CN114707401A - Fault early warning method and device for signal system equipment - Google Patents

Abstract

The invention provides a fault early warning method and a device of signal system equipment, wherein the method comprises the following steps: acquiring current working condition data of equipment to be tested in a signal system; inputting the current working condition data of the equipment to be tested into a fault prediction model to obtain the predicted fault information of the equipment to be tested; under the condition that the equipment to be tested is determined to be in fault, generating early warning information according to the current working condition data and the predicted fault information of the equipment to be tested, and sending out early warning based on the early warning information; the method and the device realize the analysis and prediction of the current working condition data of the equipment to be tested based on the fault prediction model so as to automatically and accurately acquire the predicted fault information of the equipment to be tested, and timely make early warning based on the predicted fault information and the current working condition data when a fault occurs so as to timely and quickly maintain the equipment to be tested; and the early warning information contains rich information related to equipment faults, so that an effective auxiliary effect can be provided for quickly and accurately making maintenance decisions, and the maintenance efficiency is improved.

Description

Fault warning method and device for signal system equipment

technical field

The invention relates to the technical field of rail transit, in particular to a fault warning method and device for signal system equipment.

Background technique

With the advent of the era of big data, the social environment and people's way of life have undergone earth-shaking changes and tend to develop informatization. Science and technology organizations or enterprises are committed to researching how to use technical means to excavate the deep-level relationships and laws of things hidden under big data, so that the information-based intervention methods and strategies they use are more accurate and effective. The urban rail transit signal system plays a key role in ensuring the safety of train operation, realizing the modernization of train operation command and train operation, and improving the operation efficiency. If the signal system fails, it will cause the subway or light rail to fail to operate normally, bring great trouble to passengers' daily travel, and even threaten their life safety and affect the stability and harmony of society.

In the prior art, there is a lack of relevant technologies for predicting the failure of urban rail transit signal equipment. Only after the actual failure of the signal system equipment, the maintenance personnel can monitor the failure through a large number of troubleshooting, and take corresponding countermeasures. Low efficiency and poor real-time performance. Therefore, when some core and critical signal equipment fails, it will often have a major impact on the signal system and cause huge losses to passengers.

To sum up, how to quickly and effectively give early warning of equipment failures is an important issue to be solved urgently in the industry.

SUMMARY OF THE INVENTION

The invention provides a fault early warning method and device for signal system equipment, which are used to solve the defects of the prior art that the fault can only be detected after the equipment of the signal system fails, and the maintenance efficiency is low and the real-time performance is poor, so as to realize fast and effective Provide fault early warning to equipment, so as to maintain equipment in time when fault occurs.

The present invention provides a fault warning method for signal system equipment, comprising:

Obtain the current working condition data of the equipment to be tested in the signal system;

Input the current working condition data of the equipment under test into the fault prediction model to obtain the predicted fault information of the equipment under test;

When it is determined that the equipment under test is faulty based on the predicted fault information, early warning information is generated according to the current operating condition data and predicted fault information of the equipment under test, and an early warning is issued based on the early warning information;

Wherein, the fault prediction model is obtained by training based on the historical fault logs of the sample equipment in the training data set; the historical fault logs include historical working condition data and real fault information.

According to a fault warning method for signal system equipment provided by the present invention, the predicted fault information includes predicted fault state, predicted fault type and predicted fault cause; the real fault information includes real fault state, real fault type and real fault cause .

According to a fault warning method for signal system equipment provided by the present invention, the fault prediction model is obtained by training based on the following steps:

Input the historical working condition data of the sample equipment into the fault prediction model to obtain the predicted fault state, predicted fault type and predicted fault cause of the sample equipment;

determining a first loss function of the failure prediction model based on the predicted failure state and the actual failure state of the sample device;

determining a second loss function of the fault prediction model based on the predicted fault type and the actual fault type of the sample device;

determining a third loss function of the failure prediction model based on the predicted failure cause and the actual failure cause of the sample device;

The fault prediction model is trained based on the first loss function, the second loss function and the third loss function.

According to a fault warning method for signal system equipment provided by the present invention, the training of the fault prediction model based on the first loss function, the second loss function and the third loss function includes:

Integrating the first loss function, the second loss function and the third loss function to obtain the total loss function of the fault prediction model;

The fault prediction model is trained based on the total loss function.

According to a fault early warning method for signal system equipment provided by the present invention, the inputting the historical working condition data of the sample equipment into the fault prediction model includes:

Preprocessing the historical working condition data of the sample equipment;

Wherein, the preprocessing includes one or more of normalization processing, missing value processing and noise reduction processing;

Input the preprocessed historical working condition data into the fault prediction model.

A fault warning method for signal system equipment provided according to the present invention further includes:

The training data set is extended based on the current operating condition data and predicted fault information of the device under test.

The present invention also provides a fault warning device for signal system equipment, comprising:

The acquisition module is used to acquire the current working condition data of the equipment to be tested in the signal system;

a prediction module, configured to input the current working condition data of the equipment under test into a fault prediction model to obtain predicted fault information of the equipment under test;

an early warning module, configured to generate early warning information according to the current operating condition data and predicted fault information of the equipment to be tested when it is determined that the equipment to be tested is faulty based on the predicted fault information, and based on the early warning information issue an early warning;

Wherein, the fault prediction model is obtained by training based on the historical fault logs of the sample equipment in the training data set; the historical fault logs include historical working condition data and real fault information.

The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, the processor implements any one of the signal system devices described above when the processor executes the program The steps of the fault early warning method.

The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, implements the steps of any one of the above-mentioned fault warning methods for signal system equipment.

The present invention also provides a computer program product, comprising a computer program, which, when executed by a processor, implements the steps of any one of the above-mentioned fault warning methods for signal system equipment.

The fault early warning method and device for signal system equipment provided by the present invention monitors the current working condition data of the equipment to be tested in the signal system in real time, and analyzes and predicts the current working condition data of the equipment to be tested based on a fault prediction model, so as to automatically and accurately predict the current working condition data of the equipment to be tested. Obtain the predicted fault information of the equipment under test, and determine whether the equipment under test is faulty according to the predicted fault information. , and make early warnings based on the early warning information in a timely manner; not only can the equipment under test be maintained in a timely and rapid manner when a fault occurs; and the early warning information contains rich information related to equipment faults, which can provide rapid and accurate maintenance decisions. Effective auxiliary function, thereby improving maintenance efficiency.

Description of drawings

In order to explain the present invention or the technical solutions in the prior art more clearly, the following will briefly introduce the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are the For some embodiments of the invention, for those of ordinary skill in the art, other drawings can also be obtained according to these drawings without any creative effort.

Fig. 1 is one of the schematic flow charts of the fault early warning method of the signal system equipment provided by the present invention;

Fig. 2 is the second schematic flow chart of the fault early warning method of the signal system equipment provided by the present invention;

Fig. 3 is the third schematic flow chart of the fault early warning method of the signal system equipment provided by the present invention;

Fig. 4 is the structural schematic diagram of the fault warning device of the signal system equipment provided by the present invention;

FIG. 5 is a schematic structural diagram of an electronic device provided by the present invention.

Detailed ways

In order to make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention. , not all examples. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

In the prior art, maintenance personnel usually take a large number of troubleshooting measures to detect and analyze the failure of the equipment after the equipment of the signal system actually fails, and take corresponding countermeasures to maintain the equipment, resulting in low maintenance efficiency and poor real-time performance. The problem;

In addition, in the prior art, the equipment of the signal system is also regularly maintained or replaced to reduce the possibility of equipment failure; Replacement costs are high.

In view of the above problems, the present embodiment provides a fault early warning method for signal system equipment, which can perform real-time monitoring on each equipment in the signal system and collect working condition data, fault prediction and fault early warning, so as to realize the detection of various equipment in the urban rail transit signal system. It can make early warning and early response to the possible failure of the equipment, reduce the probability of accidents, provide great help for the operation and maintenance of the signal system, and provide a strong guarantee for the safety and stability of the signal system.

It should be noted that the execution body of the fault early warning method may be a controller built in or external to the signal system, including but not limited to servers, terminal devices, and smart chips, which are not specifically limited in this embodiment.

The fault early warning method of the signal system equipment of the present invention is described below in conjunction with Fig. 1, and the method includes:

Step 101, obtaining current working condition data of the device to be tested in the signal system;

Optionally, the fault early warning method in this embodiment can be applied to various lines and signal systems of various standards, and has strong scalability and wide application range.

The equipment to be tested is equipment that needs to perform fault prediction and fault warning, including equipment in the automatic train operation control system and equipment in the depot signal control system, which is not specifically limited in this embodiment.

The current working condition data is various working condition parameters of the device to be tested in the current working environment, including but not limited to the running time and running parameters, which are not specifically limited in this embodiment.

Optionally, when it is necessary to perform fault prediction on the device under test, the current working condition data of the device under test is acquired in real time based on the collector.

Step 102: Input the current working condition data of the equipment under test into a fault prediction model, and obtain the predicted fault information of the equipment under test; wherein, the fault prediction model is trained based on the historical fault logs of the sample equipment in the training data set Obtained; the historical fault log includes historical working condition data and real fault information;

The fault prediction model is constructed and generated based on a machine learning model, including but not limited to a random forest algorithm, a convolutional neural network, a recurrent neural network, and a deep residual network, which are not specifically limited in this embodiment.

The fault information may be one or more, such as one or more combinations of fault status, fault type, and fault cause, which are not specifically limited in this embodiment.

Among them, the sample device and the device to be tested are the same type of device in the signal system.

The fault prediction model is obtained by training based on the historical fault logs of the sample devices in the training data set. The historical fault log is generated based on the historical operating data and historical fault information of the sample equipment.

As shown in FIG. 2 , when step 102 is performed, the fault prediction model needs to be pre-trained first. The specific training steps include: first, collecting historical fault logs of multiple sample devices to form a training data set; wherein, the historical fault logs can be Collect from the database.

Then, the historical working condition data of each sample device is used as input information, and the real fault information is used as label value, and the fault prediction model is iteratively trained until the fault prediction model converges, reaches the maximum number of iterations, or the evaluation result of the fault prediction model satisfies the prediction. Set the conditions to obtain a fault prediction model that can accurately predict the fault information of the equipment according to the current working condition data of the equipment.

Among them, during the fault prediction model process, the sklearn tool can be used to train the machine learning model to create a fault prediction model. For the created failure prediction model, the metrics (evaluation) module in sklearn can be used to evaluate the failure prediction model to ensure that the failure prediction model with good performance is obtained. By using the sklearn tool in the python language when modeling based on the training data set, the efficiency of data modeling can be greatly improved.

Once the fault prediction model is trained, it can be used repeatedly; therefore, the fault prediction method in this example has the characteristics of low technical cost and strong feasibility, and has a good application prospect.

Optionally, when the equipment under test needs to be predicted for fault information, the current operating condition data of the equipment under test can be directly input into the training fault information prediction, and the predicted fault information of the equipment under test can be obtained.

Step 103, in the case of determining that the equipment under test is faulty based on the predicted fault information, generate early warning information according to the current operating condition data and predicted fault information of the equipment under test, and issue an early warning based on the early warning information .

Optionally, after obtaining the predicted failure information of the device under test, it may be determined whether the device under test is faulty according to the predicted failure information;

For example, when the predicted fault information includes a fault state and no fault occurs, the fault state is 0, and when a fault occurs, the fault state is 1, then when the fault state of the device under test is 1, it is determined that the device under test is faulty .

When the predicted fault information includes the fault type and no fault occurs, the fault type is 0, and when a fault occurs, the fault status is a non-zero value such as 1, 2 or 3, then when the fault status of the device under test is a non-zero value , it is determined that the device under test is faulty.

When it is determined that the equipment under test is faulty, the current working condition data of the equipment under test and the predicted failure information are combined to generate early warning information, and the early warning information is sent to the maintenance center or maintenance terminal in time to maintain the equipment under test in time.

Among them, the maintenance method can be to prompt the maintenance personnel through the maintenance terminal, so that the maintenance personnel can perform timely maintenance of the equipment under test according to the warning information; or directly, the maintenance center can maintain the equipment under test according to the maintenance strategy corresponding to the warning information. The example does not specifically limit this.

When it is determined that the equipment under test is not faulty, continue to monitor the working condition data of the equipment under test at the next moment, and repeat the above steps to continue to predict and warn the equipment under test for fault information.

In this embodiment, the current working condition data of the equipment under test in the signal system is monitored in real time, and the current working condition data of the equipment under test is analyzed and predicted based on the fault prediction model, so as to automatically and accurately obtain the predicted fault information of the equipment under test, and according to Predict the fault information in real time to determine whether the equipment to be tested is faulty. When it is determined that the equipment to be tested is faulty, early warning information can be generated in real time according to the current working condition data and predicted fault information of the equipment to be tested, and based on the early warning information, early warnings can be made quickly and in a timely manner; Not only can the equipment under test be maintained in a timely and rapid manner when a fault occurs, but also the early warning information contains rich information related to equipment faults, which can provide effective assistance for making maintenance decisions quickly and accurately, thereby improving maintenance efficiency.

Based on the above embodiments, the predicted fault information in this embodiment includes predicted fault status, predicted fault type and predicted fault cause; the real fault information includes real fault status, real fault type and real fault cause.

Among them, the fault state is used to characterize whether the equipment fails, including failure and no failure;

The fault type is used to represent the type of equipment failure, and may include multiple types, such as no fault, a first fault type, a second fault type, etc., which are not specifically limited in this embodiment. The fault type may be classified according to the location of the fault or the fault level, etc., which is not specifically limited in this embodiment.

The failure cause is used to characterize the root cause of the device failure, and may include multiple causes. The specific number and specific cause may be set according to actual requirements, which are not specifically limited in this embodiment.

In this embodiment, the predicted fault information includes a variety of key information used to characterize the fault, thereby making the fault early warning result more accurate, and providing more comprehensive and effective reference data for maintenance decision-making. Make maintenance decisions to improve the efficiency and timeliness of maintenance.

On the basis of the above embodiment, the fault prediction model in this embodiment is obtained by training based on the following steps: input the historical working condition data of the sample equipment into the fault prediction model, and obtain the predicted fault state of the sample equipment , predicted failure type and predicted failure cause; based on the predicted failure state and real failure state of the sample device, determine the first loss function of the failure prediction model; based on the predicted failure type and real failure type of the sample device, determine the second loss function of the failure prediction model; based on the predicted failure cause and the real failure cause of the sample device, determine the third loss function of the failure prediction model; based on the first loss function, the second loss function and the The third loss function is used to train the fault prediction model.

Optionally, when the predicted fault information includes the predicted fault state, the predicted fault type, and the predicted fault cause, the fault prediction model may be a multi-task prediction model, or may be multiple single-task learning models, which are not specified in this embodiment. limited.

The fault early warning method in this embodiment is described below by taking the fault prediction model as a multi-task prediction model as an example.

Among them, the fault prediction model includes three output layers, which are respectively used to output the predicted fault state, the predicted fault type and the predicted fault cause.

Each loss function may be a cross-entropy loss function or a squared error loss function, etc., which is not specifically limited in this embodiment.

Optionally, the fault prediction model is specifically obtained by training based on the following steps:

First, directly input the historical working condition data of the sample equipment into the fault prediction model, or preprocess the historical working condition data and then input it into the fault prediction model to obtain the predicted fault status, predicted fault type and prediction of the sample equipment output by the fault prediction model. cause of issue.

Then, obtain the first loss function determined according to the predicted failure state and the real failure state of the sample device, the second loss function determined according to the predicted failure type and the real failure type of the sample device, and the predicted failure cause and real failure of the sample device. Reason, the determined third loss function;

According to the first loss function, the second loss function and the third loss function, iterative optimization training is performed on the fault prediction model; the optimization training method may be a stochastic gradient descent method, an adaptive gradient optimization algorithm, or the like.

The training method can be combined with the first loss function, the second loss function and the third loss function to perform overall training on the parameters of the fault prediction model;

The parameters of the fault prediction model can also be iteratively trained in series according to the first loss function, the second loss function and the third loss function. For example, in each iterative training process, the fault prediction model is firstly trained based on the first loss function. Fine-tuning, based on the fine-tuning of the first loss function, the fault prediction model is fine-tuned again based on the second loss function, and then based on the fine-tuning of the second loss function, the parameters of the fault prediction model are adjusted based on the third loss function. Fine tune again etc. This embodiment does not specifically limit the training method of the fault prediction model.

When training the predicted fault information, combining multiple loss functions to train the fault prediction model can quickly and effectively obtain a fault prediction model with good performance.

On the basis of the above-mentioned embodiment, the training of the fault prediction model based on the first loss function, the second loss function and the third loss function in this embodiment includes: applying the first loss function to , the second loss function and the third loss function are fused to obtain the total loss function of the fault prediction model; based on the total loss function, the fault prediction model is trained.

Optionally, this embodiment combines the first loss function determined based on the predicted fault state and the real fault state, the second loss function determined based on the predicted fault type and the real fault type, and the first loss function determined based on the predicted fault cause and the real fault cause. Three loss functions are used to optimize the training of the fault prediction model, so that the trained fault prediction model can accurately predict both the fault state and the fault type and cause.

Optionally, the step of training the fault prediction model includes: fusing the first loss function, the second loss function and the third loss function of the fault prediction model to obtain a total loss function of the fault prediction model; wherein the fusion method includes: However, it is not limited to direct addition or averaging, weighted addition or weighted average, etc., which are not specifically limited in this embodiment.

Then, based on the total loss function, the parameters of the fault prediction model are optimized as a whole until the convergence conditions of the fault prediction model are satisfied, so as to obtain a fault prediction model that can accurately predict the fault state, fault type and fault cause.

On the basis of the above embodiment, in this embodiment, inputting the historical operating condition data of the sample equipment into the fault prediction model includes: preprocessing the historical operating condition data of the sample equipment; wherein, The preprocessing includes one or more of normalization processing, missing value processing and noise reduction processing; and inputting the preprocessed historical working condition data into the fault prediction model.

Optionally, the historical working condition data of the sample device can be obtained through a collector; errors and failures of the collector, as well as human factors, etc., may cause the collected historical working condition data to be missing or have a large amount of noise. Therefore, in order to obtain a high-performance fault prediction model quickly and accurately, it is necessary to preprocess the historical working condition data of the sample equipment before training the fault prediction model.

Among them, there are multiple historical working condition data; if the difference between the multiple historical working condition data is too large, the failure prediction model will be difficult to converge during the training process, or even difficult to converge. In order to prevent the smaller value in the historical working condition data from being overwhelmed by the larger value and the influence of the dimension, the historical working condition data needs to be normalized, and the input information is normalized to a small interval, such as in the range [-1,1].

When there are missing values in the historical working condition data, it will also affect the training of the fault prediction model; the historical working condition data with missing values can be eliminated, and the missing values in the historical working condition data can also be complemented , such as completion based on the K-nearest neighbor completion algorithm, etc. This example does not specifically limit the processing of missing values.

When there is noise data in the historical working condition data, it will also affect the training of the fault prediction model. Therefore, the historical operating condition data can be denoised to eliminate the influence of noise.

After preprocessing the historical working condition data, more effective historical working condition data can be obtained; the fault prediction model can be trained based on the preprocessed historical working condition data, so as to obtain a fault prediction model with good performance quickly and accurately.

Among them, the data preprocessing can be based on the preproccessing (preprocessing) library in the sklearn tool to preprocess the historical working condition data.

Among them, the sklearn tool is scikit-learn (machine learning library), which is a common third-party module of python and is often used in the field of machine learning, which encapsulates various machine learning algorithms.

On the basis of the foregoing embodiments, this embodiment further includes: expanding the training data set based on the current operating condition data and predicted fault information of the device under test.

Optionally, after obtaining the predicted failure information of the device under test, the current operating condition data and predicted failure information of the device under test can be used as sample data and added to the training data set to expand the samples of the training data set, and then It makes the performance of the trained fault prediction model better.

It should be noted that, after the working condition data and fault information of other sample equipment are updated, the updated working condition data and fault information of other sample equipment can also be added to the training data set as samples.

As shown in FIG. 3, it is a schematic diagram of the overall flow of the fault early warning method of the signal system equipment in this embodiment, which specifically includes:

Step 1, obtaining a training data set; the specific obtaining module obtains the historical fault logs of various sample devices to form a training data set;

Step 2, data modeling; specifically includes data preprocessing and model creation; wherein, in the data processing process, the corresponding python library is called to preprocess the historical working condition data in the training data set; in the model creation process, combined with historical faults The real fault information (label) in the log is trained by using the supervised learning algorithm on the preprocessed historical working condition data to obtain the fault prediction model;

Step 3, fault prediction; specifically includes model deployment, information collection and fault prediction; wherein, in the model deployment process, the collector that collects the working condition data of the equipment to be tested is connected to the fault prediction model; in the information collection process, Use the collector to collect the current working condition data of the equipment under test, and input the current working condition data of the equipment under test into the fault prediction model; in the process of model prediction, use the fault prediction model to predict the state of the equipment under test, and get the fault Information prediction results.

The fault early warning device of the signal system equipment provided by the present invention is described below. The fault early warning device of the signal system equipment described below and the fault early warning method of the signal system equipment described above can be referred to each other correspondingly.

As shown in FIG. 4 , this embodiment provides a fault warning device for signal system equipment, the device includes: an acquisition module 401 , a prediction module 402 and an early warning module 403 , wherein:

The acquisition module 401 is used to acquire the current working condition data of the equipment to be tested in the signal system;

Optionally, the fault early warning method in this embodiment can be applied to various lines and signal systems of various standards, and has strong scalability and wide application range.

The equipment to be tested is equipment that needs to perform fault prediction and fault warning, including equipment in the automatic train operation control system and equipment in the depot signal control system, which is not specifically limited in this embodiment.

The current working condition data is various working condition parameters of the device to be tested in the current working environment, including but not limited to the running time and running parameters, which are not specifically limited in this embodiment.

Optionally, when it is necessary to perform fault prediction on the device under test, the current working condition data of the device under test is acquired in real time based on the collector.

The prediction module 402 is used for inputting the current working condition data of the equipment under test into the fault prediction model to obtain the predicted fault information of the equipment under test; wherein, the fault prediction model is based on the historical faults of the sample equipment in the training data set log training; the historical fault log includes historical working condition data and real fault information;

The fault prediction model is constructed and generated based on a machine learning model, including but not limited to a random forest algorithm, a convolutional neural network, a recurrent neural network, and a deep residual network, which are not specifically limited in this embodiment.

The fault information may be one or more, such as one or more combinations of fault status, fault type, and fault cause, which are not specifically limited in this embodiment.

Among them, the sample device and the device to be tested are the same type of device in the signal system.

The fault prediction model is obtained by training based on the historical fault logs of the sample devices in the training data set. The historical fault log is generated based on the historical operating data and historical fault information of the sample equipment.

As shown in FIG. 2 , when step 102 is performed, the fault prediction model needs to be pre-trained first. The specific training steps include: first, collecting historical fault logs of multiple sample devices to form a training data set; wherein, the historical fault logs can be Collect from the database.

Then, the historical working condition data of each sample device is used as input information, and the real fault information is used as label value, and the fault prediction model is iteratively trained until the fault prediction model converges, reaches the maximum number of iterations, or the evaluation result of the fault prediction model satisfies the prediction. Set the conditions to obtain a fault prediction model that can accurately predict the fault information of the equipment according to the current working condition data of the equipment.

Among them, during the fault prediction model process, the sklearn tool can be used to train the machine learning model to create a fault prediction model. For the created failure prediction model, the metrics module in sklearn can be used to evaluate the failure prediction model to ensure that the failure prediction model with good performance is obtained. By using the sklearn tool in the python language when modeling based on the training data set, the efficiency of data modeling can be greatly improved.

Once the fault prediction model is trained, it can be used repeatedly; therefore, the fault prediction method in this example has the characteristics of low technical cost and strong feasibility, and has a good application prospect.

Optionally, when the equipment under test needs to be predicted for fault information, the current operating condition data of the equipment under test can be directly input into the training fault information prediction, and the predicted fault information of the equipment under test can be obtained.

The early warning module 403 is configured to generate early warning information according to the current operating condition data and predicted fault information of the equipment to be tested when it is determined that the equipment to be tested is faulty based on the predicted fault information, and based on the early warning information Issue an early warning.

Optionally, after obtaining the predicted failure information of the device under test, it may be determined whether the device under test is faulty according to the predicted failure information;

When it is determined that the equipment under test is faulty, the current working condition data of the equipment under test and the predicted failure information are combined to generate early warning information, and the early warning information is sent to the maintenance center or maintenance terminal in time to maintain the equipment under test in time.

Among them, the maintenance method may be to prompt the maintenance personnel through the maintenance terminal, so that the maintenance personnel can perform timely maintenance of the equipment under test according to the warning information; or directly, the maintenance center can maintain the equipment under test according to the maintenance strategy corresponding to the warning information. The example does not specifically limit this.

In the case that it is determined that the equipment under test is not faulty, continue to monitor the working condition data of the equipment under test at the next moment, and repeat the above steps to continue to predict and give early warning of fault information of the equipment under test.

In this embodiment, the current working condition data of the equipment under test in the signal system is monitored in real time, and the current working condition data of the equipment under test is analyzed and predicted based on the fault prediction model, so as to automatically and accurately obtain the predicted fault information of the equipment under test, and according to Predict the fault information in real time to determine whether the equipment to be tested is faulty. When it is determined that the equipment to be tested is faulty, early warning information can be generated in real time according to the current working condition data and predicted fault information of the equipment to be tested, and based on the early warning information, early warnings can be made quickly and in a timely manner; Not only can the equipment under test be maintained in a timely and rapid manner when a fault occurs, but also the early warning information contains rich information related to equipment faults, which can provide effective assistance for making maintenance decisions quickly and accurately, thereby improving maintenance efficiency.

Based on the above embodiments, the predicted fault information in this embodiment includes predicted fault status, predicted fault type and predicted fault cause; the real fault information includes real fault status, real fault type and real fault cause.

On the basis of the above-mentioned embodiment, this embodiment further includes a data modeling module, which is used for: inputting the historical working condition data of the sample equipment into the fault prediction model to obtain the predicted fault state, Predict the failure type and the predicted failure cause; determine the first loss function of the failure prediction model based on the predicted failure state and the actual failure state of the sample device; determine the predicted failure type and the real failure type of the sample device. the second loss function of the fault prediction model; the third loss function of the fault prediction model is determined based on the predicted failure cause and the real failure cause of the sample device; based on the first loss function, the second loss function and the third loss function Three loss functions for training the fault prediction model.

On the basis of the above embodiment, the data modeling module in this embodiment is specifically configured to: fuse the first loss function, the second loss function and the third loss function to obtain the total loss of the fault prediction model function; based on the total loss function, the fault prediction model is trained.

On the basis of the above embodiment, the preprocessing module in the data modeling module in this embodiment is specifically used to: preprocess the historical working condition data of the sample equipment; wherein, the preprocessing includes normalization One or more of processing, missing value processing and noise reduction processing; inputting the preprocessed historical working condition data into the fault prediction model.

On the basis of the above embodiments, this embodiment further includes an expansion module, which is specifically configured to: expand the training data set based on the current operating condition data and predicted fault information of the device under test.

FIG. 5 illustrates a schematic diagram of the physical structure of an electronic device. As shown in FIG. 5 , the electronic device may include: a processor (processor) 501, a communication interface (Communications Interface) 502, a memory (memory) 503, and a communication bus 504, The processor 501 , the communication interface 502 , and the memory 503 communicate with each other through the communication bus 504 . The processor 501 can call the logic instruction in the memory 503 to execute the fault warning method of the equipment of the signal system, the method includes: acquiring the current working condition data of the equipment to be tested in the signal system; Input the fault prediction model to obtain the predicted fault information of the equipment under test; when it is determined that the equipment under test is faulty based on the predicted fault information, according to the current operating condition data and prediction of the equipment under test The fault information generates early warning information, and an early warning is issued based on the early warning information; wherein, the fault prediction model is obtained by training based on the historical fault logs of the sample equipment in the training data set; the historical fault logs include historical working condition data and real fault information .

In addition, the above-mentioned logic instructions in the memory 503 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product in essence, or the part that contributes to the prior art or the part of the technical solution. The computer software product is stored in a storage medium, including Several instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, Read-Only Memory (ROM, Read-Only Memory), Random Access Memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program codes .

In another aspect, the present invention also provides a computer program product, the computer program product includes a computer program, the computer program can be stored on a non-transitory computer-readable storage medium, and when the computer program is executed by a processor, the computer can Execute the fault early warning method of the signal system equipment provided by the above methods, the method includes: obtaining the current working condition data of the equipment to be tested in the signal system; inputting the current working condition data of the equipment to be tested into the fault prediction model to obtain The predicted fault information of the equipment under test; in the case of determining that the equipment under test is faulty based on the predicted fault information, generate early warning information according to the current operating condition data and predicted fault information of the equipment under test, and An early warning is issued based on the early warning information; wherein, the fault prediction model is obtained by training based on the historical fault logs of the sample devices in the training data set; the historical fault logs include historical working condition data and real fault information.

In another aspect, the present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, it is implemented to execute the method for early warning of a signal system device provided by the above methods, The method includes: acquiring current working condition data of an equipment to be tested in a signal system; inputting the current working condition data of the equipment to be tested into a fault prediction model to obtain predicted fault information of the equipment to be tested; Fault information, when it is determined that the equipment to be tested is faulty, early warning information is generated according to the current operating condition data and predicted fault information of the equipment to be tested, and an early warning is issued based on the early warning information; wherein, the fault prediction model , obtained by training based on the historical fault logs of the sample equipment in the training data set; the historical fault logs include historical operating condition data and real fault information.

The device embodiments described above are only illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in One place, or it can be distributed over multiple network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution in this embodiment. Those of ordinary skill in the art can understand and implement it without creative effort.

From the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and certainly can also be implemented by hardware. Based on this understanding, the above-mentioned technical solutions can be embodied in the form of software products in essence or the parts that make contributions to the prior art, and the computer software products can be stored in computer-readable storage media, such as ROM/RAM, magnetic A disc, an optical disc, etc., includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform the methods described in various embodiments or some parts of the embodiments.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, but not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that it can still be The technical solutions described in the foregoing embodiments are modified, or some technical features thereof are equivalently replaced; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims (10)

1. a fault warning method of signal system equipment, is characterized in that, comprises: Obtain the current working condition data of the equipment under test in the signal system; Input the current working condition data of the equipment under test into the fault prediction model to obtain the predicted fault information of the equipment under test; When it is determined that the equipment under test is faulty based on the predicted fault information, early warning information is generated according to the current operating condition data and predicted fault information of the equipment under test, and an early warning is issued based on the early warning information; Wherein, the fault prediction model is obtained by training based on the historical fault logs of the sample equipment in the training data set; the historical fault logs include historical working condition data and real fault information. 2. The fault early warning method for signal system equipment according to claim 1, wherein the predicted fault information comprises a predicted fault state, a predicted fault type and a predicted fault cause; the real fault information includes a real fault state, a real fault The type of failure and the actual cause of the failure. 3. The fault early warning method of signal system equipment according to claim 2, wherein the fault prediction model is obtained by training based on the following steps: Input the historical working condition data of the sample equipment into the fault prediction model to obtain the predicted fault state, predicted fault type and predicted fault cause of the sample equipment; determining a first loss function of the failure prediction model based on the predicted failure state and the actual failure state of the sample device; determining a second loss function of the fault prediction model based on the predicted fault type and the actual fault type of the sample device; determining a third loss function of the failure prediction model based on the predicted failure cause and the actual failure cause of the sample device; The fault prediction model is trained based on the first loss function, the second loss function and the third loss function. 4. The fault early warning method for signal system equipment according to claim 3, wherein the fault prediction model is trained based on the first loss function, the second loss function and the third loss function, include: Integrating the first loss function, the second loss function and the third loss function to obtain the total loss function of the fault prediction model; The fault prediction model is trained based on the total loss function. 5. The fault early warning method for signal system equipment according to claim 3, wherein the inputting the historical operating condition data of the sample equipment into the fault prediction model comprises: Preprocessing the historical working condition data of the sample equipment; Wherein, the preprocessing includes one or more of normalization processing, missing value processing and noise reduction processing; Input the preprocessed historical working condition data into the fault prediction model. 6. The fault early warning method for signal system equipment according to any one of claims 1-5, characterized in that, further comprising: The training data set is extended based on the current operating condition data and predicted fault information of the device under test. 7. A fault warning device for signal system equipment, characterized in that, comprising: The acquisition module is used to acquire the current working condition data of the equipment to be tested in the signal system; a prediction module, used for inputting the current working condition data of the equipment under test into a fault prediction model to obtain predicted fault information of the equipment under test; an early warning module, configured to generate early warning information according to the current operating condition data and predicted fault information of the equipment to be tested when it is determined that the equipment to be tested is faulty based on the predicted fault information, and based on the early warning information issue an early warning; Wherein, the fault prediction model is obtained by training based on the historical fault logs of the sample equipment in the training data set; the historical fault logs include historical working condition data and real fault information. 8. An electronic device comprising a memory, a processor and a computer program stored on the memory and running on the processor, wherein the processor implements the program as claimed in claim 1 when executing the program Steps of any one of the steps of the fault warning method for the signal system equipment described in to 6. 9. A non-transitory computer-readable storage medium on which a computer program is stored, characterized in that, when the computer program is executed by a processor, the failure of the signal system device according to any one of claims 1 to 6 is realized The steps of the early warning method. 10 . A computer program product, comprising a computer program, characterized in that, when the computer program is executed by a processor, the steps of the fault warning method for a signal system device according to any one of claims 1 to 6 are implemented.

Images