High-voltage interlocking system and device applied to railway vehicle

Abstract

The invention relates to the technical field of rail high-voltage interlocking control, in particular to a high-voltage interlocking system and device applied to a rail vehicle. The operation state and the gas path state of a gas path interlocking box in a rail train system are monitored in advance, operation data are calculated, and a time sequence model is established according to the operation data so as to predict the operation data of the gas path interlocking box in the time when the gas path interlocking box needs to be continuously unlocked in the future, and whether the high-voltage interlocking system can safely respond to control or not is judged. The safety control precision of the high-voltage interlocking control system is improved; and when whether the gas circuit interlocking box can be continuously unlocked in the future is predicted, the operation state and the fault severity of the high-voltage related equipment are fused so as to accurately evaluate the overhaul disposal time of the high-voltage related equipment, and the overhaul disposal time is combined with the time sequence model so as to accurately evaluate the overhaul disposal time of the high-voltage related equipment. Whether the gas circuit interlocking box can be safely unlocked and high-voltage related equipment can be overhauled or not can be accurately judged, and the personal safety of maintainers is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of rail high-voltage interlocking control, and in particular to a high-voltage interlocking system and device applied to rail vehicles. Background Art

[0002] The key interlocking mechanism is a physical safety measure in the rail high-voltage interlocking system. It works together with circuit interlocking (such as circuit breaker interlocking) to ensure the safety and standardization of high-voltage equipment operation. Through the transmission and interlocking of physical keys, the operating procedures of high-voltage equipment are strictly controlled to prevent misoperation and dangerous conditions.

[0003] At present, in the rail high-voltage interlocking system, the gas circuit interlock box is unlocked by a key, and the subsequent high-voltage related equipment is unlocked again based on the unlocked gas circuit interlock box, and the high-voltage equipment components in the rail train system are repaired and disposed of by unlocking layer by layer. Therefore, when unlocking the gas circuit interlock box and the subsequent high-voltage related equipment, there is a lack of status monitoring of the gas circuit interlock box and the subsequent high-voltage related equipment. If the gas circuit interlock box is abnormal and directly unlocking it when its operating status cannot be known, the abnormal situation of the gas circuit interlock box will be aggravated. In this case, if the subsequent high-voltage related equipment is unlocked and repaired, there will inevitably be certain safety hazards. Therefore, how to realize the status monitoring of the gas circuit interlock box and the subsequent high-voltage related equipment, and build a safer rail high-voltage interlocking system, which is conducive to the maintenance of high-voltage related equipment and ensures the personal safety of maintenance personnel is an urgent problem to be solved. Summary of the invention

[0004] In view of the above-mentioned shortcomings of the prior art, the present invention provides a high-voltage interlocking system and device for rail vehicles, which can effectively solve the problem that the prior art lacks real-time status monitoring of the air circuit interlocking box and high-voltage related equipment, and cannot improve the safety of the high-voltage interlocking control system.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0006] The present invention provides a high-voltage interlocking system for rail vehicles, comprising at least:

[0007] The first interlocking monitoring module is used to monitor the operating status of the gas circuit interlocking box in the rail high-voltage interlocking system, wherein:

[0008] The operation state is judged by monitoring the operation state of the gas circuit interlock box and the gas circuit state, and the operation data is determined based on the operation state;

[0009] By obtaining the operation data of the gas interlock box at different time points and constructing time series data, a time series model is established based on the time series data, and the time series model is combined with the time value at which the gas interlock box needs to be unlocked at a future time to determine the future operation data of the gas interlock box, so as to determine whether the first-level key is allowed to unlock the gas interlock box;

[0010] and, a second interlock monitoring module for determining an estimated travel time based on the second-level key and high-voltage related equipment;

[0011] By acquiring the time series data of high-voltage related equipment to generate a time-frequency diagram, the local time-frequency features are extracted to form a high-dimensional feature vector. The high-dimensional feature vector is input into the SVM classifier to determine the fault severity of the high-voltage related equipment to determine the corresponding maintenance and treatment time. The time value is constructed based on the maintenance and treatment time and the estimated travel time to determine whether the gas circuit interlock box is unlocked.

[0012] Furthermore, the method for determining the operating data of the gas circuit interlock box is as follows:

[0013] Operation status monitoring of gas circuit interlock box,

[0014]

[0015] Among them, Q FS Indicates the operating status data of the gas circuit interlock box, M run Indicates the current operating torque of the three-way valve, M sat Indicates the initial operating torque of the three-way valve, G cut Indicates the current mechanical clearance of the three-way valve, G sat Indicates the initial mechanical clearance of the three-way valve, F cut Indicates the current operating torque of the lock cylinder, F sat Indicates the initial operating torque of the lock cylinder, R cut Indicates the current rotation resistance of the lock cylinder, R sat represents the initial rotation resistance of the lock core, GF represents the vibration time of the three-way valve, FR represents the vibration time of the lock core, α1, α2, α3, α4, α5 and α6 are the corresponding weight coefficients respectively;

[0016] Gas circuit status monitoring of gas circuit interlock box,

[0017]

[0018] Among them, P FS Indicates the gas circuit status data of the gas circuit interlock box, M eas Indicates the current gas line pressure, M sat Indicates the initial gas line pressure, F eas Indicates the current gas flow rate, F satrepresents the initial gas flow rate of the gas path, MF represents the calculation item, and β1 and β2 represent the corresponding weight coefficients;

[0019] Gas circuit status data Q based on gas circuit interlock box FS and operation status data P FS Perform weighted summation to obtain the operating data QP of the gas circuit interlock box at the current time point t .

[0020] Furthermore, the method for determining the future operation data of the gas circuit interlock box is as follows:

[0021] Get the operating data QP of the gas circuit interlock box at multiple time points t , to construct time series data;

[0022] To build a time series model, we have

[0023] (1-α 1 B-α 2 B 2 -···-α p B p )(1-B) d QP t =μ+(1+θ 1 B+θ 2 B 2 +···+θ q B q )εt

[0024] Among them, QP t represents time series data, B represents the lag operator, α 1 , α 2 ···α p Represent the coefficients of the autoregressive part, θ 1 ,θ 2 ···θ q They represent the coefficients of the sliding average part, p represents the order of the autoregressive part, q represents the order of the sliding average part, d represents the difference order, μ represents the constant term, and ε t represents the error term;

[0025] Predict the operation data QP of the gas circuit interlock box at a future time t+h , then there is,

[0026]

[0027] Among them, α i represents the coefficient of the autoregressive part, θ j represents the coefficient of the sliding average part, ε t+h-j represents the error term of the past j time steps, QP t+h-iIt represents the running data at time t+hi, and h represents the prediction step size.

[0028] Furthermore, the estimated journey time is determined as follows:

[0029] Determine the coordinates of the second-level key and record them as the second coordinates, determine the coordinates of the high-voltage related equipment and record them as the high-voltage equipment coordinates, determine the distance based on the second coordinates and the high-voltage equipment coordinates and calculate the estimated travel time SS t .

[0030] Furthermore, the expression for generating the time-frequency diagram of the time series data of high-voltage related equipment is:

[0031] Perform short-time Fourier transform on the time series data to generate the time-frequency diagram S(f,τ):

[0032]

[0033] Among them, ω(t-τ) represents the short-time window function, τ represents the center position of the short-time window function, represents the position of the time axis, f represents the frequency variable, e -j2πft represents the complex exponential basis function, and t represents the time variable.

[0034] Furthermore, the method of forming a high-dimensional feature vector is:

[0035] Input the time-frequency graph S(f,τ) and extract the local time-frequency features, then we have:

[0036]

[0037] in, Represents the multi-channel feature map of the output of the k-th convolution kernel in the l-th layer, W l,k,K，m,n represents the weight matrix of the kth convolution kernel of the lth layer with size K, S m,n Represents the pixel value at the mth and nth positions in the input feature map, b k represents the bias top of the kth convolution kernel in the convolution layer, and σ represents the activation function;

[0038] By reducing the feature dimension through the pooling layer, we have

[0039]

[0040] Among them, P i,j is the output result after the pooling operation, (m,n)∈window represents the position within the pooling window, window represents the window that defines the size of the pooling area, and F l,k,m,n Represents the value of the kth output feature map at position m,n in convolutional layer l;

[0041] By outputting a high-dimensional feature vector through the fully connected layer, we have

[0042] h=σ(W f ·X+b f )

[0043] Among them, h represents the high-dimensional feature vector extracted by CNN, W f represents the weight matrix of the fully connected layer, X represents the feature input of the previous layer, and b f represents the bias top of the fully connected layer.

[0044] Furthermore, the method for determining the fault severity of high voltage related equipment by using the SVM classifier is:

[0045] Define the SVM objective function,

[0046]

[0047] Among them, ω represents the weight vector of the classifier, b represents the bias, ξ i is a slack variable, C represents the penalty coefficient, and N represents the total number of samples in the sample data set;

[0048] Define the SVM constraints. The constraints are in the following form:

[0049] y i (ω·h i +b)≥1-ξ i ,ξ i ≥0

[0050] Among them, y i Represents the true category label of the i-th sample, and y i Set to the category label representing the severity of the fault, h i Represents the high-dimensional feature vector of the i-th sample, 1-ξ i The right side of the constraint is related to the slack variable, ξ i≥0 denotes the non-negativity of the slack variable.

[0051] Furthermore, the time value is determined by:

[0052] Determine the high-voltage related equipment controlled by the gas circuit interlock box in the current rail train system;

[0053] Clearly define the maintenance and treatment time P corresponding to the severity of the fault of the high-voltage related equipment under control GS ;

[0054] The maintenance treatment time P GS Estimated travel time SS t Combined with the build time value U j .

[0055] A rail vehicle high-voltage interlocking device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements any one of the above-mentioned systems when executing the computer program.

[0056] A computer-readable storage medium stores a computer program, wherein the computer program implements any one of the above-mentioned systems when executed by a processor.

[0057] Compared with the known prior art, the technical solution provided by the present invention has the following beneficial effects:

[0058] By monitoring the operating status and gas circuit status of the gas circuit interlock box in the rail train system in advance and calculating the operating data, a time series model is established based on the operating data to predict the operating data of the gas circuit interlock box when it needs to be continuously unlocked in the future, to determine whether the high-voltage interlock system can respond safely and control, and to improve the safety control accuracy of the high-voltage interlock control system;

[0059] In addition, when predicting whether the gas interlock box can continue to be unlocked in the future, the operating status of the high-voltage related equipment and the severity of the fault are incorporated to accurately evaluate the maintenance and disposal time of the high-voltage related equipment. By combining the maintenance and disposal time with the time series model, it is accurately determined whether the gas interlock box can be safely and correctly unlocked and the high-voltage related equipment can be repaired, thereby ensuring the personal safety of the maintenance personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0061] Figure 1 It is the overall module block diagram of the present invention. DETAILED DESCRIPTION

[0062] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0063] Current rail high voltage interlocking system, example:

[0064] Step 1: Operate in the driver's cab and use the first-level key (A key, blue);

[0065] Operation content: Turn the electric grounding key switch on the fault switch panel in the driver's cab to the locked position and remove the first-level key (A key);

[0066] The circuit interlock in the driver's cab ensures that the pantograph raising and main circuit breaker closing operations cannot be performed at this time, and the high voltage power supply is cut off;

[0067] After the grounding switch is turned to the grounding position, the high-voltage system of the rail vehicle is in a grounded state to ensure safety;

[0068] Among them, the removal of the first-level key marks the entry of the operation into the physical key interlocking process. Only after the grounding switch is locked can the first-level key be obtained and the subsequent steps be entered.

[0069] Step 2: The first-level key unlocks the gas circuit interlock box and cuts off the gas circuit;

[0070] Operation content: Insert the first-level key into the bow raising gas circuit interlock box, unlock the three-way valve operating handle, cut off the bow raising gas circuit, lock the three-way valve handle, and take out the second-level key (B key, yellow).

[0071] High voltage interlock function:

[0072] This step ensures that the air path for raising the bow is cut off through physical operation, physically eliminating the possibility of raising the bow and further enhancing safety.

[0073] The second-level key can only be taken out after the gas line is cut off and the three-way valve is locked, ensuring the mandatory interlock between gas line cutoff and subsequent high-pressure operation.

[0074] The acquisition of the second-level key is strictly limited by the previous operation, ensuring that the operation can only be continued after the gas line cut-off operation is completed;

[0075] The extraction of the second-level key heralds the start of the next phase of the operation, which is unlocking high-voltage related equipment;

[0076] The second-level key unlocks the traction converter, auxiliary converter, medium-voltage busbar box and external power supply to facilitate operation and maintenance of high-voltage related equipment.

[0077] It should be noted that high-voltage related equipment is equipped with physical interlocks (first-level keys and second-level keys), and can only be opened and unlocked by obtaining the corresponding keys.

[0078] Make sure that high-voltage related equipment can only be accessed after all previous operations are completed (such as grounding status confirmation, pantograph cutting, etc.) to prevent accidental contact with high voltage during operation.

[0079] Therefore, in order to further improve the safety of the above-mentioned high-voltage interlocking system, the present invention is further described below in conjunction with embodiments.

[0080] Example 1 (see Figure 1 ): A high voltage interlocking system for rail vehicles, comprising:

[0081] The first interlock monitoring module is used to collect relevant data of the gas circuit interlock box and monitor its operating status after the electric grounding key switch on the fault switch panel in the driver's cab is turned to the locked position and the first-level key is removed. The details are as follows:

[0082] Monitor the operating status of each gas interlocking box in the rail high-voltage interlocking system, including:

[0083] Operation status monitoring of gas circuit interlock box,

[0084]

[0085] Among them, Q FS Indicates the operating status data of the gas circuit interlock box, M run Indicates the current operating torque of the three-way valve, M sat Indicates the initial operating torque of the three-way valve, G cut Indicates the current mechanical clearance (physical distance) of the three-way valve, G sat Indicates the initial mechanical clearance of the three-way valve, F cut Indicates the current operating torque of the lock cylinder, F sat Indicates the initial operating torque of the lock cylinder, R cut Indicates the current rotation resistance of the lock cylinder, R sa t represents the initial rotation resistance of the lock core, GF represents the vibration time of the three-way valve (monitored by the built-in vibration sensor), FR represents the vibration time of the lock core, and α1, α2, α3, α4, α5 and α6 are the corresponding weight coefficients respectively;

[0086] Gas circuit status monitoring of gas circuit interlock box,

[0087]

[0088] Among them, P FS Indicates the gas circuit status data of the gas circuit interlock box, M eas Indicates the current gas line pressure, M sat Indicates the initial gas line pressure, F eas Indicates the current gas flow rate, F sat represents the initial gas flow rate of the gas path, MF represents the calculation item, which is preset to 2, and β1 and β2 represent the corresponding weight coefficients;

[0089] Therefore, based on the gas circuit state data Q of the gas circuit interlock box FSand operation status data P FS Perform weighted summation (give weight coefficients corresponding to gas circuit status data and operation status data respectively and add them up) to obtain the operation data QP of the gas circuit interlock box at the current time point t Based on the above expression, the operating data QP of the gas circuit interlock box at each time point (the time point can be constructed according to hours) is calculated in the same way t , by obtaining the operating data QP of the gas circuit interlock box at multiple time points t , to construct time series data, establish a time series model based on the time series data, and determine the parameters of the model based on the time series model (α 1 , α 2 ···α p ,θ 1 ,θ 2 ···θ q ) in order to predict the operating data QP at future time t+h , where the expression of the time series model is:

[0090] (1-α 1 B-α 2 B 2 -···-α p B p )(1-B) d QP t =μ+(1+θ 1 B+θ 2 B 2 +···+θ q B q )εt

[0091] Among them, QP t represents time series data, B represents the lag operator, α 1 , α 2 ···α p Represent the coefficients of the autoregressive part, θ 1 ,θ 2 ···θ q They represent the coefficients of the sliding average part, p represents the order of the autoregressive part, which describes the use of the values ​​of the previous p moments to predict the current moment, q represents the order of the sliding average part, which describes the use of the previous q error terms to correct the predicted value of the current moment, d represents the difference order, which describes the number of times the time series data is differentiated to ensure the stability of the time series data, μ represents the constant term, and ε t represents the error term, describing the unexplained part of the time series data; where, for α 1 , α 2 ···α p With θ 1 ,θ 2 ···θq The solution is obtained by the least square method or the maximum likelihood estimation (MLE), which is a conventional method for solving the problem at present and will not be described in detail in the present invention.

[0092] Therefore, based on the prediction formula, the operation data QP of the gas circuit interlock box at a future time is predicted t+h , then there is,

[0093]

[0094] Among them, αi represents the coefficient of the autoregressive part, θ j represents the coefficient of the sliding average part, ε t+h-j represents the error term of the past j time steps, QP t+h-i represents the operating data at time t+hi, h represents the prediction step, and describes the future time point to be predicted. Therefore, based on the above model expression, the operating data QP of the gas circuit interlock box at the future time can be predicted. t+h ;

[0095] Then, obtain the time value U required to operate and unlock the gas circuit interlock box in this rail vehicle j , refers to the estimated time required to operate and unlock the gas circuit interlock box this time, then the time value U j Substitute the above prediction formula to obtain the operation data QP of the gas circuit interlock box at a future time t+h (i.e., the time when the operation is completed, such as the end time of the operation of rail vehicle maintenance, inspection and disposal, etc., describing the time from the current time to t+h (i.e., the current time and U j The operation data of the gas interlock box under the time of (combination) can support continuous unlocking) and then, based on the future operation data QP t+h Determine whether it exceeds the preset gas line interlock box operation risk threshold:

[0096] If it does not exceed, the first-level key is allowed to be placed in the gas circuit interlock box and unlocked, so that the three-way valve operating handle is unlocked, the bow lifting gas circuit is cut off, the three-way valve handle is locked, and the second-level key is taken out;

[0097] If it exceeds, a prohibition unlocking instruction is generated, that is, the first-level key is prohibited from being placed in the gas circuit interlock box and unlocking (that is, providing safety restriction conditions for the subsequent unlocking of the high-voltage interlock control system);

[0098] In this way, it is possible to avoid unlocking and using the gas circuit interlock box when there is an operational risk, thereby reducing the probability of aggravated failure of the gas circuit interlock box. Therefore, when a prohibition of unlocking instruction is generated, it is synchronously output to the rail train system (management and control center) for early warning, and indicates that the operator should handle the fault of the gas circuit interlock box to ensure that the high-voltage interlocking system of the rail vehicle as a whole is in a safe state before allowing the high-voltage related equipment in the rail train system to be unlocked and repaired. That is, in the high-voltage interlocking system, it is determined that the initial operation of the gas circuit interlock box is safe before the subsequent handling steps of the high-voltage related equipment are allowed (in the rail train system, the unlocking and closing safety of the equipment needs to be strictly controlled), thereby greatly ensuring the personal safety of the operator.

[0099] Generally speaking, when the second-level key is needed to unlock the high-voltage related equipment, it can reflect that the high-voltage related equipment of the current rail vehicle is abnormal and needs to be repaired and disposed. In order to make a specific judgment on the status of the high-voltage related equipment and clarify the severity of the current fault of the high-voltage related equipment, so as to determine the subsequent repair and disposal time, it also includes:

[0100] The second interlock monitoring module is used to determine the coordinates of the second-level key and record them as the second coordinates, determine the coordinates of the high-voltage related equipment and record them as the high-voltage equipment coordinates, determine the distance based on the second coordinates and the high-voltage equipment coordinates and calculate the estimated travel time SS t , obtained according to the ratio of distance and speed, no further explanation is given (the gas circuit interlock box can be unlocked to take out the second-level key, then it is determined that the first-level key and the second-level key are at the same coordinate position. Even if the gas circuit interlock box is not unlocked, the distance can still be determined based on the second coordinate and the high-voltage equipment coordinate and the estimated travel time SS can be calculated. t and the following steps).

[0101] Subsequently, the status of high-voltage related equipment is monitored and identified, wherein the high-voltage related equipment includes: traction converter, auxiliary converter, medium-voltage busbar and external power supply;

[0102] For the state determination of traction converter:

[0103] Collect time series data x(t) from high-voltage related equipment (traction converter, auxiliary converter, medium-voltage bus or external power supply), including temperature data (used to monitor whether the high-voltage related equipment is overheated, especially when the load is high, the temperature signal is an important indicator for evaluating faults), vibration data (reflecting mechanical failures, looseness, collision or other abnormal conditions), and current data (reflecting current fluctuations, especially high-frequency current fluctuations can reveal electrical failures or abnormalities of electronic components). The temperature data, vibration data and current data can capture information on the operating status of key components of high-voltage related equipment, and then perform normalization processing on them;

[0104] Then, a short-time Fourier transform is performed on each time series data to generate a time-frequency graph S(f,τ):

[0105]

[0106] Among them, ω(t-τ) represents the short-time window function (including Hamming window or Hanning window), τ represents the center position of the short-time window function, represents the position of the time axis, f represents the frequency variable, e -j2πft Represents the complex exponential basis function, t represents the time variable. By converting the time series data into a two-dimensional image, the time information is retained and the frequency component is introduced, which facilitates the subsequent CNN to extract time-frequency features and capture the complex dynamic characteristics of high-voltage related equipment during operation.

[0107] Furthermore, through CNN model training, including:

[0108] Input time-frequency graph S(f,τ);

[0109] The convolution layer extracts local time-frequency features, then we have,

[0110]

[0111] in, Represents the multi-channel feature map of the output of the k-th convolution kernel in the l-th layer, W l,k,K，m,n S represents the weight matrix of the kth convolution kernel of the lth layer with size K, describing the weight value of the mth and nth positions in the convolution kernel. m,n Represents the pixel value at the mth and nth positions in the input feature map, b k Represents the bias top of the kth convolution kernel in the convolution layer, σ represents the activation function, such as ReLU, etc. Therefore, through multiple convolution kernels and multi-scale convolution operations, features are extracted from the input data, that is, from the time-frequency diagram.

[0112] Furthermore, by reducing the feature dimension through the pooling layer and extracting the main features, we have:

[0113]

[0114] Among them, P i,j It is the result of the pooling operation, indicating the pooling result corresponding to position i, j in the pooling window. (m,n)∈window indicates a position in the pooling window. Window indicates the window that defines the size of the pooling area, usually a rectangular area of ​​fixed size. F l,k,m,n represents the value of the kth output feature map at position m,n in the convolution layer l. Therefore, by selecting the maximum value F in each window within the pooling window, l,k,m,n And take it as the pooling result P i,j, to reduce the size of the feature map and extract the most important features.

[0115] And, the fully connected layer is used to output high-dimensional feature vectors, then we have,

[0116] h=σ(W f ·X+b f )

[0117] Among them, h represents the high-dimensional feature vector extracted by CNN, W f represents the weight matrix of the fully connected layer, X represents the feature input of the previous layer (i.e., the feature map P after pooling i,j The result of flattening), b f represents the bias top of the fully connected layer;

[0118] Therefore, the above scheme extracts local features from the input time-frequency diagram based on the convolution operation, and further simplifies the feature dimension with the pooling layer to retain important information and reduce noise interference. Then, the feature output of the convolution layer is integrated through the fully connected layer to form a high-dimensional feature vector, which is convenient for subsequent SVM classifier processing.

[0119] Furthermore, the SVM classifier is trained, and the high-dimensional feature vector h extracted by CNN is used as input to classify the severity of the fault of high-voltage related equipment. Assuming that the category of fault severity is a continuous or discrete level, it is defined as different category labels, including normal, minor fault, moderate fault and severe fault, then,

[0120] Define the SVM objective function,

[0121]

[0122] Among them, ω represents the weight vector of the classifier, which determines the direction and position of the classification hyperplane, b represents the bias top, which is used to determine the position of the classification hyperplane, and ξ i is a slack variable, which represents the classification error of the i-th sample. If a sample is misclassified or close to the hyperplane, the slack variable will be a positive value. C represents the penalty coefficient, which weighs the classification interval and error. N represents the total number of samples in the sample data set. Therefore, the regularization term is used to avoid overfitting and make the constructed classifier have good generalization ability. The penalty term of the slack variable is used to deal with samples that cannot be completely correctly classified.

[0123] Define SVM constraints. The constraints ensure that the model can correctly classify the training data while allowing a certain degree of misclassification. The constraints are in the following form:

[0124] y i (ω·h i +b)≥1-ξ i ,ξi ≥0

[0125] Among them, y i Represents the true category label of the i-th sample, and y i is set to the category label representing the severity of the fault (e.g., 0, indicating normal; 1, indicating a minor fault; 2, indicating a moderate fault; and 3, indicating a severe fault). i Represents the high-dimensional feature vector of the i-th sample, 1-ξ i The right side of the constraint is related to the slack variable, ξ i ≥0 indicates the non-negativity of the slack variable. Therefore, by training the SVM and optimizing the weight vector ω and the bias top b, the classifier can be trained based on the high-dimensional feature vector h i Classify the severity of the faults of high-voltage related equipment, and judge the severity of the faults based on the classification results, such as:

[0126] 0 means normal; 1 means a minor fault; 2 means a moderate fault; 3 means a serious fault.

[0127] In the above, when the label is 0, it is normal, and the temperature, vibration and current indicators corresponding to the high-voltage related equipment are within the normal range; when the labels are 1, 2 and 3, they respectively indicate the degree to which the temperature, vibration and current indicators corresponding to the high-voltage related equipment exceed the normal range, which is gradually increasing here. By setting a reasonable threshold range, it can be classified as above, and this classification method is also a conventional method, that is, it will not be further elaborated. Therefore, by marking the fault data of high-voltage related equipment and using SVM for multi-category classification, the fault state can be identified according to the sensor data and classified into the above four levels to clarify whether the current high-voltage related equipment is faulty and the current severity of the fault.

[0128] Therefore, when the fault severity of the high-voltage related equipment is determined, the maintenance processing time P corresponding to the preset fault severity is obtained. GS , and thus, the corresponding maintenance treatment time P of the high-voltage related equipment with faults is obtained GS , where it should be noted that in the rail train system, there are usually the following situations, which have time values ​​U j The corresponding determination methods include:

[0129] First: In the rail train system, a gas circuit interlock box synchronously controls multiple high-voltage related equipment (traction converter, auxiliary converter, medium-voltage bus and external power supply):

[0130] In this case, if multiple high-voltage related devices with faults are identified at the same time, it is only necessary to take the longest maintenance and treatment time P among the multiple high-voltage related devices (different types of high-voltage related devices have different maintenance and treatment times preset). GS The corresponding high-voltage related equipment is taken as the target maintenance treatment time P GS′ (Because multiple faulty high-voltage related equipment need to be repaired and handled simultaneously), this is because even if other high-voltage related equipment is repaired, it is still necessary to repair and handle the longest time P GS The rail train system can only be restarted after the corresponding high-voltage related equipment is overhauled, that is, the gas circuit interlock box can be allowed to be closed.

[0131] U j =P GS′ +SS t

[0132] Thus, the time value U required to operate and unlock the gas circuit interlock box in the above scheme is realized. j Determine in order to obtain the future unlocking completion time U of the gas circuit interlock box jw , and judge at the unlocking completion time U jw Future operation data OP W , to judge whether the gas circuit interlock box can be within the time value U j The first-level key is placed into the gas circuit interlock box and unlocked. Therefore, this solution not only realizes the fault supervision of high-voltage related equipment, but also simultaneously supervises the safety status of the gas circuit interlock box to ensure the safety of the entire high-voltage interlock system.

[0133] Second: In the rail train system, one gas circuit interlock box controls one high-voltage related device respectively (one gas circuit interlock box is interlocked with one of the traction converter, auxiliary converter, medium-voltage bus or external power supply and is independently controlled)

[0134] Therefore, we have,

[0135] U j =P GS +SS t

[0136] Synchronously obtain the time value U required to operate and unlock the gas circuit interlock box j In the above formula, the gas circuit interlock box is interlocked with the corresponding high-voltage related equipment. Therefore, it is only necessary to obtain the maintenance and disposal time P of the high-voltage related equipment corresponding to the gas circuit interlock box. GS That's it.

[0137] In addition, the present invention further provides (the following specific reference to the above system is sufficient and will not be repeated here):

[0138] A rail vehicle high-voltage interlocking device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements any one of the above-mentioned systems when executing the computer program.

[0139] A computer-readable storage medium stores a computer program, wherein the computer program implements any one of the above-mentioned systems when executed by a processor.

[0140] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the protection scope of the technical solutions of the embodiments of the present invention.

Claims

1. A high voltage interlocking system for rail vehicles, characterized in that: include: The first interlocking monitoring module is used to monitor the operating status of the gas circuit interlocking box in the rail high-voltage interlocking system, wherein: The operation state is judged by monitoring the operation state of the gas circuit interlock box and the gas circuit state, and the operation data is determined based on the operation state; By obtaining the operation data of the gas circuit interlock box at different time points and constructing time series data, a time series model is established based on the time series data, and the time series model is combined with the time value of the gas circuit interlock box being unlocked at a future time to determine the operation data of the gas circuit interlock box at a future time, so as to judge whether the first-level key is allowed to unlock the gas circuit interlock box; and, a second interlock monitoring module for determining an estimated travel time based on the second-level key and high-voltage related equipment; By acquiring the time series data of high-voltage related equipment to generate a time-frequency diagram, the local time-frequency features are extracted to form a high-dimensional feature vector. The high-dimensional feature vector is input into the SVM classifier to determine the fault severity of the high-voltage related equipment to determine the corresponding maintenance and treatment time. The time value is constructed based on the maintenance and treatment time and the estimated travel time to determine whether the gas circuit interlock box is unlocked.

2. A high voltage interlocking system for rail vehicles according to claim 1, characterized in that: The method for determining the operating data of the gas circuit interlock box is as follows: Operation status monitoring of gas circuit interlock box, Among them, Q FS Indicates the operating status data of the gas circuit interlock box, M run Indicates the current operating torque of the three-way valve, M sat Indicates the initial operating torque of the three-way valve, G cut Indicates the current mechanical clearance of the three-way valve, G sat Indicates the initial mechanical clearance of the three-way valve, F cut Indicates the current operating torque of the lock cylinder, F sat Indicates the initial operating torque of the lock cylinder, R cut Indicates the current rotation resistance of the lock cylinder, R sat represents the initial rotation resistance of the lock core, GF represents the vibration time of the three-way valve, FR represents the vibration time of the lock core, α1, α2, α3, α4, α5 and α6 are the corresponding weight coefficients respectively; Gas circuit status monitoring of gas circuit interlock box, Among them, P FS Indicates the gas circuit status data of the gas circuit interlock box, M eas Indicates the current gas line pressure, M sat Indicates the initial gas line pressure, F eas Indicates the current gas flow rate, F sat represents the initial gas flow rate of the gas path, MF represents the calculation item, and β1 and β2 represent the corresponding weight coefficients; Gas circuit status data Q based on gas circuit interlock box FS and operation status data P FS Perform weighted summation to obtain the operating data QP of the gas circuit interlock box at the current time point t .

3. The high voltage interlocking system for rail vehicles according to claim 1, characterized in that: The method for determining the future operation data of the gas circuit interlock box is: Get the operating data QP of the gas circuit interlock box at multiple time points t , to construct time series data; To build a time series model, we have (1-α1B-α2B 2 -···-a p B p (1-B) d QP t =μ+(1+θ1B+θ2B 2 +···+θ q B q )et Among them, QP t represents time series data, B represents the lag operator, α1, α2, α p They represent the coefficients of the autoregressive part, θ1, θ2, θ q They represent the coefficients of the sliding average part, p represents the order of the autoregressive part, q represents the order of the sliding average part, d represents the difference order, μ represents the constant term, and ε t represents the error term; Predict the operation data QP of the gas circuit interlock box at a future time t+h , then there is, Among them, α i represents the coefficient of the autoregressive part, θ j represents the coefficient of the sliding average part, ε t+h-j Indicates the past j The error term for time steps, QP t+h-i It represents the running data at time t+hi, and h represents the prediction step size.

4. The high-voltage interlocking system for rail vehicles according to claim 1, characterized in that: The method for determining the estimated journey time is: Determine the coordinates of the second-level key and record them as the second coordinates, determine the coordinates of the high-voltage related equipment and record them as the high-voltage equipment coordinates, determine the distance based on the second coordinates and the high-voltage equipment coordinates and calculate the estimated travel time SS t .

5. The high voltage interlocking system for rail vehicles according to claim 1, characterized in that: The expression for generating the time-frequency diagram of the time series data of the high-voltage related equipment is: Perform short-time Fourier transform on the time series data to generate the time-frequency diagram S(f,τ): Among them, ω(t-τ) represents the short-time window function, τ represents the center position of the short-time window function, represents the position of the time axis, f represents the frequency variable, e -j2πft represents the complex exponential basis function, and t represents the time variable.

6. A high voltage interlocking system for rail vehicles according to claim 5, characterized in that: The method for forming a high-dimensional feature vector is: Input the time-frequency graph S(f,τ) and extract the local time-frequency features, then we have: in, Represents the multi-channel feature map of the output of the k-th convolution kernel in the l-th layer, W l,k,K，m,n represents the weight matrix of the kth convolution kernel of the lth layer with size K, S m,n Represents the pixel value at the mth and nth positions in the input feature map, b k represents the bias top of the kth convolution kernel in the convolution layer, and σ represents the activation function; By reducing the feature dimension through the pooling layer, we have Among them, P i,j is the output result after the pooling operation, (m,n)∈window represents the position within the pooling window, window represents the window that defines the size of the pooling area, and F l,k,m,n Represents the value of the kth output feature map at position m,n in convolutional layer l; By outputting a high-dimensional feature vector through the fully connected layer, we have h=σ(W f ·X+b f ) Among them, h represents the high-dimensional feature vector extracted by CNN, W f represents the weight matrix of the fully connected layer, X represents the feature input of the previous layer, and b f represents the bias top of the fully connected layer.

7. A high voltage interlocking system for rail vehicles according to claim 6, characterized in that: The method for determining the fault severity of high-voltage related equipment through SVM classifier is: Define the SVM objective function, Among them, ω represents the weight vector of the classifier, b represents the bias top, ξ represents the set of slack variables of all samples, ξ represents the set of slack variables of all samples, i is the slack variable of the i-th sample, C represents the penalty coefficient, and N represents the total number of samples in the sample data set; Define the SVM constraints. The constraints are in the following form: y i (ω·h i +b)≥1-ξ i ,x i ≥0 Among them, y i Represents the true category label of the i-th sample, and y i Set to the category label representing the severity of the fault, h i Represents the high-dimensional feature vector of the i-th sample, 1-ξ i The right side of the constraint is related to the slack variable, ξ i ≥0 indicates the non-negativity of the slack variable.

8. A high voltage interlocking system for rail vehicles according to claim 7, characterized in that: The method for determining the time value is: Determine the high-voltage related equipment controlled by the gas circuit interlock box in the current rail train system; Clearly define the maintenance and treatment time P corresponding to the severity of the fault of the high-voltage related equipment under control GS ; The maintenance treatment time P GS Estimated travel time SS t Combined with the build time value U j .

9. A rail vehicle high-voltage interlocking device, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the system according to any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the system according to any one of claims 1 to 8 is implemented.

Images