qdiag is less than the first diagnostic threshold F1, the health level is divided into level 1 and the system is executed normally without any intervention; When the first diagnostic threshold F1 ≤ comprehensive diagnostic score Qdiag ≤ second diagnostic threshold F2, the health level is divided into level 2, and the current state is marked. If it is triggered continuously for more than three times, an alarm is triggered; When the second diagnostic threshold F2 < comprehensive diagnostic score Qdiag ≤ third diagnostic threshold F3, the health level is divided into three levels, and a real-time alarm is issued, the opening and closing speed is limited to 80%, and the health log is updated; When the comprehensive diagnostic score Qdiag is greater than the third diagnostic threshold F3, the health level is divided into level four, at which time an alarm signal is issued, the solenoid valve control is locked, and a maintenance operation is prompted.

[0015] A solenoid valve opening and closing state measurement system includes an electromagnetic disturbance analysis module, a spectrum disturbance analysis module and a comprehensive state measurement and control module; The electromagnetic disturbance analysis module collects the dynamic inductance L after the electromagnetic valve control signal is triggered, and calculates the product of the second-order derivative and the first-order derivative of the dynamic inductance L to obtain the inductance mutation factor index Cimp; The spectrum disturbance analysis module compares the inductance mutation factor index Cimp with the preset mutation threshold Cthr, and triggers the disturbance analysis mechanism based on the comparison result. It collects the instantaneous acoustic signal S during the operation of the solenoid valve, performs Fourier transform on the instantaneous acoustic signal S, extracts the target frequency band set Bres, and calculates the spectrum disturbance deviation value Fvar; The comprehensive state measurement and control module calculates a comprehensive diagnostic score Qdiag by inputting the inductance mutation factor index Cimp and the spectrum disturbance deviation value Fvar into the comprehensive opening and closing state scoring model, and compares the comprehensive diagnostic score Qdiag with a preset diagnostic level range to determine the opening and closing state of the solenoid valve, and issues an alarm based on the comparison result, while restricting the execution of subsequent control instructions.

[0016] The present invention provides a system and method for measuring the opening and closing status of a solenoid valve. It has the following beneficial effects: (1) After the solenoid valve control signal is triggered, this method uses the voltage and current sampling module integrated in the drive circuit, combined with the inductance analysis module in the local microcontroller MCU, to reversely calculate the dynamic inductance L at each time t in real time. In the edge computing module, the dynamic inductance L is Gaussian filtered and derivatives are calculated to extract the first-order derivative and second-order derivative features, and the inductance mutation factor index Cimp is calculated. This index can comprehensively reflect the abnormal conditions such as "stuck, rebound, and discontinuity" that occur during the opening and closing of the solenoid valve. By setting the mutation threshold Cthr and comparing them, a dynamic opening and closing behavior diagnosis mechanism with high resolution and strong adaptability is constructed. Compared with the traditional method of switch detection or overall current envelope recognition, this method can complete high-frequency dynamic response extraction and abnormal trend quantification within 100ms after the control signal is triggered, and has stronger early warning capabilities and non-invasive compatibility.

[0017] (2) After detecting that the inductance mutation factor index Cimp exceeds the mutation threshold Cthr, the present invention automatically triggers the disturbance analysis mechanism. With the help of a MEMS microphone module installed on the outside of the solenoid valve housing, the acoustic signal S is collected within the first 50ms after the start of the driving action. The acoustic signal S is subjected to short-time Fourier transform by the local control module. The target frequency band set Bres is formed by combining the preset frequency ranges of 800Hz–1.2kHz and 2kHz–3.5kHz. The spectrum intensity value is then extracted and compared with the standard spectrum line of the healthy sample. The spectrum disturbance deviation value Fvar is calculated by the normalized relative square deviation model. This value can accurately reflect the structural oscillation characteristics caused by valve core sticking or impact, and has the advantages of strong robustness, wide adaptability, and good consistency across devices. It provides a high-confidence acoustic reference signal for subsequent abnormal state diagnosis, significantly improving the accuracy and completeness of fault identification.

[0018] (3) This method constructs a comprehensive diagnostic score Qdiag by inputting the inductance mutation factor index Cimp and the spectrum disturbance deviation value Fvar into the comprehensive opening and closing state scoring model, and introducing an improved normalization mapping and square weighting mechanism; the scoring model is derived from the Euclidean distance paradigm of the multimodal diagnostic space. After dimensionless normalization, the two indicators are integrated into a unified evaluation system to ensure logical closure and mathematical continuity. At the same time, based on statistical modeling, a diagnostic level range is established, including the first diagnostic threshold F1, the second diagnostic threshold F2, and the third diagnostic threshold F3, and based on this, a four-level division of health status and corresponding control strategy response are achieved: for example, the first state does not require intervention, the second state triggers soft alarm logic, the third state implements speed limit and log update, and the fourth state directly controls locking and maintenance prompts. This mechanism can realize continuous quantitative health management of the entire process of solenoid valve opening and closing, build a closed-loop intelligent control system of "diagnosis, evaluation and response", and greatly enhance the stability and safety of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A schematic diagram of the steps of a method for measuring the opening and closing state of a solenoid valve according to the present invention; Figure 2 This is a flow chart of a solenoid valve opening and closing state measurement system according to the present invention; Figure 3 It is the dynamic inductance L curve. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0021] Example 1, please refer to Figure 1 The present invention provides a method for measuring the opening and closing state of a solenoid valve. To achieve the above purpose, the present invention is implemented through the following technical solutions: comprising the following steps: S1. After the solenoid valve control signal is triggered, the dynamic inductance L is collected, and the product of the second-order derivative and the first-order derivative of the dynamic inductance L is calculated to obtain the inductance mutation factor index Cimp; S2. Compare the inductance mutation factor index Cimp with the preset mutation threshold Cthr. Based on the comparison result, trigger the disturbance analysis mechanism to collect the instantaneous acoustic signal S during the solenoid valve operation. Perform Fourier transform on the instantaneous acoustic signal S, extract the target frequency band set Bres, and calculate the spectrum disturbance deviation value Fvar. S3. Input the inductance mutation factor index Cimp and the spectrum disturbance deviation value Fvar into the comprehensive opening and closing state scoring model to calculate the comprehensive diagnostic score Qdiag. The comprehensive diagnostic score Qdiag is compared with the preset diagnostic level range to determine the opening and closing state of the solenoid valve, and an alarm is issued based on the comparison result, while limiting the execution of subsequent control instructions.

[0022] In this embodiment, the method collects the dynamic inductance L after the solenoid valve control signal is triggered, and calculates the inductance mutation factor index Cimp by combining the product structure of the first-order derivative and the second-order derivative. The purpose is to use the dynamic change trend of the inductance during the opening and closing process to reflect the stability of the valve core movement. The dynamic inductance can sensitively respond to the nonlinear displacement and mutation behavior of the valve core, especially when there is sticking, rebound or discontinuous opening and closing, which manifests as a significant derivative mutation, which is far better than the signal hysteresis reflected by the simple current curve. Furthermore, by combining the inductance mutation factor index Cimp with the mutation threshold Cthr generated by historical health sample statistics, the inductance mutation factor index Cimp is calculated. For comparison, the subsequent acoustic disturbance analysis process is triggered only when the inductance mutation factor index Cimp is significantly higher than the mutation threshold Cthr. This design effectively avoids unnecessary sound spectrum calculations under normal conditions, helps improve processing efficiency and reduces the probability of false alarms. After the disturbance analysis mechanism is triggered, the acoustic acquisition unit collects the instantaneous acoustic signal S at a frequency of 20kHz within a 50ms time window before the action, and focuses on the analysis of the two sensitive frequency bands of 800Hz-1.2kHz and 2kHz-3.5kHz through Fourier transform. These two frequency bands are experimentally verified to be the acoustic disturbances caused by valve core impact and friction. The most concentrated frequency region can effectively eliminate the influence of background noise. The spectrum disturbance deviation value Fvar is then obtained through relative deviation calculation to characterize the existence and intensity of the structural anomaly. Finally, in S3, the inductance mutation factor index Cimp and the spectrum disturbance deviation value Fvar are input into the comprehensive opening and closing state scoring model for normalization and merging, outputting a comprehensive diagnostic score Qdiag. A higher score indicates a stronger degree of anomaly. The scoring model uses an L2 norm structure to ensure that any anomaly indicator can improve the overall score, preventing weakly correlated features from being ignored. Simultaneously, the dimensionless scoring facilitates cross-device and cross-environmental interoperability. The comprehensive diagnostic score Qdiag is compared with the set multi-level diagnostic thresholds F1, F2, and F3 to determine the solenoid valve opening and closing status level and link the control strategy. When the score is in different level ranges, marking, alarming, speed limiting, and locking control are respectively executed. This enables early detection, quantitative judgment, and intelligent control response of anomalies in the solenoid valve opening and closing process through electromagnetic induction and acoustic sensing without changing the hardware structure. This effectively improves the operational stability and safety warning capabilities of the solenoid valve system. It is particularly suitable for environments with extremely high real-time and reliability requirements such as rail transit and pneumatic actuators.

[0023] Example 2, please refer to Figure 1 and Figure 3 , specifically: S1 includes S11; S11. By integrating a sampling circuit module into the solenoid valve control drive circuit of the rail transit train air compressor system and setting a collection strategy, the voltage signal and current signal at both ends of the solenoid valve coil are collected in real time; The voltage and current signals are transmitted to an inductance analysis module that is synchronously input into a local microcontroller MCU. The inductance analysis module uses the voltage and current signals combined with the known PWM duty cycle to reversely calculate the dynamic inductance L at each time t. The acquisition strategy is set to a sampling frequency of no less than 5kHz, and the sampling time window covers the interval from 0ms to 100ms after the control signal is triggered; The circuit module is arranged on a PWM control signal path in a solenoid valve control drive circuit.

[0024] S1 also includes S12; S12. The dynamic inductance L at each moment t is integrated into an inductance sequence in chronological order, and the inductance sequence is uploaded to the edge computing module embedded in the control host for analysis and processing. The edge computing module performs a filtering preprocessing on all the dynamic inductances L in the inductance sequence through the set denoising processor. The filtering preprocessing smoothes all the dynamic inductances L in the inductance sequence by using a Gaussian convolution kernel to weaken the derivative deviation caused by high-frequency electromagnetic interference and mechanical jitter. The inductance sequence after the smoothing operation is then passed to the derivative calculation module for feature extraction to obtain the first-order derivative and second-order derivative of the dynamic inductance L.

[0025] S1 and S13; S13. Based on the first-order derivative and second-order derivative of the dynamic inductance L, calculate the mutation response factor at any time t within the entire sampling time window, then extract the maximum value of the mutation response factor at all times t to obtain the inductance mutation factor index Cimp, which is used to measure the maximum nonlinear disturbance intensity reflected by the solenoid valve during the opening and closing process; The inductance mutation factor index Cimp is calculated and output by the following algorithm formula; ; Throughout, max represents the maximum value function, d represents the integral function, L(t) represents the dynamic inductance L at time t, represents the slope enhancement adjustment factor, ranging from 0.6 to 1.2; In order to comprehensively consider the coupling effects of the speed change during the valve core movement, that is, stability, and the sudden acceleration, that is, disturbance; in: is the second-order derivative, is the first-order derivative; Formula structure explanation: The product structure selects the product structure of the first-order derivative and the second-order derivative, which comprehensively incorporates the degree of mutation and the intensity of the change trend into the calculation dimension; introduces the slope enhancement adjustment factor It is used to enhance or suppress the sensitivity of a certain dimension, especially the weight of the first-order derivative, to the final indicator. The maximum value extraction strategy is not based on integral averaging, but rather selects the maximum response moment within the entire sampling interval to reflect the most extreme mutation location. This formula is an empirical physics-inspired improved model. It is a targeted design introduced based on the conventional second-order eigenvalue extraction strategy to consider the continuous changes in solenoid valve motion. Physical dimensional consistency of the formula: The units of the first-order derivative and the second-order derivative are both in inductance units, so the unit dimensions on both sides are consistent. At the same time, this formula is used to construct the relative anomaly factor and is not used for engineering unit calculation output. Therefore, there is no need to achieve absolute physical dimensional consistency, only to ensure mathematical logical closure; This example shows that the dynamic inductance L of a solenoid valve is collected within 0-100ms after the control signal is triggered. After filtering and derivative calculation, it is found that the derivative changes sharply at 67ms. The calculated Cimp = 0.032, which is higher than the empirical threshold Cthr = 0.025. It is judged that there is slightly abnormal behavior, triggering the subsequent sound spectrum recognition mechanism and recording this opening and closing as potentially unstable.

[0026] In this embodiment, by embedding a sampling circuit module in the control drive circuit of the solenoid valve of the rail transit train air compressor system, setting a sampling frequency higher than 5kHz and covering a time window of 0ms~100ms, the nonlinear response characteristics of the initial opening and closing can be fully captured, avoiding response distortion or feature omission caused by low-frequency sampling; based on the real-time voltage and current reverse driving dynamic inductance L, the electromagnetic characteristic changes caused by the physical displacement of the valve core can be more intuitively reflected, replacing the traditional indirect method based on driving current inference, and avoiding distortion caused by driving waveform disturbance; subsequent preprocessing is performed through Gaussian kernel filtering to suppress local high-frequency noise caused by system electromagnetic interference and microstructure jitter, ensuring the smoothness and differentiability of the derivative calculation process, and then obtaining the first-order and second-order derivatives of the dynamic inductance L in the time dimension; the derivative product structure can combine the speed change trend with the acceleration mutation amplitude By jointly incorporating them into the index calculation, the maximum value strategy is used to extract the strongest abnormal response in the entire process, effectively avoiding the average strategy from masking sudden abnormal problems; for example, when the valve core is blocked, the first-order derivative may decay rapidly, and the second-order derivative will show a strong reverse jump. The product of the two can amplify the diagnostic characteristics of such key nodes, and accurately locate the "opening and closing mutation points" from high-density data; the proposed inductance mutation factor index Cimp not only maintains physical dimension consistency, but also has cross-device generalization capabilities through normalization processing, thereby providing a stable premise for the subsequent acoustic spectrum abnormality triggering mechanism; in summary, this process constructs a structured path from low-cost inductance acquisition-high-fidelity filtering-derivative modeling-nonlinear mutation identification, which improves the accuracy and robustness of early abnormality detection of solenoid valve opening and closing, and is particularly suitable for fault precursor identification in high-speed opening and closing or micro-stuck scenarios.

[0027] Example 3, please refer to Figure 1 , specifically: S2 includes S21; S21. Select N groups of opening and closing inductance behavior samples under the same temperature, pressure, and driving conditions, where the N groups of opening and closing inductance behavior samples are the inductance mutation factor indicators Cimp of multiple groups of solenoid valves in a historical healthy state, and then set the 95th percentile of the inductance mutation factor indicators Cimp of the multiple groups of solenoid valves as the mutation threshold Cthr; The inductance mutation factor index Cimp obtained in real time is compared with the mutation threshold Cthr to determine the inductance disturbance during the opening and closing process of the solenoid valve; whether the solenoid valve spool has "nonlinear mutation behavior" during the opening and closing process, such as the following abnormal conditions: valve spool stuck, valve spool rebound, and discontinuous opening and closing; The inductance curve corresponding to valve core sticking is characterized by a slowdown or sudden stop in inductance change in a local section, which will cause delays in opening and closing actions and the risk of startup control failure. The inductance curve corresponding to the valve core rebound shows that the inductance first rises and then falls to form a local oscillation, which will lead to incomplete movement, impact fatigue, and shortened service life. The inductance curve corresponding to the discontinuous opening and closing is characterized by a sudden change in the second-order conduction plus a zero trend in the first-order conduction, which will lead to failure in the drive signal response and a sign of mechanical structure failure. The specific comparison contents are as follows; If the inductance mutation factor index Cimp in the current opening and closing cycle exceeds the mutation threshold Cthr, it is judged that the current opening and closing state is abnormal, and the subsequent disturbance analysis mechanism is automatically triggered, and the time of the abnormal event is recorded. If the inductance mutation factor index Cimp in the current opening and closing cycle is less than or equal to Cthr, the current opening and closing state is judged to be normal, and it is automatically recorded as a healthy opening and closing cycle, and the disturbance analysis mechanism is skipped, and the next control cycle is continued.

[0028] S2 also includes S22; S22. After the disturbance analysis mechanism is triggered, an acoustic signal S is collected by a MEMS micro-microphone module disposed outside the solenoid valve housing. The microphone module collects sound pressure at a sampling frequency of 20 kHz within the first 50 ms after the solenoid valve drive signal is triggered, obtaining the acoustic signal S at each time instant t. After the acquisition is completed, the local control module performs short-time Fourier transform analysis on the acoustic signal S at each time t, and combines the preset target frequency range judgment rules to screen out abnormal frequency bands from the complete spectrum to form a target frequency band set Bres. The target frequency range includes two sub-intervals: 800Hz–1.2kHz and 2kHz–3.5kHz, which are used to capture the acoustic spectrum disturbance characteristics caused by sticking friction and micro-impact behavior.

[0029] S2 also includes S23; S23, extracting the amplitude-frequency response of the acoustic signal S at each time t in the frequency domain to obtain the current frame sound spectrum intensity value of each target frequency point; Call the benchmark spectrum curve that matches the current ambient temperature and opening and closing conditions from the historical health sample set built into the device to obtain the standard average spectrum line under the target frequency band set Bres; Compare the relative deviation between the current sound spectrum intensity value and the corresponding reference spectrum value at each frequency point, and sum the squares of the deviations to obtain the spectrum disturbance deviation value Fvar, which is used to quantitatively reflect the intensity and concentration of acoustic anomalies during the current opening and closing process. The specific calculation form is: ; Where k represents the frequency point index number, fk represents the kth frequency point, S(fk) represents the amplitude-frequency response of the current acoustic signal S at the kth frequency point fk, represents the amplitude-frequency response of the healthy sample acoustic signal S at the kth frequency point fk; The formula for the spectrum disturbance deviation value Fvar is derived from the frequency domain statistical anomaly detection model. It is a relative deviation measurement method for comparing the "current frame spectrum" with the "reference spectrum" in the classic signal processing field. It is widely used in fault diagnosis, voiceprint analysis, and modal recognition. This formula is based on the idea of ​​normalized spectrum deviation and has been improved in the following aspects: The key target frequency band set Bres is introduced to selectively weight the abnormally sensitive frequency range; the relative deviation square sum is used instead of the absolute difference to strengthen the central judgment of the deviation; the amplitude-frequency response of the healthy sample acoustic signal S at the kth frequency point fk is used Perform normalization to make the results more consistent across different devices or backgrounds; Simplified calculation example: Assumptions: After short-time Fourier transform analysis, 10 frequency points are obtained in the target frequency band set Bres band; The corresponding healthy spectrum is [100, 102, 98, …]; The current frame spectrum is [110, 115, 97, …]; Calculate the square of the relative deviation of each point and sum it up to get Fvar=0.064; When the empirical threshold Fthr=0.040, it is determined that there is a structural disturbance anomaly.

[0030] In this embodiment, the method first constructs multiple sets of baseline behavior samples by collecting historical healthy solenoid valve inductance mutation factor indicators (Cimp) under identical temperature, pressure, and actuation conditions. Using the 95th percentile statistical strategy to set the mutation threshold (Cthr), this method effectively mitigates perturbations of the threshold caused by extreme values ​​or occasional fluctuations, improving the robustness and universality of the threshold and providing a clear healthy reference baseline for subsequent comparisons of Cimp and Cthr. For example, manufacturing tolerances or microstructural variations may exist between batches of solenoid valves. By learning from a large sample under the same operating conditions and taking the 95th percentile as Cthr, this method can fully accommodate normal fluctuations and ensure that abnormality detection is not falsely positive. Once the real-time collected inductance mutation factor indicator (Cimp) exceeds the mutation threshold (Cthr), it can be inferred that the valve core motion trajectory may exhibit nonlinear perturbations, such as sticking, rebound, and intermittent opening and closing, among other complex mechanical anomalies. Sticking is often caused by factors such as valve chamber contamination and seal aging, while rebound is often caused by valve core inertia impact or elastic component reaction. Discontinuous opening and closing may be caused by interrupted drive logic or a disconnected mechanism. If such behavior is not promptly identified, it will directly affect the solenoid valve's operating accuracy, leading to delays in train air pressure control and even the risk of loss of control. To further confirm whether the aforementioned inductance mutation is associated with a structural physical anomaly, an acoustic disturbance analysis mechanism is automatically triggered when Cimp exceeds the limit. This mechanism uses a MEMS microphone module deployed on the exterior of the solenoid valve housing to collect acoustic signals S with a high-frequency accuracy of 20kHz within a 50ms time window before opening and closing, thereby capturing the acoustic characteristics of events such as friction, impact, and sticking during high-speed operation. A Fourier transform is used to convert the time signal into the frequency domain. A target frequency band set Bres is constructed by combining two abnormality-sensitive frequency bands—800Hz–1.2kHz and 2kHz–3.5kHz—to enhance the identification of high-probability anomalies such as structural resonance and friction oscillation. The system then compares the amplitude of the current frame's spectrum at the target frequency with the baseline spectrum of a historical healthy sample at the same frequency, frequency by frequency. The spectral disturbance deviation value Fvar is calculated by summing the squared relative deviations. The method uses squared relative deviations rather than absolute differences. This not only strengthens the influence of points with large deviations, highlighting areas of concentrated anomalies, but also eliminates positive and negative offset errors, improving overall anomaly sensitivity. In particular, when a frequency point significantly exceeds the baseline spectrum in the current frame, its squared deviation will nonlinearly amplify Fvar, thereby increasing the diagnostic sensitivity to features such as stuck shock. For example, if the actual sound spectrum at 2.8kHz rises by 15% compared to the baseline, while other points have smaller deviations, this frequency point will dominate the final Fvar value, directly driving the system to identify it as a structural disturbance and triggering an anomaly flag. This strategy outperforms traditional models based on total energy or absolute differences in identifying weak but critical frequency band anomalies. In summary, the combined diagnosis of Cimp and Fvar achieves cross-domain fusion monitoring from non-contact electromagnetic behavior to structural acoustic response.Its design not only significantly improves the dimension and accuracy of the solenoid valve opening and closing health status assessment under the premise of low cost and non-invasiveness, but also ensures good robustness under complex working conditions. It is especially suitable for operating scenarios with extremely high real-time and safety requirements in high-speed train air compressor control systems.

[0031] Example 4, please refer to Figure 1 ,Specifically: S3 also includes S31; S31. The inductance mutation factor index Cimp and the spectrum disturbance deviation value Fvar are used as input feature vectors, mapped to the standard scoring space through a nonlinear normalization function, and input into a preset comprehensive opening and closing state scoring model. The comprehensive opening and closing state scoring model establishes a boundary interval in the scoring space through normal state samples and abnormal state samples in the training data, and outputs a comprehensive diagnostic score Qdiag as the health index of the current opening and closing state of the solenoid valve, where a lower score indicates a higher abnormality risk. The comprehensive opening and closing state scoring model constructs an opening and closing state distribution fitting network by adopting an improved radial basis kernel function, where: the inductance mutation factor index Cimp is used to characterize the nonlinear abnormal trend of the valve core mechanical dynamics, and the spectrum disturbance deviation value Fvar is used to characterize the energy offset or stuck harmonic anomaly of the structural acoustics; The comprehensive diagnostic score Qdiag is calculated using the following comprehensive opening and closing state scoring model; ; Where Fthr represents the spectrum disturbance threshold, which is obtained by the statistical mean of the normal sample distribution; The comprehensive diagnostic score Qdiag is built on the output of two innovative modules: the inductance mutation factor index Cimp, which is used to measure the discontinuity or mutation trend of the valve core motion trajectory; and the spectrum disturbance deviation value Fvar, which is used to identify potential sticking or friction oscillation disturbances in the acoustic frequency band. The original physical foundation formulas come from: the derivative operation principle of the relationship between inductance derivative and time, the basic law of electromagnetic induction; the amplitude offset and energy distribution calculation based on Fourier transform to extract frequency domain response, signal processing principles; the anomaly scoring method based on normalized deviation value calculation in Euclidean space, and the multimodal fusion scoring model; The sum-of-squares structure is derived from the Euclidean distance and L2 norm model concepts and is often used to jointly evaluate two or more orthogonal or weakly correlated feature variables. The square operation amplifies deviations; stronger anomalies have a greater impact; eliminates negative interference to ensure a positive total score; maintains mathematical differentiability and continuity to facilitate subsequent fitting analysis. The additive structure indicates that the two indicators are parallel factors; any anomaly in one will increase the overall score; and avoids the problem of the multiplication structure causing the entire score to be invalid if one item is zero. Verification of dimensional consistency: Both terms in the fraction are dimensionless ratios, and the square operation retains the dimensionlessness. Therefore, the comprehensive diagnostic score Qdiag is a dimensionless scalar, which is suitable for threshold comparison. The dimensions on both sides of the formula are consistent, and the formula structure is reasonable.

[0032] S3 also includes S32; S32. Based on the healthy samples and the early fault samples, statistical threshold values ​​are calculated to preset a diagnostic level range. The diagnostic level range includes a first diagnostic threshold F1, a second diagnostic threshold F2, and a third diagnostic threshold F3. The comprehensive diagnostic score Qdiag is compared with the diagnostic level range to determine the opening and closing status of the solenoid valve. An alarm is issued based on the comparison result, and the execution of subsequent control instructions is restricted. The specific comparison content is as follows: When the comprehensive diagnostic score Qdiag is less than the first diagnostic threshold F1, the health level is divided into level 1, which means normal execution without any intervention, the signal has disturbances but the opening and closing states have no abnormal levels, and the opening and closing are stable; When the first diagnostic threshold F1 ≤ comprehensive diagnostic score Qdiag ≤ second diagnostic threshold F2, the health level is classified as level 2 and the current state is marked. If the alarm is triggered three or more times in a row, an alarm is triggered. To avoid occasional misjudgments or occasional abnormal triggering, soft decision logic is used. When the second diagnostic threshold F2 is less than the comprehensive diagnostic score Qdiag and less than the third diagnostic threshold F3, the health level is divided into three levels. At this time, a real-time alarm is issued, the opening and closing speed is limited to 80%, and the health log is updated, indicating that there is structural viscosity or harmonic anomaly, and the risk is increased; When the comprehensive diagnostic score Qdiag is greater than the third diagnostic threshold F3, the health level is divided into level 4. At this time, an alarm signal is issued, the solenoid valve control is locked, and maintenance operations are prompted. The valve core may have structural damage or a high probability of being stuck, and intervention is necessary; Among them: the first diagnostic threshold F1 is used as the healthy confidence boundary, the second diagnostic threshold F2 is the slight fault trigger threshold, and the third diagnostic threshold F3 is the obvious stuck abnormality sign.

[0033] In this embodiment, by inputting the inductance mutation factor index Cimp and the spectrum disturbance deviation value Fvar into the comprehensive opening and closing state scoring model, the two characteristic indicators are firstly subjected to nonlinear normalization respectively, with the aim of eliminating the scoring bias caused by the difference in physical magnitude and distribution, and ensuring that the diagnostic factors from different sources can be integrated and compared in the same standardized space; the comprehensive scoring model using the sum of squares structure can amplify the contribution of abnormal points to the overall score, and is particularly suitable for high-sensitivity identification of early weak faults. At the same time, a nonlinear mapping relationship between the feature space and the opening and closing health state is established through the improved radial basis kernel function, so that the model not only has good fitting accuracy, but also has a certain abnormal generalization ability, and can still maintain effective judgment when facing unknown or boundary-type behaviors. The strategy of setting a three-level diagnostic level range can realize multi-level health state judgment and avoid classifying all deviations as faults. For example, if the system's comprehensive diagnostic score Qdiag consistently remains below the first diagnostic threshold F1, the solenoid valve is in a completely healthy state and requires no intervention. However, if the score is between F1 and F2, marginal disturbances or sporadic interference may be present. By implementing a soft-decision mechanism that triggers an alarm only after three consecutive triggers, false positives caused by transient ambient noise or fluctuating operating conditions can be effectively avoided. Once the comprehensive diagnostic score Qdiag exceeds the second diagnostic threshold F2, indicating possible mechanical sticking or structural harmonic anomalies, speed limiting and logging mechanisms are immediately activated to prevent further escalation of the fault. If the score further exceeds the third diagnostic threshold F3, a critical fault is identified, prompting immediate control interruption and maintenance notification to ensure equipment and personnel safety. The overall strategy utilizes a multimodal input and multi-level response mechanism to dynamically detect and respond to the solenoid valve's opening and closing status, from normal to faulty, with a graded response. This significantly improves the real-time and accuracy of fault detection, while effectively reducing false alarms and the risk of missed detections. This is particularly applicable to scenarios such as rail transit, where response speed and safety are paramount.

[0034] Example 5, please refer to Figure 1 and Figure 2 , a solenoid valve opening and closing state measurement system, including an electromagnetic disturbance analysis module, a spectrum disturbance analysis module and a comprehensive state measurement and control module; The electromagnetic disturbance analysis module collects the dynamic inductance L after the electromagnetic valve control signal is triggered, and calculates the product of the second-order derivative and the first-order derivative of the dynamic inductance L to obtain the inductance mutation factor index Cimp; The spectrum disturbance analysis module compares the inductance mutation factor index Cimp with the preset mutation threshold Cthr, and triggers the disturbance analysis mechanism based on the comparison result. It collects the instantaneous acoustic signal S during the operation of the solenoid valve, performs Fourier transform on the instantaneous acoustic signal S, extracts the target frequency band set Bres, and calculates the spectrum disturbance deviation value Fvar; The comprehensive state measurement and control module calculates a comprehensive diagnostic score Qdiag by inputting the inductance mutation factor index Cimp and the spectrum disturbance deviation value Fvar into the comprehensive opening and closing state scoring model, and compares the comprehensive diagnostic score Qdiag with a preset diagnostic level range to determine the opening and closing state of the solenoid valve, and issues an alarm based on the comparison result, while restricting the execution of subsequent control instructions.

[0035] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. A method for measuring the opening and closing state of a solenoid valve, characterized by: The following steps are involved: S1. After the solenoid valve control signal is triggered, the dynamic inductance L is collected, and the product of the second-order derivative and the first-order derivative of the dynamic inductance L is calculated to obtain the inductance mutation factor index Cimp; S2. Compare the inductance mutation factor index Cimp with the preset mutation threshold Cthr. Based on the comparison result, trigger the disturbance analysis mechanism to collect the instantaneous acoustic signal S during the solenoid valve operation. Perform Fourier transform on the instantaneous acoustic signal S, extract the target frequency band set Bres, and calculate the spectrum disturbance deviation value Fvar. S3. Input the inductance mutation factor index Cimp and the spectrum disturbance deviation value Fvar into the comprehensive opening and closing state scoring model to calculate the comprehensive diagnostic score Qdiag. Compare the comprehensive diagnostic score Qdiag with the preset diagnostic level range to determine the opening and closing state of the solenoid valve, issue an alarm based on the comparison result, and limit the execution of subsequent control instructions.

2. A method for measuring the opening and closing state of a solenoid valve according to claim 1, characterized in that: Said S1 includes S11; S11. By integrating a sampling circuit module into the solenoid valve control drive circuit of the rail transit train air compressor system and setting a collection strategy, the voltage signal and current signal at both ends of the solenoid valve coil are collected in real time; The voltage and current signals are transmitted to an inductance analysis module that is synchronously input into a local microcontroller MCU. The inductance analysis module uses the voltage and current signals combined with the known PWM duty cycle to reversely calculate the dynamic inductance L at each time t. The acquisition strategy is set to have a sampling frequency of no less than 5kHz, and a sampling time window covering the interval from 0ms to 100ms after the control signal is triggered; The circuit module is arranged on a PWM control signal path in a solenoid valve control drive circuit.

3. A method for measuring the opening and closing state of a solenoid valve according to claim 2, characterized in that: Said S1 also includes S12; S12. The dynamic inductance L at each moment t is integrated into an inductance sequence in chronological order, and the inductance sequence is uploaded to the edge computing module embedded in the control host for analysis and processing. The edge computing module performs a filtering preprocessing on all the dynamic inductances L in the inductance sequence through a set denoising processor. The filtering preprocessing smoothes all the dynamic inductances L in the inductance sequence by using a Gaussian convolution kernel to weaken the derivative deviation caused by high-frequency electromagnetic interference and mechanical jitter. The smoothed inductance sequence is then passed to the derivative calculation module for feature extraction to obtain the first-order derivative and second-order derivative of the dynamic inductance L.

4. A method for measuring the opening and closing state of a solenoid valve according to claim 3, characterized in that: Said S1 and S13; S13. Based on the first-order derivative and second-order derivative of the dynamic inductance L, the mutation response factor is calculated for any time t within the entire sampling time window, and then the maximum value of the mutation response factor at all times t is extracted to obtain the inductance mutation factor index Cimp, which is used to measure the maximum nonlinear disturbance intensity reflected by the solenoid valve during the opening and closing process.

5. A method for measuring the opening and closing state of a solenoid valve according to claim 4, characterized in that: Said S2 includes S21; S21. Select N groups of opening and closing inductance behavior samples under the same temperature, pressure, and driving conditions, where the N groups of opening and closing inductance behavior samples are the inductance mutation factor indicators Cimp of multiple groups of solenoid valves in a historical healthy state, and then set the 95th percentile of the inductance mutation factor indicators Cimp of the multiple groups of solenoid valves as the mutation threshold Cthr; The inductance mutation factor index Cimp obtained in real time is compared with the mutation threshold Cthr to determine the inductance disturbance during the opening and closing process of the solenoid valve; The specific comparison contents are as follows; If the inductance mutation factor index Cimp in the current opening and closing cycle exceeds the mutation threshold Cthr, it is judged that the current opening and closing state is abnormal, and the subsequent disturbance analysis mechanism is automatically triggered, and the time when the abnormal event occurs is recorded; If the inductance mutation factor index Cimp in the current opening and closing cycle is less than or equal to Cthr, the current opening and closing state is judged to be normal, and it is automatically recorded as a healthy opening and closing cycle, and the disturbance analysis mechanism is skipped, and the next control cycle is continued.

6. The method for measuring the opening and closing state of a solenoid valve according to claim 1, characterized in that: Said S2 also includes S22; S22. After the disturbance analysis mechanism is triggered, an acoustic signal S is collected by a MEMS micro-microphone module disposed outside the solenoid valve housing. The microphone module collects sound pressure at a sampling frequency of 20 kHz within the first 50 ms after the solenoid valve drive signal is triggered, obtaining the acoustic signal S at each time instant t. After the acquisition is completed, the local control module performs short-time Fourier transform analysis on the acoustic signal S at each time t, and combines the preset target frequency range judgment rules to screen out abnormal frequency bands from the complete spectrum to form a target frequency band set Bres. The target frequency range includes two sub-intervals: 800Hz–1.2kHz and 2kHz–3.5kHz, which are used to capture the acoustic spectrum disturbance characteristics caused by sticking friction and micro-impact behavior.

7. A method for measuring the opening and closing state of a solenoid valve according to claim 6, characterized in that: Said S2 also includes S23; S23, extracting the amplitude-frequency response of the acoustic signal S at each time t in the frequency domain to obtain the current frame sound spectrum intensity value of each target frequency point; Call the benchmark spectrum curve that matches the current ambient temperature and opening and closing conditions from the historical health sample set built into the device to obtain the standard average spectrum line under the target frequency band set Bres; The relative deviation between the current sound spectrum intensity value and the corresponding reference spectrum value is compared frequency point by frequency point, and the square of the deviation is summed to obtain the spectrum disturbance deviation value Fvar, which quantitatively reflects the intensity and concentration of the acoustic anomaly during the current opening and closing process.

8. The method for measuring the opening and closing state of a solenoid valve according to claim 6, characterized in that: Said S3 also includes S31; S31. The inductance mutation factor index Cimp and the spectrum disturbance deviation value Fvar are used as input feature vectors, mapped to the standard scoring space respectively through a nonlinear normalization function, and input into a preset comprehensive opening and closing state scoring model. The comprehensive opening and closing state scoring model establishes a boundary interval in the scoring space through normal state samples and abnormal state samples in the training data, and outputs the comprehensive diagnostic score Qdiag as the health index of the current opening and closing state of the solenoid valve.

9. A method for measuring the opening and closing state of a solenoid valve according to claim 8, characterized in that: Said S3 also includes S32; S32. Based on the healthy samples and the early fault samples, statistical threshold values ​​are calculated to preset a diagnostic level range. The diagnostic level range includes a first diagnostic threshold F1, a second diagnostic threshold F2, and a third diagnostic threshold F3. The comprehensive diagnostic score Qdiag is compared with the diagnostic level range to determine the opening and closing status of the solenoid valve. An alarm is issued based on the comparison result, and the execution of subsequent control instructions is restricted. The specific comparison content is as follows: When the comprehensive diagnostic score Qdiag is less than the first diagnostic threshold F1, the health level is divided into level 1 and the system is executed normally without any intervention; When the first diagnostic threshold F1 ≤ comprehensive diagnostic score Qdiag ≤ second diagnostic threshold F2, the health level is divided into level 2, and the current state is marked. If it is triggered continuously for more than three times, an alarm is triggered; When the second diagnostic threshold F2 < comprehensive diagnostic score Qdiag ≤ third diagnostic threshold F3, the health level is divided into three levels, and a real-time alarm is issued, the opening and closing speed is limited to 80%, and the health log is updated; When the comprehensive diagnostic score Qdiag is greater than the third diagnostic threshold F3, the health level is divided into level four, at which time an alarm signal is issued, the solenoid valve control is locked, and a maintenance operation is prompted.

10. A solenoid valve opening and closing state measurement system, applied to a solenoid valve opening and closing state measurement method according to any one of claims 1 to 9, characterized in that: Including electromagnetic disturbance analysis module, spectrum disturbance analysis module and comprehensive state measurement and control module; The electromagnetic disturbance analysis module collects the dynamic inductance L after the electromagnetic valve control signal is triggered, and calculates the product of the second-order derivative and the first-order derivative of the dynamic inductance L to obtain the inductance mutation factor index Cimp; The spectrum disturbance analysis module compares the inductance mutation factor index Cimp with the preset mutation threshold Cthr, and triggers the disturbance analysis mechanism based on the comparison result. It collects the instantaneous acoustic signal S during the operation of the solenoid valve, performs Fourier transform on the instantaneous acoustic signal S, extracts the target frequency band set Bres, and calculates the spectrum disturbance deviation value Fvar; The comprehensive state measurement and control module calculates a comprehensive diagnostic score Qdiag by inputting the inductance mutation factor index Cimp and the spectrum disturbance deviation value Fvar into the comprehensive opening and closing state scoring model, and compares the comprehensive diagnostic score Qdiag with a preset diagnostic level range to determine the opening and closing state of the solenoid valve, and issues an alarm based on the comparison result, while restricting the execution of subsequent control instructions.

Images