A platform screen door for rail transit
By using monitoring sensors and fault warning devices in rail transit station shield doors, combined with prediction model and fault monitoring model, accurate prediction of shield door faults is achieved, the problem of fault monitoring error in the existing technology is solved, and operation reliability is improved sex.
Patent Information
- Application Number
- CN202411296324.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-18
- Publication Date
- 2024-12-27
- Estimated Expiration
- 2044-09-18
AI Technical Summary
The fault monitoring method of existing rail transit station shielded doors cannot accurately predict the fault time and type, resulting in maintenance errors and affecting operational reliability.
The monitoring sensor and fault warning device are used to obtain real-time operation data, and multiple rounds of cycle segmentation prediction are performed through the pre-constructed prediction model. Combined with the fault monitoring model, the fault type monitoring results are output to accurately predict potential faults.
Accurate and efficient fault monitoring, reduce maintenance errors, and improve the operating reliability and stability of shield doors.
Smart Images
Figure CN118810834B_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present application relate to the technical field of platform screen doors, and in particular to a platform screen door for a rail transit station. Background Art
[0002] At present, during the operation of the platform screen door of rail transit, the control module is used to realize the automatic opening and closing of the screen door body. In the screen door operation and maintenance scenario of the platform screen door, its maintenance mode is mainly to monitor the operating status of the screen door and then perform fault repair or overhaul operations. Among them, in order to give early warning of the screen door failure, the fault prediction model will be used to analyze the equipment reliability according to the status information of the screen door. When the equipment reliability drops to the set value, the equipment operation failure will be warned, and potential operation failures will be monitored in time to ensure the stability of the screen door operation.
[0003] However, the current fault monitoring method of the shielding door based on reliability analysis cannot accurately give the time and type of the shielding door failure, resulting in a certain error in the shielding door fault monitoring, and maintenance personnel cannot accurately determine the fault type and the time node of preventive maintenance based on the fault warning. Its maintenance auxiliary effect on the shielding door is poor, which easily leads to excessive maintenance or delayed maintenance, thereby affecting the reliability of the shielding door operation. Summary of the invention
[0004] The embodiment of the present application provides a rail transit platform screen door, which can accurately predict its own operating failures, timely monitor potential operating failures, and solve the problem of fault monitoring errors of the screen door.
[0005] In a first aspect, an embodiment of the present application provides a rail transit platform screen door comprising:
[0006] A controller, used for controlling the shielding door body to perform opening and closing operations;
[0007] Monitoring sensors are used to collect real-time operation data of the platform shielding door based on the opening and closing operations of the platform shielding door body;
[0008] A fault early warning device is used to obtain the real-time operation data, input the real-time operation data into a pre-built prediction model, perform multiple rounds of cyclic segmented prediction on the real-time operation data based on the prediction model, and output the operation prediction results of each time node in the target time period. In the multiple rounds of cyclic segmented prediction process, the real-time operation data is used as the initial model input data of the prediction model, and each round of prediction outputs a segmented prediction data of the target time period, and the predicted segmented prediction data and the real-time operation data are used as the model input data of the next round of prediction, and the operation prediction results are input into a pre-built fault monitoring model, and the fault type monitoring results of the corresponding time node are output based on the fault monitoring model. The fault monitoring model is pre-trained based on the fault operation data of the shielding door, and the fault operation data is pre-labeled with the corresponding fault type label;
[0009] The training process of the prediction model includes:
[0010] Collect historical operation data of the platform shielding door to construct training samples, and split the training samples into a first sequence and a second sequence according to time sequence;
[0011] Inputting the first sequence into the prediction model to perform multiple rounds of cyclic segmented prediction to obtain a corresponding prediction sequence;
[0012] According to the calculated loss function of the predicted sequence and the second sequence, the model parameters of the prediction model are iteratively adjusted based on the loss function until the prediction model converges.
[0013] Furthermore, the historical operation data includes shielding door switch data, door displacement data and door speed data collected at different historical time nodes.
[0014] Further, the prediction model includes a first multi-layer perception module, a decoder network, and a second multi-layer perception module connected in sequence;
[0015] The first multi-layer perception module is used to convert the model input data of the prediction model into a vector of a set shape;
[0016] The decoder network is used to generate an initial prediction sequence based on the vector of the set shape;
[0017] The second multi-layer perception module is used to convert the initial prediction sequence into the segmented prediction data according to the shape of the model input data.
[0018] Furthermore, the calculation formula of the first multi-layer perception module is expressed as:
[0019]
[0020] in,w represents the weights of the hidden layers of the first multi-layer perception module, b represents the bias term of the hidden layer, relu is a nonlinear activation function, x out is the output of the first multi-layer perception module, x It is the input of the first multi-layer perception module.
[0021] Furthermore, before inputting the real-time operation data into the pre-built prediction model, the method further includes:
[0022] The real-time operation data is processed by filling missing values and filtering out abnormal values, and the processed real-time operation data is standardized.
[0023] Furthermore, the standardization process is used to standardize the real-time operation data into sequence data with a mean of 0 and a standard deviation of 1.
[0024] Furthermore, the monitoring sensor is a monitoring camera, which performs target detection and positioning on the platform shielding door body and determines the real-time operation data of the platform shielding door based on the detected position of the platform shielding door body.
[0025] In a second aspect, an embodiment of the present application provides a fault monitoring method for a platform screen door, which is applied to the rail transit platform screen door as described in the first aspect, comprising:
[0026] Acquire real-time operation data of the platform shielding door, and input the real-time operation data into a pre-built prediction model;
[0027] Based on the prediction model, multiple rounds of cyclic segmented prediction are performed on the real-time operation data, and the operation prediction results of each time node in the target time period are output. In the multiple rounds of cyclic segmented prediction process, the real-time operation data is used as the initial model input data of the prediction model, and each round of prediction outputs a segmented prediction data of the target time period, and the predicted segmented prediction data and the real-time operation data are used as the model input data of the next round of prediction;
[0028] Input the operation prediction result into a pre-built fault monitoring model, and output the fault type monitoring result of the corresponding time node based on the fault monitoring model, wherein the fault monitoring model is pre-trained based on the fault operation data of the shielding door, and the fault operation data is pre-labeled with the corresponding fault type label;
[0029] The training process of the prediction model includes:
[0030] Collect historical operation data of the platform shielding door to construct training samples, and split the training samples into a first sequence and a second sequence according to time sequence;
[0031] Inputting the first sequence into the prediction model to perform multiple rounds of cyclic segmented prediction to obtain a corresponding prediction sequence;
[0032] According to the calculated loss function of the predicted sequence and the second sequence, the model parameters of the prediction model are iteratively adjusted based on the loss function until the prediction model converges.
[0033] In a third aspect, an embodiment of the present application provides a fault monitoring device for a shielding door, comprising:
[0034] An input module, used for acquiring real-time operation data of the platform shielding door, and inputting the real-time operation data into a pre-built prediction model;
[0035] A data prediction module is used to perform multiple rounds of cyclic segmented prediction on the real-time operation data based on the prediction model, and output the operation prediction results of each time node in the target time period. In the multiple rounds of cyclic segmented prediction process, the real-time operation data is used as the initial model input data of the prediction model, and each round of prediction outputs a segmented prediction data of the target time period, and the predicted segmented prediction data and the real-time operation data are used as the model input data of the next round of prediction;
[0036] A fault monitoring module, used for inputting the operation prediction result into a pre-built fault monitoring model, and outputting the fault type monitoring result of the corresponding time node based on the fault monitoring model, wherein the fault monitoring model is pre-trained based on the fault operation data of the shielding door, and the fault operation data is pre-labeled with the corresponding fault type label;
[0037] The training process of the prediction model includes:
[0038] Collect historical operation data of the platform shielding door to construct training samples, and split the training samples into a first sequence and a second sequence according to time sequence;
[0039] Inputting the first sequence into the prediction model to perform multiple rounds of cyclic segmented prediction to obtain a corresponding prediction sequence;
[0040] According to the calculated loss function of the predicted sequence and the second sequence, the model parameters of the prediction model are iteratively adjusted based on the loss function until the prediction model converges.
[0041] In a fourth aspect, an embodiment of the present application provides a storage medium comprising computer executable instructions, which, when executed by a computer processor, are used to execute the fault monitoring method for a shielding door as described in the second aspect.
[0042] The embodiment of the present application obtains the real-time operation data of the shielding door by setting the monitoring sensor and fault alarm of the shielding door, and inputs the real-time operation data into the pre-built prediction model; performs multiple rounds of cyclic segmented prediction on the real-time operation data based on the prediction model, and outputs the operation prediction results of each time node in the target period. In the process of multiple rounds of cyclic segmented prediction, the real-time operation data is used as the initial model input data of the prediction model, and each round of prediction outputs a segmented prediction data of the target period, and the predicted segmented prediction data and the real-time operation data are used as the model input data of the next round of prediction; the operation prediction results are input into the pre-built fault monitoring model, and the fault type monitoring results of the corresponding time node are output based on the fault monitoring model. The fault monitoring model is pre-trained based on the fault operation data of the shielding door, and the fault operation data is pre-labeled with the corresponding fault type label. The above technical means are adopted to predict the operation prediction results of the shielding door at each time node in the future target period through cyclic segmented prediction, and then perform fault monitoring based on the operation prediction results, timely discover the potential fault type of the corresponding time node, realize accurate and efficient fault monitoring, and thus improve the reliability and stability of the shielding door operation. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] Figure 1 This is a schematic diagram of the structural connection of a rail transit platform screen door provided in Example 1 of the present application;
[0044] Figure 2 It is a flow chart of a fault monitoring method of a shielding door provided in the first embodiment of the present application;
[0045] Figure 3 is a training flow chart of the prediction model in Example 1 of the present application;
[0046] Figure 4 is a schematic diagram of sequence prediction of the prediction model in Example 1 of the present application;
[0047] Figure 5 is the fault monitoring flow chart in the first embodiment of the present application;
[0048] Figure 6 It is a structural schematic diagram of a fault monitoring device for a shielding door provided in the second embodiment of the present application;
[0049] Figure 7 It is a structural schematic diagram of an electronic device provided in Example 3 of the present application. DETAILED DESCRIPTION
[0050] In order to make the purpose, technical scheme and advantages of the present application clearer, the specific embodiments of the present application are further described in detail below in conjunction with the accompanying drawings. It is understood that the specific embodiments described herein are only used to explain the present application, rather than to limit the present application. It should also be noted that, for the convenience of description, only part of the present application is shown in the accompanying drawings, but not all of the content. Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow chart describes each operation (or step) as a sequential process, many of the operations therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but it can also have additional steps not included in the accompanying drawings. The process can correspond to a method, a function, a procedure, a subroutine, a subprogram, etc.
[0051] Embodiment 1:
[0052] Figure 1 A rail transit platform screen door provided in Example 1 of the present application is given, referring to Figure 1 The rail transit platform screen door includes:
[0053] The controller 11 is used to control the shielding door body 10 to perform opening and closing operations;
[0054] A monitoring sensor 12, used for collecting real-time operation data of the shielding door based on the opening and closing operations of the shielding door body 10;
[0055] The fault warning device 13 is used to obtain real-time operation data, input the real-time operation data into a pre-built prediction model, perform multiple rounds of cyclic segmented prediction on the real-time operation data based on the prediction model, and output the operation prediction results of each time node in the target time period. In the process of multiple rounds of cyclic segmented prediction, the real-time operation data is used as the initial model input data of the prediction model, and each round of prediction outputs a segmented prediction data of the target time period, and the predicted segmented prediction data and the real-time operation data are used as the model input data of the next round of prediction, and the operation prediction results are input into a pre-built fault monitoring model, and the fault type monitoring results of the corresponding time node are output based on the fault monitoring model. The fault monitoring model is pre-trained based on the fault operation data of the shielding door, and the fault operation data is pre-labeled with the corresponding fault type label.
[0056] By setting monitoring sensors and fault alarms on the corresponding shielding door bodies, the operating status of the shielding door can be monitored in real time, and an alarm can be issued when a fault is detected. During the process of the controller 11 controlling the opening and closing operations of the shielding door body, the monitoring sensor 12 detects the opening and closing operations of the shielding door body, and then obtains the corresponding real-time operating data of the shielding door. The real-time operating data is accurately analyzed by the fault alarm 13 to achieve the effect of fault alarm.
[0057] Specifically, Figure 2 A flow chart of a fault monitoring method for a shielding door provided in the first embodiment of the present application is given. The fault monitoring method for a shielding door provided in this embodiment can be executed by a fault early warning device for the shielding door.
[0058] The following description is made by taking the fault early warning device of the shielding door as the main body of the shielding door fault monitoring method as an example. Figure 2 , the fault monitoring method of the shielding door specifically includes:
[0059] S110 , obtaining real-time operation data of the platform shielding door, and inputting the real-time operation data into a pre-built prediction model.
[0060] When performing fault monitoring of the shielding door, the present application collects real-time operating data of the shielding door, and then uses the real-time operating data to predict the operating status of the shielding door in the future time period through a pre-built prediction model, so as to perform accurate and timely fault monitoring according to the predicted operating status of the shielding door.
[0061] Among them, the monitoring sensor 12 is a monitoring camera. The monitoring camera detects and locates the shielding door body and determines the real-time operation data of the shielding door based on the detected shielding door body position. The real-time operation data can be shielding door switch data, door body displacement data, door body speed data, etc. The monitoring camera is installed at a position where the movement trajectory of the shielding door body can be clearly captured, such as the side or top of the shielding door. According to the monitoring requirements, the focal length, exposure, frame rate and other parameters of the camera are adjusted to ensure that the image is clear and stable, and reduce noise interference. The monitoring camera collects the video stream by corresponding to the shielding door body in real time. The video stream captured by the camera is first processed by image processing, including denoising, contrast enhancement, edge detection and other steps to improve the image quality and facilitate subsequent analysis. Then, computer vision technology (such as convolutional neural network CNN in deep learning) is used to detect the target of the processed image and identify the position of the shielding door body. Prior to this, a target detection model is trained so that the model can recognize specific objects (i.e., shielding door bodies) in the image. After the platform shielding door body is detected, the position of the platform shielding door body in the three-dimensional space is calculated based on the pixel coordinates in the image and the geometric parameters of the camera (such as focal length, viewing angle, etc.), and then based on camera calibration and three-dimensional reconstruction technology.
[0062] By monitoring the position change of the door body, it is possible to determine whether the shielding door is in the open or closed state. When the door body moves from the closed position to the open position (or vice versa), it can be recorded as a switch operation; by continuously monitoring the change of the door body position, the displacement of the door body in a specific time period can be calculated; by comparing the change of the door body position in two consecutive frames of images, combined with the time interval, the instantaneous speed of the door body can be calculated. In this way, the real-time operation data such as the shielding door switch data, door body displacement data, and door body speed data are obtained, and the real-time operation data processed by the monitoring camera is transmitted to the fault warning device in real time.
[0063] Optionally, in the application, the operation data of the shielding door can also be collected in real time through nodes such as infrared sensors, displacement sensors, and processors of the shielding door. The shielding door switch data can be motor current data, voltage data, infrared sensor data, etc. of the door switch. According to the actual fault monitoring requirements, the real-time operation data can also be parameters such as door position, door opening and closing time, current and voltage, temperature, vibration, etc. This application does not impose fixed restrictions on the types of specific real-time operation data, and will not be elaborated here.
[0064] The collected real-time operation data is input into the pre-trained prediction model to predict the operation status of the platform shielding door in the future. Prior to this, the historical operation data of the platform shielding door is collected to train the prediction model based on the historical operation data.
[0065] Reference Figure 3, the training process of the prediction model includes:
[0066] S1001, collecting historical operation data of the platform shielding door to construct training samples, and splitting the training samples into a first sequence and a second sequence according to time sequence;
[0067] S1002, inputting the first sequence into the prediction model to perform multiple rounds of cyclic segmented prediction to obtain a corresponding prediction sequence;
[0068] S1003. According to the calculated loss function of the prediction sequence and the second sequence, iteratively adjust the model parameters of the prediction model based on the loss function until the prediction model converges.
[0069] By collecting the actual historical operation data of the shielding door, this part of the historical operation data corresponds to the type of the above-mentioned real-time operation data, and the data type is set according to actual needs. Optionally, the historical operation data includes shielding door switch data, door displacement data and door speed data collected at different historical time nodes. These historical operation data are preprocessed according to the order of the historical time nodes collected, such as data cleaning, data formatting (unified data format, dimension normalization) and data partitioning (dividing the data set into training set, verification set and test set). Then, the historical operation data is divided into multiple sample pairs according to the time sequence, each sample pair contains multiple groups of operation data sorted by time nodes (each group of data records shielding door switch data, door displacement data and door speed data once), and each sample is divided into an input sequence (first sequence) and a target sequence (second sequence). The first sequence is used for model prediction, and the second sequence is used to calculate the model prediction error.
[0070] After completing the construction of the above training samples, initialize the prediction model, set the initial parameters of the model, and then use the training samples to perform multiple rounds of cyclic segmented prediction. Among them, for each sample pair of the training samples, the first sequence is input into the prediction model, and then multiple rounds of cyclic segmented prediction are performed inside the model. Each round of prediction outputs the prediction results for a time period, and then each round of prediction uses the predicted results as the input part of the next round of prediction, and splices them with the first sequence as the model input for the next round. Repeat the above process until the prediction of the entire sequence is completed and a complete prediction sequence is obtained.
[0071] Based on the prediction sequence, the set loss function (such as mean square error, cross entropy, etc.) is used to calculate the difference between the prediction sequence and the second sequence. The difference reflects the accuracy of the model prediction. According to the value of the loss function, the corresponding optimization algorithm can be used to iteratively adjust the parameters of the prediction model. After each prediction of a set of sample pairs is completed, the loss function is calculated and the model parameters are adjusted. By monitoring the changes in the loss function, when the loss function value reaches the set value, or reaches the preset training rounds, time and other conditions, it is determined that the prediction model has converged and the training process is completed.
[0072] Optionally, during the actual training process, the validation set and test set can also be used to evaluate the performance of the trained prediction model. The validation set is used to adjust the model hyperparameters during the training process to prevent overfitting; the test set is used to finally evaluate the generalization ability of the model. By training a prediction model that can accurately predict the future operating status of the shielding door, the operating status of the shielding door in the future period can be accurately predicted, and then fault monitoring can be performed based on the operating status of the shielding door in the future period to timely discover potential fault types.
[0073] Specifically, the prediction model of the present application includes a first multi-layer perception module, a decoder network, and a second multi-layer perception module connected in sequence;
[0074] The first multi-layer perception module is used to convert the model input data of the prediction model into a vector of a set shape;
[0075] The decoder network is used to generate an initial prediction sequence based on a vector of a given shape,
[0076] The second multi-layer perception module is used to convert the initial prediction sequence into segmented prediction data according to the shape of the model input data.
[0077] Among them, refer to Figure 4 By adding a multi-layer perception module on the basis of the decoder network, the model can adapt to different input sequence shapes. The model uses multiple cycles to generate the final prediction sequence. Since each input sequence will splice the last predicted sequence at the back, the length of the sequence in the second dimension will be increased. Therefore, in order to solve the problem of inconsistent input sequence shapes, the application first processes the data through the first multi-layer perception module before inputting the data into the model, so that the data shape input to the decoder network is consistent.
[0078] The first multi-layer perception module includes an input layer, an output layer and multiple hidden layers, which converts the first sequence of the input model into a set shape through nonlinear transformation. The second multi-layer perception module converts the output of the decoder network into the shape of the original sequence, which is structurally opposite to the first multi-layer perception module. The calculation formula of the first multi-layer perception module is expressed as:
[0079]
[0080] in, w represents the weights of the hidden layer of the first multi-layer perception module, b represents the bias term of the hidden layer, relu is a nonlinear activation function, x out is the output of the first multi-layer perception module, x It is the input of the first multi-layer perception module.
[0081] like Figure 4 As shown, the shape of the input sequence of the first multi-layer perception module is ,in is the length of a single sequence, m The operating data of the initial input sequence, such as the number of opening and closing operations of the shielding door, k For the predicted operating data, n is the number of cycles in a prediction. K The run data is divided into n In the first stage of prediction, the initial sequence (the first sequence at this time) is input into the model to obtain the following Run the data for the second stage of prediction, add the initial sequence to the predicted k The data from the previous runs are fed into the model together to predict the This cycle repeats until the prediction is complete. K Run the data. Each time the input sequence shape of the model is ,…, Using the above method, the prediction model only needs to cycle n times to complete all predictions, which can not only ensure the accuracy of long-term predictions, but also improve the reasoning speed.
[0082] Based on the above-mentioned trained prediction model, during the process of platform shielding door fault monitoring, the acquired real-time operation data of the platform shielding door is input into the prediction model, and the operation status of the platform shielding door in the future period can be determined through model reasoning.
[0083] Optionally, before feeding the live operational data into the pre-built predictive model, also include:
[0084] The real-time operation data is processed to fill missing values and filter out outliers, and the processed real-time operation data is standardized.
[0085] It is understandable that during the collection of operating data, due to sensor failure or communication failure, a very small amount of abnormal data and missing data will be collected. This part of the data will seriously affect the performance of the model and needs to be processed. Based on this, the present application uses data filling to process the missing values in the operating data. The filling value can be a historical mean or a set value. For example, the average value of the 10 values before and after the missing position is filled into the missing position. For outliers, the entire sequence containing the outliers is deleted by deletion.
[0086] After the processing is completed, the real-time running data is standardized to meet the input and reasoning requirements of the model. The standardization process is used to standardize the real-time running data into sequence data with a mean of 0 and a standard deviation of 1.
[0087] Standardization refers to standardizing a sequence of data into a standard normal distribution with a mean of 0 and a standard deviation of 1. The formula for standardization is as follows:
[0088]
[0089] in x 0 is the original value of the running data, μ is the mean, σ is the standard deviation, x ´ is the value after normalization. After completing the above processing, the real-time operation data can be input into the prediction model for model inference.
[0090] It should be noted that in the process of predictive model training, in order to accelerate the convergence of the model in the subsequent training model, the historical operation data also needs to be filled with missing values, deleted outliers and standardized. This application does not impose fixed restrictions on the preprocessing process of the operation data, and will not be elaborated here.
[0091] S120. Perform multiple rounds of cyclic segmented prediction on the real-time operation data based on the prediction model, and output the operation prediction results of each time node in the target time period. In the multiple rounds of cyclic segmented prediction process, the real-time operation data is used as the initial model input data of the prediction model. Each round of prediction outputs a segmented prediction data of the target time period, and the predicted segmented prediction data and the real-time operation data are used as the model input data for the next round of prediction.
[0092] Furthermore, during the inference process, the prediction model refers to the segmented prediction process during the above-mentioned model training to perform multiple rounds of cyclic segmented prediction on the input real-time operation data. Define the future period to be predicted as the target period. Using the real-time operation data as the initial input sequence, the first stage predicts the operation data of the first segment of the target period, which is defined as the first segmented prediction data; the first segmented prediction data and the real-time operation data are spliced as the input sequence for the second stage prediction, and the second stage predicts the second segmented prediction data for the target period; the first segmented prediction data, the second segmented prediction data and the real-time operation data are spliced as the input sequence for the third stage prediction, and the third stage prediction reasoning is performed. And so on, until the prediction of all segmented prediction data of the target period is completed, all segmented prediction data are spliced in time sequence, and the operation prediction results of each time node of the target period are obtained.
[0093] For example, if the operating status of the shielding door in the next 24 hours (target period) is to be predicted, the prediction is performed by inputting real-time operating data (such as real-time operating data within the current hour). The target period is divided into multiple time periods of equal length, each of which has a fixed length (such as one period per hour). In the multi-round segmented prediction process of the prediction model, the first round of prediction uses the initial real-time operating data as input, and the operating data of the first hour in the future is predicted to obtain the prediction result of the time period. This prediction result is saved and used as part of the input in the next round of prediction. In the second round of prediction, the prediction result of the first time period obtained by the first round of prediction is used together with the real-time operating data as input to predict the prediction result of the second hour. Repeat the above process, each time using the predicted prediction result and real-time operating data as input, and perform the next round of prediction until the prediction of the operating data of the 24th hour is completed, and all the predicted operating data for the next 24 hours are output, thereby completing the prediction of the operating status of the shielding door.
[0094] S130, inputting the operation prediction result into a pre-built fault monitoring model, and outputting the fault type monitoring result of the corresponding time node based on the fault monitoring model, the fault monitoring model is pre-trained based on the fault operation data of the shielding door, and the fault operation data is pre-labeled with the corresponding fault type label.
[0095] Finally, the above operation prediction results are input into the pre-built fault monitoring model to diagnose the potential fault types that may occur at the corresponding time nodes in the future period. The fault monitoring model can be built using a CNN convolutional neural network. CNN can effectively extract local features in the data and accelerate model training through local receptive fields and parameter sharing mechanisms. By inputting the operation prediction results into the fault monitoring model, the model processes through convolutional layers, pooling layers, and fully connected layers, and finally outputs the fault monitoring results of the shielding door at each time node. If a fault is monitored, the fault monitoring results contain the corresponding fault type.
[0096] Prior to this, the equipment was manually adjusted to normal and multiple fault states, and the operating data during this period was collected to obtain the operating data of the shield door in each fault type state. The collected fault operation data was then labeled, that is, each fault operation data was assigned a corresponding fault type label. The labeled fault operation data was then used as a training set to train the fault monitoring model. During the training process, the model will learn to extract features from the fault operation data and establish a mapping relationship between features and fault types.
[0097] After the model training is completed, based on the operation prediction results of different time nodes in the target period of the input fault monitoring model, the fault monitoring model performs fault diagnosis on the operation prediction results of each time node in time sequence to determine whether there is an operation fault at the time node. If so, the corresponding fault type is output. By analyzing the results output by the fault monitoring model, it is determined whether there is a potential fault risk. If the model predicts that there is a risk of a corresponding fault type at a certain time node, further attention is required. For example, according to the analysis results, formulate corresponding fault response measures. Including arranging maintenance plans in advance, increasing monitoring frequency, adjusting operating parameters, etc. If the predicted fault risk is very high, emergency measures can be taken immediately to prevent the fault from occurring. In this way, by timely discovering the potential operation faults of the shielding door in the future period, timely response measures can be taken to ensure the reliability and stability of the shielding door operation.
[0098] Reference Figure 5, by using unlabeled data to build a prediction model data set, build a prediction model for model training, and output the operation prediction results based on the input real-time operation data. On the other hand, by using labeled data to build a fault monitoring model data set. In this process, by simulating the operation status of the shielding door under various fault conditions, accurate and high-availability operation data can be collected to build a fault monitoring model data set, and only a small amount of data is needed for model training. Based on the trained fault monitoring model, by inputting the operation prediction results output by the above prediction model, the fault monitoring model is used to diagnose whether the shielding door will fail in the future period and the possible fault type. A predictive maintenance report is formed based on the results of prediction and diagnosis, which can assist the staff to complete the inspection and maintenance of the shielding door. By predicting in advance, the number of times the shielding door is shut down for maintenance is reduced. And by predicting the fault type, the difficulty of fault screening can be reduced, effectively improving work efficiency.
[0099] In the above, by setting the monitoring sensor and fault alarm of the shielding door, the real-time operation data of the shielding door is obtained, and the real-time operation data is input into the pre-built prediction model; based on the prediction model, the real-time operation data is predicted in multiple rounds of cyclic segmentation, and the operation prediction results of each time node in the target period are output. In the process of the multiple rounds of cyclic segmentation prediction, the real-time operation data is used as the initial model input data of the prediction model, and each round of prediction outputs a segmented prediction data of the target period, and the predicted segmented prediction data and the real-time operation data are used as the model input data of the next round of prediction; the operation prediction result is input into the pre-built fault monitoring model, and the fault type monitoring result of the corresponding time node is output based on the fault monitoring model. The fault monitoring model is pre-trained based on the fault operation data of the shielding door, and the fault operation data is pre-labeled with the corresponding fault type label. The above technical means are adopted to predict the operation prediction results of the shielding door at each time node in the future target period by cyclic segmentation, and then perform fault monitoring based on the operation prediction results, timely discover the potential fault type of the corresponding time node, realize accurate and timely fault monitoring, and thus improve the reliability and stability of the shielding door operation.
[0100] Embodiment 2:
[0101] Based on the above embodiments, Figure 6 This is a schematic diagram of the structure of a fault monitoring device for a shielding door provided in the second embodiment of the present application. Figure 6 The fault monitoring device for the shielding door provided in this embodiment specifically includes: an input module 21, a data prediction module 22 and a fault monitoring module 23.
[0102] The input module 21 is used to obtain the real-time operation data of the shielding door and input the real-time operation data into the pre-built prediction model;
[0103] The data prediction module 22 is used to perform multiple rounds of cyclic segmented prediction on the real-time operation data based on the prediction model, and output the operation prediction results of each time node in the target period. In the multiple rounds of cyclic segmented prediction process, the real-time operation data is used as the initial model input data of the prediction model, and each round of prediction outputs a segmented prediction data of the target period, and the predicted segmented prediction data and the real-time operation data are used as the model input data of the next round of prediction;
[0104] The fault monitoring module 23 is used to input the operation prediction results into a pre-built fault monitoring model, and output the fault type monitoring results of the corresponding time node based on the fault monitoring model. The fault monitoring model is pre-trained based on the fault operation data of the shielding door, and the fault operation data is pre-labeled with the corresponding fault type label.
[0105] Specifically, the training process of the prediction model includes:
[0106] The historical operation data of the platform shielding door is collected to construct training samples, and the training samples are divided into a first sequence and a second sequence according to the time sequence;
[0107] The first sequence is input into the prediction model for multiple rounds of cyclic segmented prediction to obtain the corresponding prediction sequence;
[0108] According to the calculated loss function of the prediction sequence and the second sequence, the model parameters of the prediction model are iteratively adjusted based on the loss function until the prediction model converges.
[0109] Specifically, the historical operation data includes shielding door switch data, door displacement data and door speed data collected at different historical time nodes.
[0110] Specifically, the prediction model includes a first multi-layer perception module, a decoder network, and a second multi-layer perception module connected in sequence;
[0111] The first multi-layer perception module is used to convert the model input data of the prediction model into a vector of a set shape;
[0112] The decoder network is used to generate an initial prediction sequence based on a vector of a set shape;
[0113] The second multi-layer perception module is used to convert the initial prediction sequence into segmented prediction data according to the shape of the model input data.
[0114] Specifically, the calculation formula of the first multi-layer perception module is expressed as:
[0115]
[0116] in, w represents the weights of the hidden layer of the first multi-layer perception module,b represents the bias term of the hidden layer, relu is a nonlinear activation function, x out is the output of the first multi-layer perception module, x It is the input of the first multi-layer perception module.
[0117] Specifically, before real-time operational data is fed into a pre-built predictive model, it also includes:
[0118] The real-time operation data is processed to fill missing values and filter out outliers, and the processed real-time operation data is standardized.
[0119] Specifically, the standardization process is used to standardize the real-time running data into sequence data with a mean of 0 and a standard deviation of 1.
[0120] In the above, by setting the monitoring sensor and fault alarm of the shielding door, the real-time operation data of the shielding door is obtained, and the real-time operation data is input into the pre-built prediction model; based on the prediction model, the real-time operation data is predicted in multiple rounds of cyclic segmentation, and the operation prediction results of each time node in the target period are output. In the process of the multiple rounds of cyclic segmentation prediction, the real-time operation data is used as the initial model input data of the prediction model, and each round of prediction outputs a segmented prediction data of the target period, and the predicted segmented prediction data and the real-time operation data are used as the model input data of the next round of prediction; the operation prediction result is input into the pre-built fault monitoring model, and the fault type monitoring result of the corresponding time node is output based on the fault monitoring model. The fault monitoring model is pre-trained based on the fault operation data of the shielding door, and the fault operation data is pre-labeled with the corresponding fault type label. The above technical means are adopted to predict the operation prediction results of the shielding door at each time node in the future target period by cyclic segmentation, and then perform fault monitoring based on the operation prediction results, timely discover the potential fault type of the corresponding time node, realize accurate and timely fault monitoring, and thus improve the reliability and stability of the shielding door operation.
[0121] The fault monitoring device for the shielding door provided in the second embodiment of the present application can be used to execute the fault monitoring method for the shielding door provided in the first embodiment above, and has corresponding functions and beneficial effects.
[0122] Embodiment three:
[0123] Embodiment 3 of the present application provides an electronic device, referring to Figure 7The electronic device includes: a processor 31, a memory 32, a communication module 33, an input device 34 and an output device 35. The number of processors in the electronic device can be one or more, and the number of memories in the electronic device can be one or more. The processor, memory, communication module, input device and output device of the electronic device can be connected via a bus or other means.
[0124] As a computer-readable storage medium, the memory can be used to store software programs, computer executable programs and modules, such as the program instructions / modules corresponding to the fault monitoring method of the shielding door described in any embodiment of the present application (for example, the input module, data prediction module and fault monitoring module in the fault monitoring device of the shielding door). The memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the device, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network and a combination thereof.
[0125] The communication module is used for data transmission.
[0126] The processor executes various functional applications and data processing of the device by running the software programs, instructions and modules stored in the memory, that is, realizes the above-mentioned fault monitoring method of the shielding door.
[0127] The input device can be used to receive input digital or character information and generate key signal input related to user settings and function control of the device. The output device can include display devices such as display screens.
[0128] The electronic device provided above can be used to execute the fault monitoring method of the shielding door provided in the above embodiment 1, and has corresponding functions and beneficial effects.
[0129] Embodiment 4:
[0130] The embodiment of the present application also provides a storage medium containing computer executable instructions, which are used to execute a fault monitoring method for a shielding door when executed by a computer processor. The fault monitoring method for the shielding door includes: obtaining real-time operation data of the shielding door, and inputting the real-time operation data into a pre-built prediction model; performing multiple rounds of cyclic segmented prediction on the real-time operation data based on the prediction model, and outputting operation prediction results for each time node in the target time period. In the multiple rounds of cyclic segmented prediction process, the real-time operation data is used as the initial model input data of the prediction model, and each round of prediction outputs a segmented prediction data for the target time period, and the predicted segmented prediction data and the real-time operation data are used as the model input data for the next round of prediction; the operation prediction result is input into a pre-built fault monitoring model, and the fault type monitoring result of the corresponding time node is output based on the fault monitoring model. The fault monitoring model is pre-trained based on the fault operation data of the shielding door, and the fault operation data is pre-labeled with the corresponding fault type label.
[0131] Storage medium - any of various types of memory devices or storage devices. The term "storage medium" is intended to include: installation media, such as CD-ROM, floppy disk or tape device; computer system memory or random access memory, such as DRAM, DDR RAM, SRAM, EDO RAM, Rambus RAM, etc.; non-volatile memory, such as flash memory, magnetic media (such as hard disk or optical storage); registers or other similar types of memory elements, etc. Storage media may also include other types of memory or combinations thereof. In addition, the storage medium may be located in the first computer system in which the program is executed, or may be located in a different second computer system, which is connected to the first computer system via a network (such as the Internet). The second computer system can provide program instructions to the first computer for execution. The term "storage medium" may include two or more storage media residing in different locations (for example, in different computer systems connected by a network). The storage medium may store program instructions (for example, embodied as a computer program) that can be executed by one or more processors.
[0132] Of course, the storage medium containing computer executable instructions provided in an embodiment of the present application is not limited to the fault monitoring method of the shielding door as described above, and can also execute related operations in the fault monitoring method of the shielding door provided in any embodiment of the present application.
[0133] The fault monitoring device, storage medium and electronic device of the shielding door provided in the above embodiments can execute the fault monitoring method of the shielding door provided in any embodiment of the present application. For technical details not described in detail in the above embodiments, please refer to the fault monitoring method of the shielding door provided in any embodiment of the present application.
[0134] The above are only preferred embodiments of the present application and the technical principles used. The present application is not limited to the specific embodiments described herein, and various obvious changes, readjustments and substitutions that can be made by those skilled in the art will not deviate from the scope of protection of the present application. Therefore, although the present application is described in more detail through the above embodiments, the present application is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present application, and the scope of the present application is determined by the scope of the claims.
Claims
1. A rail transit platform screen door, characterized in that: include: A controller, used for controlling the shielding door body to perform opening and closing operations; Monitoring sensors are used to collect real-time operation data of the platform shielding door based on the opening and closing operations of the platform shielding door body; A fault early warning device is used to obtain the real-time operation data, input the real-time operation data into a pre-built prediction model, perform multiple rounds of cyclic segmented prediction on the real-time operation data based on the prediction model, and output the operation prediction results of each time node in the target time period. In the multiple rounds of cyclic segmented prediction process, the real-time operation data is used as the initial model input data of the prediction model, and each round of prediction outputs a segmented prediction data of the target time period, and the predicted segmented prediction data and the real-time operation data are used as the model input data of the next round of prediction, and the operation prediction results are input into a pre-built fault monitoring model, and the fault type monitoring results of the corresponding time node are output based on the fault monitoring model. The fault monitoring model is pre-trained based on the fault operation data of the shielding door, and the fault operation data is pre-labeled with the corresponding fault type label; The training process of the prediction model includes: Collect historical operation data of the platform shielding door to construct training samples, and split the training samples into a first sequence and a second sequence according to time sequence; Inputting the first sequence into the prediction model to perform multiple rounds of cyclic segmented prediction to obtain a corresponding prediction sequence; A loss function is calculated according to the prediction sequence and the second sequence, and model parameters of the prediction model are iteratively adjusted based on the loss function until the prediction model converges.
2. The rail transit platform screen door according to claim 1, characterized in that: The historical operation data includes shielding door switch data, door displacement data and door speed data collected at different historical time nodes.
3. The rail transit platform screen door according to claim 1, characterized in that: The prediction model includes a first multi-layer perception module, a decoder network, and a second multi-layer perception module connected in sequence; The first multi-layer perception module is used to convert the model input data of the prediction model into a vector of a set shape; The decoder network is used to generate an initial prediction sequence based on the vector of the set shape; The second multi-layer perception module is used to convert the initial prediction sequence into the segmented prediction data according to the shape of the model input data.
4. The rail transit platform screen door according to claim 3, characterized in that: The calculation formula of the first multi-layer perception module is expressed as: in, w represents the weights of the hidden layers of the first multi-layer perception module, b represents the bias term of the hidden layer, relu is a nonlinear activation function, x out is the output of the first multi-layer perception module, x It is the input of the first multi-layer perception module.
5. The rail transit platform screen door according to claim 1, characterized in that: Before inputting the real-time operation data into the pre-built prediction model, the fault early warning device further includes: The real-time operation data is processed by filling missing values and filtering out abnormal values, and the processed real-time operation data is standardized.
6. The rail transit platform screen door according to claim 5, characterized in that: The standardization process is used to standardize the real-time operation data into sequence data with a mean of 0 and a standard deviation of 1.
7. The rail transit platform screen door according to claim 1, characterized in that: The monitoring sensor is a monitoring camera, which performs target detection and positioning on the shielding door body and determines the real-time operation data of the shielding door based on the detected position of the shielding door body.
8. A fault monitoring method for a platform screen door, applied to a rail transit platform screen door as claimed in any one of claims 1 to 7, characterized in that: include: Acquire real-time operation data of the platform shielding door, and input the real-time operation data into a pre-built prediction model; Based on the prediction model, multiple rounds of cyclic segmented prediction are performed on the real-time operation data, and the operation prediction results of each time node in the target time period are output. In the multiple rounds of cyclic segmented prediction process, the real-time operation data is used as the initial model input data of the prediction model, and each round of prediction outputs a segmented prediction data of the target time period, and the predicted segmented prediction data and the real-time operation data are used as the model input data of the next round of prediction; Input the operation prediction result into a pre-built fault monitoring model, and output the fault type monitoring result of the corresponding time node based on the fault monitoring model, wherein the fault monitoring model is pre-trained based on the fault operation data of the shielding door, and the fault operation data is pre-labeled with the corresponding fault type label; The training process of the prediction model includes: Collect historical operation data of the platform shielding door to construct training samples, and split the training samples into a first sequence and a second sequence according to time sequence; Inputting the first sequence into the prediction model to perform multiple rounds of cyclic segmented prediction to obtain a corresponding prediction sequence; A loss function is calculated according to the prediction sequence and the second sequence, and model parameters of the prediction model are iteratively adjusted based on the loss function until the prediction model converges.
9. A fault monitoring device for a shielding door, characterized in that: include: An input module, used for acquiring real-time operation data of the shielding door, and inputting the real-time operation data into a pre-built prediction model; A data prediction module is used to perform multiple rounds of cyclic segmented prediction on the real-time operation data based on the prediction model, and output the operation prediction results of each time node in the target time period. In the multiple rounds of cyclic segmented prediction process, the real-time operation data is used as the initial model input data of the prediction model, and each round of prediction outputs a segmented prediction data of the target time period, and the predicted segmented prediction data and the real-time operation data are used as the model input data of the next round of prediction; A fault monitoring module, used for inputting the operation prediction result into a pre-built fault monitoring model, and outputting the fault type monitoring result of the corresponding time node based on the fault monitoring model, wherein the fault monitoring model is pre-trained based on the fault operation data of the shielding door, and the fault operation data is pre-labeled with the corresponding fault type label; The training process of the prediction model includes: Collect historical operation data of the platform shielding door to construct training samples, and split the training samples into a first sequence and a second sequence according to time sequence; Inputting the first sequence into the prediction model to perform multiple rounds of cyclic segmented prediction to obtain a corresponding prediction sequence; A loss function is calculated according to the prediction sequence and the second sequence, and model parameters of the prediction model are iteratively adjusted based on the loss function until the prediction model converges.
10. A storage medium containing computer executable instructions, characterized in that: The computer executable instructions are used to perform the fault monitoring method for the shielding door according to claim 8 when executed by a computer processor.