Method and system for identifying grounding insulation state of on-train power supply system of train

Abstract

The present invention provides a method for identifying the grounding insulation state of a train on-vehicle power supply system, which relates to the technical field of train on-vehicle power supply systems. The method includes, under different grounding fault positions, constructing characteristic variables based on the relationship between the obtained DC bus voltage and the grounding detection voltage, and respectively constructing a characteristic vector set and a classification enabling mechanism according to the characteristic variables; judging whether the train on-vehicle power supply system is in a fault state based on the classification enabling mechanism, and constructing a total data set of the characteristic vectors as the input of the grounding fault classification model to output the occurrence position of the grounding fault; selecting the corresponding voltage signal as the input of the real-time prediction model of the fault grounding resistance to output the real-time prediction result of the fault grounding resistance; and identifying the grounding insulation state of the train on-vehicle power supply system according to the real-time prediction result of the fault grounding resistance. The present invention effectively improves the anti-interference ability and the fault diagnosis accuracy, reduces the resistance prediction error, and ensures the accuracy of the grounding insulation state identification.

Description

Technical Field

[0001] The present invention relates to the technical field of train on-vehicle power supply systems, and particularly relates to a method and system for identifying the grounding insulation state of a train on-vehicle power supply system. Background Art

[0002] The train on-vehicle power supply system provides power for various electrical loads on the train. As the service time increases, the train on-vehicle power supply system is prone to being affected by equipment aging and environmental factors, resulting in electrical faults. Such faults may lead to a poor riding experience and even pose safety problems. In order to ensure the safe and reliable operation of the train and provide passengers with a comfortable riding experience, an efficient and reliable fault diagnosis method becomes particularly important.

[0003] In recent years, the research on grounding fault diagnosis has mainly focused on the grounding faults of transmission lines in power transmission systems. Different from this, the train on-vehicle power supply system is directly connected to multiple converter modules with almost no transmission distance. Its grounding faults are mainly caused by abnormal grounding due to insulation deterioration at different positions in the converter system. The research on accurately and quickly identifying and locating grounding faults in the train on-vehicle power supply system is not yet mature.

[0004] The prior art discloses a method for identifying and locating grounding faults in the power supply system of a diesel locomotive based on time series characteristics. This method uses historical system signal data to construct a feature vector with fault location discrimination, extracts feature indicators within a sliding window, attaches time series characteristics, improves the distinguishability of fault types, reduces noise interference, constructs a grounding fault location model through machine learning, and explores the non-linear relationship between features and fault types to achieve fault location. However, this solution fails to further obtain the magnitude of the grounding insulation resistance to evaluate the severity of the system grounding fault. The prior art also discloses a method for evaluating the grounding insulation state. This method first obtains operating parameters and states, and determines whether there is a grounding fault based on the operating parameters and states; if so, obtains feature indicator data and operating condition information based on the operating parameters and states; obtains the fault type based on the feature indicators and the operating condition information; obtains the grounding insulation resistance based on the fault type; and evaluates the grounding insulation state based on the grounding insulation resistance. However, this method uses a decision tree based on inequality relationships to make judgments during the process of fault classification and judgment, which is difficult to cope with noise and uncertainty, lacks adaptability, and has low real-time fault diagnosis accuracy; at the same time, the obtained grounding insulation resistance is only calculated through a physical model, which is greatly affected by signal fluctuations and interference, resulting in a large resistance prediction error, thus affecting the accuracy of grounding insulation state identification. Summary of the Invention

[0005] To solve the problems in the above-mentioned existing technologies, such as poor anti-interference ability, low real-time fault diagnosis accuracy, large resistance prediction error, and poor accuracy in identifying the grounding insulation state, the present invention proposes a method and system for identifying the grounding insulation state of a train on-board power supply system, which effectively improves the anti-interference ability and fault diagnosis accuracy, reduces the resistance prediction error, and ensures the accuracy of identifying the grounding insulation state.

[0006] To achieve the above technical effects, the technical solution of the present invention is as follows:

[0007] A method for identifying the grounding insulation state of a train on-board power supply system, the method comprising the following steps:

[0008] S1. Obtain the relationship between the DC bus voltage and the grounding detection voltage under the normal state of the train on-board power supply system;

[0009] S2. Under different grounding fault positions, based on the relationship between the DC bus voltage and the grounding detection voltage, construct characteristic variables, and respectively construct a characteristic vector set and a classification enabling mechanism according to the characteristic variables;

[0010] S3. Based on the classification enabling mechanism, judge whether the train on-board power supply system is in a fault state. If so, execute S4; if not, it is in a normal state;

[0011] S4. Use the characteristic vector set to construct a total data set of characteristic vectors under different grounding fault states;

[0012] S5. Take the total data set as the input of a preset grounding fault classification model, and output the grounding fault occurrence position by the grounding fault classification model;

[0013] S6. Based on the grounding fault occurrence position, select the voltage signal corresponding to the grounding fault occurrence position as the input of a preset real-time prediction model for the fault grounding resistance, and output the real-time prediction result of the fault grounding resistance;

[0014] S7. Identify the grounding insulation state of the train on-board power supply system according to the real-time prediction result of the fault grounding resistance.

[0015] Preferably, for S1, the calculation expression for obtaining the relationship between the DC bus voltage U dc and the grounding detection voltage U jd is as follows:

[0016]

[0017] S1 further includes obtaining the relationship between the fault grounding resistance and the DC bus voltage U dc , the grounding detection voltage U jd under different grounding fault positions.

[0018] Preferably, based on the relationship between the DC bus voltage and the grounding detection voltage in S2, a characteristic variable is constructed The calculation expression is as follows:

[0019]

[0020] According to the characteristic variable The process of constructing the characteristic vector set includes:

[0021]

[0022] wherein, represents the mean value of the characteristic variable F within a sliding window period at the k-th moment, N represents the total number of sampling points within a sliding window, , represents the variance of the characteristic variable F within a sliding window period at the k-th moment, represents the complex spectrum value of the grounding detection voltage U jd at the frequency f, represents the value of the i-th sampling point of the grounding detection voltage U jd , represents the harmonic amplitude, represents the complex spectrum value of the grounding detection voltage U jd at the m-fold frequency f, represents the total harmonic distortion.

[0023] Preferably, the classification enabling mechanism includes:

[0024] Based on the characteristic variable F, calculate the cumulative deviation S k of the characteristic variable F as follows:

[0025]

[0026] wherein, E represents the mean value of the characteristic variable F, and V represents the variance of the characteristic variable F;

[0027] Judge whether the cumulative deviation S k is greater than the fault threshold h. If so, the on-train power supply system of the train is in a fault state; if not, the on-train power supply system of the train is in a normal state.

[0028] Preferably, the calculation expression for constructing the total data set D of the characteristic vectors in different grounding fault states by using the characteristic vector set is as follows:

[0029]

[0030] wherein, Represents the set of characteristic vectors for the positive pole grounding fault in the DC link within the sliding window period, Represents the mean value of the characteristic variable F for the positive pole grounding fault in the DC link within the sliding window period, Represents the variance of the characteristic variable F for the positive pole grounding fault in the DC link within the sliding window period, Represents the harmonic amplitude of the positive pole grounding fault in the DC link, Represents the total harmonic distortion of the positive pole grounding fault in the DC link; Represents the set of characteristic vectors for the negative pole grounding fault in the DC link within the sliding window period, Represents the mean value of the characteristic variable F for the negative pole grounding fault in the DC link within the sliding window period, Represents the variance of the characteristic variable F for the negative pole grounding fault in the DC link within the sliding window period, Represents the harmonic amplitude of the negative pole grounding fault in the DC link, Represents the total harmonic distortion of the negative pole grounding fault in the DC link; Represents the set of characteristic vectors for the grounding fault at the front end of the reactor within the sliding window period, Represents the mean value of the characteristic variable F for the grounding fault at the front end of the reactor within the sliding window period, Represents the variance of the characteristic variable F for the grounding fault at the front end of the reactor within the sliding window period, Represents the harmonic amplitude of the grounding fault at the front end of the reactor, Represents the total harmonic distortion of the grounding fault at the front end of the reactor; Represents the set of characteristic vectors for the positive pole grounding fault of the RC input within the sliding window period, Represents the mean value of the characteristic variable F for the positive pole grounding fault of the RC input within the sliding window period, Represents the variance of the characteristic variable F for the positive pole grounding fault of the RC input within the sliding window period, Represents the harmonic amplitude of the positive pole grounding fault of the RC input, Represents the total harmonic distortion of the positive pole grounding fault of the RC input; Represents the set of characteristic vectors for the negative pole grounding fault of the RC input within the sliding window period, Represents the mean value of the characteristic variable F for the negative pole grounding fault of the RC input within the sliding window period, Represents the variance of the characteristic variable F for the negative pole grounding fault of the RC input within the sliding window period, Represents the harmonic amplitude of the negative pole grounding fault of the RC input, Represents the total harmonic distortion of the negative pole grounding fault of the RC input.

[0031] Preferably, the ground fault classification model includes an input layer, a convolutional layer, a pooling layer, a first fully connected layer, and an output layer connected in sequence. The convolutional kernel size of the convolutional layer is 3×33, the pooling layer is a max pooling layer, the pooling window size is 2×22, the number of nodes in the first fully connected layer is 50, the output layer is 6-dimensional, and the activation function of the output layer is softmax.

[0032] Preferably, the ground fault classification model is trained using backpropagation and gradient descent methods until the number of training rounds reaches a preset threshold or the loss function converges, then the training of the ground fault classification model ends, and a trained ground fault classification model is obtained.

[0033] Preferably, the fault grounding resistance real-time prediction model includes a plurality of second fully connected layers, each of the second fully connected layers has 64 neurons, and uses the ReLU activation function for output.

[0034] Preferably, the total loss function used by the fault grounding resistance real-time prediction model during training is as follows:

[0035]

[0036] where, represents the total loss function, represents the data error, represents the first physical constraint error, represents the weight coefficient of the first physical constraint error, represents the second physical constraint error, represents the weight coefficient of the second physical constraint error, represents the third physical constraint error, represents the weight coefficient of the third physical constraint error, represents the fourth physical constraint error, represents the weight coefficient of the fourth physical constraint error, represents the fifth physical constraint error, represents the weight coefficient of the fifth physical constraint error.

[0037] The present invention also proposes a ground insulation state recognition system for a train on-vehicle power supply system. The ground insulation state recognition system is implemented based on the ground insulation state recognition method for the train on-vehicle power supply system, and includes:

[0038] An acquisition module, configured to acquire the relationship between the DC bus voltage and the ground detection voltage under the normal state of the train on-vehicle power supply system;

[0039] A first construction module, configured to construct characteristic variables based on the relationship between the DC bus voltage and the grounding detection voltage under different grounding fault positions, and respectively construct a characteristic vector set and a classification enabling mechanism according to the characteristic variables;

[0040] A judgment module, configured to judge whether the on-train power supply system is in a fault state based on the classification enabling mechanism;

[0041] A second construction module, configured to construct a total data set of characteristic vectors in different grounding fault states by using the characteristic vector set when the on-train power supply system is in a fault state;

[0042] A grounding fault classification module, configured to use the total data set as the input of a preset grounding fault classification model, and output the grounding fault occurrence position by the grounding fault classification model;

[0043] A fault grounding resistance real-time prediction module, configured to select the voltage signal corresponding to the grounding fault occurrence position as the input of a preset fault grounding resistance real-time prediction model based on the grounding fault occurrence position, and output the real-time prediction result of the fault grounding resistance;

[0044] A grounding insulation state identification module, configured to identify the grounding insulation state of the on-train power supply system according to the real-time prediction result of the fault grounding resistance.

[0045] Compared with the prior art, the beneficial effects of the technical solution of the present invention are:

[0046] The present invention provides a method and system for identifying the grounding insulation state of an on-train power supply system. First, by obtaining the relationship between the DC bus voltage and the grounding detection voltage of the on-train power supply system in a normal state, and combining the extraction of multi-dimensional characteristic variables under different grounding fault positions to construct a characteristic vector set and a classification enabling mechanism, the anti-interference ability of the system is significantly improved; then, after confirming the fault of the on-train power supply system through the classification enabling mechanism, a total data set of characteristic vectors in different grounding fault states is constructed, and the total data set is used as the input of a preset grounding fault classification model to output the grounding fault occurrence position, realizing data-driven fault classification and effectively improving the real-time diagnosis accuracy of the grounding fault occurrence position; then, by selecting the voltage signal corresponding to the grounding fault occurrence position as the input of a preset fault grounding resistance real-time prediction model to output the real-time prediction result of the fault grounding resistance, the prediction error of the resistance is effectively reduced, ensuring the accuracy of the real-time prediction result of the fault grounding resistance; further, according to the real-time prediction result of the fault grounding resistance, the grounding insulation state of the on-train power supply system is identified, effectively improving the accuracy of the grounding insulation state identification. Description of the Drawings

[0047] Figure 1Flow chart showing a method for identifying the grounding insulation state of a train on - vehicle power supply system proposed in an embodiment of the present invention;

[0048] Figure 2 Circuit structure diagram of a train on - vehicle power supply system proposed in an embodiment of the present invention;

[0049] Figure 3 Principle block diagram for identifying the grounding insulation state of a train on - vehicle power supply system proposed in an embodiment of the present invention;

[0050] Figure 4 Structure block diagram of a system for identifying the grounding insulation state of a train on - vehicle power supply system proposed in an embodiment of the present invention. Detailed implementation manners

[0051] The drawings are only for illustrative purposes and should not be construed as a limitation of this patent;

[0052] For those skilled in the art, it is understandable that some well - known content descriptions in the drawings may be omitted;

[0053] The technical solutions of the present invention will be further described below with reference to the drawings and embodiments.

[0054] Embodiment 1

[0055] As Figure 1 shown, this embodiment proposes a method for identifying the grounding insulation state of a train on - vehicle power supply system, and the method includes the following steps:

[0056] S1. Obtain the relationship between the DC bus voltage and the grounding detection voltage under the normal state of the train on - vehicle power supply system;

[0057] In S1, referring to Figure 2 , the train on - vehicle power supply system includes a vacuum circuit breaker, a traction transformer, a synchronous transformer TV, a current transformer TA, a first grounding resistor , a second grounding resistor , a third grounding resistor , a fourth grounding resistor , a fifth grounding resistor , a first diode VD1, a second diode VD2, a first thyristor VT3, a second thyristor VT4, an inductor L, a capacitor C, a first detection resistor , a second detection resistor , a third detection resistor , a DC bus voltage sensor , a grounding detection circuit voltage sensor and a control system, one end of the vacuum circuit breaker is connected to the external power supply network, the other end of the vacuum circuit breaker is connected to the primary side of the traction transformer, one end of the secondary side of the traction transformer is respectively connected to one end of the synchronous transformer TV and the current transformer TA, and the other end of the secondary side of the traction transformer is respectively connected to the other end of the synchronous transformer TV, one end of the fifth grounding resistor , the positive electrode end of the first thyristor VT3 and the negative electrode end of the second thyristor VT4, and the other end of the fifth grounding resistor is grounded. The other end of the current transformer TA is respectively connected to one end of the fourth grounding resistor , the positive electrode end of the first diode VD1 and the negative electrode end of the second diode VD2. The other end of the fourth grounding resistor is grounded. The negative electrode end of the first diode VD1 is respectively connected to the negative electrode end of the first thyristor VT3, one end of the third grounding resistor and one end of the inductor L. The other end of the third grounding resistor is grounded. The other end of the inductor L is respectively connected to one end of the capacitor C, one end of the first grounding resistor and one end of the first detection resistor . The other end of the first grounding resistor is grounded. The other end of the first detection resistor is respectively connected to one end of the second detection resistor and one end of the third detection resistor . The other end of the third detection resistor is grounded. The positive electrode end of the second diode VD2 is respectively connected to the positive electrode end of the second thyristor VT4, the other end of the capacitor C, the other end of the second grounding resistor and the other end of the second detection resistor ; The DC bus voltage sensor is connected in parallel on both sides of the circuit composed of the first detection resistor and the third detection resistor ; The grounding detection circuit voltage sensor is connected in parallel on both sides of the circuit composed of the first detection resistor and the second detection resistor ; The control system is electrically connected to the synchronous transformer TV, the current transformer TA, the DC bus voltage sensor and the grounding detection circuit voltage sensor .

[0058] Among them, the first detection resistor , the second detection resistor , the third detection resistor , the DC bus voltage sensor and the grounding detection circuit voltage sensor The ground detection circuit is composed of R1 equal to R2, the inductor L and the capacitor C form an intermediate DC circuit, the first diode VD1, the second diode VD2, the first thyristor VT3 and the second thyristor VT4 form a single-phase half-controlled rectifier; DC bus voltage sensor The collected voltage is the DC bus voltage U dc , ground detection circuit voltage sensor The collected voltage is the ground detection voltage U jd ;

[0059] The working principle of the train onboard power supply system is as follows:

[0060] The train onboard power supply system obtains single-phase 25kV AC power from the external power supply network and converts it into 600V DC power to power the electrical loads in the train compartment. The train onboard power supply system consists of a traction transformer, a single-phase half-controlled rectifier, an intermediate DC circuit and a ground detection circuit. The single-phase half-controlled rectifier can convert AC power into DC power. In order to obtain a stable 600V DC voltage, the control system collects the secondary side AC voltage U s (sampled by synchronous transformer TV), secondary side AC current I s (measured by current transformer TA) and DC bus voltage U dc (measured by DC bus voltage sensor VS1) for closed loop control. At the same time, the parallel reactance L and supporting capacitor C are used for filtering. The first grounding resistor at each location , Second grounding resistance , the third grounding resistance , the fourth grounding resistor and the fifth grounding resistor The range is "megohm level". Due to the harsh working environment, the resistance value of the grounding resistor at each location will always deteriorate. In engineering practice, when the resistance value is greater than 20000Ω, the grounding is considered normal.

[0061] S1 obtains the DC bus voltage U of the train onboard power supply system in a normal state dc With ground detection voltage U jd The calculation expression of the relationship is as follows:

[0062]

[0063] S1 also includes obtaining the fault grounding resistance and DC bus voltage U under different grounding fault positions dc , ground detection voltage U jd The relationship includes:

[0064] See also Figure 2 At point ①, when the first grounding resistance When deterioration causes the resistance value to decrease, a DC link positive pole grounding fault occurs in the train on-vehicle power supply system, where U jd is a DC quantity proportional to U dc . As the degree of grounding insulation damage intensifies, when the first grounding resistance R g1 gradually decreases to 0, U jd will gradually decrease from 0.5U dc to 0; According to circuit theory, the relationship between R g1 and U dc and U jd is expressed as:

[0065]

[0066] See Figure 2 at ② in g2 . When deterioration of the second grounding resistance R jd causes the resistance value to decrease, a DC link negative pole grounding fault occurs in the train on-vehicle power supply system, where U dc is also a DC quantity proportional to U g2 . As the degree of grounding insulation damage intensifies, when the second grounding resistance R jd gradually decreases to 0, U dc will gradually increase from 0.5U dc to U g2 ; According to circuit theory, the relationship between R dc and U jd is expressed as:

[0067]

[0068] See Figure 2 at ③ in g3 . When deterioration of R s causes the resistance value to decrease, a grounding fault occurs at the front end of the reactor in the train on-vehicle power supply system, where Us is the secondary side AC voltage of the traction transformer, I ab is the secondary side AC current of the traction transformer, and U s is the first AC component strongly correlated with the secondary side AC voltage Us and the secondary side AC current I g3 . When the grounding resistance R jd is a fixed value, only when U ab is also an AC quantity can it cancel out with U g3 ; According to circuit theory, the relationship between R dc and U jd is expressed as:

[0069]

[0070] The first AC component U abThe calculation expression is:

[0071]

[0072] See Figure 2 at ④ in g4 When the deterioration causes the resistance value to decrease, an RC input positive pole grounding fault occurs in the train on-vehicle power supply system. Among them, U cb1 is the second AC component strongly related to the secondary side AC voltage U s and the secondary side AC current I s When R g4 is a fixed value, only when U jd is also an AC quantity can it cancel out with U cb1 According to the circuit principle, the relationship between R g4 and U dc and U jd is expressed as:

[0073]

[0074] The calculation expression of the second AC component U cb1 is:

[0075]

[0076] See Figure 2 at ⑤ in g5 When the deterioration causes the resistance value to decrease, an RC input negative pole grounding fault occurs in the train on-vehicle power supply system. Among them, U cb2 is the third AC component strongly related to the secondary side AC voltage U s and the secondary side AC current I s When R g5 is a fixed value, only when U jd is also an AC quantity can it cancel out with U cb2 According to the circuit principle, the relationship between R g5 and U dc and U jd can be expressed as:

[0077] According to the circuit principle, the relationship between U dc and U jd is expressed as:

[0078]

[0079] The third AC component The calculation expression is:

[0080]

[0081] S2. Based on the relationship between the DC bus voltage and the grounding detection voltage at different grounding fault positions, construct characteristic variables, and respectively construct a feature vector set and a classification enabling mechanism according to the characteristic variables;

[0082] In S2, analyze the grounding fault mechanism at different grounding fault positions, and construct the characteristic variables based on the relationship between the DC bus voltage and the grounding detection voltage The calculation expression is as follows:

[0083]

[0084] According to the characteristic variables Construct the feature vector set The calculation process includes:

[0085] For the normal operating state, DC link positive pole grounding fault, and DC link negative pole grounding fault types, distinguish them through the mean E(k) and variance V(k) of a sliding window F at the k-th moment. The calculation expressions of the mean E(k) and variance V(k) are as follows:

[0086]

[0087] For the three types of grounding faults at the front end of the reactor, RC input positive pole grounding fault, and RC input negative pole grounding fault, because its U jd contains DC components and AC components, distinguish them through the mean E and variance V of F, and in combination with the harmonic amplitude H m and the total harmonic distortion THD. Perform a fast Fourier transform on U jd to extract each harmonic component, and calculate the m-th harmonic amplitude H m and the total harmonic distortion THD as follows:

[0088]

[0089] According to the characteristic variables Construct the feature vector set , FV is an m + 3-dimensional feature vector set composed of time-domain and frequency-domain features. The feature vector set The mathematical expression is as follows:

[0090]

[0091] Among them, represents the mean of the characteristic variable F within a sliding window period at the k-th moment, N represents the total number of sampling points within a sliding window, , represents the variance of the characteristic variable F within a sliding window period at the k-th moment, and k is the index at a certain moment. Indicates the grounding detection voltage U jd The complex spectrum value at frequency f, which contains the amplitude and phase information of this frequency component; Indicates the grounding detection voltage U jd The value of the i-th sampling point of, representing the sampling data of the original sampling signal within the sliding window; Indicates the harmonic amplitude, Indicates the grounding detection voltage U jd The complex spectrum value at m times the frequency f, Indicates the total harmonic distortion.

[0092] S3. Based on the classification enabling mechanism, determine whether the train on-vehicle power supply system is in a fault state. If so, execute S4; if not, it is in a normal state;

[0093] In S3, the classification enabling mechanism includes:

[0094] The characteristic variable F can be used to detect whether a fault occurs in the train on-vehicle power supply system to determine whether to enable the subsequent grounding fault classification model; according to formula (7), under normal circumstances, the value of F is 0. However, due to the existence of on-site noise and measurement errors, the characteristic variable F may fluctuate near 0, affecting fault detection; to eliminate the interference of noise on fault detection, based on the characteristic variable F, the CUSUM control chart method is used to judge whether a fault occurs, and the cumulative deviation S of the characteristic variable F is calculated k The following formula:

[0095]

[0096] where, E represents the mean value of the characteristic variable F, V represents the variance of the characteristic variable F; is used as a smoothing constant;

[0097] Judge the cumulative deviation S k Whether it is greater than the fault threshold h. If so, the train on-vehicle power supply system is in a fault state and the subsequent grounding fault classification model is started; if not, the train on-vehicle power supply system is in a normal state.

[0098] S4. Use the feature vector set to construct the total data set of feature vectors under different grounding fault states;

[0099] In S4, through MATLAB-Simulink, fault simulation of the train on-vehicle power supply system is carried out, and the obtained data of the system in normal state and faulty state are used as the original data set. The original data set includes the normal state and five types of fault states, a total of six groups of data; extract the feature vectors FV1 to FV5 of the five types of fault states within each sliding window period from the feature vector set FV, and the calculation expression for constructing the total data set D of feature vectors under different grounding fault states is as follows:

[0100]

[0101] Among them, represents the set of feature vectors of the positive pole grounding fault in the DC link within the sliding window period, represents the mean value of the feature variable F of the positive pole grounding fault in the DC link within the sliding window period, represents the variance of the feature variable F of the positive pole grounding fault in the DC link within the sliding window period, represents the harmonic amplitude of the positive pole grounding fault in the DC link, represents the total harmonic distortion of the positive pole grounding fault in the DC link; represents the set of feature vectors of the negative pole grounding fault in the DC link within the sliding window period, represents the mean value of the feature variable F of the negative pole grounding fault in the DC link within the sliding window period, represents the variance of the feature variable F of the negative pole grounding fault in the DC link within the sliding window period, represents the harmonic amplitude of the negative pole grounding fault in the DC link, represents the total harmonic distortion of the negative pole grounding fault in the DC link; represents the set of feature vectors of the grounding fault at the front end of the reactor within the sliding window period, represents the mean value of the feature variable F of the grounding fault at the front end of the reactor within the sliding window period, represents the variance of the feature variable F of the grounding fault at the front end of the reactor within the sliding window period, represents the harmonic amplitude of the grounding fault at the front end of the reactor, represents the total harmonic distortion of the grounding fault at the front end of the reactor; represents the set of feature vectors of the positive pole grounding fault of the RC input within the sliding window period, represents the mean value of the feature variable F of the positive pole grounding fault of the RC input within the sliding window period, represents the variance of the feature variable F of the positive pole grounding fault of the RC input within the sliding window period, represents the harmonic amplitude of the positive pole grounding fault of the RC input, represents the total harmonic distortion of the positive pole grounding fault of the RC input; represents the set of feature vectors of the negative pole grounding fault of the RC input within the sliding window period, represents the mean value of the feature variable F of the negative pole grounding fault of the RC input within the sliding window period, represents the variance of the feature variable F of the negative pole grounding fault of the RC input within the sliding window period, represents the harmonic amplitude of the negative pole grounding fault of the RC input, Represents the total harmonic distortion of the RC input negative terminal grounding fault; the total data set D contains five types of fault data feature vector indicators, namely FV1 to FV5. The total data set D describes the fault characteristics of different types, provides the signal characteristics of the system under different faults, and provides training samples for the subsequent grounding fault classification model.

[0102] S5. Use the total data set as the input of the preset grounding fault classification model, and the grounding fault classification model outputs the grounding fault occurrence location.

[0103] S6. Based on the grounding fault occurrence location, select the voltage signal corresponding to the grounding fault occurrence location as the input of the preset fault grounding resistance real-time prediction model, and output the real-time prediction result of the fault grounding resistance.

[0104] S7. Identify the grounding insulation state of the train on-board power supply system according to the real-time prediction result of the fault grounding resistance.

[0105] In this embodiment, first, by obtaining the relationship between the DC bus voltage and the grounding detection voltage of the train on-board power supply system in the normal state, and combining the extraction of multi-dimensional characteristic variables under different grounding fault positions to construct a feature vector set and a classification enabling mechanism, the anti-interference ability of the system is significantly improved; then, after confirming the fault of the train on-board power supply system through the classification enabling mechanism, construct the total data set of feature vectors under different grounding fault states, use the total data set as the input of the preset grounding fault classification model, and output the grounding fault occurrence location, realizing data-driven fault classification and effectively improving the real-time diagnosis accuracy of the grounding fault occurrence location; then, by selecting the voltage signal corresponding to the grounding fault occurrence location as the input of the preset fault grounding resistance real-time prediction model, and outputting the real-time prediction result of the fault grounding resistance, the prediction error of the resistance is effectively reduced, ensuring the accuracy of the real-time prediction result of the fault grounding resistance; further, according to the real-time prediction result of the fault grounding resistance, identify the grounding insulation state of the train on-board power supply system, effectively improving the accuracy of the grounding insulation state identification.

[0106] Embodiment 2

[0107] See Figure 1 And Figure 3 This embodiment further describes the grounding fault classification model described in S5. The grounding fault classification model described in S5 is constructed using a convolutional neural network (CNN). The grounding fault classification model includes an input layer, a convolutional layer, a pooling layer, a first fully connected layer, and an output layer connected in sequence. Use the feature indicators in the total data set D constructed in step S4 as the input of the preset grounding fault classification model. The feature vector FV of each fault state iIt contains m + 3 features; the convolution kernel size of the convolutional layer is 3×33, the pooling layer is the max pooling, the pooling window size is 2×22, the number of nodes in the first fully connected layer is 50, the output layer is 6-dimensional, and the activation function of the output layer is softmax. Cross entropy is used as the loss function, the Adam optimizer is used, the batch size is 32 during the training process, and the number of training epochs is 100.

[0108] The training set and test set of the ground fault classification model come from the total data set D constructed in step S4. The ground fault classification model is trained using backpropagation and gradient descent methods. During the training process of the ground fault classification model, the weights of the convolution kernel and the fully connected layer are optimized through backpropagation and gradient descent methods until the number of training epochs reaches the preset threshold or the loss function converges. Then the training of the ground fault classification model ends, and a trained ground fault classification model is obtained. The trained ground fault classification model is used to accurately identify different types of ground faults. After training, the trained ground fault classification model is verified using the test set to evaluate the accuracy and robustness of the trained ground fault classification model in predicting the occurrence location of the ground fault.

[0109] Embodiment 3

[0110] See Figure 1 and Figure 3 , this embodiment further describes the ground fault classification model described in S6. The fault ground resistance real-time prediction model described in S6 is constructed based on the physics-informed neural network (PINN), and the fault ground resistance real-time prediction model is used to perform real-time prediction on the fault ground resistance R g . S6 combines the fault ground resistance real-time prediction model constructed based on the physics-informed neural network (PINN) with the voltage signal corresponding to the occurrence location of the ground fault, and can accurately output the real-time prediction result of the fault ground resistance, thereby measuring the degree of the ground fault.

[0111] Train the fault ground resistance real-time prediction model constructed based on the physics-informed neural network (PINN)

[0112] Use the ground fault classification model described in step S5 to identify and output the occurrence location of the ground fault. Based on the occurrence location of the ground fault, select the voltage signal corresponding to the occurrence location of the ground fault (such as U dc , U jd , U ab , U cb1 , U cb2 ) as the input of the preset fault ground resistance real-time prediction model, and the corresponding fault ground resistance R g as the output of the fault ground resistance real-time prediction model.

[0113] The real-time prediction model of the fault grounding resistance includes multiple second fully connected layers, each of which has 64 neurons and uses the ReLU activation function for output;

[0114] Divide the original dataset described in S4 of the above embodiment into the training set and the dataset of the real-time prediction model of the fault grounding resistance, and use optimization algorithms including but not limited to backpropagation algorithm and gradient descent method to set the network weight coefficients , and minimize the total loss function. Use the training set of the real-time prediction model of the fault grounding resistance to train the real-time prediction model of the fault grounding resistance until the total loss function converges and the training ends, so that the real-time prediction model of the fault grounding resistance can reasonably learn physical laws according to different situations and obtain accurate real-time prediction results of the fault grounding resistance; input the test set of the real-time prediction model of the fault grounding resistance into the trained grounding fault location identification model for verification to obtain the trained grounding fault location identification model;

[0115] The total loss function used in the training process of the real-time prediction model of the fault grounding resistance combines data error and physical constraint error. The data error is used to measure the difference between the output of the real-time prediction model of the fault grounding resistance and the actual grounding resistance value, and the physical constraint error is used to ensure that the output of the real-time prediction model of the fault grounding resistance satisfies the physical formula. For the function of physical constraint, for the fault grounding resistance

[0116] and the DC bus voltage U dc , grounding detection voltage U jd 's relationship, introduce five physical constraints as formulas (2), (3), (4), (5), (6) respectively, corresponding to five grounding fault situations; in the training process of the real-time prediction model of the fault grounding resistance, for each physical constraint situation, the real-time prediction model of the fault grounding resistance calculates the corresponding fault grounding resistance R g according to the input voltage signal through the physical formula, and compares it with the estimated value to obtain the physical constraint error. Therefore, the total loss function used in the training process of the real-time prediction model of the fault grounding resistance is as follows:

[0117]

[0118] Among them, represents the total loss function, represents the data error, represents the first physical constraint error, represents the weight coefficient of the first physical constraint error, represents the second physical constraint error, represents the weight coefficient of the second physical constraint error, represents the third physical constraint error, The weight coefficient representing the third physical constraint error represents the fourth physical constraint error The weight coefficient representing the fourth physical constraint error represents the fifth physical constraint error The weight coefficient representing the fifth physical constraint error

[0119] After the trained grounding fault location and identification model, the fault grounding resistance R can be estimated according to the grounding fault occurrence location identified by the grounding fault classification model described in step S5 and the voltage signal corresponding to the grounding fault occurrence location g . By statistically analyzing the off-line test and maintenance experience data, the grounding resistance thresholds for different health status levels (normal, deteriorated, abnormal, protected) are obtained, and compared with the real-time prediction result of the fault grounding resistance R g to estimate the health status and formulate corresponding fault handling strategies

[0120] Herein, the present invention proposes a grounding insulation state identification based on a data-driven algorithm. By constructing a grounding fault classification model and a real-time prediction model of the fault grounding resistance through an efficient data-driven algorithm, the accurate identification of the grounding insulation state of the train on-board power supply system is realized; the present invention generates multi-state simulation data and extracts features, uses MATLAB-Simulink simulation technology to generate sample data under normal state and five types of fault states, and combines the sliding window method to extract key feature variables to ensure the comprehensiveness and high quality of the training data of the grounding fault classification model and the real-time prediction model of the fault grounding resistance; the present invention also establishes a grounding fault classification and grounding insulation state identification mechanism, and optimizes the algorithm selection and parameter adjustment of the grounding fault classification model based on the feature vector set constructed based on the feature vectors to achieve a high-precision estimation of the grounding fault category and state

[0121] It should be particularly noted that the existing grounding diagnosis of train on-board power supply systems mostly adopts a model-based method, and most of them can only diagnose the inverter overcurrent of a single fault source type. However, the train on-board power supply system is complex and difficult to be represented by a single mathematical model, and a small interference will have a great impact on the diagnosis result. The existing calculation of the grounding resistance of the train on-board power supply system mostly adopts a calculation method based on model-derived formulas, which is established under a series of simplified assumptions and idealized conditions. When the actual situation is quite different from the assumed conditions, the calculation method based on model-derived formulas may not accurately reflect the true behavior of the system, resulting in a deviation between the calculation result and the actual situation

[0122] Aiming at the above-mentioned disadvantages in the prior art, the purpose of the present invention is to improve the existing train power supply system fault location and grounding resistance estimation methods in the following ways

[0123] First, improve the adaptability to complex interference; the present invention proposes a grounding fault location framework for a train on-board power supply system based on data-driven, and this framework includes a classification enabling mechanism and a grounding fault classification model constructed based on a convolutional neural network. Under normal circumstances, only the classification enabling mechanism runs, and after enabling, the grounding fault classification model based on feature vector learning runs, which improves the computing efficiency and the anti-interference ability.

[0124] Second, combine the physical information and the data-driven real-time prediction model of the fault grounding resistance to improve the accuracy of resistance estimation; the present invention uses a real-time prediction model of the fault grounding resistance constructed based on the physics-informed neural network (PINN) to estimate the fault grounding resistance. The physics-informed neural network (PINN) combines physical laws and data-driven learning. By taking physical formulas as constraint conditions, the real-time prediction model of the fault grounding resistance can ensure that the estimation result conforms to physical laws while learning the data pattern. This method can more accurately reflect the behavior of the actual system and overcome the limitations of the assumptions and idealized conditions in the traditional model derivation formula.

[0125] Third, provide more accurate and real-time fault diagnosis results; through the fault classification of the data-driven grounding fault classification model, the present invention can accurately judge the fault type of the train on-board power supply system, and further combine the real-time prediction model of the fault grounding resistance to estimate the fault grounding resistance. The combination of these technical means improves the accuracy and real-time performance of fault diagnosis, ensures that the system can respond quickly during actual operation, discovers potential faults in time, and avoids major safety hazards.

[0126] The grounding insulation state recognition method for a train on-board power supply system proposed by the present invention also has the following advantages:

[0127] 1. Based on data-driven grounding insulation state recognition, the advantage of the present invention is to automatically evaluate the grounding faults of the train on-board power supply system using data-driven algorithms. By training and learning the system data in different states, the faults can be accurately detected and classified. The system can timely discover grounding faults and provide early warnings by real-time analyzing the operation data of the electric traction system, effectively improving the safety and reliability of the system.

[0128] 2. Adopt multi-dimensional feature fusion and data augmentation techniques; the present invention combines time-domain and frequency-domain feature extraction techniques to extract features with significant discrimination from the original data of the train on-board power supply system, and uses data augmentation techniques to enhance the robustness of the data-driven grounding fault classification model. By fusing multiple features, a comprehensive feature vector set FV is constructed, which improves the accuracy and generalization ability of the grounding fault classification model for fault detection, enabling it to accurately identify the location of the grounding fault in the complex and dynamically changing train on-board circuit.

[0129] Embodiment 4

[0130] Refer to Figure 4 , this embodiment proposes a grounding insulation state recognition system for a train on-vehicle power supply system, and the grounding insulation state recognition system is implemented by the grounding insulation state recognition method for the train on-vehicle power supply system described in the above embodiment, including:

[0131] An acquisition module, configured to acquire the relationship between the DC bus voltage and the grounding detection voltage under the normal state of the train on-vehicle power supply system;

[0132] A first construction module, configured to construct characteristic variables based on the relationship between the DC bus voltage and the grounding detection voltage under different grounding fault positions, and respectively construct a characteristic vector set and a classification enabling mechanism according to the characteristic variables;

[0133] A judgment module, configured to judge whether the train on-vehicle power supply system is in a fault state based on the classification enabling mechanism;

[0134] A second construction module, configured to construct a total data set of characteristic vectors in different grounding fault states by using the characteristic vector set when the train on-vehicle power supply system is in a fault state;

[0135] A grounding fault classification module, configured to use the total data set as the input of a preset grounding fault classification model, and output the grounding fault occurrence position by the grounding fault classification model;

[0136] A fault grounding resistance real-time prediction module, configured to select the voltage signal corresponding to the grounding fault occurrence position as the input of a preset fault grounding resistance real-time prediction model based on the grounding fault occurrence position, and output the real-time prediction result of the fault grounding resistance;

[0137] A grounding insulation state recognition module, configured to recognize the grounding insulation state of the train on-vehicle power supply system according to the real-time prediction result of the fault grounding resistance.

[0138] In this embodiment, first, by obtaining the relationship between the DC bus voltage and the ground detection voltage of the train on-vehicle power supply system in the normal state, and combining the extraction of multi-dimensional characteristic variables under different ground fault positions to construct a feature vector set and a classification enabling mechanism, the anti-interference ability of the system is significantly improved. Then, after confirming the fault of the train on-vehicle power supply system through the classification enabling mechanism, a total data set of feature vectors in different ground fault states is constructed, and the total data set is used as the input of a preset ground fault classification model to output the occurrence position of the ground fault, realizing data-driven fault classification and effectively improving the real-time diagnosis accuracy of the occurrence position of the ground fault. Next, by selecting the voltage signal corresponding to the occurrence position of the ground fault as the input of a preset real-time prediction model of the fault ground resistance, the real-time prediction result of the fault ground resistance is output, effectively reducing the prediction error of the resistance and ensuring the accuracy of the real-time prediction result of the fault ground resistance. Further, according to the real-time prediction result of the fault ground resistance, the ground insulation state of the train on-vehicle power supply system is identified, effectively improving the accuracy of the ground insulation state identification.

[0139] Obviously, the above embodiments of the present invention are only examples for clearly explaining the present invention, rather than limiting the implementation manners of the present invention. For those of ordinary skill in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to enumerate all the implementation manners here. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included in the protection scope of the claims of the present invention.

Claims

1. A method for identifying the grounding insulation state of a train onboard power supply system, characterized in that: The method comprises the following steps: S1. Obtain the relationship between the DC bus voltage and the ground detection voltage of the train onboard power supply system under normal conditions; S1 obtains the DC bus voltage U of the train onboard power supply system under normal conditions dc With ground detection voltage U jd The calculation expression of the relationship is as follows: S1 also includes obtaining the fault grounding resistance and DC bus voltage U under different grounding fault positions dc , ground detection voltage U jd relationship; S2. Under different ground fault positions, based on the relationship between the DC bus voltage and the ground detection voltage, construct characteristic variables, and construct characteristic vector sets and classification enabling mechanisms according to the characteristic variables; S2: constructing a characteristic variable based on the relationship between the DC bus voltage and the ground detection voltage The calculation expression is as follows: According to the characteristic variables Construct the feature vector set The calculation process includes: in, represents the mean value of the characteristic variable F within a sliding window period at the kth moment, N represents the total number of sampling points in a sliding window, , represents the variance of the feature variable F within a sliding window period at the kth moment, Indicates the ground detection voltage U jd The complex spectrum value at frequency f, Indicates the ground detection voltage U jd The value of the i-th sampling point, represents the harmonic amplitude, Indicates the ground detection voltage U jd The complex spectrum value at m times the frequency f, It represents total harmonic distortion; S3. Based on the classification enabling mechanism, determine whether the train onboard power supply system is in a fault state. If so, execute S4; if not, it is in a normal state; S4. Using the feature vector set, construct a total data set of feature vectors under different ground fault conditions; S5. Using the total data set as the input of a preset ground fault classification model, and outputting the ground fault location by the ground fault classification model; S6. Based on the location of the ground fault, select the voltage signal corresponding to the location of the ground fault as the input of the preset real-time prediction model of the fault ground resistance, and output the real-time prediction result of the fault ground resistance; S7. Identify the grounding insulation status of the train's onboard power supply system based on the real-time prediction result of the fault grounding resistance.

2. The method for identifying the grounding insulation status of a train onboard power supply system according to claim 1, characterized in that: The classification enabling mechanism includes: Based on the characteristic variable F, the cumulative deviation S of the characteristic variable F is calculated k The following formula: Among them, E represents the mean of the feature variable F, and V represents the variance of the feature variable F; Determine the cumulative deviation S k Is it greater than the fault threshold h? If so, the train's onboard power supply system is in a fault state; if not, the train's onboard power supply system is in a normal state.

3. The method for identifying the grounding insulation state of a train onboard power supply system according to claim 2, characterized in that: The calculation expression for constructing the total data set D of the feature vectors under different ground fault states by using the feature vector set is as follows: in, The characteristic vector set representing the DC link positive pole grounding fault within the sliding window period, represents the mean value of the characteristic variable F of the DC link positive pole grounding fault within the sliding window period, represents the variance of the characteristic variable F of the DC link positive pole grounding fault within the sliding window period, Indicates the harmonic amplitude of the DC link positive pole grounding fault, Indicates the total harmonic distortion of the DC link positive pole grounding fault; The characteristic vector set representing the negative pole grounding fault of the DC link within the sliding window period, represents the mean value of the characteristic variable F of the negative pole grounding fault in the DC link within the sliding window period, represents the variance of the characteristic variable F of the negative pole grounding fault in the DC link within the sliding window period, Indicates the harmonic amplitude of the negative ground fault in the DC link. Indicates the total harmonic distortion of the negative pole grounding fault in the DC link; The characteristic vector set representing the ground fault at the front end of the reactor within the sliding window period, represents the mean value of the characteristic variable F of the reactor front-end ground fault within the sliding window period, The variance of the characteristic variable F of the ground fault at the front end of the reactor within the sliding window period, Indicates the harmonic amplitude of the ground fault at the front end of the reactor. Indicates the total harmonic distortion of the ground fault at the front end of the reactor; The feature vector set representing the RC input positive ground fault within the sliding window period, represents the mean value of the characteristic variable F of the RC input positive pole grounding fault within the sliding window period, represents the variance of the characteristic variable F of the RC input positive pole grounding fault within the sliding window period, Indicates the harmonic amplitude of the RC input positive ground fault, Indicates the total harmonic distortion of the RC input positive pole grounding fault; The set of feature vectors representing the RC input negative ground fault within the sliding window period, represents the mean value of the characteristic variable F of the RC input negative pole grounding fault within the sliding window period, represents the variance of the characteristic variable F of the RC input negative pole grounding fault within the sliding window period, Indicates the harmonic amplitude of the RC input negative ground fault, Indicates the total harmonic distortion of the RC input negative pole grounding fault.

4. The method for identifying the grounding insulation state of a train onboard power supply system according to claim 1, characterized in that: The ground fault classification model includes an input layer, a convolution layer, a pooling layer, a first fully connected layer and an output layer connected in sequence, the convolution kernel size of the convolution layer is 3×33, the pooling layer is maximum pooling, the pooling window size is 2×22, the number of nodes in the first fully connected layer is 50, the output layer is 6-dimensional, and the activation function of the output layer is softmax.

5. The method for identifying the grounding insulation state of a train onboard power supply system according to claim 4, characterized in that: The ground fault classification model is trained by using back propagation and gradient descent method until the number of training rounds reaches a preset threshold or the loss function converges, then the training of the ground fault classification model is completed and a trained ground fault classification model is obtained.

6. The method for identifying the grounding insulation state of a train onboard power supply system according to claim 1, characterized in that: The real-time prediction model for fault grounding resistance includes multiple second fully connected layers, each of which is provided with 64 neurons and uses a ReLU activation function for output.

7. The method for identifying the grounding insulation state of a train onboard power supply system according to claim 5, characterized in that: The total loss function used by the real-time prediction model for fault grounding resistance in the training process is as follows: in, represents the total loss function, Indicates data error, represents the first physical constraint error, represents the weight coefficient of the first physical constraint error, represents the second physical constraint error, represents the weight coefficient of the second physical constraint error, represents the third physical constraint error, represents the weight coefficient of the third physical constraint error, represents the fourth physical constraint error, represents the weight coefficient of the fourth physical constraint error, represents the fifth physical constraint error, Represents the weight coefficient of the fifth physical constraint error.

8. A grounding insulation state identification system for a train onboard power supply system, the grounding insulation state identification system is implemented based on the grounding insulation state identification method for a train onboard power supply system according to any one of claims 1 to 7, characterized in that: include: An acquisition module is used to obtain the relationship between the DC bus voltage and the ground detection voltage under the normal state of the train onboard power supply system; A first construction module is used to construct characteristic variables based on the relationship between the DC bus voltage and the ground detection voltage at different ground fault locations, and to respectively construct a characteristic vector set and a classification enabling mechanism according to the characteristic variables; A judgment module, used to judge whether the train onboard power supply system is in a fault state based on the classification enabling mechanism; The second construction module is used to construct a total data set of feature vectors under different ground fault states using the feature vector set when the on-board power supply system of the train is in a fault state; A ground fault classification module, used for taking the total data set as an input of a preset ground fault classification model, and the ground fault classification model outputs a ground fault occurrence location; A fault grounding resistance real-time prediction module is used to select a voltage signal corresponding to the grounding fault location as an input of a preset fault grounding resistance real-time prediction model based on the grounding fault location, and output a real-time prediction result of the fault grounding resistance; The grounding insulation status identification module is used to identify the grounding insulation status of the train onboard power supply system according to the real-time prediction result of the fault grounding resistance.

Images