Method, system and equipment for determining health state of electro-hydraulic proportional servo valve, medium and product

Abstract

The invention discloses an electro-hydraulic proportional servo valve health state determination method, system and device, a medium and a product, and relates to the field of electro-hydraulic proportional servo valve health state evaluation.The method comprises the steps that working data of a to-be-detected electro-hydraulic proportional servo valve is obtained; preprocessing the working data of the electro-hydraulic proportional servo valve to be detected to obtain processed working data; the preprocessing comprises noise reduction processing, feature extraction and feature screening; according to the processed working data, the health state of the electro-hydraulic proportional servo valve to be detected is determined through an electro-hydraulic proportional servo valve health state judgment model; the electro-hydraulic proportional servo valve health state judgment model is obtained by training a Transform model by using a training data set; the training data set is obtained through an electro-hydraulic proportional servo valve fault simulation experiment. The health state of the proportional electromagnet of the electro-hydraulic proportional servo valve can be automatically and accurately judged.

Description

Technical Field

[0001] The present application relates to the field of health status assessment of electro-hydraulic proportional servo valves, and particularly to a method, system, device, medium and product for determining the health status of electro-hydraulic proportional servo valves. Background Technique

[0002] The electro-hydraulic proportional servo valve is a key actuator that converts an electrical signal into a hydraulic output and is widely used in fields such as industrial automation, aerospace, and defense equipment. It is particularly indispensable in occasions that require high-precision position and force control. The core component, the proportional solenoid, generates an electromagnetic force acting on the spool through current regulation to achieve precise control and regulation of hydraulic oil. However, under long-term working conditions, the coil of the proportional solenoid is affected by various factors such as external temperature, current frequency, and voltage, resulting in insulation material degradation caused by stress and temperature changes. Creep and deterioration of the inter-turn insulation layer may lead to local inter-turn short circuits, ultimately affecting the inductance and electromagnetic force characteristics of the coil, and causing the control performance of the electro-hydraulic proportional servo valve to gradually decline. Currently, in the diagnosis of the performance degradation of electro-hydraulic proportional servo valves, it mainly relies on threshold judgment methods based on statistics or experience. However, these methods usually cannot fully capture the degradation modes of proportional solenoids under multi-field coupling effects and are difficult to accurately diagnose complex degradation phenomena such as insulation material degradation and inter-turn short circuits inside the solenoid. Secondly, the feature extraction process of traditional methods relies on manual design, lacks automation and systematization, is prone to ignoring potential complex feature patterns in multi-dimensional data, and is difficult to comprehensively and accurately describe the performance degradation process of electro-hydraulic proportional servo valves. Summary of the Invention

[0003] The purpose of the present application is to provide a method, system, device, medium and product for determining the health status of electro-hydraulic proportional servo valves, which can accurately judge the health status of electro-hydraulic proportional servo valves.

[0004] To achieve the above purpose, the present application provides the following solutions:

[0005] In the first aspect, the present application provides a method for determining the health status of an electro-hydraulic proportional servo valve, including:

[0006] Obtaining the working data of the electro-hydraulic proportional servo valve to be detected; the working data includes spool position signal, spool position feedback signal, actuator position signal, current signal, hydraulic signal at port A, and hydraulic signal at port B;

[0007] Preprocessing the working data of the electro-hydraulic proportional servo valve to be detected to obtain the processed working data; the preprocessing includes noise reduction processing, feature extraction, and feature screening;

[0008] Based on the processed working data, use the electro-hydraulic proportional servo valve health status determination model to determine the health status of the electro-hydraulic proportional servo valve to be detected; wherein, the electro-hydraulic proportional servo valve health status determination model is obtained by training a Transformer model using a training data set; the training data set is obtained through an electro-hydraulic proportional servo valve fault simulation experiment.

[0009] Optionally, preprocess the working data of the electro-hydraulic proportional servo valve to be detected to obtain processed working data, which specifically includes:

[0010] Use the wavelet transform method to perform noise reduction processing on the working data of the electro-hydraulic proportional servo valve to be detected to obtain noise-reduced working data;

[0011] Use the tsfresh toolbox to extract multi-information domain features from the noise-reduced working data to obtain the information domain features of the working data;

[0012] Use the maximum correlation minimum redundancy algorithm to perform feature screening on the information domain features of the working data to obtain processed working data.

[0013] Optionally, the multi-information domain includes the time domain, the frequency domain, and the time-frequency domain.

[0014] Optionally, based on the processed working data, use the electro-hydraulic proportional servo valve health status determination model to determine the health status of the electro-hydraulic proportional servo valve to be detected, which specifically includes:

[0015] Perform position encoding on the processed working data to obtain working data containing sequence position information;

[0016] Based on the working data containing sequence position information, use the electro-hydraulic proportional servo valve health status determination model to determine the health status of the electro-hydraulic proportional servo valve to be detected.

[0017] Optionally, obtain the training data set through an electro-hydraulic proportional servo valve fault simulation experiment, which specifically includes:

[0018] Based on Newton's second law, establish a mathematical model of the electro-hydraulic proportional servo valve;

[0019] Based on the working principle of the proportional solenoid in the electro-hydraulic proportional servo valve, establish a mathematical model of the proportional solenoid;

[0020] According to the working environment of the electro-hydraulic proportional servo valve, analyze the time when the inter-turn distance of the electromagnetic coil of the proportional solenoid in the electro-hydraulic proportional servo valve degenerates with environmental factors, determine the time of inter-turn short circuit, and establish a performance degradation model of the electro-hydraulic proportional servo valve over time;

[0021] Based on the mathematical model of the electro-hydraulic proportional servo valve, the mathematical model of the proportional solenoid valve, the performance degradation model of the electro-hydraulic proportional servo valve over time, and the electro-hydraulic proportional servo valve fault simulation test bench, different degrees of degradation of the proportional solenoid valve are simulated, and the working data corresponding to different degrees of degradation are collected;

[0022] Based on the working data corresponding to different degrees of degradation and the corresponding true fault states, a training data set is established.

[0023] Optionally, the Transformer model is trained using the training data set, specifically including:

[0024] Preprocess the training data set to obtain a processed training data set;

[0025] Input the processed working data in the processed data set into the current Transformer model to obtain a predicted fault state;

[0026] Determine the loss function value according to the predicted fault state and the corresponding true fault state;

[0027] Judge whether the end training condition is satisfied; the end training condition is to reach the maximum number of iterations or the loss function value is less than the set loss function threshold;

[0028] If so, use the current Transformer model as the electro-hydraulic proportional servo valve health state determination model;

[0029] If not, adjust the hyperparameters of the current Transformer model according to the loss function value, and return "Input the processed working data in the processed data set into the current Transformer model to obtain a predicted fault state".

[0030] In a second aspect, the present application provides a system for determining the health state of an electro-hydraulic proportional servo valve, including:

[0031] A data acquisition module for acquiring the working data of the electro-hydraulic proportional servo valve to be detected; the working data includes spool position signal, spool position feedback signal, actuator position signal, current signal, hydraulic signal at port A, and hydraulic signal at port B;

[0032] A preprocessing module for preprocessing the working data of the electro-hydraulic proportional servo valve to be detected to obtain processed working data; the preprocessing includes noise reduction processing, feature extraction, and feature screening;

[0033] A health status determination module, configured to determine the health status of the to-be-detected electro-hydraulic proportional servo valve according to the processed working data by using an electro-hydraulic proportional servo valve health status determination model; wherein, the electro-hydraulic proportional servo valve health status determination model is obtained by training a Transformer model by using a training data set; and the training data set is obtained through an electro-hydraulic proportional servo valve fault simulation experiment.

[0034] In a third aspect, the present application provides a computer device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program to implement the electro-hydraulic proportional servo valve health status determination method described in any one of the above.

[0035] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the electro-hydraulic proportional servo valve health status determination method described in any one of the above is implemented.

[0036] In a fifth aspect, the present application provides a computer program product, including a computer program, and when the computer program is executed by a processor, the electro-hydraulic proportional servo valve health status determination method described in any one of the above is implemented.

[0037] According to the specific embodiments provided by the present application, the present application has the following technical effects:

[0038] The present application provides an electro-hydraulic proportional servo valve health status determination method, system, device, medium, and product, which obtain the working data of the to-be-detected electro-hydraulic proportional servo valve; the working data includes a spool position signal, a spool position feedback signal, an actuator position signal, a current signal, an A-port hydraulic signal, and a B-port hydraulic signal; preprocess the working data of the to-be-detected electro-hydraulic proportional servo valve to obtain the processed working data; the preprocessing includes noise reduction processing, feature extraction, and feature screening; determine the health status of the to-be-detected electro-hydraulic proportional servo valve according to the processed working data by using an electro-hydraulic proportional servo valve health status determination model; wherein, the electro-hydraulic proportional servo valve health status determination model is obtained by training a Transformer model by using a training data set; and the training data set is obtained through an electro-hydraulic proportional servo valve fault simulation experiment. The present application constructs an electro-hydraulic proportional servo valve health status determination model based on a Transformer model, and judges the health status of the current electro-hydraulic proportional servo valve according to relevant features. The present application can automatically and accurately judge the health status of the proportional electromagnet of the electro-hydraulic proportional servo valve. Description of the Drawings

[0039] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0040] Figure 1 It is a schematic flowchart of a method for determining the health state of an electro-hydraulic proportional servo valve provided in an embodiment of the present application;

[0041] Figure 2 It is a flowchart of an algorithm for evaluating the health state of the electro-hydraulic proportional servo valve of the present application;

[0042] Figure 3 It is a schematic diagram of the appearance and structure of a typical electro-hydraulic proportional servo valve;

[0043] Figure 4 It is a schematic diagram of the structure of a proportional solenoid;

[0044] Figure 5 It is a layout diagram of the electromagnetic coils of a proportional solenoid;

[0045] Figure 6 It is a curve graph of the change in the number of shorted turns under different environments;

[0046] Figure 7 It is a schematic diagram of a method for simulating the degradation of the proportional solenoid of an electro-hydraulic proportional servo valve;

[0047] Figure 8 It is a schematic diagram of a fault simulation test bench for an electro-hydraulic proportional servo valve;

[0048] Figure 9 It is a schematic diagram of overlapping sampling;

[0049] Figure 10 It is a schematic flowchart of a health state evaluation process based on a Transformer model;

[0050] Figure 11 It is a loss curve graph during the training process of the Transformer model;

[0051] Figure 12 It is a curve graph of the change in accuracy during the training process of the Transformer model;

[0052] Figure 13 It is a loss curve graph during the verification process of the Transformer model;

[0053] Figure 14 It is a curve graph of the change in accuracy during the verification process of the Transformer model;

[0054] Figure 15 Schematic diagram of the confusion matrix of the evaluation results of this application under multiple groups;

[0055] Figure 16 Bar chart of statistical results under different groups;

[0056] Figure 17 Schematic diagram of the structure of a computer device provided by an embodiment of this application. Detailed implementation manners

[0057] Next, the technical solutions in the embodiments of this application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this application.

[0058] To make the above objects, features, and advantages of this application more obvious and understandable, the following further details this application in conjunction with the accompanying drawings and specific implementation manners.

[0059] In an exemplary embodiment, as Figure 1 shown, a method for determining the health status of an electro-hydraulic proportional servo valve is provided, including the following steps:

[0060] S1: Obtain the working data of the electro-hydraulic proportional servo valve to be detected; the working data includes spool position signal, spool position feedback signal, actuator position signal, current signal, hydraulic signal at port A, and hydraulic signal at port B.

[0061] S2: Preprocess the working data of the electro-hydraulic proportional servo valve to be detected to obtain the preprocessed working data; the preprocessing includes noise reduction processing, feature extraction, and feature screening.

[0062] As an optional implementation manner, S2 specifically includes:

[0063] S21: Use the wavelet transform method to perform noise reduction processing on the working data of the electro-hydraulic proportional servo valve to be detected to obtain the noise-reduced working data.

[0064] S22: Use the tsfresh toolbox to perform multi-information domain feature extraction on the noise-reduced working data to obtain the information domain features of the working data. The multi-information domain includes the time domain, frequency domain, and time-frequency domain.

[0065] S23: Use the maximum correlation minimum redundancy algorithm to perform feature screening on the information domain features of the working data to obtain the preprocessed working data.

[0066] S3: Based on the processed working data, use the electro-hydraulic proportional servo valve health status determination model to determine the health status of the electro-hydraulic proportional servo valve to be detected; wherein, the electro-hydraulic proportional servo valve health status determination model is obtained by training a Transformer model using a training data set; the training data set is obtained through an electro-hydraulic proportional servo valve fault simulation experiment.

[0067] As an alternative implementation, S3 specifically includes:

[0068] S31: Perform position encoding on the processed working data to obtain working data containing sequence position information.

[0069] S32: Based on the working data containing sequence position information, use the electro-hydraulic proportional servo valve health status determination model to determine the health status of the electro-hydraulic proportional servo valve to be detected.

[0070] As an alternative implementation, obtaining the training data set through an electro-hydraulic proportional servo valve fault simulation experiment specifically includes:

[0071] Establish a mathematical model of the electro-hydraulic proportional servo valve according to Newton's second law.

[0072] Establish a mathematical model of the proportional solenoid according to the working principle of the proportional solenoid in the electro-hydraulic proportional servo valve.

[0073] According to the working environment of the electro-hydraulic proportional servo valve, analyze the time when the inter-turn distance of the electromagnetic coil of the proportional solenoid in the electro-hydraulic proportional servo valve degrades with environmental factors, determine the time of inter-turn short circuit, and establish a performance degradation model of the electro-hydraulic proportional servo valve over time.

[0074] Based on the electro-hydraulic proportional servo valve mathematical model, the proportional solenoid mathematical model, the electro-hydraulic proportional servo valve performance degradation model over time, and the electro-hydraulic proportional servo valve fault simulation test bench, simulate different degrees of degradation of the proportional solenoid and collect the working data corresponding to different degrees of degradation.

[0075] Based on the working data corresponding to different degrees of degradation and the corresponding true fault states, establish a training data set.

[0076] In practical applications, such as Figure 2As shown, the method for determining the health status of the electro-hydraulic proportional servo valve in this application mainly includes the following six steps. In step 1, the working mathematical model of the electro-hydraulic proportional servo valve is established. In step 2, the mathematical model of the proportional solenoid valve is clarified. In step 3, the performance degradation model of the proportional solenoid valve is constructed. According to the performance degradation model in step 3, the degradation degree of the proportional solenoid valve over time can be obtained. After obtaining the degradation degree, the output force performance degradation of the proportional solenoid valve can be obtained according to the mathematical model in step 2. According to the performance degradation of the proportional solenoid valve, the degradation degree of the electro-hydraulic proportional servo valve can be obtained in step 1. Combining steps 1-3, each degradation stage of the electro-hydraulic proportional servo valve can be given at the model level. In step 4, a fault degradation simulation test bench for the electro-hydraulic proportional servo valve is established. In step 5, the characteristics of the data are determined. In step 6, a feature judgment method based on the Transformer model is determined to judge the degradation state of the current proportional servo valve. Its main judgment logic is as follows: through the degradation data extracted by the proportional servo valve degradation simulation test bench mentioned in step 4, the data characteristics in the current sensor are obtained according to the feature extraction method in step 5, and then the current characteristics are matched with the data characteristics in the model-based method according to the method in step 6 to determine the degradation stage of the electro-hydraulic proportional servo valve in the actual system. Combining the six steps, the degradation state of the electro-hydraulic proportional servo valve in the actual system is obtained.

[0077] Step 1: Establish a mathematical model of the electro-hydraulic proportional servo valve according to its working principle.

[0078] Step 1.1: According to Newton's second law, the mathematical model of the electro-hydraulic proportional servo valve is constructed as:

[0079]

[0080] Where, is the spool movement acceleration, m e is the spool mass, x c is the spool displacement, F c is the output force of the proportional solenoid valve. is the friction force, B e is the friction coefficient, is the spool velocity. K me x c is the spring force, K me is the spring stiffness coefficient. F d is a superimposed parameter used to consider unmodeled effects and external disturbances, such as flow forces.

[0081] Step 1.2: According to the law of electromagnetic induction, the voltage equation of the proportional solenoid valve is:

[0082]

[0083] Among them, u(t) is the voltage across the electromagnetic coil, i(t) is the current passing through the coil, and R is the resistance of the coil. The magnetic flux linkage λ(i(t), x(t)) is a function related to the current and the armature position, and x c (t) is the spool displacement. The structure of the electro-hydraulic proportional servo valve is as Figure 3 shown.

[0084] Step 1.3: For the proportional solenoid. Part of the amplifier output voltage is consumed on the coil resistance, and the other part is used to overcome the back electromotive force generated in the coil due to the change in the armature magnetic flux.

[0085] Step 2: Establish a mathematical model of the proportional solenoid according to the working principle of the proportional solenoid in the electro-hydraulic proportional servo valve.

[0086] Step 2.1: The proportional solenoid is as Figure 4 shown. After the coil is energized to generate a magnetic field, the magnetic conductive element in the magnetic field is magnetized to generate a magnetic circuit. Due to the existence of the magnetic isolation ring, the magnetic isolation ring divides the magnetic circuit into two parts. The first magnetic circuit φ1 passes through the axial non-working gap, pole shoe, housing and armature along the basin shape to form a closed magnetic circuit, generating an axial thrust. The second magnetic circuit φ2 passes through the radial working air gap, pole shoe, housing and armature along the edge of the basin to generate an axial additional force. The electromagnetic force of the proportional solenoid is obtained by combining the two:

[0087] Step 2.2: The axial electromagnetic force received by the armature is:

[0088]

[0089] Step 3: According to the working environment of the electro-hydraulic proportional servo valve, analyze the time when the inter-turn distance of the electromagnetic coil of the proportional solenoid degenerates with multiple environmental factors, estimate the time of inter-turn short circuit, and further construct a performance degradation model of the electro-hydraulic proportional servo valve over time.

[0090] Step 3.1: The creep degradation of the coil is related to multiple environmental factors. After the temperature rises, the stress between the coil turns increases significantly. The coil arrangement is as Figure 5 shown. The degree of polymerization of the insulating material decreases, resulting in electrical connection between two adjacent turns in the winding, causing inter-turn short circuit.

[0091] The decrease in the degree of polymerization of the inter-turn insulating material of the solenoid is significantly related to the temperature. The change in its degree of polymerization P follows the Arrhenius formula:

[0092]

[0093] Among them, P(t) is the degree of polymerization at time t; P0 is the degree of polymerization at the initial time; k(t) is the reaction rate at time t, and its value is determined by the environment:

[0094]

[0095] Among them, A is the process constant, E is the molar activation energy of the degradation process, R is the universal gas constant, and T is the temperature.

[0096] At time t, the internal probability density with degree of polymerization x is:

[0097]

[0098] Among them, μ is the failure threshold; σ is the variance.

[0099] Then the failure probability at time t is:

[0100] p th (x) = ∫p dp (x, t)dx.

[0101] Then the reliability of a single turn at time t is:

[0102] R(t) = 1 - ∫p th (x)p dp (x, t)dx.

[0103] The reliability of a coil with N turns at time t is:

[0104] R (i,j) = (1 - R j )R (i,j-1) + R j R (i-1,j-1) .

[0105] Among them, R (i,j) is the reliability that at least i turns are still working in the j - turn coil; R j is the reliability of the j - th coil; R (i,j-1) is the reliability that at least i - 1 turns are still working in the j - turn coil; R (i-1,j-1) is the reliability that at least i - 1 turns are still working in the (j - 1) - turn coil.

[0106] Then the coil turn number change curve is obtained as:

[0107] N(t) = N0·R (i,j) .

[0108] Among them, N0 is the initial turn number.

[0109] The change curve of N at different temperatures is as Figure 6 shown.

[0110] Step 4: Control the current of the proportional solenoid in the electro - hydraulic proportional servo valve by shunting with a parallel resistor, construct an electro - hydraulic proportional servo valve fault simulation test bench, and simulate different degrees of degradation of the proportional solenoid.

[0111] Step 4.1: The main manifestation of the aging of the proportional solenoid valve coil is the short circuit between turns of the coil, resulting in a decrease in the number of coil turns and further leading to a decrease in the output force. Therefore, when the same current signal is input, the output force of the proportional solenoid valve decreases. As Figure 7 shown, by connecting a resistor in series in the proportional solenoid valve coil circuit, the aging fault of the proportional solenoid valve coil is simulated. By connecting a rheostat in parallel in the proportional solenoid valve coil circuit, the performance degradation of the proportional solenoid valve is simulated by changing the current magnitude through changing the resistance value.

[0112] According to the degradation simulation analysis of the electro-hydraulic proportional servo valve, a fault simulation test bench for the electro-hydraulic proportional servo valve is constructed. The electro-hydraulic proportional servo valve to be measured is a four-way three-position valve and is installed on the valve test platform; the LVDT displacement sensor is installed inside the proportional solenoid valve to measure the spool position signal and the spool position feedback; the Hall current sensor is installed between the proportional solenoid valve coil in the valve and the valve controller to measure the current signal; the pressure sensors are respectively installed at port A and port B of the valve to obtain the hydraulic signal at port A of the oil and the hydraulic signal at port B of the oil in different health states of the valve. The spool drives the oil cylinder to reciprocate, and the sensor installed on the oil cylinder obtains the actuator position signal. The data acquisition frequency of each sensor is 100 Hz.

[0113] The hydraulic system mainly includes a platform support, an oil tank, a hydraulic circuit (oil pump, driving motor, hydraulic cylinder, electro-hydraulic proportional servo valve and other various valves, pipelines and joints). The electric control system mainly includes a power supply, a touch screen-industrial computer, a hydraulic axis motion controller, a data acquisition and control module, a high-frequency response proportional servo valve drive controller, a proportional relief valve amplifier, an electromagnet current transmitter, a motor start and protection component, a test control software, etc. The fault simulation test bench for the electro-hydraulic proportional servo valve is as Figure 8 shown. After the test bench is powered on, the electric control system controls the voltage and current to be injected into the proportional solenoid valve according to the system signal, and the proportional solenoid valve generates a magnetic force to drive the spool of the electro-hydraulic proportional servo valve to move.

[0114] Step 4.2: The data set used in this application is the data collected on the electro-hydraulic proportional servo valve fault simulation test bench. By connecting rheostats with different resistance values in parallel, the resistance values are 100 ohms, 9 ohms, 6 ohms, and 3 ohms respectively, and four different proportional solenoid valve fault states are obtained: normal, mild, moderate, and severe.

[0115] As an optional implementation manner, the Transformer model is trained by using the training data set, specifically including:

[0116] Preprocess the training data set to obtain the processed training data set.

[0117] Input the processed working data in the processed dataset into the current Transformer model to obtain the predicted fault status.

[0118] Determine the loss function value according to the predicted fault status and the corresponding true fault status.

[0119] Judge whether the training termination condition is satisfied; the training termination condition is reaching the maximum number of iterations or the loss function value is less than the set loss function threshold.

[0120] If so, use the current Transformer model as the electro-hydraulic proportional servo valve health status determination model.

[0121] If not, adjust the hyperparameters of the current Transformer model according to the loss function value, and return "Input the processed working data in the processed dataset into the current Transformer model to obtain the predicted fault status".

[0122] During the training process, the process of preprocessing the dataset is step 5.

[0123] Step 5: Process the data with different fault degrees collected from the electro-hydraulic proportional servo valve fault simulation test bench, use wavelet transform for noise reduction, further perform feature extraction to obtain the features at different degradation stages, and perform feature dimensionality reduction.

[0124] Step 5.1: To ensure that a single sample contains more fault information while reducing data redundancy, each fault status class intercepts 10 actuation cycles as a sample, samples the data with an overlap rate of 20%, and collects multiple samples for each fault degree, as Figure 9 shown. Each sample contains 6 types of signal data, namely spool position signal, spool position feedback, actuator position signal, current signal, hydraulic signal at port A, and hydraulic signal at port B.

[0125] Step 5.2: Since the signals collected under complex working conditions contain relatively large noise, which affects the accuracy of feature extraction, it is necessary to perform noise reduction on the collected signals to improve the noise reduction ratio of information. In this application, wavelet transform is used to eliminate the noise of the signals, and the dB4 wavelet is selected for three-layer noise reduction.

[0126] After selecting the samples, use the tsfresh toolbox to perform multi-information domain feature extraction on each signal of the samples, including time domain, frequency domain, and time-frequency domain features and other signals. 2550 features can be extracted for each sample.

[0127] Step 5.4: Excessive feature dimensions can easily affect the judgment of the current state. Therefore, to compress the feature dimensions, the maximum relevance minimum redundancy algorithm is used to screen the features. Feature screening can be divided into two processes: measuring feature relevance and solving redundancy between features. Mutual information is an index that can measure the correlation between two variables and also takes into account the non-linear relationship between features. The mutual information between variable X and variable Y is defined as:

[0128]

[0129] where X = {x 0:n}, Y = {y 0:n} represent two types of sensor signals respectively, which are a set of data. x j , y i are the specific values of sensor signals X and Y at a certain moment respectively. p(x j , y i ) is the joint distribution, p(x j ) and p(y i ) are the marginal distributions, and the mutual information I(X; Y) is the relative quotient of the joint distribution and the marginal distribution.

[0130] It makes a trade-off between correlation and redundancy in different ways and uses mutual information as the calculation criterion to measure the redundancy between features and the correlation between features and the target variable. Feature selection is carried out by maximizing the correlation between features and the target variable and minimizing the redundancy between features, ultimately making the differences between features very large and the correlation with the target variable also very large.

[0131] Maximize the mutual information between features and classes to select the features most relevant to class prediction. The objective function is as follows:

[0132]

[0133] where S represents the feature subset, c is the class label, and f i is the i-th feature belonging to the feature subset S. In this experiment, the features extracted from each sensor data form a feature subset.

[0134] The minimum redundancy condition aims to find a feature subset that minimizes the overlapping information between features. By minimizing the mutual information between features, it can be ensured that the selected features do not contain any redundant information. The objective function is as follows:

[0135]

[0136] Combining the two conditions of maximum relevance and minimum redundancy, the objective function of the MRMR (Maximal Relevance and Minimal Redundancy) algorithm can be obtained as follows:

[0137]

[0138] Finally, after removing redundant features, each sensor has 425 remaining features.

[0139] Step 6: Utilize the advantages of the encoder-decoder and attention mechanism of the Transformer model to complete the evaluation of the health state of the electro-hydraulic proportional servo valve. Its structure is different from that of the recurrent neural network and can process data in parallel. Compared with the recurrent neural network, the Transformer model can improve the training efficiency; in addition, it can embed position information into the input features, ensure the preservation of element order information during training, and effectively capture long-distance dependencies. According to the data performance, it can match the degradation situation of the proportional solenoid valve, and then estimate the degradation situation of the electro-hydraulic proportional servo valve, which can effectively avoid the occurrence of catastrophic accidents.

[0140] Step 6.1: The Transformer model is a network model with an encoder-decoder structure based on the attention mechanism, mainly composed of multiple encoders and decoders stacked together. Among them, the encoder is mainly used to encode the input feature sequence into an intermediate vector; the decoder is mainly used to decode the intermediate vector encoded by the encoder into an output label column. The steps used in this application are as Figure 10 shown.

[0141] After extracting the sample features, each sample is 6 groups of sensor data, and each group of sensors contains 425 features. Therefore, the data dimension is 6×425. To indicate the type of each data sample, a classification vector is added to the vector for learning category information during training, and then the sample size changes to 7×425.

[0142] The Transformer model requires position embedding to encode the position information of each group of vectors. This is mainly due to the perturbation invariance (Permutation-invariant) of the self-attention mechanism, that is, shuffling the order of each vector will not change the result. The core of the Transformer model is the attention mechanism, but the attention mechanism cannot learn sequence position information, so sequence position information is added through position encoding. The position encoding alternately encodes the even and odd dimensions of the input feature samples through trigonometric functions. Its encoding method is:

[0143]

[0144] Among them, pos represents the position index of the elements in the input sequence, i represents the encoding dimension, and d represents the dimension of the embedding vector in the model. For multi-sensors, there are significant physical relationships in the changes among the sensors. Therefore, there are obvious positional relationships in the data features among the channels. Therefore, positional encoding is added. Positional encoding can be understood as a table with a total of N rows, where the size of N is the same as the length of the input sequence. Each row represents a vector, and the dimension of the vector is the same as the dimension of the input sequence. After adding the positional encoding information, the dimension is still 7×425. After positional encoding, the sequence is input into the encoder.

[0145] The attention mechanism used in the Transformer model is called "Scaled Dot-Product Attention". The self-attention mechanism is a variant of the attention mechanism, which reduces the dependence on external information and is better at capturing the correlations inside the input feature sequence. The input of this module, in addition to the input sequence X with a dimension of 7×425, includes three vectors: the query vector Q, the key vector K, and the value vector V. Its output feature sequence is calculated as follows:

[0146] Q = W q X

[0147] K = W k X.

[0148] V = W v X

[0149]

[0150] Among them, W q , W k , W v are the weight matrices corresponding to Q, K, and V, and d model is the dimension of the model.

[0151] The multi-head self-attention mechanism is the concatenation and linear transformation of multiple self-attention mechanisms. The multi-head self-attention mechanism extends the ability to focus on different positions compared to a single self-attention mechanism. Its calculation method is as follows:

[0152]

[0153] Among them, H represents the output of each head, with a total of n. In the multi-head attention mechanism, the input query, key, and value in the model are linearly transformed into multiple different subspaces respectively. Each head has its own attention mechanism. In this experimental model, there are 4. H i represents the self-attention output of the i-th head, respectively represent the linear transformation weight matrices of the i-th head, which help it capture different features.

[0154] After passing through MSA (Multi-Head Self-Attention), and then through another layer of normalization, the dimension of the sequence remains 7×425.

[0155] Another important structure in the Transformer architecture is the MLP (Multilayer Perceptron). The multilayer perceptron consists of an input layer, an output layer, and hidden layers. Neurons in each hidden layer of the network can receive information transmitted from all neurons in the adjacent previous hidden layer, and after processing, output the information to all neurons in the adjacent subsequent hidden layer. In the multilayer perceptron, neurons in adjacent layers are usually connected in a "fully connected" manner. The multilayer perceptron simulates the function of complex non-linear functions in the encoder of the Transformer.

[0156] Step 6.2: According to the data features obtained in Step 5, initialize the model parameters, set the training dataset, validation dataset, and test dataset in a ratio of 7∶2∶1, and use multiple groups for testing to further verify the evaluation ability of the model. The changes in the loss function and accuracy of the training results are as Figure 11 and 12 shown. The loss results and accuracy of the validation dataset are as Figure 13 and Figure 14 shown. The confusion matrices obtained from the tests of each group are as Figure 15 shown.

[0157] Step 6.3: In the experiment, accuracy, precision, recall, and F1-score are introduced to evaluate in four aspects, and at the same time, the confusion matrix is used to evaluate the overall performance of the model. The bar charts of the statistical results under different groups are as Figure 16 shown.

[0158] The digital model of the electro-hydraulic proportional servo valve provided in this application is a mathematical model constructed based on the actual physical situation. By fully analyzing the multi-field coupling degradation mechanism of the proportional solenoid valve, a degradation model of the proportional solenoid valve over time under different environmental factors is established. The proposed method for simulating and reproducing the aging faults of the proportional solenoid valve coil can more accurately simulate the states of the electro-hydraulic proportional servo valve under different fault conditions. Training the Transformer model with the constructed training dataset can automatically and accurately identify the health state of the electro-hydraulic proportional servo valve.

[0159] Based on the same inventive concept, an embodiment of the present application further provides an electro-hydraulic proportional servo valve health state determination system for implementing the electro-hydraulic proportional servo valve health state determination method involved above. The implementation solution provided by this system to solve the problem is similar to the implementation solution described in the above method. Therefore, the specific limitations in the embodiments of the electro-hydraulic proportional servo valve health state determination system provided below can refer to the limitations on the electro-hydraulic proportional servo valve health state determination method in the above text, and will not be elaborated here.

[0160] In an exemplary embodiment, an electro-hydraulic proportional servo valve health state determination system is provided, including:

[0161] A data acquisition module, configured to acquire the working data of the electro-hydraulic proportional servo valve to be detected; the working data includes a spool position signal, a spool position feedback signal, an actuator position signal, a current signal, a hydraulic signal at port A, and a hydraulic signal at port B.

[0162] A preprocessing module, configured to preprocess the working data of the electro-hydraulic proportional servo valve to be detected to obtain the preprocessed working data; the preprocessing includes noise reduction processing, feature extraction, and feature screening.

[0163] A health state determination module, configured to determine the health state of the electro-hydraulic proportional servo valve to be detected according to the preprocessed working data by using an electro-hydraulic proportional servo valve health state determination model; wherein, the electro-hydraulic proportional servo valve health state determination model is obtained by training a Transformer model using a training data set; the training data set is obtained through an electro-hydraulic proportional servo valve fault simulation experiment.

[0164] In an exemplary embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the above-mentioned electro-hydraulic proportional servo valve health state determination method is implemented.

[0165] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the above-mentioned electro-hydraulic proportional servo valve health state determination method is implemented.

[0166] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the above-mentioned electro-hydraulic proportional servo valve health state determination method is implemented.

[0167] In an exemplary embodiment, a computer device is provided. The computer device can be a server or a terminal, and its internal structure diagram can be as Figure 17As shown in the figure. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it realizes a method for determining the health state of an electro-hydraulic proportional servo valve.

[0168] Those skilled in the art can understand that Figure 17 the structure shown in the figure is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0169] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0170] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memory can include Read-Only Memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0171] The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0172] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0173] Specific examples are used in this article to elaborate on the principles and implementation manners of the present application. The descriptions of the above embodiments are only used to help understand the method and its core idea of the present application; at the same time, for those of ordinary skill in the art, according to the idea of the present application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present application.

Claims

1. A method for determining the health state of an electro-hydraulic proportional servo valve, characterized in that, Including: Obtain the working data of the electro-hydraulic proportional servo valve to be detected; the working data includes spool position signal, spool position feedback signal, actuator position signal, current signal, hydraulic signal at port A, and hydraulic signal at port B; Preprocess the working data of the electro-hydraulic proportional servo valve to be detected to obtain the processed working data; the preprocessing includes noise reduction processing, feature extraction, and feature screening; According to the processed working data, use the electro-hydraulic proportional servo valve health status determination model to determine the health status of the electro-hydraulic proportional servo valve to be detected; wherein, the electro-hydraulic proportional servo valve health status determination model is obtained by training a Transformer model using a training dataset; the training dataset is obtained through an electro-hydraulic proportional servo valve fault simulation experiment.

2. The method for determining the health state of the electro-hydraulic proportional servo valve according to claim 1, wherein Preprocess the working data of the electro-hydraulic proportional servo valve to be detected to obtain the processed working data, specifically including: Use the wavelet transform method to perform noise reduction processing on the working data of the electro-hydraulic proportional servo valve to be detected to obtain the noise-reduced working data; Use the tsfresh toolbox to perform multi-information domain feature extraction on the noise-reduced working data to obtain the information domain features of the working data; Use the maximum correlation minimum redundancy algorithm to perform feature screening on the information domain features of the working data to obtain the processed working data.

3. The method for determining the health state of the electro-hydraulic proportional servo valve according to claim 2, wherein The multi-information domain includes the time domain, frequency domain, and time-frequency domain.

4. The method for determining the health state of the electro-hydraulic proportional servo valve according to claim 1, characterized in that According to the processed working data, use the electro-hydraulic proportional servo valve health status determination model to determine the health status of the electro-hydraulic proportional servo valve to be detected, specifically including: Perform position encoding on the processed working data to obtain the working data containing sequence position information; According to the working data containing sequence position information, use the electro-hydraulic proportional servo valve health status determination model to determine the health status of the electro-hydraulic proportional servo valve to be detected.

5. The method for determining the health state of the electro-hydraulic proportional servo valve according to claim 1, characterized in that, Obtain the training dataset through an electro-hydraulic proportional servo valve fault simulation experiment, specifically including: Establish a mathematical model of the electro-hydraulic proportional servo valve according to Newton's second law; Establish a mathematical model of the proportional solenoid according to the working principle of the proportional solenoid in the electro-hydraulic proportional servo valve; According to the working environment of the electro-hydraulic proportional servo valve, analyze the time of the degradation of the inter-turn distance of the electromagnetic coil of the proportional solenoid in the electro-hydraulic proportional servo valve with respect to environmental factors, determine the time of inter-turn short circuit, and establish a performance degradation model of the electro-hydraulic proportional servo valve over time; Based on the electro-hydraulic proportional servo valve mathematical model, the proportional solenoid mathematical model, the electro-hydraulic proportional servo valve performance degradation model over time, and the electro-hydraulic proportional servo valve fault simulation test bench, simulate different degrees of degradation of the proportional solenoid and collect the working data corresponding to different degrees of degradation; Based on the working data corresponding to different degrees of degradation and the corresponding true fault states, establish a training dataset.

6. The method for determining the health state of the electro-hydraulic proportional servo valve according to claim 1, wherein Use the training dataset to train the Transformer model, specifically including: Preprocess the training dataset to obtain the processed training dataset; Input the processed working data in the processed dataset into the current Transformer model to obtain the predicted fault state; Determine the loss function value according to the predicted fault state and the corresponding true fault state; Judge whether the end training condition is satisfied; the end training condition is reaching the maximum number of iterations or the loss function value is less than the set loss function threshold; If so, use the current Transformer model as the electro-hydraulic proportional servo valve health state determination model; If not, adjust the hyperparameters of the current Transformer model according to the loss function value, and return "Input the processed working data in the processed data set into the current Transformer model to obtain the predicted fault state".

7. An electro-hydraulic proportional servo valve health status determination system, characterized in that Including: A data acquisition module, configured to acquire the working data of the electro-hydraulic proportional servo valve to be detected; the working data includes a spool position signal, a spool position feedback signal, an actuator position signal, a current signal, an A-port hydraulic signal, and a B-port hydraulic signal; A preprocessing module, configured to preprocess the working data of the electro-hydraulic proportional servo valve to be detected to obtain the processed working data; the preprocessing includes noise reduction processing, feature extraction, and feature screening; A health state determination module, configured to determine the health state of the electro-hydraulic proportional servo valve to be detected according to the processed working data by using the electro-hydraulic proportional servo valve health state determination model; wherein, the electro-hydraulic proportional servo valve health state determination model is obtained by training a Transformer model using a training data set; the training data set is obtained through an electro-hydraulic proportional servo valve fault simulation experiment.

8. A computer device, comprising: A memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to implement the electro-hydraulic proportional servo valve health state determination method according to any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the electro-hydraulic proportional servo valve health state determination method according to any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the electro-hydraulic proportional servo valve health state determination method according to any one of claims 1-6.

Images