Rail transit system operation trend prediction method and device, equipment and medium

Abstract

The invention provides an operation trend prediction method and device of a rail transit system, equipment and a medium, and the method comprises the steps: determining first index data corresponding to a plurality of index types at a current moment and second index data corresponding to a plurality of index types at a historical moment from operation data corresponding to the rail transit system; according to the first index data corresponding to the multiple index types and the second index data corresponding to the multiple index types, stable sequences corresponding to the multiple index types are determined; and inputting all the stationary sequences into a trend prediction model to obtain a target running trend output by the trend prediction model, the trend prediction model being obtained by training based on stationary sequence samples and running trend samples. According to the method, a trend prediction model is adopted to predict a plurality of stable sequences corresponding to operation data, and a target operation trend corresponding to a rail transit system is accurately obtained, so that conversion from passive response to active management is realized, and potential faults can be automatically prevented or avoided subsequently.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method, device, equipment and medium for predicting the operation trend of a rail transit system. Background Art

[0002] To ensure the safety and smooth operation of rail transit systems, it is crucial to monitor and predict the operating status of servers. At present, the industry widely uses open source tools to build efficient full-stack monitoring systems to capture and display the operating status of various indicators in rail transit systems in real time, such as real-time operating status and historical operating status. However, this full-stack monitoring system mainly responds to operational problems that have occurred, but cannot prevent or avoid potential failures, which prolongs the time to handle emergencies and increases the risk of failure expansion. Summary of the invention

[0003] The present invention provides an operation trend prediction method, device, equipment and medium for a rail transit system, which are used to solve the defect that the existing full-stack monitoring system cannot prevent or avoid potential faults in the rail transit system, thereby prolonging the time for handling emergencies and increasing the risk of fault expansion. After the stable sequences corresponding to multiple indicator types are determined according to the operation data of the rail transit system, a trend prediction model can be used to predict multiple stable sequences to accurately obtain the target operation trend corresponding to the rail transit system. The whole process aims to achieve a transition from passive response to active management, so as to automatically prevent or avoid potential faults in the future, so as to reduce the processing time of emergencies and the risk of fault expansion.

[0004] The present invention provides a method for predicting the operation trend of a rail transit system, comprising the following steps.

[0005] From the operation data corresponding to the rail transit system, first indicator data corresponding to each of the multiple indicator types at the current moment and second indicator data corresponding to each of the multiple indicator types at the historical moment are determined.

[0006] According to the first indicator data and the second indicator data corresponding to the multiple indicator types, the stationary sequences corresponding to the multiple indicator types are determined.

[0007] All stationary sequences are input into a trend prediction model to obtain a target operation trend of the rail transit system output by the trend prediction model, wherein the trend prediction model is trained based on stationary sequence samples and operation trend samples.

[0008] According to a method for predicting the operation trend of a rail transit system provided by the present invention, the stationary sequence corresponding to each of the multiple indicator types is determined according to the first indicator data corresponding to each of the multiple indicator types and the second indicator data corresponding to each of the multiple indicator types, including: performing the following operations on the first indicator data and the second indicator data corresponding to each indicator type: preprocessing the first indicator data to obtain first target indicator data, and preprocessing the second indicator data to obtain second target indicator data; differentiating a first feature corresponding to the first target indicator data and a second feature corresponding to the second target indicator data to obtain a stationary sequence corresponding to the indicator type; wherein the preprocessing includes: data cleaning, removing outliers and filling missing values.

[0009] According to a method for predicting an operation trend of a rail transit system provided by the present invention, the method also includes: inputting the stationary sequence samples into a long short-term memory network to obtain a first predicted operation trend output by the long short-term memory network; and inputting the stationary sequence samples into a state space model to obtain a second predicted operation trend output by the state space model; determining a first evaluation result corresponding to the long short-term memory network based on the first predicted operation trend and the operation trend samples; and determining a second evaluation result corresponding to the state space model based on the second predicted operation trend and the operation trend samples; and determining the trend prediction model from the long short-term memory network and the state space model based on the first evaluation result and the second evaluation result.

[0010] According to a method for predicting an operation trend of a rail transit system provided by the present invention, the first feature corresponding to the first target indicator data and the second feature corresponding to the second target indicator data are differentiated to obtain a stationary sequence corresponding to the indicator type, including: normalizing the first target indicator data to obtain third target indicator data, and normalizing the second target indicator data to obtain fourth target indicator data; extracting features from a multidimensional vector corresponding to the third target indicator data to obtain the first feature, and extracting features from a multidimensional vector corresponding to the fourth target indicator data to obtain the second feature; and differentiating the first feature and the second feature to obtain a stationary sequence corresponding to the indicator type.

[0011] According to a method for predicting the operation trend of a rail transit system provided by the present invention, the trend prediction model is obtained based on the following steps: inputting the stationary sequence samples into the initial trend prediction model to obtain the current predicted operation trend output by the initial trend prediction model; updating the parameters of the initial trend prediction model according to the target evaluation results corresponding to the current predicted operation trend and the operation trend samples to obtain a trained trend prediction model.

[0012] According to a method for predicting an operation trend of a rail transit system provided by the present invention, the method further includes: generating early warning information when the target operation trend reaches a preset threshold, wherein the early warning information is used to indicate that an abnormality exists in the rail transit system.

[0013] According to a method for predicting an operation trend of a rail transit system provided by the present invention, all first indicator data and all second indicator data are determined from the operation data through a Prometheus monitoring engine, and the method further includes: setting a data source in a Grafana data visualization tool and connecting the Prometheus monitoring engine; creating a dashboard in the Grafana data visualization tool, and determining the metrics corresponding to the target operation trend in the dashboard; and displaying the metrics using a constructed Prometheus query language statement.

[0014] The present invention also provides an operation trend prediction device for a rail transit system, comprising the following modules: a data processing module and a trend prediction module.

[0015] The data processing module is used to determine the first indicator data corresponding to each of the multiple indicator types at the current moment and the second indicator data corresponding to each of the multiple indicator types at the historical moment from the operation data corresponding to the rail transit system; and determine the stationary sequence corresponding to each of the multiple indicator types based on the first indicator data corresponding to each of the multiple indicator types and the second indicator data corresponding to each of the multiple indicator types.

[0016] The trend prediction module is used to input all stationary sequences into the trend prediction model to obtain the target operation trend of the rail transit system output by the trend prediction model. The trend prediction model is trained based on stationary sequence samples and operation trend samples.

[0017] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a method for predicting an operation trend of a rail transit system as described in any one of the above-mentioned methods is implemented.

[0018] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for predicting the operation trend of a rail transit system as described in any one of the above-mentioned methods is implemented.

[0019] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for predicting the operation trend of a rail transit system as described in any one of the above is implemented.

[0020] The operation trend prediction method, device, equipment and medium of the rail transit system provided by the present invention determine the first indicator data corresponding to each of the multiple indicator types at the current moment and the second indicator data corresponding to each of the multiple indicator types at the historical moment from the operation data corresponding to the rail transit system; determine the stationary sequence corresponding to each of the multiple indicator types according to the first indicator data corresponding to each of the multiple indicator types and the second indicator data corresponding to each of the multiple indicator types; input all the stationary sequences into the trend prediction model to obtain the target operation trend corresponding to the rail transit system output by the trend prediction model, and the trend prediction model is obtained by training based on stationary sequence samples and operation trend samples. After determining the stationary sequences corresponding to each of the multiple indicator types according to the operation data of the rail transit system, the method can use the trend prediction model to predict multiple stationary sequences and accurately obtain the target operation trend corresponding to the rail transit system. The whole process aims to achieve a transition from passive response to active management, so as to automatically prevent or avoid potential faults in the future, so as to reduce the processing time of emergencies and the risk of fault expansion. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0022] Figure 1 It is a flow chart of the operation trend prediction method of the rail transit system provided by the present invention.

[0023] Figure 2 It is a structural schematic diagram of the operation trend prediction device of the rail transit system provided by the present invention.

[0024] Figure 3 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0026] In order to better understand the embodiments of the present invention, the background technology is first described in detail: In the prior art, the rail transit system in the rail transit system has functional limitations, which are specifically reflected in the following two aspects.

[0027] The first aspect: the limitation of rail transit system relying on single indicator data. Specifically, most existing rail transit systems rely on single indicator data to evaluate their operation status. Since the whole process fails to fully consider the interaction and synergy between multiple indicator data, and in the actual rail transit system, the indicator data are often interrelated, and the change of one indicator data may have an impact on other indicator data, the whole process may not be able to fully reflect the actual operation status of the rail transit system, thereby limiting the system's ability to identify and respond to potential faults in a timely manner. In addition, the monitoring method of single indicator data may also ignore some important nonlinear relationships and dynamic change characteristics, resulting in limited accuracy of the prediction results.

[0028] Secondly, there are limitations in data visualization of rail transit systems. Specifically, the existing rail transit system provides basic displays of real-time data and historical data, as well as some simple historical data curves. This display method lacks the ability to conduct in-depth analysis and visualization of multiple indicator data, and cannot intuitively reveal the complex interrelationships and evolution trends among multiple indicator data, which limits the operation and maintenance personnel's in-depth understanding of the system status and effective decision-making.

[0029] In order to solve the above technical problems, the present invention provides a method, device, equipment and medium for predicting the operation trend of a rail transit system. The method overcomes the limitation of the existing rail transit system relying on a single indicator data. After determining the stationary sequences corresponding to the various indicator types according to the operation data of the rail transit system, a trend prediction model can be used to predict the multiple stationary sequences to obtain the target operation trend corresponding to the rail transit system. The whole process integrates the characteristics of multiple indicators, which can more comprehensively capture the changes in the system state and improve the ability to identify and respond to potential faults in a timely manner, thereby improving the accuracy and reliability of the prediction. In other words, the whole process aims to achieve a transition from passive response to active management, so as to automatically prevent or avoid potential faults in the future, so as to reduce the processing time of emergencies and the risk of fault expansion.

[0030] Secondly, this method can also intuitively display the prediction results (i.e., target operation trends) by using the Grafana data visualization tool, revealing the complex relationships and trends among various indicator data, solving the limitations of data visualization in the existing rail transit system, and providing operation and maintenance personnel with a more intuitive and rigorous monitoring and analysis method, helping them to have an in-depth understanding of the system status and make effective decisions.

[0031] In addition, the above process not only improves the reliability, safety and operational efficiency of the rail transit system, reduces equipment failures and accidents, but also extends the service life of equipment and reduces maintenance costs.

[0032] Taking electronic equipment as an example, the operation trend prediction method of the rail transit system provided by the present invention is described in detail below.

[0033] Among them, the system environment of the electronic equipment includes components such as the Prometheus monitoring engine, the Internet of Things Database (IoTDB) time series database, the alarm management (AlertManager) component, the Ansible automated deployment tool, and the Grafana data visualization tool.

[0034] It should be noted that electronic equipment needs to ensure the normal operation of each component to provide support for monitoring, analysis and prediction.

[0035] Figure 1 FIG. 1 is a flow chart of a method for predicting the operation trend of a rail transit system provided by the present invention. Figure 1 As shown, the method includes the following: step 101-step 103.

[0036] Step 101: Determine, from the operation data corresponding to the rail transit system, first indicator data corresponding to each of the plurality of indicator types at the current moment, and second indicator data corresponding to each of the plurality of indicator types at the historical moment.

[0037] Among them, the rail transit system is a complex network, including trains, tracks, stations, signal systems and other components. It is mainly responsible for providing efficient and convenient public transportation services between cities or regions.

[0038] Operation data refers to various data generated by the rail transit system during daily operation. Optionally, the operation data may include: passenger flow, train operation status, signal system status, equipment operation status, etc.

[0039] The current moment refers to the real-time point in time that we are analyzing or focusing on.

[0040] Historical moments refer to points in the past that are used for comparative analysis with the current moment to discover trends. Historical data is the basis for predicting future trends and formulating long-term strategies.

[0041] The indicator type refers to different dimensions or aspects used to measure the operating status or performance of the rail transit system. Optionally, the indicator type may include: passenger flow indicator, train operating status indicator, signal system status indicator, and equipment operating status indicator.

[0042] Indicator data refers to the value or status of a specific indicator at a specific point in time.

[0043] The electronic device can first monitor the rail transit system through the Prometheus monitoring engine and collect the corresponding operation data of the rail transit system; the electronic device then uses the Prometheus Query Language (PromQL) statement to query all indicator data from the operation data. In order to overcome the limitation of the existing rail transit system relying on a single indicator data, the electronic device can determine the first indicator data corresponding to each of the multiple indicator types at the current moment from all the indicator data, and determine the second indicator data corresponding to each of the multiple indicator types at the historical moment, so as to prepare for the subsequent fusion of multiple indicator data.

[0044] The number of first indicator data is the same as the number of indicator types, and the number of second indicator data is the same as the number of indicator types.

[0045] Step 102: Determine the stationary sequences corresponding to the various indicator types according to the first indicator data corresponding to the various indicator types and the second indicator data corresponding to the various indicator types.

[0046] A stationary series refers to a series whose statistical properties (such as mean, variance and correlation coefficient) do not change over time. In rail transit systems, a stationary series may refer to data such as passenger flow and train running time that have stable characteristics over a period of time (such as between the current time and the historical time).

[0047] In some embodiments, the electronic device determines the stationary sequence corresponding to each of the multiple indicator types based on the first indicator data corresponding to each of the multiple indicator types and the second indicator data corresponding to each of the multiple indicator types, which may include: performing the following operations on the first indicator data and the second indicator data corresponding to each indicator type: the electronic device preprocesses the first indicator data to obtain first target indicator data, and preprocesses the second indicator data to obtain second target indicator data; the electronic device differentiates a first feature corresponding to the first target indicator data and a second feature corresponding to the second target indicator data to obtain a stationary sequence corresponding to the indicator type.

[0048] Among them, preprocessing can include: data cleaning, removing outliers and filling missing values, etc.

[0049] Data cleaning is the final step to discover and correct identifiable errors in the indicator data series, aiming to improve data quality and make the data more suitable for analysis and mining. Optionally, this data cleaning can include: checking data consistency, handling invalid values ​​and missing values, etc.

[0050] Removing outliers refers to data points in the indicator data series that are significantly different from other data points, that is, abnormal data points, where the abnormal data points may be caused by measurement errors, recording errors, data capture errors, etc.

[0051] Filling missing values ​​means that there are missing or unfilled data values ​​in the indicator data series.

[0052] In order to ensure the data instructions of the first indicator data and the second indicator data corresponding to each indicator type, the electronic device can pre-process each first indicator data and each second indicator data respectively to obtain relatively complete and accurate first target indicator data and second target indicator data; the electronic device then determines the first feature corresponding to the first target indicator data and the second feature corresponding to the second target indicator data, and then differentiates the first feature and the second feature to obtain a stationary sequence corresponding to the indicator type. The entire differentiation process can eliminate or reduce non-stationary factors such as trend and seasonality in the original sequence, so that the differentiated stationary sequence is easier to model and predict.

[0053] Based on this, the electronic device can eventually obtain as many stable sequences as there are indicator types, in order to prepare for the subsequent feature fusion of multiple indicator data.

[0054] In some embodiments, the electronic device differentiates a first feature corresponding to the first target indicator data and a second feature corresponding to the second target indicator data to obtain a stationary sequence corresponding to the indicator type, which may include: the electronic device normalizes the first target indicator data to obtain the third target indicator data, and normalizes the second target indicator data to obtain the fourth target indicator data; the electronic device extracts features from the multidimensional vector corresponding to the third target indicator data to obtain the first feature, and extracts features from the multidimensional vector corresponding to the fourth target indicator data to obtain the second feature; the electronic device differentiates the first feature and the second feature to obtain a stationary sequence corresponding to the indicator type.

[0055] In order to facilitate the subsequent feature differential calculation, after determining the first target indicator data and the second target indicator data, the electronic device may first normalize the first target indicator data and the second target indicator data respectively to obtain the third target indicator data and the fourth target indicator data. At this time, the third target indicator data and the fourth target indicator data are in the same dimension. The electronic device then determines the multidimensional vector corresponding to the third target indicator data, and then extracts a more comprehensive and accurate first feature, and determines the multidimensional vector corresponding to the fourth target index feature, and then extracts a more comprehensive and accurate second feature; finally, the electronic device differentiates the first feature and the second feature to obtain a stationary sequence corresponding to the indicator type. The accuracy of the stationary sequence is also high, which provides data support for improving the accuracy of subsequent prediction results.

[0056] Optionally, after step 102, the method may further include: the electronic device stores all stationary sequences in the IoTDB time series database for subsequent data query.

[0057] Among them, the above-mentioned IoTDB time series database can also store stable sequence samples and operating trend samples for subsequent model training.

[0058] Among them, the stationary sequence sample is a kind of historical data, which records the sequence samples of the statistical characteristics of the rail transit system in the past period of time, and the statistical characteristics do not change with time.

[0059] Operation trend samples are also a kind of historical data, which record the operation status and change trend of the rail transit system over a period of time in the past.

[0060] Step 103: input all stationary sequences into the trend prediction model to obtain the target operation trend of the rail transit system output by the trend prediction model. The trend prediction model is trained based on stationary sequence samples and operation trend samples.

[0061] Among them, the target operation trend is usually expressed as a curve, which can intuitively show the operating status of the rail transit system in the future.

[0062] Since the trend prediction model is pre-trained, after determining all the stationary sequences, the electronic equipment can input all the stationary sequences as input data into the trend prediction model to obtain the target operation trend of the rail transit system output by the trend prediction model. The whole process can achieve effective active management and provide data support for the subsequent determination of targeted preventive measures. In this way, it is possible to achieve accurate prediction and timely intervention of the operating status of the rail transit system, thereby improving the reliability, safety and operational efficiency of the entire system.

[0063] In some embodiments, before step 103, the method may further include: the electronic device inputs the stationary sequence samples into a long short-term memory network (LSTM) to obtain a first predicted operating trend output by the LSTM network; and inputs the stationary sequence samples into a state space model to obtain a second predicted operating trend output by the state space model; the electronic device determines a first evaluation result corresponding to the LSTM network based on the first predicted operating trend and the operating trend samples; and determines a second evaluation result corresponding to the state space model based on the second predicted operating trend and the operating trend samples; the electronic device determines a trend prediction model from the LSTM network and the state space model based on the first evaluation result and the second evaluation result.

[0064] Among them, LSTM is a powerful recurrent neural network architecture that is particularly suitable for processing and predicting time series data. The main advantage of this LSTM is that it can capture long-term dependencies, which is crucial for server health analysis.

[0065] The state space model is a mathematical model used to describe the change of the state of a dynamic system over time. The state space model can model the current operating state of the rail transit system and predict the future operating state of the rail transit system. The advantage of the state space model is that it can process data containing noise and uncertainty, thereby improving the robustness of the prediction results.

[0066] The first evaluation indicator is used to evaluate the performance of LSTM; the second evaluation indicator is used to evaluate the performance of the state space model. Optionally, the first evaluation indicator can be mean square error (MSE) or root mean square error (RMSE), and the second evaluation indicator can be mean square error or root mean square error, which is not specifically limited here. However, the indicator type of the first evaluation indicator is the same as the indicator type of the second evaluation indicator, that is, the first evaluation indicator and the second evaluation indicator are both mean square error or root mean square error.

[0067] The electronic device can use the stationary sequence samples as training sets, and input them into the LSTM and state-space models respectively, to obtain the corresponding predicted operating trends; for the first predicted operating trend corresponding to the LSTM, the electronic device can determine the first evaluation result corresponding to the LSTM based on the first predicted operating trend and the operating trend samples; for the second predicted operating trend corresponding to the state-space model, the electronic device can determine the second evaluation result corresponding to the state-space model based on the second predicted operating trend and the operating trend samples; since the first evaluation result and the second evaluation result can determine the quality of the model, the electronic device can analyze the first evaluation result and the second evaluation result, and determine the best model among the LSTM and the state-space model as the trend prediction model.

[0068] In some embodiments, the trend prediction model is obtained based on the following steps: inputting the stationary sequence samples into the initial trend prediction model to obtain the current predicted operating trend output by the initial trend prediction model; updating the parameters of the initial trend prediction model according to the target evaluation results corresponding to the current predicted operating trend and the operating trend samples to obtain the trained trend prediction model.

[0069] When the trend prediction model is LSTM, the target evaluation result is the first evaluation result; when the trend prediction model is a state space model, the target evaluation result is the second evaluation result.

[0070] In the process of training the trend prediction model, the electronic device can first input the stationary sequence samples into the initial trend prediction model, and use the initial trend prediction model to predict the operation trend of the rail transit system based on the stationary sequence samples to obtain the current predicted operation trend. At this time, the current predicted operation trend is a relatively accurate prediction result; then, the electronic device determines the loss value between the current predicted operation trend and the operation trend sample, that is, the target evaluation result, and uses the target evaluation result to update the parameters of the initial trend prediction model to obtain a trained trend prediction model, which has high generalization ability and robustness.

[0071] It should be noted that during the model training process, cross-validation can be performed in the training set to evaluate the performance and generalization ability of the trend prediction model. Among them, cross-validation is used to evaluate the robustness and accuracy of the trend prediction model. By training and validating the trend prediction model multiple times, the generalization ability of the trend prediction model on unknown indicator data is ensured, which helps to ensure the reliability and effectiveness of the established trend prediction model, thereby improving the accuracy of the prediction results.

[0072] Optionally, the electronic device may periodically evaluate and retrain the trend prediction model to adapt to changes in the distribution of the indicator data.

[0073] Optionally, after the electronic device determines the trend prediction model, the method may further include: the electronic device deploys the trend prediction model to a system environment corresponding to the electronic device through the Ansible automated deployment tool.

[0074] In this way, electronic devices can load real-time data based on the deployed trend prediction models.

[0075] In some embodiments, after step 103, the method may further include: the electronic device generates warning information when the target operation trend reaches a preset threshold.

[0076] Among them, the early warning information is used to indicate abnormalities in the rail transit system.

[0077] Optionally, the preset threshold may be set before the electronic device leaves the factory, or may be user-defined, which is not specifically limited here.

[0078] After determining the target operation trend, the electronic device can compare the target operation trend with the preset threshold through the AlertManager component: if the target operation trend does not reach the preset threshold, it means that the rail transit system is in a normal state; if the target operation trend reaches the preset threshold, it means that the rail transit system is in an abnormal state. At this time, the electronic device can generate an early warning message to indicate that the rail transit system is abnormal. In addition, the early warning message may also carry the type of abnormal information. In other words, the electronic device can implement timely alarms based on the preset threshold, prompting possible health problems so that operation and maintenance personnel can identify and avoid potential faults in a timely manner.

[0079] In some embodiments, all first indicator data and all second indicator data are determined from the operation data by the Prometheus monitoring engine. After step 103, the method may further include: the electronic device sets the data source (Data Sources) in the Grafana data visualization tool and connects to the Prometheus monitoring engine; the electronic device creates a dashboard Dashboard in the Grafana data visualization tool, and determines the metrics corresponding to the target operation trend in the dashboard Dashboard; the electronic device uses the constructed Prometheus query language statement to display the metrics Metrics.

[0080] In order to visualize the target operation trend, the electronic device can set the data source in the Grafana data visualization tool, set the type of the data source to Prometheus, and connect the Prometheus monitoring engine; then, the electronic device creates a dashboard in the Grafana data visualization tool, the dashboard can be named "System Monitoring", and selects a panel type suitable for displaying the target operation trend in the dashboard, and then modifies the metrics in the panel to determine the metrics corresponding to the target operation trend. In addition, the title, color theme, legend, etc. of the panel can be adjusted to better display the target operation trend; finally, the electronic device can construct a PromQL statement to achieve an intuitive display of the target operation trend corresponding to a variety of indicator data. The whole process can timely and intuitively see the system operation status, prevent or avoid potential failures in advance, shorten the processing time of emergencies, and reduce the risk of failure expansion.

[0081] In an embodiment of the present invention, the first indicator data corresponding to each of the multiple indicator types at the current moment and the second indicator data corresponding to each of the multiple indicator types at the historical moment are determined from the operation data corresponding to the rail transit system; the stationary sequences corresponding to each of the multiple indicator types are determined based on the first indicator data corresponding to each of the multiple indicator types and the second indicator data corresponding to each of the multiple indicator types; all the stationary sequences are input into the trend prediction model to obtain the target operation trend corresponding to the rail transit system output by the trend prediction model, and the trend prediction model is obtained by training based on stationary sequence samples and operation trend samples. After determining the stationary sequences corresponding to each of the multiple indicator types based on the operation data of the rail transit system, the method can use the trend prediction model to predict multiple stationary sequences and accurately obtain the target operation trend corresponding to the rail transit system. The whole process aims to achieve a transition from passive response to active management, so as to automatically prevent or avoid potential failures in the future, so as to reduce the processing time of emergencies and the risk of failure expansion.

[0082] The operation trend prediction device of the rail transit system provided by the present invention is described below. The operation trend prediction device of the rail transit system described below and the operation trend prediction method of the rail transit system described above can be referenced to each other.

[0083] Figure 2 Schematic diagram of the structure of the operation trend prediction device of the rail transit system provided by the present invention. Figure 2 As shown, the device includes the following: a data processing module 201 and a trend prediction module 202.

[0084] The data processing module 201 is used to determine, from the operation data corresponding to the rail transit system, first indicator data corresponding to each of the multiple indicator types at the current moment, and second indicator data corresponding to each of the multiple indicator types at the historical moment; and to determine the stationary sequence corresponding to each of the multiple indicator types according to the first indicator data corresponding to each of the multiple indicator types and the second indicator data corresponding to each of the multiple indicator types; The trend prediction module 202 is used to input all stationary sequences into a trend prediction model to obtain a target operation trend of the rail transit system output by the trend prediction model. The trend prediction model is trained based on stationary sequence samples and operation trend samples.

[0085] Optionally, the data processing module 201 is specifically used to perform the following operations on the first indicator data and the second indicator data corresponding to each indicator type: preprocess the first indicator data to obtain first target indicator data, and perform the preprocessing on the second indicator data to obtain second target indicator data; differentiate the first feature corresponding to the first target indicator data and the second feature corresponding to the second target indicator data to obtain a stationary sequence corresponding to the indicator type; wherein the preprocessing includes: data cleaning, removal of outliers and filling in missing values.

[0086] Optionally, the data processing module 201 is specifically used to input the stationary sequence sample into the long short-term memory network to obtain a first predicted operating trend output by the long short-term memory network; and input the stationary sequence sample into the state space model to obtain a second predicted operating trend output by the state space model; determine a first evaluation result corresponding to the long short-term memory network based on the first predicted operating trend and the operating trend sample; and determine a second evaluation result corresponding to the state space model based on the second predicted operating trend and the operating trend sample; and determine the trend prediction model from the long short-term memory network and the state space model based on the first evaluation result and the second evaluation result.

[0087] Optionally, the data processing module 201 is specifically used to normalize the first target indicator data to obtain the third target indicator data, and to normalize the second target indicator data to obtain the fourth target indicator data; to extract features from the multidimensional vector corresponding to the third target indicator data to obtain the first feature, and to extract features from the multidimensional vector corresponding to the fourth target indicator data to obtain the second feature; and to differentiate the first feature and the second feature to obtain a stationary sequence corresponding to the indicator type.

[0088] Optionally, the trend prediction model is obtained based on the following steps: inputting the stationary sequence sample into the initial trend prediction model to obtain the current predicted operating trend output by the initial trend prediction model; updating the parameters of the initial trend prediction model according to the target evaluation results corresponding to the current predicted operating trend and the operating trend sample to obtain a trained trend prediction model.

[0089] Optionally, the data processing module 201 is further used to generate warning information when the target operation trend reaches a preset threshold, and the warning information is used to indicate that an abnormality exists in the rail transit system.

[0090] Optionally, all first indicator data and all second indicator data are determined from the operation data through the Prometheus monitoring engine. The data processing module 201 is also used to set the data source in the Grafana data visualization tool and connect to the Prometheus monitoring engine; create a dashboard in the Grafana data visualization tool, and determine the metrics corresponding to the target operation trend in the dashboard; and use the constructed Prometheus query language statement to display the metrics.

[0091] Figure 3 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 3 As shown, the electronic device may include: a processor 310, a communication interface 320, a memory 330 and a communication bus 340, wherein the processor 310, the communication interface 320 and the memory 330 communicate with each other through the communication bus 340. The processor 310 may call the logic instructions in the memory 330 to execute the operation trend prediction method of the rail transit system, the method comprising: determining the first indicator data corresponding to each of the multiple indicator types at the current moment and the second indicator data corresponding to each of the multiple indicator types at the historical moment from the operation data corresponding to the rail transit system; determining the stationary sequence corresponding to each of the multiple indicator types according to the first indicator data corresponding to each of the multiple indicator types and the second indicator data corresponding to each of the multiple indicator types; inputting all the stationary sequences into the trend prediction model to obtain the target operation trend corresponding to the rail transit system output by the trend prediction model, wherein the trend prediction model is obtained by training based on stationary sequence samples and operation trend samples.

[0092] In addition, the logic instructions in the above-mentioned memory 330 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment 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), disk or optical disk, etc. Various media that can store program codes.

[0093] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the operation trend prediction method of the rail transit system provided by the above-mentioned methods, and the method includes: determining, from the operation data corresponding to the rail transit system, the first indicator data corresponding to each of the multiple indicator types at the current moment, and the second indicator data corresponding to each of the multiple indicator types at the historical moment; determining the stationary sequence corresponding to each of the multiple indicator types according to the first indicator data corresponding to each of the multiple indicator types and the second indicator data corresponding to each of the multiple indicator types; inputting all stationary sequences into a trend prediction model to obtain the target operation trend corresponding to the rail transit system output by the trend prediction model, and the trend prediction model is trained based on stationary sequence samples and operation trend samples.

[0094] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the operation trend prediction method of the rail transit system provided by the above-mentioned methods, the method comprising: determining, from the operation data corresponding to the rail transit system, first indicator data corresponding to each of the multiple indicator types at the current moment, and second indicator data corresponding to each of the multiple indicator types at the historical moment; determining, based on the first indicator data corresponding to each of the multiple indicator types and the second indicator data corresponding to each of the multiple indicator types, a stationary sequence corresponding to each of the multiple indicator types; inputting all stationary sequences into a trend prediction model to obtain the target operation trend corresponding to the rail transit system output by the trend prediction model, wherein the trend prediction model is trained based on stationary sequence samples and operation trend samples.

[0095] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0096] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0097] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting the operation trend of a rail transit system, characterized in that: include: Determine, from the operation data corresponding to the rail transit system, first indicator data corresponding to each of the plurality of indicator types at the current moment, and second indicator data corresponding to each of the plurality of indicator types at the historical moment; Determine, according to the first indicator data and the second indicator data corresponding to the plurality of indicator types, the stationary sequences corresponding to the plurality of indicator types; All stationary sequences are input into a trend prediction model to obtain a target operation trend of the rail transit system output by the trend prediction model, wherein the trend prediction model is trained based on stationary sequence samples and operation trend samples.

2. The method for predicting the operation trend of a rail transit system according to claim 1, characterized in that: The determining, according to the first indicator data and the second indicator data corresponding to the multiple indicator types, the stationary sequences corresponding to the multiple indicator types respectively include: The following operations are performed for the first indicator data and the second indicator data corresponding to each indicator type: Preprocessing the first indicator data to obtain first target indicator data, and preprocessing the second indicator data to obtain second target indicator data; Differentiating a first feature corresponding to the first target indicator data and a second feature corresponding to the second target indicator data to obtain a stationary sequence corresponding to the indicator type; The preprocessing includes: data cleaning, removing outliers and filling missing values.

3. The method for predicting the operation trend of a rail transit system according to claim 1, characterized in that: The method further comprises: Inputting the stationary sequence samples into a long short-term memory network to obtain a first predicted operating trend output by the long short-term memory network; and inputting the stationary sequence samples into a state space model to obtain a second predicted operating trend output by the state space model; Determine a first evaluation result corresponding to the long short-term memory network according to the first predicted operating trend and the operating trend sample; and determine a second evaluation result corresponding to the state space model according to the second predicted operating trend and the operating trend sample; According to the first evaluation result and the second evaluation result, the trend prediction model is determined from the long short-term memory network and the state space model.

4. The method for predicting the operation trend of a rail transit system according to claim 2, characterized in that: The step of differentiating the first feature corresponding to the first target indicator data and the second feature corresponding to the second target indicator data to obtain a stationary sequence corresponding to the indicator type includes: Normalizing the first target indicator data to obtain third target indicator data, and normalizing the second target indicator data to obtain fourth target indicator data; Performing feature extraction on the multidimensional vector corresponding to the third target indicator data to obtain the first feature, and performing feature extraction on the multidimensional vector corresponding to the fourth target indicator data to obtain the second feature; The first feature and the second feature are differentiated to obtain a stationary sequence corresponding to the indicator type.

5. The method for predicting the operation trend of a rail transit system according to claim 1, characterized in that: The trend prediction model is obtained based on the following steps: Inputting the stationary sequence samples into the initial trend prediction model to obtain the current predicted operating trend output by the initial trend prediction model; According to the target evaluation results corresponding to the current predicted operating trend and the operating trend samples, the parameters of the initial trend prediction model are updated to obtain a trained trend prediction model.

6. The method for predicting the operation trend of a rail transit system according to any one of claims 1 to 5, characterized in that: The method further comprises: When the target operation trend reaches a preset threshold, a warning message is generated, and the warning message is used to prompt that an abnormality exists in the rail transit system.

7. The method for predicting the operation trend of a rail transit system according to any one of claims 1 to 5, characterized in that: All first indicator data and all second indicator data are determined from the operation data by a Prometheus monitoring engine, and the method further includes: Set up the data source in the Grafana data visualization tool and connect to the Prometheus monitoring engine; Create a dashboard in the Grafana data visualization tool, and determine the metrics corresponding to the target operation trend in the dashboard; The constructed Prometheus query language statement is used to display the metrics.

8. A device for predicting the operation trend of a rail transit system, characterized in that: include: A data processing module is used to determine, from the operation data corresponding to the rail transit system, first indicator data corresponding to each of the multiple indicator types at the current moment, and second indicator data corresponding to each of the multiple indicator types at the historical moment; and to determine the stationary sequence corresponding to each of the multiple indicator types according to the first indicator data corresponding to each of the multiple indicator types and the second indicator data corresponding to each of the multiple indicator types; The trend prediction module is used to input all stationary sequences into the trend prediction model to obtain the target operation trend of the rail transit system output by the trend prediction model. The trend prediction model is trained based on stationary sequence samples and operation trend samples.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the operation trend prediction method of the rail transit system as described in any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the operation trend prediction method of the rail transit system as claimed in any one of claims 1 to 7 is implemented.

Images