Intelligent prediction method for train receiving and sending route occupation time of railway technical station

Abstract

The application discloses a kind of railway technical station train receiving and sending route occupation time intelligent prediction method, for the railway technical station train receiving and sending operation process, using the massive data accumulated in daily production process, the route required occupation time in different operation links of train is calculated scientifically and accurately, and train command personnel can know the future execution progress of each operation in advance, which can reduce labor intensity and improve operation efficiency.

Description

Technical Field

[0001] This invention relates to the field of rail transit technology, and in particular to an intelligent prediction method for train arrival and departure route occupancy time at railway technical stations. Background Technology

[0002] As a key node in the railway transportation network, railway technical stations are responsible for the arrival, departure, shunting, and locomotive swapping of freight trains. These operations are extremely busy, and the station's route resources are very limited. Within the same throat area of ​​the technical station, train routes, shunting routes, and locomotive routes intersect and overlap, causing conflicts in time and space that affect operational efficiency and traffic safety.

[0003] In actual production operations, freight trains arriving and departing at the technical station are often delayed due to factors such as uneven train flow, equipment failure, construction work, weather, and human factors. Currently, experienced train dispatchers can only rely on manual intervention by identifying opportunities based on station conditions, operational status, and interlocking relationships to select and arrange routes for execution. The prolonged periods of high concentration required by train dispatchers can easily lead to missed or incorrect operations, potentially causing safety accidents and hindering the overall transportation organization and turnover efficiency of the railway freight hub.

[0004] Chinese invention patent application No. 201911077494.2, entitled "A Method for Predicting the Arrival Time of High-Speed ​​Railway Trains Based on a Dispatch and Command System," discloses a method for predicting the arrival time of high-speed railway trains based on a dispatch and command system. The method includes: standardizing historical data from the dispatch and command system; constructing a spatiotemporal state data processing model based on block sections and discretized train trajectories; extracting and standardizing the spatiotemporal data of train operation; estimating the block section running time parameters for various types of trains in the actual operating environment by analyzing the standardized historical train operation data; using the estimated block section running time parameters, combined with real-time operating environment and train status, predicting the signal status for a future period; and ultimately predicting the train arrival time through the mutual constraints and interactions between train and signal status.

[0005] However, the patent application provides a method for predicting the arrival time of high-speed railway trains. High-speed railways are characterized by scheduled train operation and stable operation, which is quite different from the flexible operation scenario of railway technical stations. The factors affecting the operation of high-speed railways and railway technical stations are different. Therefore, the solution in this patent application cannot be applied to predict the occupancy time of train arrival and departure routes at railway technical stations.

[0006] In view of this, the present invention is hereby proposed. Summary of the Invention

[0007] The purpose of this invention is to provide an intelligent prediction method for the occupancy time of train arrival and departure routes at railway technical stations. This method accurately calculates the occupancy time of train arrival and departure routes by combining massive amounts of daily data during the train arrival and departure operations at railway technical stations.

[0008] The objective of this invention is achieved through the following technical solution:

[0009] A method for intelligently predicting the occupancy time of train arrival and departure routes at railway technical stations includes:

[0010] Analyze the train reception and dispatch operations at railway technical stations to determine the factors affecting the time occupied by reception and dispatch routes;

[0011] Data is collected using identified influencing factors as feature parameters, and after processing, training and test sets are obtained.

[0012] The training set is used to build an intelligent prediction model, which is then used to predict the test set. The prediction results are then used to optimize the model parameters.

[0013] After optimization, the feature values ​​of the current operation route instruction are extracted based on the identified influencing factors and input into the optimized intelligent prediction model to obtain the predicted occupancy time for each route.

[0014] As can be seen from the technical solution provided by this invention, by utilizing the massive amounts of data accumulated in daily production processes, the time required for routes in different operational stages of a train can be scientifically and accurately calculated. Train dispatchers can then know the future progress of each operation in advance, reducing their workload and improving operational efficiency. Simultaneously, it provides more accurate time parameters for station transportation production organization decisions, improving station capacity utilization and laying the foundation for the future intelligent transformation of railway technical stations. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A flowchart illustrating an intelligent prediction method for train arrival and departure route occupancy time at railway technical stations, provided in an embodiment of the present invention;

[0017] Figure 2 This invention provides an analysis diagram of the operation process of a technical station via trains and terminating trains, as provided in an embodiment of the invention.

[0018] Figure 3This invention provides an analysis diagram of the technical station originating train operation process in an embodiment of the invention.

[0019] Figure 4 This is a flowchart of the random forest regression prediction model provided in an embodiment of the present invention;

[0020] Figure 5 This is a schematic diagram illustrating the application scheme of the route occupancy time prediction model provided in an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.

[0022] First, the following explanations are provided for the terms that may be used in this article:

[0023] The terms “including,” “comprising,” “containing,” “having,” or other similar semantic descriptions should be interpreted as non-exclusive inclusion. For example, “including a technical feature element (such as raw material, component, ingredient, carrier, dosage form, material, size, part, component, mechanism, device, step, process, method, reaction conditions, processing conditions, parameter, algorithm, signal, data, product or article of manufacture, etc.)” should be interpreted as including not only the expressly listed technical feature element, but also other technical feature elements that are not expressly listed and are well-known in the art.

[0024] The term "composed of" excludes any technical features not expressly listed. When used in a claim, it closes the claim to exclude all technical features other than those expressly listed, except for associated conventional impurities. If the term appears only in a clause of a claim, it limits the claim to the elements expressly listed in that clause; elements recited in other clauses are not excluded from the overall claim.

[0025] The following is a detailed description of the intelligent prediction method for train arrival and departure route occupancy time at railway technical stations provided by this invention. Contents not described in detail in the embodiments of this invention belong to prior art known to those skilled in the art. Where specific conditions are not specified in the embodiments of this invention, they shall be performed according to conventional conditions in the art or conditions recommended by the manufacturer.

[0026] like Figure 1The diagram shows a flowchart of an intelligent prediction method for train arrival and departure route occupancy time at a railway technical station, provided by an embodiment of the present invention. This method can be abstracted as follows: for the same route, based on known influencing factors of different operational processes, the occupancy time of the route is predicted. The input is a series of numerical or categorical feature data, and the output is the occupancy time range for the entire route. The overall steps are as follows:

[0027] (1) Analyze the train reception and dispatch process at railway technical stations and determine the factors affecting the time occupied by the reception and dispatch routes.

[0028] (2) Collect relevant data by using the identified influencing factors as feature parameters, and obtain training and test sets after processing.

[0029] (3) Use the training set to build an intelligent prediction model, use the intelligent prediction model to predict the test set, and use the prediction results to optimize the model parameters;

[0030] (4) After optimization, the feature values ​​of the current operation route instruction (real-time data on site) are extracted based on the determined influencing factors and input into the optimized intelligent prediction model to obtain the predicted occupancy time of each route.

[0031] To more clearly demonstrate the technical solution and its effects provided by the present invention, the method provided by the embodiments of the present invention will be described in detail below with reference to specific examples.

[0032] 1. Analyze the work process and identify influencing factors.

[0033] In actual operation, train dispatchers promptly process train arrival and departure procedures and open arrival and departure routes according to the phased plan requirements issued by the train dispatcher at the dispatching center to ensure train safety and punctuality. Trains within the technical station can be categorized into three types based on their arrival and departure methods: terminating trains, originating trains, and through trains. The factors influencing route occupancy time differ for each type of train arrival and departure operation. Therefore, for each type, the arrival and departure operation process is analyzed separately to determine the factors affecting route occupancy time.

[0034] 1. The final train.

[0035] In this embodiment of the invention, for the type of terminating train, the influencing factors of the occupancy time of the receiving route are determined in conjunction with the train receiving operation process. Specifically, for the type of terminating train, the entire train receiving operation process is from the time the station agrees to the completion of departure procedures at the neighboring station A until the train is completely stopped on the track. This includes: the departure procedures at station A, the opening of the departure route at station A, the train passing through the departure signal at station A, the train passing through the approach section at this station, the opening of the receiving route at this station, and the train completely entering the track. Taking the completion of departure procedures at station A as the starting point for calculation, and the other operation processes as key adjustment points, the operation time between all pairs of adjacent key adjustment points constitutes the occupancy time of the terminating train receiving route at this station (from the opening of the route at this station to the complete entry of the train onto the track). The operation time between each pair of adjacent key adjustment points is analyzed separately to obtain the corresponding influencing factors.

[0036] like Figure 2 The diagram shown is an analysis of the operational processes of the technical station for passing trains and arriving trains. For the arriving train, the seven operational processes involved correspond to the following time intervals between adjacent key adjustment points: Figure 2 The values ​​t1, t2, T2, T3, t3, and t5 are given.

[0037] The main factors influencing the time required for each task are as follows:

[0038] Factors affecting t1: the operator's operating habits.

[0039] Factors affecting t2: arrival and departure type of trains at adjacent stations, train type, total weight of train, number of cars in the train, weather conditions, and speed limit conditions.

[0040] Factors affecting T2 include: route, speed limit conditions, train type, total train weight, and number of cars.

[0041] Factors affecting T3: train type, total train weight, number of cars, and speed limit conditions.

[0042] t3 Influencing factors: Timing of route opening.

[0043] Factors affecting T5: route, speed limit conditions, train type, train arrival and departure type, total train weight, and number of cars.

[0044] 2. Originating train.

[0045] In this embodiment of the invention, for the originating train type, the influencing factors of the departure route occupancy time are determined in conjunction with the departure operation process. Specifically, for the originating train type, the entire departure operation process of the originating train is from the time the station applies for the departure notice to the time the train completely enters the departure section, including: the station handles the departure procedures, the station opens the departure route, the train turns off the station's departure signal, and the train turns red in the station's departure section. Taking the station handling the departure procedures as the starting point for calculation, other operation processes are taken as key adjustment points. The operation time between all pairs of adjacent key adjustment points constitutes the occupancy time of the originating train's departure route at the station (from the opening of the station's departure route to the time the train completely enters the departure section). The operation time between each pair of adjacent key adjustment points is analyzed separately to obtain the corresponding influencing factors.

[0046] like Figure 3 The diagram shown is an analysis of the operation process for a train originating from a technical station. For the four operational processes involved in the originating train, the operation time between any two adjacent critical adjustment points corresponds to... Figure 3 t6, t7, and T9.

[0047] The main factors influencing the time required for each task are as follows:

[0048] t6 Influencing factors: On-duty operator's operating habits.

[0049] Factors affecting T7: train type, total train weight, number of cars, weather conditions, and speed limit conditions.

[0050] Factors affecting T9: route, speed limit conditions, train type, total train weight, and number of cars.

[0051] 3. By train.

[0052] In this embodiment of the invention, for through train types, the factors influencing the time occupied by the train receiving and dispatching routes are determined in conjunction with the train receiving and dispatching operation process. Specifically, for through train types, the entire train receiving and dispatching operation process is from the time the station agrees to the completion of departure procedures at the neighboring station A until the train completely enters the departure section of this station. This includes: station A completing departure procedures, station A opening its departure route, the train passing through the departure signal of station A, the train passing through the departure section of station A, the train passing through the approach section of this station, the station opening its receiving route, the station completing departure procedures, the station opening its departure route, the train completely entering the track, the train passing through the departure signal of this station, and the train passing through the departure section of this station. Taking the completion of departure procedures at station A as the starting point for calculation, and the other operation processes as key adjustment points, the operation time between all pairs of adjacent key adjustment points constitutes the time occupied by the train receiving and dispatching routes at this station. The operation time between each pair of adjacent key adjustment points is analyzed separately to obtain the corresponding influencing factors.

[0053] like Figure 2 As shown, for the eleven operational processes involved in the originating train, the operational times between any two adjacent key adjustment points are t1, t2, T2, T3, t3, t4, t6, t7, T6, and T7, respectively.

[0054] By analyzing the factors affecting the time required for each operation process, apart from the already mentioned operation times for terminating and originating trains, the main factors affecting operation time that were not mentioned are as follows:

[0055] Factors affecting T4: route, speed limit conditions, train type, train arrival and departure type, total train weight, and number of cars.

[0056] Factors affecting T6: train type, total train weight, number of cars, weather conditions, and speed limits.

[0057] Factors affecting T7: route, speed limit conditions, train type, total train weight, number of cars.

[0058] It should be noted that the above provides examples of the influencing factors for each task time. In practical applications, users can also adjust the influencing factors according to the actual situation or experience.

[0059] II. Massive data collection, processing, and feature variable selection.

[0060] In this embodiment of the invention, station display information is collected, and train number window information is combined to search for work routes. Occupancy and clearance information of sections within the route is collected to form tracking result information containing the route, which is then stored in a database. Simultaneously, the tracking result information and the executed work plan are correlated to obtain relevant train information. All of the above information belongs to historical data. Determined influencing factors are used as feature parameters, and corresponding data are collected from the above historical data. For influencing factors not included in the historical data, data is collected using the corresponding influencing factor as the primary key. Finally, data corresponding to all determined influencing factors are obtained.

[0061] In this embodiment of the invention, the data is standardized and normalized according to the characteristics of different influencing factors, dividing the influencing factors into numerical feature influencing factors and categorical feature influencing factors. For the collected data belonging to numerical feature influencing factors, normalization is performed. Specifically, numerical features are features whose values ​​are numbers. To eliminate the influence of differences in units and scales between features and to treat each feature dimension equally, feature normalization is required. Normalization is performed using standard deviation, that is, scaling the feature values ​​to a mean of 0 and a variance of 1, which can reduce the impact of features with excessively large values ​​on the accuracy of the training model. For the collected data belonging to categorical feature influencing factors, one-hot feature encoding is performed. Specifically, categorical features are features whose values ​​are different categories. Different categorical features typically have values ​​that are characters with different meanings. To enable the training model to recognize their meanings and perform processing operations, one-hot feature encoding is required.

[0062] Next, the processed data is divided into training and test sets according to a set ratio. For example, the Bootstrap sampling method can be used to divide the processed data into training and test sets in a 2:1 ratio.

[0063] Taking the prediction of the time occupied by the route for receiving the final train as an example, assuming that time t is the time when the procedures for the departure of the train from station A are completed at the adjacent station, the time range occupied by the route is let be a = t1 + t2 + T2 + T3 + t3, that is, the prediction target is the time range of [t + a, t + a + t5]. The influencing factors are summarized in Table 1.

[0064] Table 1: Summary of Influencing Factors

[0065]

[0066] The function of filtering predictor variables was used to perform Spearman correlation analysis on nine feature variables and the target value. Feature variables that are significantly correlated with the target value were selected as input variables for the prediction model, and a usable dataset of multiple feature variables and target values ​​was obtained, as shown in Table 2.

[0067] Table 2: Available Datasets

[0068]

[0069] III. Construction and Training of Intelligent Prediction Models

[0070] In this embodiment of the invention, the model type can be selected according to actual needs, such as a random forest model, a recurrent neural network model, a convolutional neural network model, a support vector machine model, etc. The random forest model will be used as an example below.

[0071] The training set obtained through the aforementioned method is used to train the model, and the test set is used to measure the model's predictive performance. Utilizing the idea of ​​random forest regression, a random forest-based intelligent prediction model is constructed to obtain the final predicted value, such as... Figure 4 As shown, the main steps are:

[0072] (1) Load the training set for model training.

[0073] (2) Random sampling. A number of samples are randomly selected from the training set to form a sample set. The size of the sample set is the same as that of the training set. This sampling method is called random sampling. Usually, data preprocessing has been completed before forming the training set. Sampling with replacement can ensure that the number of samples in the sample set is the same as that in the training set.

[0074] (3) Constructing decision trees. For each sample set, a decision tree is constructed using a random feature selection method, resulting in multiple decision trees. The Gini index is used for feature selection and node partitioning when constructing the decision trees.

[0075] (4) Ensemble decision trees. For multiple constructed decision trees, they are combined into a random forest model; for example, classification can be performed by calculating the average or votes of the classification results of each sample in each decision tree.

[0076] (5) Result prediction. The constructed random forest model is used to perform classification or regression prediction on the test set to obtain the result of each decision tree. The average value is the model output result, and the training is finally completed.

[0077] (6) Evaluate model performance. Evaluate the performance of the random forest model based on the model output results, and adjust the model parameters according to the evaluation results. For example, cross-validation and other methods can be used to evaluate the performance of the random forest model and perform parameter tuning.

[0078] IV. Route Occupancy Time Prediction.

[0079] Based on the above scheme, a practically applicable intelligent prediction model is obtained. Real-time on-site data is input into the intelligent prediction model to obtain accurate route occupancy time prediction results, which are then provided to the station control system for necessary autonomous calculations and safety control. Preferably, after train operation is completed, the actual train operation time is recorded and compared with the corresponding route occupancy time prediction value. The error rate is calculated, and it is determined whether the error rate is within the set range. If not, the data is updated to the dataset to retrain the intelligent prediction model.

[0080] like Figure 5 The diagram illustrates the application process of the route occupancy time prediction model, which mainly includes:

[0081] (1) After the duty officer selects the train operation task on the station control system and starts the operation, the system decomposes the train operation plan and obtains the operation route instruction.

[0082] (2) Pass the feature value of the task to the intelligent prediction model to obtain the occupancy time prediction value of each route.

[0083] (3) When the route arrangement is completed, the corresponding operations are carried out, and after the train operation is completed, the actual train operation time is recorded, compared with the predicted value, and the error rate is calculated (the acceptable error rate range is set by experts) to determine whether the error is acceptable.

[0084] (4) If the error rate is acceptable, the prediction ends; if the error rate requirement is not met, the data involved in the prediction process is updated to the training set to retrain the prediction model, resulting in an optimized new prediction model. The data involved in the prediction process mentioned here mainly include: feature values ​​(influencing factors), actual train operation time, and may also include predicted occupancy time.

[0085] The above-mentioned solution provided by the embodiments of the present invention accurately calculates the time occupied by the arrival and departure routes of trains waiting to be operated by combining massive amounts of daily data, based on the operational process of train dispatchers. It can provide train dispatchers with auxiliary decision-making and adjustments, which is of great significance for improving the level of railway transportation dispatch and operation control, strengthening the overall throughput capacity of stations, optimizing scheduling, improving transportation efficiency and ensuring train operation safety.

[0086] Through the above description of the embodiments, those skilled in the art can clearly understand that the above embodiments can be implemented by software, or by using software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solutions of the above embodiments can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.), including several instructions to cause a computer device (such as a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0087] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for intelligently predicting the occupancy time of train arrival and departure routes at railway technical stations, characterized in that, include: Analyze the train reception and dispatch operations at railway technical stations to determine the factors affecting the time occupied by reception and dispatch routes; Data is collected using identified influencing factors as feature parameters, and after processing, training and test sets are obtained. The training set is used to build an intelligent prediction model, which is then used to predict the test set. The prediction results are then used to optimize the model parameters. After optimization, the feature values ​​of the current operation route instruction are extracted based on the identified influencing factors and input into the optimized intelligent prediction model to obtain the predicted occupancy time for each route. Among them, the factors affecting the time occupied by the train receiving and dispatching operations at railway technical stations, as analyzed, include: Railway technical stations classify train arrival and departure types into three categories: terminating trains, originating trains, and through trains. For terminating trains, the entire process of receiving a terminating train, from the moment the station agrees to the completion of departure procedures at the neighboring station A until the train has completely entered the track and come to a complete stop, includes: the departure procedures at station A, the opening of the departure route at station A, the train passing through the departure signal at station A, the train passing through the red section of the departure area at station A, the train passing through the red section of the approach area at this station, the opening of the receiving route at this station, and the train completely entering the track. Taking the completion of departure procedures at station A as the starting point for calculation, and the other operational processes as key adjustment points, the operational time between all pairs of adjacent key adjustment points constitutes the time occupied by the receiving route of the terminating train at this station. For originating trains, the entire departure process begins from the time the station applies for departure notice and ends when the train has fully entered the departure section. This includes: processing departure procedures at the station, opening the departure route at the station, the train passing through the station's departure signal, and the train passing through the station's departure section. Taking the processing of departure procedures at the station as the starting point, other operational processes are considered as key adjustment points. The operational time between all two adjacent key adjustment points constitutes the time the originating train's departure route occupies at the station. For through trains, the entire process of receiving and dispatching a through train begins from the moment the train agrees to the completion of departure procedures at the neighboring station A, until the train has completely entered the departure section of this station. This includes: the train completing departure procedures at station A, the train opening its departure route at station A, the train passing through the departure section of station A, the train passing through the approach section of this station, the train opening its receiving route at this station, the train completing departure procedures at this station, the train opening its departure route at this station, the train completely entering the track, the train passing through the departure signal of this station, and the train passing through the departure section of this station. Taking the completion of departure procedures at station A as the starting point, and the other operational processes as key adjustment points, the operational time between all pairs of adjacent key adjustment points constitutes the time occupied by the receiving and departure routes of the through train at this station. The operation time between each pair of adjacent critical adjustment points is analyzed to obtain the corresponding influencing factors.

2. The intelligent prediction method for train arrival and departure route occupancy time at railway technical stations according to claim 1, characterized in that, The process of collecting relevant data using identified influencing factors as characteristic parameters includes: By collecting station display information and combining it with train number window information, the operation route is searched. The occupancy and clearance information of the section within the route is collected, and the tracking result information containing the route is generated and stored in the database. At the same time, the tracking result information and the executed operation plan are correlated to obtain relevant train information. All of the above information belongs to historical data. Based on the identified influencing factors as feature parameters, corresponding data is collected from the above historical data. For influencing factors not included in historical data, data are collected using the corresponding influencing factors as the primary key. Ultimately, data corresponding to all identified influencing factors are obtained.

3. The intelligent prediction method for train arrival and departure route occupancy time at railway technical stations according to claim 1 or 2, characterized in that, The processed training and test sets include: The influencing factors are divided into numerical feature influencing factors and categorical feature influencing factors; the collected data belonging to numerical feature influencing factors are normalized; and the collected data belonging to categorical feature influencing factors are one-hot feature encoded. The processed data is divided into training and testing sets according to a set ratio.

4. The intelligent prediction method for train arrival and departure route occupancy time at railway technical stations according to claim 1, characterized in that, The steps of establishing an intelligent prediction model using the training set, predicting the test set using the intelligent prediction model, and optimizing the model parameters using the prediction results include: Load the training set; Randomly select a sample set from the training set; For each sample set, a decision tree is constructed using a random feature selection method, resulting in multiple decision trees. For the multiple decision trees that have been constructed, they are combined into a random forest model as an intelligent prediction model; Use the constructed random forest model to perform classification or regression prediction on the test set and obtain the model output results; The performance of the random forest model is evaluated using the model output results, and the model parameters are adjusted based on the evaluation results.

5. The intelligent prediction method for train arrival and departure route occupancy time at railway technical stations according to claim 1, characterized in that, After obtaining the predicted occupancy time for each route, the following is also included: After the train operation is completed, the actual train operation time is recorded and compared with the predicted occupancy time of the corresponding route. The error rate is calculated and it is determined whether the error rate is within the set range. If not, the data involved in the prediction process will be updated in the training set to retrain the intelligent prediction model.

Images