A method, device, equipment and storage medium for locating turnout fault

Abstract

The present invention discloses a method, device, equipment, and storage medium for locating a turnout fault. The method comprises: obtaining motion curve data corresponding to at least one conversion of a target turnout; for each motion curve data, dividing the current motion curve data into stages according to the stages experienced by the target turnout during the conversion process to obtain target motion curve data; based on the similarity between the motion curve data corresponding to the current conversion and the historical motion curve data corresponding to each conversion before the current conversion, determining whether there is a fault in the current conversion; if so, inputting the motion curve data corresponding to the current conversion into a pre-trained turnout fault identification model to determine the fault type corresponding to the current conversion; and by comparing the motion curve data corresponding to the current target turnout conversion with the historical motion curve data of the target turnout, accurately locating the fault occurrence point, improving the efficiency of fault identification while reducing the cost of fault identification.

Description

Technical Field

[0001] The present invention relates to the technical field of turnout fault locating technology, and in particular to a turnout fault locating method, device, equipment and storage medium. Background Art

[0002] The turnout switching system is a crucial component in maintaining safe railway operation. However, due to its mechanical complexity and the diversity of its application environments, it presents potential fault hazards. Therefore, it is necessary to implement intelligent fault location for turnout switching.

[0003] Traditionally, turnout fault diagnosis mainly relies on staff regularly browsing the turnout operation data collected by computer monitoring, and comparing the monitored operating current, power and other curve data with the normal operating turnout monitoring data. If there is an abnormality, technical personnel will investigate and repair it.

[0004] At present, this manual fault diagnosis method not only requires a lot of manpower and material resources, but is also inefficient and prone to misjudgment and missed judgment. Summary of the Invention

[0005] The present invention provides a method, device, equipment and storage medium for locating turnout faults, so as to achieve preliminary judgment and accurate location of possible faults in the target turnout switching process, thereby improving the efficiency of fault identification and reducing the cost of fault identification.

[0006] In a first aspect, an embodiment of the present invention provides a method for locating a turnout fault, the method comprising:

[0007] Acquiring action curve data corresponding to at least one conversion of a target turnout, wherein the action curve data includes at least one of rod displacement curve data, current curve data, and conversion force curve data corresponding to the conversion of the target turnout;

[0008] For each action curve data, the current action curve data is divided into stages according to the stages experienced by the target turnout during the switching process to obtain target action curve data;

[0009] Based on the similarity between the action curve data corresponding to the current moment conversion and the historical action curve data corresponding to each conversion before the current moment, it is determined whether there is a fault in the current moment conversion;

[0010] If so, the action curve data corresponding to the current moment conversion is input into a pre-trained turnout fault identification model to determine the fault type corresponding to the current moment conversion.

[0011] In a second aspect, an embodiment of the present invention further provides a device for locating a turnout fault, the device comprising:

[0012] a data acquisition module for acquiring action curve data corresponding to at least one conversion of a target turnout, wherein the action curve data includes at least one of rod displacement curve data, current curve data, and conversion force curve data corresponding to the conversion of the target turnout;

[0013] a stage division module for dividing the current action curve data into stages according to the stages experienced by the target turnout during the switching process, thereby obtaining target action curve data;

[0014] A fault judgment module judges whether there is a fault in the current moment conversion based on the similarity between the action curve data corresponding to the current moment conversion and the historical action curve data corresponding to each conversion before the current moment;

[0015] The fault identification module is used to input the action curve data corresponding to the current moment conversion into a pre-trained turnout fault identification model to determine the fault type corresponding to the current moment conversion.

[0016] In a third aspect, the present invention further provides an electronic device, comprising:

[0017] at least one processor; and

[0018] a memory communicatively connected to the at least one processor; wherein,

[0019] The memory stores a computer program executable by the at least one processor. The computer program is executed by the at least one processor so that the at least one processor can execute the method for locating a turnout fault according to any embodiment of the present invention.

[0020] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method for locating a turnout fault according to any embodiment of the present invention when executed.

[0021] The technical solution of the embodiment of the present invention obtains action curve data corresponding to at least one conversion of a target turnout, wherein the action curve data includes at least one of rod displacement curve data, current curve data, and conversion force curve data corresponding to the conversion of the target turnout; for each action curve data, the current action curve data is stage-divided according to the various stages experienced by the target turnout during the conversion process to obtain target action curve data; based on the similarity between the action curve data corresponding to the conversion at the current moment and the historical action curve data corresponding to each conversion before the current moment, it is determined whether there is a fault in the conversion at the current moment; if so, the action curve data corresponding to the conversion at the current moment is input into a pre-trained turnout fault recognition model to determine the fault type corresponding to the conversion at the current moment, and by comparing the action curve data corresponding to the current target turnout conversion with the historical action curve data of the target turnout, it is determined whether there is a fault in the current target turnout conversion and the fault point is located. This solves the problem that manual fault diagnosis methods consume a lot of manpower and material resources, are inefficient, and have the problem of misjudgment and missed judgment, and achieves accurate positioning of the fault occurrence point, improves the efficiency of fault identification and fault location, and reduces the manpower cost of fault location.

[0022] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the present invention will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0024] Figure 1 This is a flow chart of a method for locating a turnout fault according to the first embodiment of the present invention;

[0025] Figure 2 This is an action curve data diagram provided according to the first embodiment of the present invention;

[0026] Figure 3 is a target action curve data graph provided according to an embodiment of the present invention;

[0027] Figure 4 is a flow chart of a training method for a turnout fault identification model according to an embodiment of the present invention;

[0028] Figure 5 2. It is a structural diagram of a turnout fault identification model to be trained according to an embodiment of the present invention;

[0029] Figure 6 This is a flow chart of a method for locating a turnout fault according to a second embodiment of the present invention;

[0030] Figure 7 is a target current curve data diagram provided according to an embodiment of the present invention;

[0031] Figure 8 This is a flow chart of a method for locating a turnout fault according to a third embodiment of the present invention;

[0032] Figure 9 2 is a schematic structural diagram of a device for locating a turnout fault according to a third embodiment of the present invention;

[0033] Figure 10 The figure is a schematic structural diagram of an electronic device for implementing the method for locating a turnout fault according to an embodiment of the present invention. DETAILED DESCRIPTION

[0034] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions 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 embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0035] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0036] Before introducing this technical solution, we can first introduce the turnout and the process of turnout conversion: the turnout is a line connection device that allows locomotives to transfer from one track to another on the railway line. It consists of three units: a switch, a connecting part, a switch and a guard rail. The switch includes a base rail, a point rail and a switch mechanism. The conversion and locking of the turnout need to be performed by the turnout conversion equipment (including a switch machine, an external locking device and an installation device, etc.). The turnout conversion equipment receives the command of the interlocking system to perform turnout conversion. When the locomotive vehicle wants to transfer from track A to track B, the switch machine controls the point rail to move the position, unlocks and disengages the close-fitting point rail on track A from the base rail, and makes the point rail close to and locks the base rail of track B to complete the switch from track A to track B.

[0037] Taking the switching process of a split-motion externally locked turnout as an example, the switch switching process is divided into unlocking, switching and locking. Unlocking includes unlocking inside the machine and unlocking outside the machine, and locking includes locking inside the machine and locking outside the machine. Unlocking inside the machine means that the lock between the point rail and the stock rail is released inside the switch machine; unlocking outside the machine means that the external locking device outside the switch machine releases the lock between the point rail and the stock rail; locking inside the machine means that the lock between the point rail and the stock rail is applied inside the switch machine, and locking outside the machine means that the point rail and the stock rail are locked together by the external locking device outside the switch machine. When the switch machine receives the switch pulling command from the interlocking system, the switch machine starts to move and enters the unlocking process. The switch machine starts to unlock inside and drives the external locking device outside the machine to unlock. After the switch machine and the external locking device are unlocked, the switch machine drives the point rail to start moving and switches the positions of the two point rails. When the close-contact side point rail is moved to the terminal position, that is, the point rail and the base rail are in close contact, the external locking device starts to lock. After the external locking device is locked, the switch machine completes the locking inside, and the indicating rod connected to the switch outside the switch machine reaches the terminal position. The indicating circuit is connected through the switch machine. At this point, the switch completes the conversion.

[0038] Example 1

[0039] Figure 1 A flowchart of a method for locating a turnout fault is provided for the first embodiment of the present invention. This embodiment is applicable to situations where fault diagnosis and locating of a turnout is performed. The method can be performed by a device for locating a turnout fault, which can be implemented in the form of hardware and / or software. The device for locating a turnout fault can be configured in a computer.

[0040] like Figure 1 As shown, the method includes:

[0041] S110. Acquire action curve data corresponding to at least one conversion of the target turnout, wherein the action curve data includes at least one of rod displacement curve data, current curve data, and conversion force curve data corresponding to the conversion of the target turnout.

[0042] The target turnout refers to the turnout for which possible faults during the switching process are monitored in real time. The action curve data refers to the discrete data such as current, pressure, displacement, etc. generated by each rod and switch circuit during the switching process of the target turnout. Figure 2 As shown, the action curve data may include: indicating rod displacement curve data, current curve data and conversion force curve data. Among them, the indicating rod refers to the rod used in the switch machine to check whether the point rail is tightly attached and whether the target switch is in the positioning or reverse position. The indicating rod includes a positioning indicating rod and a reverse indicating rod. The indicating rod displacement curve data refers to curve data that can reflect the displacement amount of the positioning indicating rod and the reverse indicating rod at each sampling time during the switch switching process. The current curve data refers to the curve data of the current value corresponding to each sampling time during the switch switching process inside the switch machine. The conversion force curve data refers to the curve data of the pressure value corresponding to each sampling time during the switch switching process of the action rod that pushes the point rail conversion. In this embodiment, the sampling frequencies of the indicating rod displacement curve data, the current curve data and the conversion force curve data can be the same or different, and this embodiment does not limit this.

[0043] Specifically, the host computer obtains the current switching of the target turnout and the switching before the current switching, and the action curve data corresponding to the switching collected by the current sensor, displacement sensor and pressure sensor inside the switch machine.

[0044] It should be noted that in order to improve the accuracy of turnout fault location, as much action curve data as possible should be obtained.

[0045] S120 , for each action curve data, divide the current action curve data into stages according to the stages that the target turnout goes through during the switching process to obtain target action curve data.

[0046] Among them, the stage refers to the stage that the target turnout goes through in chronological order during the conversion process. Figure 3As shown, the stages include the internal unlocking stage t1-t2, the external unlocking stage t2-t3, the conversion stage t3-t4, the external locking stage t4-t5, the internal locking stage t5-t6, and the indication stage t6-t7. The internal unlocking stage refers to the time range corresponding to the internal release of the lock between the close-fitting switch rail and the stock rail; the external unlocking stage refers to the time range corresponding to the external locking device releasing the lock between the close-fitting switch rail and the stock rail; the conversion stage refers to the time range from the completion of unlocking to the entry into the locking stage, when the two switch rails undergo position conversion; the external locking stage refers to the time range when the external locking device locks the close-fitting switch rail and the stock rail; the internal locking stage refers to the time range when the close-fitting switch rail and the stock rail are locked by the internal locking device; and the indication stage refers to the time range when the switch rail reaches the terminal position, the switch machine connects the indication circuit, and the interlocking system collects the indication signal.

[0047] Furthermore, the switch is switched between the positioning and reverse positions through the switching device. Positioning refers to the position where the target switch is always open, and reverse position refers to the position where the target switch needs to be temporarily changed. The target action curve data refers to the action curve data divided into stages, see Figure 3 .

[0048] Specifically, the start and end times of the positioning and reverse positioning levers can be used to determine the boundary between the internal unlocking and external unlocking phases (t2), the boundary between the external unlocking and conversion phases (t3), and the boundary between the conversion phase and the external locking phase (t4). Because the current exhibits corresponding variations at the start and end of each phase, the start time t1 of the internal unlocking phase, the boundary between the external locking and internal locking phases (t5), the boundary between the internal locking and display phases (t6), and the end time t7 of the display phase can be determined based on the variation patterns of the current curve data.

[0049] S130 , judging whether there is a fault in the conversion at the current moment based on the similarity between the action curve data corresponding to the conversion at the current moment and the historical action curve data corresponding to each conversion before the current moment.

[0050] The similarity refers to the degree of similarity between the motion curve data corresponding to the current transition and the historical motion curve data corresponding to previous transitions. A similarity closer to 1 indicates a more similar motion curve data between the two transitions, while a similarity closer to 0 indicates a less similar motion curve data between the two transitions. Furthermore, the greater the number of historical motion curves whose similarity to the current motion curve data falls outside the preset similarity range, the greater the probability of a fault at the current transition. A fault refers to a problem that may occur during the switch transition process.

[0051] Specifically, the features corresponding to the current curve data, the conversion force curve data and the rod displacement curve data can be extracted to form a feature vector corresponding to the turnout conversion, and the similarity between the feature vectors corresponding to the current conversion and each historical conversion is calculated respectively. The number of feature vectors corresponding to the historical conversions that do not meet the preset similarity range is counted, and a quantity threshold is set in advance. When the historical action curve data that does not meet the preset similarity range is greater than or equal to this quantity threshold, it can be determined that there is a fault in the conversion at the current moment. When the historical action curve data that does not meet the preset similarity range is less than this quantity threshold, it can be determined that there is no fault in the conversion at the current moment.

[0052] For example, the preset similarity range is 0.8-1, and a threshold value of 3 is preset. The feature vector corresponding to the current transition is a0. The feature vectors corresponding to the historical motion curve data corresponding to each transition before the current moment are a1, a2, a3, a4, a5, a6, and a7, respectively. The similarities between a0 and a1, a2, a3, a4, a5, a6, and a7 are calculated to be 0.92, 0.94, 0.95, 0.7, 0.72, 0.65, and 0.5, respectively. Four pieces of historical motion curve data do not meet the preset similarity range, so it can be determined that a fault exists in the current transition.

[0053] S140: If yes, input the action curve data corresponding to the current moment conversion into a pre-trained turnout fault identification model to determine the fault type corresponding to the current moment conversion.

[0054] The turnout fault recognition model refers to a pre-trained neural network model that can identify specific turnout faults. The fault type refers to the type of fault corresponding to the turnout switching fault, such as external locking failure, external locking failure, foreign objects caught in the point rail, and a gap.

[0055] Specifically, the action curve data corresponding to the current conversion can be input into the turnout fault identification model, which ultimately determines the probability corresponding to each fault type and determines the fault type corresponding to the maximum probability as the fault type corresponding to the current conversion.

[0056] For example, the turnout fault identification model outputs a probability of 0.9 for external locking and not unlocking, a probability of 0.2 for external locking and not locking, a probability of 0.3 for a foreign object clamped in the point rail, and a probability of 0.6 for a card gap. It can be determined that the currently converted fault type is external locking and not unlocking.

[0057] In this embodiment, when the sampling frequencies of the sensors are different, it is necessary to resample the action curve data corresponding to the current moment conversion, and fill the resampled current action curve data with zeros to a preset length, and then input it into the pre-trained turnout fault recognition model.

[0058] Resampling refers to sampling the rod displacement curve data, current curve data, and converted force curve data at the same sampling frequency to obtain the same amount of data for each curve data. Preset length refers to the preset amount of data corresponding to the rod displacement curve data, current curve data, and converted force curve data.

[0059] Specifically, since the resampling methods for the rod displacement curve data, current curve data, and conversion force curve data are the same, the resampling of the current curve data will be explained below: if the resampling frequency is n seconds, then starting from time t1, the average can be calculated every n seconds. If the amount of current curve data does not meet the preset length, the end can be padded with zeros to the preset length. After resampling the rod displacement curve data and the conversion force curve data, an N*4 matrix can be obtained, where N is the preset length. The N*4 matrix is input into the pre-trained turnout fault identification model to determine the fault type. The advantage is that it solves the problem of inconsistent data length caused by different sampling frequencies and sampling times of each sensor.

[0060] The technical solution of the embodiment of the present invention obtains action curve data corresponding to at least one conversion of a target turnout, wherein the action curve data includes at least one of rod displacement curve data, current curve data, and conversion force curve data corresponding to the conversion of the target turnout; for each action curve data, the current action curve data is stage-divided according to the various stages experienced by the target turnout during the conversion process to obtain target action curve data; based on the similarity between the action curve data corresponding to the conversion at the current moment and the historical action curve data corresponding to each conversion before the current moment, it is determined whether there is a fault in the conversion at the current moment; if so, the action curve data corresponding to the conversion at the current moment is input into a pre-trained turnout fault recognition model to determine the fault type corresponding to the conversion at the current moment, and by comparing the action curve data corresponding to the current target turnout conversion with the historical action curve data of the target turnout, it is determined whether there is a fault in the current target turnout conversion and the fault point is located. This solves the problem that manual fault diagnosis methods consume a lot of manpower and material resources, are inefficient, and have the problem of misjudgment and missed judgment, and achieves accurate positioning of the fault occurrence point, improves the efficiency of fault identification and fault location, and reduces the manpower cost of fault location.

[0061] In this embodiment, the training method of the turnout fault identification model may include: Figure 4 :

[0062] S101. Acquire training sample data and test sample data, wherein the sample data includes action curve data corresponding to at least one conversion of the target turnout and a corresponding fault type label.

[0063] The training sample data refers to the sample data used to train the model. The test sample data refers to the sample data used to test the turnout fault identification model. The sample data may include the action curve data corresponding to at least one switch of the target turnout and the corresponding fault type label.

[0064] Specifically, the host computer can acquire as much action curve data corresponding to multiple switch transitions as possible, and manually label the corresponding fault type labels as sample data. Furthermore, the sample data also includes action curve data corresponding to switch transitions with faults and action curve data corresponding to switch transitions under normal conditions.

[0065] In this embodiment, after the sample data is obtained, the sample data is resampled and padded to a preset length to obtain the training sample data and the test sample data.

[0066] Specifically, after resampling the current curve data, the rod displacement curve data, and the conversion force curve data, an N*4 matrix can be obtained. After resampling each sample data, a number of training sample data and a number of test sample data with the same length are obtained.

[0067] S102: training the turnout fault recognition model to be trained based on the training sample data to obtain a turnout fault recognition model to be used.

[0068] The turnout fault recognition model to be trained refers to an untrained turnout fault recognition model, and the turnout fault recognition model to be used refers to a trained turnout fault recognition model.

[0069] Specifically, for each training sample data, the current training sample data is input into the turnout fault identification model to be trained to obtain the actual output result corresponding to the current training sample data; based on the expected output result and the actual output result, the loss of the turnout fault identification model to be trained is determined, and the model parameters of the prediction model to be trained are adjusted based on the loss of the turnout fault identification model to be trained to obtain the turnout fault identification model to be used.

[0070] For example, Figure 5As shown in the figure, the switch fault recognition model to be trained can be a CNN-GRU (Convolutional Neural Networks-Gated Recurrent Unit) network model, which includes an input layer, a 1D-CNN convolutional layer, a maximum pooling layer, a gated recurrent unit (GRU) layer, a dropout layer, a fully connected layer, and an output layer:

[0071] Input layer: The input data includes rod displacement curve data, current curve data, and conversion force curve data. To address the issue of inconsistent data length due to different sampling frequencies and sampling times, the input data is preprocessed. First, each set of curve data is resampled. Then, the maximum sample length is determined (set to 600 in this example). Finally, each set of curve data is padded with zeros to the maximum length, resulting in a 600×4 matrix.

[0072] The first 1D-CNN convolutional layer: uses the convolution kernel to extract features from the input data. The convolution kernel size is defined as 100, the step size is 1, the number is 150, and the activation function uses ReLU. After the first convolutional layer, the network outputs a 501×100 neuron matrix.

[0073] The second 1D-CNN convolutional layer: The convolution kernel size is defined as 100, the stride is 1, the number is 150, and the activation function uses ReLU. After the second convolutional layer, the network outputs a 402×100 neuron matrix.

[0074] Pooling layer: To reduce feature complexity and prevent overfitting, a pooling layer is typically used to perform feature selection and information filtering on the features extracted by the convolutional layer. In this embodiment, a maximum pooling layer is selected with a pooling size of 3, reducing the features to 1 / 3. The network outputs a 130×150 neuron matrix.

[0075] The third 1D-CNN convolutional layer: The convolution kernel size is defined as 50, the stride is 1, the number is 100, and the activation function uses ReLU. After the third convolutional layer, the network outputs an 81×100 neuron matrix.

[0076] The fourth 1D-CNN convolutional layer: The convolution kernel size is defined as 80, the stride is 1, the number is 100, and the activation function uses ReLU. After the fourth convolutional layer, the network outputs a 2×100 neuron matrix.

[0077] GRU layer: The GRU layer filters the incoming features through the gate function and saves the important features. The GRU hidden unit is defined as 100, and the network output is a 1×100 neuron matrix.

[0078] Dropout layer: The dropout layer randomly assigns weights of 0 to neurons in the network, making the network insensitive to small changes in the data. In this solution, the weighting ratio is set to 0.4, and the network output is a 1×100 neuron matrix.

[0079] Fully connected layer: reduces the network output to the number of fault categories that need to be identified. In this embodiment, the number of fault categories is 4. The fully connected layer uses Softmax as the excitation function, and the output value represents the possible probability of each category among all fault categories.

[0080] S103: Verify the turnout fault identification model to be used based on the test sample data, so that when the accuracy of the turnout fault identification model to be used reaches a preset accuracy threshold, the turnout fault identification model to be used is used as the target turnout fault identification model.

[0081] The accuracy of the turnout fault identification model to be used refers to the degree of consistency between the actual output of the turnout fault identification model to be used and the result of manual annotation. The preset accuracy threshold is a pre-set accuracy.

[0082] For example, an accuracy threshold of 90% is set in advance. If 100 test sample data are used to test the turnout fault identification model to be used, and 95 output results are consistent with the manual labeling results, it means that the accuracy of the turnout fault identification model to be used is 95%, and the turnout fault identification model to be used is used as the target turnout fault identification model.

[0083] The technical solution of the embodiment of the present invention is to obtain training sample data and test sample data, wherein the sample data includes action curve data corresponding to at least one conversion of the target turnout and the corresponding fault type label; based on the training sample data, a turnout fault recognition model to be trained is trained to obtain a turnout fault recognition model to be used; based on the test sample data, the turnout fault recognition model to be used is verified, so that when the accuracy of the turnout fault recognition model to be used reaches a preset accuracy threshold, the turnout fault recognition model to be used is used as the target turnout fault recognition model, and the target turnout fault recognition model is trained to identify the turnout fault type, thereby improving the efficiency of turnout fault locating, reducing the fault locating cost, and further improving the accuracy of turnout fault identification.

[0084] Example 2

[0085] Figure 6A flowchart of a method for locating a turnout fault is provided for the second embodiment of the present invention. On the basis of the aforementioned embodiment, the current action curve data can be divided into stages according to the various stages experienced by the target turnout during the conversion process for each action curve data, and the target action curve data can be further refined. The specific implementation method can be found in the detailed description of the embodiment of the present invention, wherein the technical terms that are the same as or corresponding to the above embodiment are not repeated here.

[0086] like Figure 6 As shown, the method includes:

[0087] S210. Obtain action curve data corresponding to at least one conversion of the target turnout, wherein the action curve data includes at least one of rod displacement curve data, current curve data, and conversion force curve data corresponding to the conversion of the target turnout.

[0088] S220 , pre-processing the rod displacement curve data to obtain target displacement curve data.

[0089] Preprocessing involves reducing each displacement value in the displacement curve data by a preset multiple and rounding it off. The target displacement curve data refers to the preprocessed displacement curve data. Specifically, the displacement data in the displacement curve data can be reduced by a certain multiple to obtain the target displacement curve data. This eliminates the problem of slight amplitude variations in the displacement curve data even when the turnout is not moving, caused by continuous interference from the external environment or train vibrations.

[0090] S230. Determine, based on the position where the target displacement curve data changes, the boundary time point between the internal unlocking stage and the external unlocking stage, the boundary time point between the external unlocking stage and the conversion stage, and the boundary time point between the conversion stage and the external locking stage.

[0091] Specifically, if Figure 3 As shown, at time t2, the reverse position indicator rod data shows a significant change, indicating that the reverse position indicator rod begins to move and the target turnout enters the external unlocking phase. Therefore, t2 is determined as the dividing time point between the internal unlocking phase and the external unlocking phase. At time t3, the positioning indicator rod displacement begins to change, indicating that the positioning indicator rod begins to move and the target turnout enters the switching phase. Therefore, t3 is determined as the dividing time point between the external unlocking phase and the switching phase. Since the reverse position indicator rod displacement remains essentially unchanged after time t4, it indicates that the reverse position indicator rod has moved to the designated position and the target turnout enters the external locking phase. Therefore, t4 is determined as the dividing time point between the switching phase and the external locking phase.

[0092] S240 : Preprocess the current curve data to obtain target current curve data.

[0093] The pre-processing refers to removing the redundant parts of the current curve data before and after the switch is switched. The target current curve data refers to the pre-processed current curve data.

[0094] Specifically, in order to ensure that the complete turnout current curve data is collected, the current curve data transmitted back to the host computer includes the conversion process and the unconverted current curve data before and after the conversion. Therefore, the current curve data needs to be preprocessed. The method is to set the current data adjustment threshold and adjust the unconverted current curve data before and after the turnout conversion in the current curve data to below zero to obtain the target current curve data. The target current curve data can be found in Figure 7 .

[0095] For example, taking the current curve data of the ZDJ9 switch as an example, the current data of the ZDJ9 switch 3s before and after the switch is about 0.05A-0.07A, so the current data adjustment threshold is set to 0.07, and each current value in the current data curve is subtracted by 0.07 to obtain the target current curve data.

[0096] S250. Based on the target current curve data and the switch conversion relationship, determine the starting time point of the internal unlocking phase, the boundary time point between the external locking phase and the internal locking phase, the boundary time point between the internal locking phase and the display phase, and the end time point of the display phase.

[0097] Specifically, if Figure 7 As shown, based on the conversion relationship between the target current curve data and the turnout, a small locking current is generated at the beginning of the internal locking phase, marking the target turnout entering the internal locking phase. Therefore, the starting time t5 of the locking current is determined as the demarcation point between the external locking phase and the internal locking phase. A "small step" forms during the display phase, indicating that the target turnout has entered the display phase. Therefore, the time t6 when the "small step" occurs is used as the demarcation point between the internal locking phase and the display phase. Furthermore, the first positive current value in the target current curve is used as the starting time t1 of the internal unlocking phase, and the last positive current value is used as the ending time t7 of the display phase.

[0098] S260: Obtain target action curve data based on the time points of each stage.

[0099] The target action curve data refers to the action curve data obtained by dividing the rod displacement curve data, current curve data and conversion force curve data into stages, see Figure 3 .

[0100] Specifically, if Figure 3 As shown, the determined start time and end time of each stage are mapped to the three curve data to obtain the target action curve data.

[0101] S270 , judging whether there is a fault in the conversion at the current moment based on the similarity between the action curve data corresponding to the conversion at the current moment and the historical action curve data corresponding to each conversion before the current moment.

[0102] S280: If yes, input the action curve data corresponding to the current moment conversion into a pre-trained turnout fault identification model to determine the fault type corresponding to the current moment conversion.

[0103] The technical solution of the embodiment of the present invention obtains target displacement curve data by preprocessing the representation rod displacement curve data; determines the boundary time point between the internal unlocking stage and the external unlocking stage, the boundary time point between the external unlocking stage and the conversion stage, and the boundary time point between the conversion stage and the external locking stage based on the position where the target displacement curve data changes; preprocesses the current curve data to obtain target current curve data; and determines the starting time point of the internal unlocking stage, the boundary time point between the external locking stage and the internal locking stage, the boundary time point between the internal locking stage and the representation stage, and the end time point of the representation stage based on the target current curve data and the switch conversion relationship, thereby realizing the stage division of the action curve, so as to facilitate further feature extraction of the data of each stage, and further improve the accuracy of switch conversion fault location.

[0104] Example 3

[0105] Figure 8 A flowchart of a method for locating a turnout fault is provided for embodiment three of the present invention. Based on the above embodiments, the method can further refine the judgment of whether there is a fault in the current moment conversion based on the similarity between the action curve data corresponding to the current moment conversion and the historical action curve data corresponding to each conversion before the current moment. The specific implementation method can refer to the detailed description of the embodiment of the present invention, among which the technical terms that are the same as or corresponding to the above embodiments are not repeated here.

[0106] like Figure 8 As shown, the method includes:

[0107] S301. Acquire action curve data corresponding to at least one conversion of a target turnout, wherein the action curve data includes at least one of rod displacement curve data, current curve data, and conversion force curve data corresponding to the conversion of the target turnout.

[0108] S302 : For each action curve data, the current action curve data is divided into stages according to the stages experienced by the target turnout during the switching process to obtain target action curve data.

[0109] S303: Determine the conversion duration based on the start time point of the in-device unlocking phase and the end time point of the display phase.

[0110] The conversion time refers to the time length corresponding to the target switch conversion.

[0111] Specifically, if Figure 3 As shown, the starting time point of the internal unlocking phase is t1, and the ending time point of the unlocking phase is t7, so the conversion time length T=t7-t1 can be determined.

[0112] S304: Determine a first current difference corresponding to the external unlocking stage, a second current difference corresponding to the conversion stage, and a third current difference corresponding to the external locking stage based on the maximum current and the minimum current corresponding to the external unlocking stage, the conversion stage, and the external locking stage in the current curve data.

[0113] The first current difference is the difference between the maximum and minimum currents during the external unlocking phase. The second current difference is the difference between the maximum and minimum currents during the conversion phase. The third current difference is the difference between the maximum and minimum currents during the external locking phase.

[0114] For example, since the methods for determining the first current difference, the second current difference, and the third current difference are the same, the method for determining the first current difference will be described below: Figure 3 As shown, during the external unlocking phase, i.e., the time range t1-t2, the current data are compared to determine the maximum current value Imax and the minimum current value Imin within the time range t1-t2. The first current difference value I1 = Imax - Imin. The same operation is performed on the current values during the transition phase and the external locking phase to obtain the second current difference value I2 and the third current difference value I3.

[0115] S305 : Determine, based on the conversion force curve data, a first conversion force accumulated value corresponding to the external unlocking stage, a second conversion force accumulated value corresponding to the conversion stage, and a third conversion force accumulated value corresponding to the external locking stage.

[0116] The first conversion force accumulated value refers to the accumulated value of the conversion force data at each moment during the external unlocking phase. The second conversion force accumulated value refers to the accumulated value of the conversion force data at each moment during the conversion phase. The third conversion force accumulated value refers to the accumulated value of the conversion force data at each moment during the external locking phase.

[0117] For example, since the methods for determining the first conversion force accumulated value, the second conversion force accumulated value, and the third conversion force accumulated value are the same, the method for determining the first conversion force accumulated value will be described below: Figure 3 As shown, during the external unlocking phase (i.e., the time range t1-t2), the conversion forces corresponding to each moment are accumulated to determine a first accumulated conversion force value N1. The same operation is performed on the conversion force data during the conversion phase and the external locking phase to obtain a second accumulated conversion force value N2 and a third accumulated conversion force value N3.

[0118] S306 : Determine a characteristic vector corresponding to each action curve data based on the conversion duration, the first current difference, the second current difference, the third current difference, the first conversion force accumulated value, the second conversion force accumulated value, and the third conversion force accumulated value.

[0119] Among them, the characteristic vector refers to the vector that can reflect the characteristics of current, conversion force, rod displacement, etc. corresponding to the important stages in the target turnout conversion process.

[0120] Specifically, since the method for determining the characteristic vector corresponding to each action curve data of the target turnout is the same, the method for determining the characteristic vector corresponding to one of the action curve data is now explained: for the current action curve data, its corresponding conversion time, first current difference, second current difference, third current difference, first conversion force cumulative value, second conversion force cumulative value and third conversion force cumulative value are combined to obtain the corresponding characteristic vector.

[0121] Based on the above example, since the method of determining the characteristic vector corresponding to each action curve data of the target turnout is the same, the method of determining the characteristic vector corresponding to one of the action curve data is now explained: for the current action curve data, its conversion time is T, the first current difference is I1, the second current difference is I2, the third current difference is I3, the first conversion force cumulative value is N1, the second conversion force cumulative value is N2, and the third conversion force cumulative value is N3, then the characteristic vector A1 corresponding to the current action curve data = [T, I1, I2, I3, N1, N2, N3].

[0122] S307: Determine a feature matrix based on the feature vectors corresponding to the action curve data.

[0123] The characteristic matrix refers to a matrix composed of characteristic vectors corresponding to each action curve data.

[0124] For example, if the eigenvectors corresponding to the action curve data are A1, A2, A3, A4, and A5 respectively, then the eigenmatrix is [A1, A2, A3, A4, A5] T ,Furthermore, the rows of the feature matrix represent the ,feature vectors corresponding to each action curve data, and the ,columns represent each feature.

[0125] S308: Perform dimensionality reduction and normalization processing on the feature matrix.

[0126] Dimensionality reduction refers to merging highly correlated features in a feature matrix or removing at least one of them. Normalization refers to converting the data corresponding to each feature in a feature matrix into data between 0 and 1.

[0127] Specifically, the PCA method can be used to reduce the dimension of the feature matrix, and the data in the feature matrix after the dimension reduction process can be normalized. Furthermore, the normalization method can be the maximum-minimum normalization method, and the formula is as follows:

[0128]

[0129] Where i refers to the i-th eigenvector and j refers to the j-th feature. ij It refers to the value obtained after normalizing the data corresponding to the jth feature in the i-th eigenvector. refers to the maximum value corresponding to the jth feature, It refers to the minimum value corresponding to the j-th feature.

[0130] S309: Based on the DBSCAN algorithm, determine whether there is a fault in the current conversion.

[0131] The DBSCAN algorithm is a density-based clustering algorithm used to detect outliers in a dataset. Furthermore, the DBSCAN algorithm can be used to determine whether the feature vector corresponding to the current switch transition differs significantly from the feature vectors corresponding to previous transitions, thereby determining whether the current transition is faulty.

[0132] Specifically, a feature matrix is input as sample data D, where each feature vector is a sample point. The parameters of the neighborhood radius r and the minimum sample number MinPts are set. The Euclidean distance between the current sample point corresponding to the current turnout conversion and other sample points in the feature matrix is calculated. If the number of sample points with a distance less than r from the current sample point is less than the minimum sample number MinPts, the current sample point is determined to be a noise point, that is, there is a fault in the conversion at the current moment. If the number of sample points with a distance less than r from the current sample point is greater than or equal to the minimum sample number MinPts, there is no fault in the conversion at the current moment.

[0133] S310: If yes, input the action curve data corresponding to the current moment conversion into a pre-trained turnout fault identification model to determine the fault type corresponding to the current moment conversion.

[0134] The technical solution of the embodiment of the present invention is to determine the conversion time based on the starting time point of the internal unlocking stage and the ending time point of the representation stage; determine the first current difference corresponding to the external unlocking stage, the second current difference corresponding to the conversion stage, and the third current difference corresponding to the external locking stage based on the current maximum and current minimum values corresponding to the external unlocking stage, the conversion stage, and the external locking stage in the current curve data; determine the first conversion force cumulative value corresponding to the external unlocking stage, the second conversion force cumulative value corresponding to the conversion stage, and the third conversion force cumulative value corresponding to the external locking stage based on the conversion force curve data; and determine the first conversion force cumulative value corresponding to the external unlocking stage, the second conversion force cumulative value corresponding to the conversion stage, and the third conversion force cumulative value corresponding to the external locking stage based on the conversion time, the first current difference, the second current difference, and the third current difference. The method comprises the following steps: determining the characteristic vectors corresponding to the action curve data based on the current difference, the third current difference, the first conversion force accumulated value, the second conversion force accumulated value and the third conversion force accumulated value; determining a characteristic matrix based on the characteristic vectors corresponding to the action curve data; performing dimensionality reduction and normalization processing on the characteristic matrix; judging whether there is a fault in the current conversion based on the DBSCAN algorithm, monitoring the target turnout conversion in real time, extracting the action curve data features, improving data processing efficiency, and determining whether there is a fault in the current turnout conversion based on the DBSCAN algorithm, thereby achieving a preliminary judgment on whether there is a fault in the target turnout, so as to facilitate subsequent fault location.

[0135] Example 4

[0136] Figure 9 This is a structural schematic diagram of a device for locating a turnout fault provided in accordance with a fourth embodiment of the present invention.

[0137] like Figure 9 As shown, the device includes:

[0138] A data acquisition module 410 is configured to acquire action curve data corresponding to at least one conversion of a target turnout, wherein the action curve data includes at least one of rod displacement curve data, current curve data, and conversion force curve data corresponding to the conversion of the target turnout; a stage division module 420 is configured to divide the current action curve data into stages according to the stages experienced by the target turnout during the conversion process to obtain target action curve data; a fault judgment module 430 is configured to judge whether there is a fault in the conversion at the current moment based on the similarity between the action curve data corresponding to the conversion at the current moment and the historical action curve data corresponding to each conversion before the current moment; and a fault identification module 440 is configured to, if so, input the action curve data corresponding to the conversion at the current moment into a pre-trained turnout fault identification model to determine the type of fault corresponding to the conversion at the current moment.

[0139] Optionally, the stages include: an internal unlocking stage, an external unlocking stage, a conversion stage, an external locking stage, an internal locking stage and a display stage.

[0140] Based on the above technical solutions, the stage division module specifically includes:

[0141] a displacement curve processing unit, configured to pre-process the displacement curve data of the representative rod to obtain target displacement curve data;

[0142] a first stage division unit for determining, based on a position at which the target displacement curve data changes, a boundary time point between an internal unlocking stage and an external unlocking stage, a boundary time point between the external unlocking stage and a conversion stage, and a boundary time point between the conversion stage and the external locking stage;

[0143] a current curve processing unit, configured to pre-process the current curve data to obtain target current curve data;

[0144] The second stage division unit is used to determine the starting time point of the internal unlocking stage, the boundary time point between the external locking stage and the internal locking stage, the boundary time point between the internal locking stage and the representation stage, and the end time point of the representation stage based on the target current curve data and the turnout conversion relationship.

[0145] The target action curve data determination unit is used to obtain the target action curve data based on the time points of each stage.

[0146] Based on the above technical solutions, the fault judgment module specifically includes:

[0147] a conversion duration determining unit, configured to determine the conversion duration based on the start time point of the in-machine unlocking phase and the end time point of the display phase;

[0148] a current difference determining unit, configured to determine a first current difference corresponding to the external unlocking stage, a second current difference corresponding to the conversion stage, and a third current difference corresponding to the external locking stage based on maximum current values and minimum current values corresponding to the external unlocking stage, the conversion stage, and the external locking stage in the current curve data;

[0149] a conversion force determination unit, configured to determine, based on the conversion force curve data, a first conversion force cumulative value corresponding to the external unlocking stage, a second conversion force cumulative value corresponding to the conversion stage, and a third conversion force cumulative value corresponding to the external locking stage;

[0150] a characteristic vector determining unit, configured to determine a characteristic vector corresponding to each action curve data based on a conversion time, a first current difference, a second current difference, a third current difference, a first conversion force cumulative value, a second conversion force cumulative value, and a third conversion force cumulative value;

[0151] a characteristic matrix determining unit, configured to determine a characteristic matrix based on the characteristic vectors corresponding to the respective motion curve data;

[0152] A matrix processing unit, used for performing dimensionality reduction and normalization processing on the feature matrix;

[0153] The fault judgment unit is used to judge whether there is a fault in the current conversion based on the DBSCAN algorithm.

[0154] Based on the above technical solutions, the fault identification module also includes:

[0155] The data preprocessing unit is used to resample the action curve data corresponding to the current moment conversion, fill the resampled current action curve data with zeros to a preset length, and input it into the pre-trained turnout fault recognition model.

[0156] Based on the above technical solutions, the device for locating a turnout fault further includes a model determination module, wherein the model determination module specifically includes:

[0157] A sample data acquisition unit, configured to acquire training sample data and test sample data, wherein the sample data includes action curve data corresponding to at least one conversion of the target turnout and a corresponding fault type label;

[0158] A model training module is used to train the turnout fault identification model to be trained based on the training sample data to obtain the turnout fault identification model to be used;

[0159] A model determination module is used to verify the turnout fault identification model to be used based on the test sample data, so that when the accuracy of the turnout fault identification model to be used reaches a preset accuracy threshold, the turnout fault identification model to be used is used as the target turnout fault identification model.

[0160] Based on the above technical solutions, the sample data acquisition unit can also be used to:

[0161] After obtaining the sample data, the sample data is resampled and padded to a preset length to obtain the training sample data and the test sample data.

[0162] The technical solution of the embodiment of the present invention obtains action curve data corresponding to at least one conversion of a target turnout, wherein the action curve data includes at least one of rod displacement curve data, current curve data, and conversion force curve data corresponding to the conversion of the target turnout; for each action curve data, the current action curve data is stage-divided according to the various stages experienced by the target turnout during the conversion process to obtain target action curve data; based on the similarity between the action curve data corresponding to the conversion at the current moment and the historical action curve data corresponding to each conversion before the current moment, it is determined whether there is a fault in the conversion at the current moment; if so, the action curve data corresponding to the conversion at the current moment is input into a pre-trained turnout fault recognition model to determine the fault type corresponding to the conversion at the current moment, and by comparing the action curve data corresponding to the current target turnout conversion with the historical action curve data of the target turnout, it is determined whether there is a fault in the current target turnout conversion and the fault point is located. This solves the problem that manual fault diagnosis methods consume a lot of manpower and material resources, are inefficient, and have the problem of misjudgment and missed judgment, and achieves accurate positioning of the fault occurrence point, improves the efficiency of fault identification and fault location, and reduces the manpower cost of fault location.

[0163] The device for locating a turnout fault provided by an embodiment of the present invention can execute the method for locating a turnout fault provided by any embodiment of the present invention, and has functional modules and beneficial effects corresponding to the execution method.

[0164] Example 5

[0165] Figure 10 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0166] like Figure 10As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by the at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. Various programs and data required for the operation of the electronic device 10 can also be stored in the RAM 13. The processor 11, ROM 12 and RAM 13 are connected to each other via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0167] Multiple components in the electronic device 10 are connected to the I / O interface 15, including an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0168] The processor 11 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any other suitable processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the method for locating a turnout fault.

[0169] In some embodiments, the method for locating a switch fault may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the method for locating a switch fault described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to execute the method for locating a switch fault in any other appropriate manner (e.g., via firmware).

[0170] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0171] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer program may be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0172] In the context of the present invention, computer-readable storage media can be tangible media that can contain or store a computer program for use with an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. Computer-readable storage media can include but are not limited to electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, computer-readable storage media can be machine-readable signal media. More specific examples of machine-readable storage media can include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0173] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0174] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0175] A computing system may include clients and servers. The clients and servers are typically remote from each other and typically interact via a communication network. This client-server relationship arises through computer programs running on the respective computers, creating a client-server relationship. The server may be a cloud server, also known as a cloud computing server or cloud host. This server is a hosting product within the cloud computing service ecosystem that addresses the management difficulties and limited scalability of traditional physical hosting and VPS services.

[0176] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in the present invention can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved. This is not limited herein.

[0177] The above specific embodiments do not limit the scope of protection of the present invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A method for locating a turnout fault, characterized in that: include: Acquiring action curve data corresponding to at least one conversion of a target turnout, wherein the action curve data includes at least one of rod displacement curve data, current curve data, and conversion force curve data corresponding to the conversion of the target turnout; For each action curve data, the current action curve data is divided into stages according to the stages experienced by the target turnout during the switching process to obtain the target action curve data, including: The indicator rod displacement curve data is preprocessed to obtain target displacement curve data; based on the position where the target displacement curve data changes, the boundary time point between the internal unlocking stage and the external unlocking stage, the boundary time point between the external unlocking stage and the conversion stage, and the boundary time point between the conversion stage and the external locking stage are determined; the current curve data is preprocessed to obtain target current curve data; based on the target current curve data and the switch conversion relationship, the starting time point of the internal unlocking stage, the boundary time point between the external locking stage and the internal locking stage, the boundary time point between the internal locking stage and the indication stage, and the end time point of the indication stage are determined; and based on the time points of each stage, target action curve data is obtained; Based on the similarity between the action curve data corresponding to the current moment conversion and the historical action curve data corresponding to each conversion before the current moment, it is determined whether there is a fault in the current moment conversion; If so, the action curve data corresponding to the current moment conversion is input into the pre-trained turnout fault identification model to determine the fault type corresponding to the current moment conversion; wherein the fault type is: external locking does not unlock, external locking does not lock, foreign objects are clamped in the point rail or there is a gap.

2. The method according to claim 1, characterized in that The stages include: an internal unlocking stage, an external unlocking stage, a conversion stage, an external locking stage, an internal locking stage and a display stage.

3. The method according to claim 1, characterized in that The determining whether there is a fault in the current conversion based on the similarity between the action curve data corresponding to the current conversion and the historical action curve data corresponding to each conversion before the current conversion includes: Determine the conversion duration based on the start time of the in-device unlocking phase and the end time of the presentation phase; Determine, based on the maximum current values and the minimum current values corresponding to the external unlocking stage, the conversion stage, and the external locking stage in the current curve data, a first current difference corresponding to the external unlocking stage, a second current difference corresponding to the conversion stage, and a third current difference corresponding to the external locking stage; Determining, based on the conversion force curve data, a first conversion force cumulative value corresponding to the external unlocking stage, a second conversion force cumulative value corresponding to the conversion stage, and a third conversion force cumulative value corresponding to the external locking stage; Determining a characteristic vector corresponding to each action curve data based on the conversion time, the first current difference, the second current difference, the third current difference, the first conversion force accumulated value, the second conversion force accumulated value, and the third conversion force accumulated value; Determine a characteristic matrix based on the characteristic vectors corresponding to the action curve data; Performing dimensionality reduction and normalization processing on the feature matrix; Based on the DBSCAN algorithm, determine whether there is a fault in the current conversion.

4. The method according to claim 1, wherein inputting the action curve data corresponding to the current moment conversion into a pre-trained turnout fault identification model comprises: The action curve data corresponding to the current moment conversion is resampled, the resampled current action curve data is padded with zeros to a preset length, and input into the pre-trained turnout fault recognition model.

5. The method according to claim 1, wherein Also includes: Acquire training sample data and test sample data, wherein the sample data includes action curve data corresponding to at least one conversion of the target turnout and a corresponding fault type label; Performing training processing on the turnout fault identification model to be trained based on the training sample data to obtain the turnout fault identification model to be used; The turnout fault identification model to be used is verified based on the test sample data, so that when the accuracy of the turnout fault identification model to be used reaches a preset accuracy threshold, the turnout fault identification model to be used is used as the target turnout fault identification model.

6. The method according to claim 5, characterized in that include: After obtaining the sample data, the sample data is resampled and padded to a preset length to obtain the training sample data and the test sample data.

7. A device for locating turnout faults, characterized in that: include: a data acquisition module for acquiring action curve data corresponding to at least one conversion of a target turnout, wherein the action curve data includes at least one of rod displacement curve data, current curve data, and conversion force curve data corresponding to the conversion of the target turnout; a stage division module for dividing the current action curve data into stages according to the stages that the target turnout experiences during the conversion process, thereby obtaining target action curve data; the stage division module specifically comprises: a displacement curve processing unit for pre-processing the representation rod displacement curve data to obtain target displacement curve data; a first stage division unit for determining, based on the position at which the target displacement curve data changes, the boundary time point between the internal unlocking stage and the external unlocking stage, the boundary time point between the external unlocking stage and the conversion stage, and the boundary time point between the conversion stage and the external locking stage; a current curve processing unit for pre-processing the current curve data to obtain target current curve data; a second stage division unit for determining, based on the target current curve data and the conversion relationship between the turnout and the internal unlocking stage, the boundary time point between the external locking stage and the internal locking stage, the boundary time point between the internal locking stage and the representation stage, and the end time point of the representation stage; a target action curve data determination unit for obtaining target action curve data based on the time points of each stage; A fault judgment module judges whether there is a fault in the current conversion based on the similarity between the action curve data corresponding to the current conversion and the historical action curve data corresponding to each conversion before the current conversion; A fault identification module is used to input the action curve data corresponding to the current moment conversion into a pre-trained turnout fault identification model to determine the fault type corresponding to the current moment conversion; wherein the fault type is: external locking does not unlock, external locking does not lock, foreign objects are clamped in the point rail, or there is a gap.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor. The computer program is executed by the at least one processor to enable the at least one processor to perform the method for locating a turnout fault according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method for locating a turnout fault according to any one of claims 1 to 6 when executed.

Images