Electromagnetic valve health state self-diagnosis and predictive maintenance method and system based on coil current waveform analysis

Abstract

The application discloses a kind of based on coil current waveform analysis solenoid health state self-diagnosis and predictive maintenance method and system, belong to solenoid fault diagnosis technical field.It includes: when solenoid is powered off, change the current freewheeling path of coil, transfer coil energy to energy storage element to form damped oscillation, collect first voltage waveform and carry out time-frequency transform, extract the delay length and change slope of instantaneous frequency to determine the degree of mechanical jamming;In the process of collecting first voltage waveform, monitor the tail section characteristics to confirm that the spool movement stops, apply short-time excitation that does not cause spool action and change the freewheeling path again to excite second damped oscillation, collect second voltage waveform and calculate quality factor and dynamic inductance, determine the degree of coil turn-to-turn insulation deterioration by comparing with factory reference value.The application realizes online decoupling diagnosis and quantitative evaluation of solenoid mechanical jamming and coil insulation deterioration without additional sensors.

Description

Technical Field

[0001] This invention relates to a method and system for self-diagnosis and predictive maintenance of the health status of solenoid valves based on coil current waveform analysis, belonging to the field of solenoid valve fault diagnosis technology. Background Technology

[0002] As a core actuator in process industries, automotive, aerospace and other fields, the health status of solenoid valves directly affects system safety and operational continuity.

[0003] Existing diagnostic techniques based on coil current or voltage waveforms mainly include the following approaches: the action inflection point detection method, which calculates the valve core action time by monitoring the slope change point of the rising edge of the coil current; and the back electromotive force characteristic method, which assesses the spring force decay by detecting the back electromotive force amplitude or return time during the power outage follow current period.

[0004] The common drawback of the above methods is that they only utilize the limited characteristics generated by the natural switching process. Multiple factors, such as coil resistance temperature drift, inductance changes, and insulation aging, are coupled together, making it difficult to distinguish fault types and quantify fault severity. Furthermore, existing technologies generally treat the high-frequency damped oscillations generated after power failure as harmful noise, suppressing or eliminating them through freewheeling diodes, absorption circuits, etc., without any approach to actively extracting multi-dimensional state information from this transient process.

[0005] Therefore, there is an urgent need for a method and system that can actively utilize the electrical transient processes of the coil itself to achieve online decoupling diagnosis of mechanical and electrical degradation. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a method and system for self-diagnosis and predictive maintenance of the health status of solenoid valves based on coil current waveform analysis, which realizes online decoupled diagnosis and quantitative evaluation of mechanical jamming and coil insulation degradation of solenoid valves.

[0007] The technical problem to be solved by this invention is achieved by the following technical solution: A method for self-diagnosis and predictive maintenance of solenoid valve health status based on coil current waveform analysis includes the following steps: When the solenoid valve is de-energized, the current freewheeling path of the coil is changed, so that the energy stored in the coil is transferred to an energy storage element of a specified capacity, forming a damped oscillation, and the first voltage waveform during the damped oscillation process is collected. The first voltage waveform is subjected to time-frequency transformation to extract the instantaneous frequency characteristics reflecting the valve core motion state, and the degree of mechanical jamming of the solenoid valve is determined based on the instantaneous frequency characteristics. During the acquisition of the first voltage waveform, the characteristics of the tail segment of the waveform are monitored to confirm whether the valve core movement has stopped. After confirming that the valve core movement has stopped, a short-term excitation that does not cause the valve core to move is applied to the coil, and the current freewheeling path is changed again to excite the second damped oscillation, and the second voltage waveform of the second damped oscillation is acquired; The quality factor and dynamic inductance are calculated based on the second voltage waveform, and the degree of inter-turn insulation degradation of the coil is determined based on the comparison results of the quality factor and dynamic inductance with the factory reference values.

[0008] The present invention is further configured such that the step of changing the current freewheeling path of the coil to form damped oscillation includes: After issuing the power-off command, the energy storage element is first connected in parallel to the coil circuit; The freewheeling path is then disconnected, allowing the coil current to be transferred uninterruptedly to the energy storage element, forming a damped oscillation.

[0009] The present invention is further configured such that the capacity value of the energy storage element of the specified capacity satisfies the following condition: The peak voltage of the damped oscillation does not exceed 60% of the rated withstand voltage of the coil-driven switch. Furthermore, the frequency of the damped oscillation is not less than 10 times the effective bandwidth of the valve core's mechanical action, so as to generate an oscillation waveform of at least 5 complete cycles during the valve core's movement period.

[0010] The present invention is further configured to: extract instantaneous frequency characteristics reflecting the movement state of the valve core, and determine the degree of mechanical jamming of the solenoid valve based on the instantaneous frequency characteristics, specifically including: Perform time-frequency transformation on the first voltage waveform to obtain the curve of instantaneous frequency changing with time; Identify the starting moment when the curve begins to deviate continuously from its initial stable value, and use the time difference between this starting moment and the power outage moment as the delay duration; Calculate the slope of the curve at the initial stage of sustained deviation; The degree of mechanical jamming is determined based on the increase in the delay duration and / or the decrease in the slope of the change.

[0011] The present invention is further configured to: determine the degree of mechanical jamming based on the delay duration and the change slope, specifically as follows: The extracted delay duration and the slope of change are input into the mapping relationship model between "delay duration-slope of change" and the sticking force established through offline calibration, and the current equivalent friction force is obtained by querying.

[0012] The present invention is further configured to: determine the degree of inter-turn insulation degradation of the coil based on the comparison results of the quality factor and dynamic inductance with the factory reference value, specifically including: The calculated quality factor is compared with the benchmark quality factor calibrated at the factory when the valve core movement stops and the air gap is at its maximum. If a step drop occurs that exceeds the preset threshold, it is judged as abnormal. The calculated dynamic inductance is compared with the reference dynamic inductance calibrated at the factory when the valve core movement stops and the air gap is at its maximum. If the reduction exceeds the normal drift range caused by temperature changes, it is judged as abnormal. When any of the above anomalies occurs, it is determined that there is inter-turn insulation degradation.

[0013] The present invention is further configured to include a temperature compensation step. During the energized holding phase of the solenoid valve, the voltage across the coil and the current flowing through the coil are integrated over at least one complete PWM cycle. Calculate the ratio of the voltage integral to the current integral, and use it as the DC resistance of the coil at the current temperature; The reference quality factor or reference dynamic inductance is corrected by using the DC resistance of the coil to eliminate the influence of temperature changes on the inter-turn insulation degradation judgment result.

[0014] The present invention is further configured such that the method for confirming whether the valve core movement has stopped is as follows: Monitor the instantaneous frequency exhibited by the first voltage waveform at the tail end; When the rate of change of the instantaneous frequency over time is lower than the preset static determination threshold, it is confirmed that the valve core movement has stopped.

[0015] The present invention is further configured to: after determining the degree of mechanical jamming and the degree of inter-turn insulation degradation of the coil, include: Calculate the equivalent frictional force based on the degree of mechanical jamming; The pressure before and after the valve is obtained through industrial communication to calculate the fluid pressure differential force, or the maximum working pressure differential specified on the valve nameplate is used as the fluid pressure differential force in the event of communication interruption. Based on the equivalent frictional force, fluid pressure difference force, and the correspondence between spring force and displacement, the minimum electromagnetic force required for reliable valve core operation is calculated. The minimum holding current and minimum driving voltage are derived from the minimum electromagnetic force. The driving parameters are adaptively adjusted using the minimum holding current and minimum driving voltage as lower limits. When it is determined that the adjustment of the drive parameters cannot be maintained above the lower limit of the constraint, a forced maintenance command is issued.

[0016] A solenoid valve health status self-diagnosis and predictive maintenance system based on coil current waveform analysis includes a switching circuit for driving the solenoid valve, a switching circuit for changing the freewheeling path, an energy storage element, a sampling circuit, and a controller, characterized in that the controller is configured to perform the aforementioned method.

[0017] The beneficial effects of this invention are: 1. By changing the suppression of oscillation to active excitation, without adding any physical sensors, the dual degradation information of mechanical jamming and coil insulation can be obtained in a single operation cycle by simply switching the diagnostic network, realizing online decoupled diagnosis of electromechanical degradation; 2. By performing time-frequency transformation on actively excited damped oscillations, the delay duration and slope of the instantaneous frequency are extracted to quantify the degree of mechanical jamming. Compared with existing technologies that only determine whether jamming "exists", this method can provide a quantitative assessment of the jamming force. 3. By employing a time-division secondary diagnostic strategy, the coil quality factor and dynamic inductance are independently tested under the locked baseline state where the valve core is stationary and the air gap is at its maximum. This effectively eliminates the cross-interference of valve core position changes and temperature drift on insulation diagnosis, and improves the sensitivity and reliability of inter-turn micro short circuit detection. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the system architecture of the present invention.

[0019] Figure 2 This is a flowchart of the main method of the present invention. Detailed Implementation

[0020] To facilitate a clear understanding of the technical means, creative features, objectives, and effects of this invention, the invention will be further described below in conjunction with specific illustrations.

[0021] This invention provides a self-diagnostic and predictive maintenance system for the health status of a solenoid valve based on coil current waveform analysis. This system is integrated into the solenoid valve drive controller. Please refer to [link / reference]. Figure 1The main components include: a main control unit (i.e., controller), which uses an MCU or DSP and has a built-in high-speed ADC with a sampling rate of no less than 100kSps, a PWM generation module, and a diagnostic algorithm processing engine; a drive circuit, which includes a main drive switch SWmain (i.e., the switch circuit used to drive the solenoid valve) that forms an H-bridge or high-low side drive structure, used to control the on / off state of the solenoid valve coil; a freewheeling interruption switch SWfree and an auxiliary diagnostic switch SWD (which together form the switch circuit used to change the freewheeling path), wherein SWfree is connected in series in the coil freewheeling circuit, and it remains constantly on under normal operating conditions without diagnostics to ensure the integrity of the freewheeling function, and is disconnected in diagnostic mode to forcibly cut off the low-voltage freewheeling clamping path; SWD and the diagnostic capacitor Cdiag are connected in series and are usually connected in parallel across the coil to form an active diagnostic network; a voltage sampling circuit and a current sampling circuit, used to collect the voltage across the diagnostic capacitor and the coil current, respectively; and a communication interface, used to receive process pressure data from the host computer and upload diagnostic results and early warning information.

[0022] The selection of the diagnostic capacitor Cdiag is a crucial aspect of the system's hardware design, requiring rigorous physical calculations and the simultaneous fulfillment of two constraints. The first is a voltage safety constraint: the peak voltage Vpeak of the diagnostic oscillation must not exceed 60% of the rated withstand voltage of the coil's main drive switch SWmain to ensure long-term device reliability. This peak voltage is determined by the coil current I and coil inductance L before power-off, following the formula Vpeak = I·√(L / Cdiag). The second constraint is a time-frequency analysis constraint: the frequency fosc = 1 / (2π√(LCdiag)) of the diagnostic oscillation must be at least 10 times the effective bandwidth of the valve core's mechanical movement. This ensures that at least five complete damped oscillation cycles can be generated within the millisecond window of valve core movement, providing sufficient data samples for subsequent time-frequency analysis.

[0023] As a specific and preferred engineering example, for an application scenario with a coil inductance of 100mH, an operating current of 1A, and a driving MOSFET withstand voltage of 500V, CDiag is selected as a 2.2μF high-voltage thin-film capacitor with a withstand voltage of 400V. According to calculations, its theoretical peak voltage is about 213V, which is far below the safety boundary; the oscillation frequency is about 339Hz, and about 7 complete oscillation waveforms can be generated within a 20ms mechanical response window, perfectly meeting the dual requirements of safety and time-frequency analysis.

[0024] During the smooth commutation from normal freewheeling to diagnostic oscillation, this system employs a strict timing sequence. At time t0, when the power-off command is issued, the auxiliary diagnostic switch SWD is turned on first. Since the initial voltage of capacitor CDiag is zero, there is no inrush current when it is connected to the circuit. After a short delay, at time t1, the coil main drive switch SWmain and the freewheeling blocking switch SWfree are simultaneously turned off. At this time, because SWD has been turned on in advance, the current in the coil is seamlessly and uninterruptedly transferred to the CDiag branch, thereby forming the expected RLC damped oscillation. The entire switching process ensures the continuity of the current path and eliminates the generation of dangerous voltage spikes at the source.

[0025] See Figure 2 Based on the above system, this invention also provides a method for self-diagnosis and predictive maintenance of the health status of solenoid valves based on coil current waveform analysis. This method completely changes the technical bias of prior art in suppressing power-off oscillations. By actively constructing a controlled resonant network, it transforms the power-off transient process, once considered harmful, into an analytical multi-dimensional information carrier, realizing online decoupled diagnosis and predictive maintenance of the solenoid valve's mechanical jamming degree and the degree of coil inter-turn insulation degradation. The specific implementation steps of this method in the controller are as follows.

[0026] S1. When the solenoid valve is de-energized, the current freewheeling path of the coil is changed, so that the energy stored in the coil is transferred to an energy storage element of a specified capacity to form a damped oscillation, and the first voltage waveform during the damped oscillation process is collected.

[0027] In step S1, the energy storage element is the aforementioned diagnostic capacitor CDiag. The specific operation of changing the freewheeling path involves executing the smooth commutation strategy: first, the auxiliary diagnostic switch SWD is turned on, connecting CDiag in parallel to the coil circuit without impact; then, the coil main drive switch SWmain and the freewheeling blocking switch SWfree are simultaneously turned off, allowing the coil current to seamlessly transfer to the CDiag branch, forming an RLC damped oscillation. The main control unit's ADC then continuously acquires the voltage waveform across CDiag at a sampling rate of no less than 100 kSps, with the acquisition time covering the entire valve core movement process, typically 20 to 30 ms.

[0028] S2. Perform time-frequency transformation on the first voltage waveform to extract the instantaneous frequency characteristics reflecting the valve core's motion state, and determine the degree of mechanical jamming of the solenoid valve based on the instantaneous frequency characteristics.

[0029] In step S2, the acquired first voltage waveform undergoes a time-frequency transformation, preferably using a short-time Fourier transform (STFT) or a Hilbert-Huang transform (HHT), to calculate and obtain a curve showing the instantaneous frequency f(t) changing with time. This waveform is not a simple amplitude decay signal; its frequency components are dynamically modulated by the change in coil inductance caused by the valve core's movement. The inventors have discovered that the mechanical state of the valve core is directly mapped to two key characteristics of the f(t) curve. One is the delay time td, which refers to the time elapsed from the moment the power-off command is issued until the instantaneous frequency f(t) begins to deviate significantly from its initial stable value. This parameter physically directly reflects the start-up lag time of the valve core overcoming static friction to begin movement.

[0030] The criteria for determining continuous deviation are as follows: when the deviation of the instantaneous frequency from the initial stable value continuously exceeds a preset frequency fluctuation threshold (e.g., 5% of the initial stable frequency value) for a preset number of sampling points (e.g., 5 consecutive sampling points), continuous deviation is confirmed, and the time corresponding to the first sampling point exceeding the threshold is taken as the starting time of the deviation. Secondly, the slope of change, kf, refers to the rate of change of the instantaneous frequency f(t) in the initial stage of continuous deviation, i.e., kf = df / dt. This parameter physically characterizes the speed of the valve core motion establishment process and intuitively reflects the magnitude of the dynamic friction force.

[0031] The slope of change at the initial stage of sustained deviation refers to the slope value obtained by performing linear least-squares fitting on the instantaneous frequency curve within a preset time window (e.g., 10ms) after the sustained deviation is confirmed. It should be noted that because the air gap decreases and the inductance increases after the valve core movement begins, leading to a decrease in oscillation frequency, this slope kf is usually negative. In subsequent quantification of the degree of jamming, its absolute value |kf| is used for comparison or substituted into the mapping model (model input is |kf|) to avoid confusion caused by the sign. The more severe the jamming, the smaller |kf|.

[0032] To quantify the degree of jamming, this method establishes a multidimensional mapping model between "td-kf" and the equivalent jamming force through offline calibration. The specific process of offline calibration is as follows: A solenoid valve test bench is used to apply a precisely adjustable frictional force radially onto the valve stem. Specifically, a pair of opposing miniature electric actuators or spring-loaded devices, along with a high-precision force sensor (accuracy better than ±0.1N), generate a stable additional frictional force F_friction along the valve core's movement path. This frictional force is in the opposite direction to the valve core's movement, simulating a mechanical jamming effect. During calibration, starting from 0 N, the force is gradually increased in increments of 0.5 N or 1 / 10 of the estimated maximum jamming force based on the valve type, until a critical value at which the valve core completely fails to move is reached.

[0033] For each set jamming force level, steps S1 to S2 are repeated at least 10 times. Each time, the first voltage waveform is collected from the time the power-off command is issued until the valve core movement completely stops. The delay time td and the slope of change kf are extracted after short-time Fourier transform. The median of multiple measurements is taken as the characteristic value (td_i, kf_i) for that jamming force level to eliminate random errors.

[0034] Using the jamming force F_friction as the output and (td, kf) as the input, a second-order polynomial surface fitting is employed, resulting in the model form: F_friction = a0 + a1·td + a2·kf + a3·td² + a4·kf² + a5·td·kf. The coefficient vectors a0~a5 are solved using the least squares method. The fitted coefficients are stored in the controller. For higher accuracy requirements, Gaussian process regression or support vector regression can be used to establish a nonparametric model. After calibration, during online diagnostics, the measured (td, kf) is substituted into the model to retrieve the current equivalent friction force Ff, achieving a quantitative assessment of mechanical jamming.

[0035] During online diagnosis, the main control unit automatically extracts td and kf from the f(t) curve and inputs them into the mapping model to query the current equivalent friction force Ff, thus achieving a quantitative assessment of mechanical jamming. Under normal conditions, td is usually less than 3ms and |kf| is relatively large; when severe jamming occurs, td may extend to 8 to 15ms, and |kf| will decrease significantly. This step, through these two characteristics, not only determines "whether jamming occurs" but also achieves precise quantification of "how severe the jamming is."

[0036] It should be noted that the "clamping force" (or "equivalent clamping force") mentioned in this application is physically represented by the equivalent frictional force Ff obtained by querying the mapping model, and the two have the same meaning in the context of this invention. Furthermore, in simplified application scenarios such as those with limited controller computing power, a one-dimensional clamping assessment can be performed solely based on the increase in delay time td or the decrease in the slope kf; preferably, the aforementioned two-dimensional mapping model is used to obtain higher diagnostic robustness and quantification accuracy.

[0037] S3. During the acquisition of the first voltage waveform, monitor the characteristics of the tail segment of the waveform to confirm whether the valve core movement has stopped.

[0038] In step S3, this method utilizes the waveform tail segment resources from the first diagnosis to determine the valve core movement state without the need for additional sensors. Specifically, it continuously monitors the instantaneous frequency exhibited by the first voltage waveform at its tail segment. When the rate of change of the instantaneous frequency over time is lower than a preset static determination threshold, it can be confirmed that the valve core movement has completely stopped. The static determination threshold needs to be engineered based on the actual sampling rate and the frequency resolution of the time-frequency conversion.

[0039] The specific method is as follows: First, during the factory calibration stage of the solenoid valve, the valve core is kept completely stationary (i.e., not energized). A low-energy pulse with the same parameters as S1 is applied to the coil to excite a small oscillation. The waveform is collected, and the rate of change of instantaneous frequency over time is calculated. This is repeated 100 times to obtain the root mean square value of the rate of change of frequency under stationary conditions, σ_static. Then, the stationary judgment threshold is set to 3 × σ_static. For most industrial solenoid valves, under the conditions of a sampling rate of not less than 100 kSps and a short-time Fourier transform window length of 10 ms, the measured value of σ_static is usually between 2 and 5 Hz / s. Therefore, it is recommended to use a dynamic calibration threshold of 3 × σ_static, rather than a fixed value of 0.5 Hz / s, as an alternative. If dynamic calibration cannot be performed, the threshold can be conservatively set to 15 Hz / s (this value has been tested and covers the stationary fluctuation range of more than 95% of valve types). When the instantaneous frequency change rate is below the threshold for five consecutive sampling points, it is determined that the valve core movement has completely stopped. This determination method is based entirely on signal analysis, which maximizes the reuse of diagnostic resources.

[0040] S4. After confirming that the valve core movement has stopped, apply a short-term excitation to the coil that does not cause the valve core to move, and change the current freewheeling path again to excite the second damped oscillation, and collect the second voltage waveform of the second damped oscillation.

[0041] It should be noted that during the tens of milliseconds window period of waiting for the valve core movement to stop, the energy of the first waveform damped oscillation excited in step S1 has been completely dissipated in the equivalent resistance of the coil, and the voltage across the diagnostic capacitor CDiag has naturally decayed to zero. This ensures that the subsequent second damped oscillation is safely performed in a strictly zero initial state, eliminating the risk of surge when the switch is connected due to residual capacitor voltage, and providing unbiased initial conditions for the subsequent accurate extraction of electrical characteristics.

[0042] In step S4, after confirming that the valve core is stationary, the main control unit applies a short-duration excitation pulse with extremely low energy to the coil. The upper limit of the pulse energy must strictly meet the constraint of "not causing the valve core to move". To this end, this method adopts the following quantification criteria: First, by offline calibration or by referring to the solenoid valve's manufacturer's manual, obtain the static pull-in force-current characteristic curve of the valve under the current working air gap (maximum air gap), and record the critical current I_th (usually defined as the current value when the pull-in force equals the sum of the spring return force and the maximum static friction force) at which the valve core begins to generate detectable displacement. Then, based on the coil DC resistance R_DC and the target pulse width T_pulse, determine the applied voltage pulse amplitude V_pulse, such that the peak current of the coil during the pulse I_peak = V_pulse / R_DC × (1 - e^(-T_pulse / τ)), where τ = L / R_DC is the coil electrical time constant. The constraint is: I_peak ≤ 0.5 × I_th, that is, a safety margin of more than 50% is reserved to cover the fluctuation of I_th under different temperatures and different degradation states.

[0043] Typically, for common industrial solenoid valves, I_th is in the tens to hundreds of milliamperes range. Therefore, a low-voltage pulse with a pulse width T_pulse of 10~50μs and an amplitude V_pulse of 3~12V is sufficient. If the controller has programmable output capability, it can also dynamically adjust the pulse amplitude based on the minimum electromagnetic force obtained from the most recent diagnosis, further improving the signal-to-noise ratio. Next, the aforementioned smooth commutation strategy is executed again, cutting off the freewheeling path of this weak current and transferring its energy to CDiag, exciting a second damped oscillation. The main control unit acquires the second voltage waveform of this second damped oscillation. Since the valve core is stationary, the air gap is at its maximum and constant, and the coil inductance is a fixed value, this waveform is specifically used to extract purely electrical characteristics.

[0044] S5. Calculate the quality factor and dynamic inductance based on the second voltage waveform, and determine the degree of inter-turn insulation degradation of the coil based on the comparison results of the quality factor and dynamic inductance with the factory reference value.

[0045] In step S5, the quality factor Q and dynamic inductance L are accurately extracted from the second voltage waveform through parameter fitting. Parameter fitting can preferably employ the Planckian method or the least squares fitting method to extract the damping coefficient α and the damped angular frequency ωd, and then calculate the quality factor Q = ωd / (2α). In particular, given that the key diagnostic target of this invention is inter-turn insulation degradation, the damping coefficient α will significantly increase when degradation occurs. Therefore, a rigorous physical formula must be used to calculate the dynamic inductance to eliminate theoretical errors caused by the attenuation factor. From the classical dynamic equations of the RLC resonant circuit, the accurate formula for calculating the dynamic inductance L is L = 1 / [(ωd² + α²)·Cdiag], rather than a simplified solution ignoring the damping term. These two parameters are extremely sensitive to the inter-turn insulation state of the coil. A micro-short circuit between turns is equivalent to a short-circuited secondary winding inside the coil, which will severely dissipate oscillation energy and cause a sudden change in apparent inductance.

[0046] The key to diagnosis lies in baseline locking. The calculated Q and L values ​​are compared with the reference values ​​Q0 and L0 calibrated at the factory under the exact same physical condition of "valve spool movement stopped and air gap at maximum". This operation is crucial, as it fundamentally eliminates the interference of inductance changes caused by different valve spool positions on the diagnostic results, allowing even minute inductance losses caused by inter-turn short circuits to be detected with high sensitivity.

[0047] The specific determination rule is as follows: if the Q value experiences a step drop exceeding a preset threshold, or the L value experiences a sudden decrease exceeding the normal drift range that can be explained by temperature changes, then inter-turn insulation degradation is determined to have occurred. The preset threshold for the step drop is typically set as a single sudden drop exceeding 30% of the baseline Q0, to distinguish it from slow temperature drift or aging drift. The normal drift range is calculated based on the temperature coefficient of resistance of the copper conductor (approximately 0.393% / ℃) combined with the expected temperature fluctuation range of the equipment's operating environment; sudden changes exceeding ±5% are considered abnormal. When any of the above abnormalities occurs, inter-turn insulation degradation is confirmed.

[0048] Furthermore, this method includes an online, non-disruptive temperature compensation step to eliminate the influence of temperature changes on the reference values ​​Q0 and L0. During the steady-state phase when the solenoid valve is energized, the voltage u(t) across the coil and the current i(t) flowing through the coil are synchronously integrated within a complete PWM cycle using a normal PWM drive waveform. The reason for using this ratio is that, within a complete steady-state PWM cycle, the dynamic voltage integral term (∫L·di / dt·dt) resulting from inductor energy storage is strictly zero. Therefore, this algorithm perfectly filters out the dynamic inductive interference under high-frequency PWM chopping without changing the normal operating rhythm of the equipment, extracting the pure DC resistance. Thus, the coil DC resistance RDC = ∫u(t)dt / ∫i(t)dt, which is only related to temperature, can be directly calculated. Using the temperature coefficient of resistance of copper conductors, the baseline value affected by temperature can be accurately corrected.

[0049] The specific correction method is as follows: Temperature correction for the quality factor Q. The equivalent total resistance R of the coil includes the DC resistance R_DC and the equivalent resistance of eddy current losses. At low-frequency oscillations (<1kHz), eddy current losses are very small, and Q is approximately inversely proportional to R_DC. Therefore, the corrected reference quality factor Q0_corrected = Q0 × (R_DC0 / R_DC), where R_DC0 is the reference DC resistance at the factory calibration temperature (e.g., 20℃), and R_DC is the measured DC resistance at the current temperature. If the temperature coefficient of resistance of the copper conductor is known to be α_Cu ≈ 0.00393 / ℃, it can also be directly corrected according to the temperature difference: Q0_corrected = Q0 × [1 + α_Cu·(T0 - T_current)], where T0 is the factory calibration temperature, and T_current is derived from R_DC.

[0050] Temperature correction for dynamic inductance L. The coil inductance L is primarily affected by the temperature of the core material's permeability, and its change is much smaller than that of the resistance. For most solenoid valves, the rate of change of L does not exceed ±2% within the range of -40℃ to 125℃. Therefore, no correction is necessary, or a pre-calibrated linearity coefficient β can be used: L0_corrected = L0 × [1 + β·(T_current - T0)], where β is determined through high and low temperature chamber experiments (typical values ​​are -0.0005~+0.0002 / ℃). If β is unknown, ignoring the inductance correction still ensures the reliability of the inter-turn insulation degradation assessment, because the inductance decrease caused by inter-turn short circuits typically exceeds 5%, much greater than temperature drift.

[0051] After diagnosing the degree of mechanical jamming and the degree of insulation degradation between coil turns, this method further transforms the diagnostic information into maintenance decisions with mandatory safety baselines, forming a complete closed loop from diagnosis to control.

[0052] Specifically, the equivalent frictional force Ff is calculated based on the degree of mechanical jamming. Simultaneously, to obtain the fluid pressure differential force Fp, this method employs a dual-path redundancy strategy. Path one involves communication acquisition, using an industrial communication bus to obtain the upstream and downstream pressures from the DCS or PLC in real time, and accurately calculating Fp based on the effective area of ​​the valve seat. Path two is a conservative extreme value strategy; when communication is interrupted or data is unavailable, the system automatically adopts the most unfavorable conservative strategy, directly using the maximum operating pressure differential specified on the valve nameplate to calculate Fp, ensuring that the calculated safety baseline is absolutely reliable under any operating condition.

[0053] The equivalent frictional force Ff, the fluid pressure differential force Fp, and the spring force Fspring obtained from the spring-displacement curve are added together to obtain the minimum electromagnetic force Fem required for reliable valve core operation. Subsequently, the minimum holding current Ihold_min and the minimum driving voltage Vdrive_min are derived by using the electromagnetic force-current-air gap model.

[0054] The reverse engineering process of the model is as follows: a pre-established three-dimensional mapping lookup table (MAP) of "electromagnetic force-air gap-current / voltage" is pre-loaded into the controller. There are two methods for establishing this MAP, which can be selected according to actual conditions: Before the solenoid valve leaves the factory, three-dimensional magnetic field finite element simulation software (such as Ansys Maxwell, JMAG) is used to accurately model the coil, yoke, and valve core assembly. The simulation obtains the axial electromagnetic force Fem under different air gap δ (from 0 to the maximum air gap, step size 0.05mm) and different excitation current I (from 0 to the rated current, step size 5% of the rated value). The simulation results are discretized into a three-dimensional table and stored in the controller's non-volatile memory.

[0055] If an accurate simulation model is unavailable, estimations can be made using the parameters on the solenoid valve's nameplate. The rated voltage U_rated, rated holding current I_hold_rated, maximum operating pressure difference ΔP_max, and spring return force F_spring (obtained by measuring the force required to push the valve core when not energized) are known on the nameplate. When the valve core is at its maximum air gap (the critical position before actuation), the electromagnetic force is approximately proportional to the square of the current. Therefore, a simplified model can be established: Fem(I) = k_m·I², where the proportionality coefficient k_m is derived from the condition Fem(I_hold_rated) = F_spring + ΔP_max·A (where A is the effective area of ​​the valve seat) under rated holding conditions. During online operation, Ihold_min = √(Fem / k_m) is calculated from the required Fem. The minimum drive voltage Vdrive_min must ensure that the coil current can rise to Ihold_min within a specified time during the full conduction period of the PWM. It can be estimated as Vdrive_min = Ihold_min·R_DC + L·(dI / dt), where dI / dt is the maximum allowable current change rate in the design.

[0056] After obtaining the required minimum electromagnetic force Fem and substituting it into the fixed air gap value under the known limit displacement, the minimum holding current and minimum driving voltage constraint lower limit are obtained by looking up tables and using reverse interpolation algorithms or by direct calculation.

[0057] Based on this, the drive parameters are adaptively adjusted with minimum holding current and minimum drive voltage as lower limits. When a purely mechanical jamming is detected, the drive voltage can be adaptively increased to compensate for friction while meeting safety constraints, thereby extending the valve's effective lifespan. However, when coil inter-turn insulation degradation is detected, or when it is determined that the adjustment of the drive parameters can no longer be maintained above the aforementioned lower limits, the optimization algorithm will refuse to execute any instructions that may further accelerate failure or lead to abort, and will directly issue a forced maintenance instruction, thus finding the optimal balance between maximizing equipment lifespan and ensuring absolute functional safety.

[0058] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments, and various changes and modifications can be made without departing from the spirit and scope of the invention, all of which fall within the scope of protection claimed by the present invention. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for self-diagnosis and predictive maintenance of the health status of a solenoid valve based on coil current waveform analysis, characterized in that, Includes the following steps: When the solenoid valve is de-energized, the current freewheeling path of the coil is changed, so that the energy stored in the coil is transferred to an energy storage element of a specified capacity, forming a damped oscillation, and the first voltage waveform during the damped oscillation process is collected. The first voltage waveform is subjected to time-frequency transformation to extract the instantaneous frequency characteristics reflecting the valve core motion state, and the degree of mechanical jamming of the solenoid valve is determined based on the instantaneous frequency characteristics. During the acquisition of the first voltage waveform, the characteristics of the tail segment of the waveform are monitored to confirm whether the valve core movement has stopped. After confirming that the valve core movement has stopped, a short-term excitation that does not cause the valve core to move is applied to the coil, and the current freewheeling path is changed again to excite the second damped oscillation, and the second voltage waveform of the second damped oscillation is acquired; The quality factor and dynamic inductance are calculated based on the second voltage waveform, and the degree of inter-turn insulation degradation of the coil is determined based on the comparison results of the quality factor and dynamic inductance with the factory reference values.

2. The method according to claim 1, characterized in that, The steps to change the current freewheeling path of the coil to generate damped oscillations include: After issuing the power-off command, the energy storage element is first connected in parallel to the coil circuit; The freewheeling path is then disconnected, allowing the coil current to be transferred uninterruptedly to the energy storage element, forming a damped oscillation.

3. The method according to claim 1, characterized in that, The capacity value of the specified energy storage element satisfies the following condition: The peak voltage of the damped oscillation does not exceed 60% of the rated withstand voltage of the coil-driven switch. Furthermore, the frequency of the damped oscillation is not less than 10 times the effective bandwidth of the valve core's mechanical action, so as to generate an oscillation waveform of at least 5 complete cycles during the valve core's movement period.

4. The method according to claim 1, characterized in that, Extracting instantaneous frequency characteristics reflecting the valve core's motion state, and determining the degree of mechanical jamming of the solenoid valve based on these instantaneous frequency characteristics, specifically includes: Perform time-frequency transformation on the first voltage waveform to obtain the curve of instantaneous frequency changing with time; Identify the starting moment when the curve begins to deviate continuously from its initial stable value, and use the time difference between this starting moment and the power outage moment as the delay duration; Calculate the slope of the curve at the initial stage of sustained deviation; The degree of mechanical jamming is determined based on the increase in the delay duration and / or the decrease in the slope of the change.

5. The method according to claim 4, characterized in that, The degree of mechanical jamming is determined based on the delay duration and the slope of change, specifically as follows: The extracted delay duration and the slope of change are input into the mapping relationship model between "delay duration-slope of change" and the sticking force established through offline calibration, and the current equivalent friction force is obtained by querying.

6. The method according to claim 1, characterized in that, Based on the comparison results of the quality factor and dynamic inductance with the factory reference values, the degree of inter-turn insulation degradation of the coil is determined, specifically including: The calculated quality factor is compared with the benchmark quality factor calibrated at the factory when the valve core movement stops and the air gap is at its maximum. If a step drop occurs that exceeds the preset threshold, it is judged as abnormal. The calculated dynamic inductance is compared with the reference dynamic inductance calibrated at the factory when the valve core movement stops and the air gap is at its maximum. If the reduction exceeds the normal drift range caused by temperature changes, it is judged as abnormal. When any of the above anomalies occurs, it is determined that there is inter-turn insulation degradation.

7. The method according to claim 6, characterized in that, It also includes a temperature compensation step: During the energized holding phase of the solenoid valve, the voltage across the coil and the current flowing through the coil are integrated over at least one complete PWM cycle. Calculate the ratio of the voltage integral to the current integral, and use it as the DC resistance of the coil at the current temperature; The reference quality factor or reference dynamic inductance is corrected by using the DC resistance of the coil to eliminate the influence of temperature changes on the inter-turn insulation degradation judgment result.

8. The method according to claim 1 or 6, characterized in that, The method to confirm whether the valve core movement has stopped is as follows: Monitor the instantaneous frequency exhibited by the first voltage waveform at the tail end; When the rate of change of the instantaneous frequency over time is lower than the preset static determination threshold, it is confirmed that the valve core movement has stopped.

9. The method according to claim 1, characterized in that, After determining the degree of mechanical jamming and the degree of inter-turn insulation degradation of the coil, the following is also included: Calculate the equivalent frictional force based on the degree of mechanical jamming; The pressure before and after the valve is obtained through industrial communication to calculate the fluid pressure differential force, or the maximum working pressure differential specified on the valve nameplate is used as the fluid pressure differential force in the event of communication interruption. Based on the equivalent frictional force, fluid pressure difference force, and the correspondence between spring force and displacement, the minimum electromagnetic force required for reliable valve core operation is calculated. The minimum holding current and minimum driving voltage are derived from the minimum electromagnetic force. The driving parameters are adaptively adjusted using the minimum holding current and minimum driving voltage as lower limits. When it is determined that the adjustment of the drive parameters cannot be maintained above the lower limit of the constraint, a forced maintenance command is issued.

10. A self-diagnostic and predictive maintenance system for the health status of a solenoid valve based on coil current waveform analysis, comprising a switching circuit for driving the solenoid valve, a switching circuit for changing the freewheeling path, an energy storage element, a sampling circuit, and a controller, characterized in that, The controller is configured to perform the method according to any one of claims 1 to 9.

Images