WO2022178865A1

WO2022178865A1 - Method and device for monitoring and predicting traction power supply system of rail transit

Info

Publication number: WO2022178865A1
Application number: PCT/CN2021/078261
Authority: WO
Inventors: 杜峰; 吴剑强; 朱义鹏
Original assignee: 西门子股份公司; 西门子（中国）有限公司
Priority date: 2021-02-26
Filing date: 2021-02-26
Publication date: 2022-09-01
Also published as: CN116324639A

Abstract

Disclosed are a method for monitoring and predicting a traction power supply system of rail transit, comprising: determining at least one operating scene of a traction power supply system of rail transit; acquiring, on the basis of the at least one operating scene, simulated electrical data of at least one virtual scene model corresponding to the traction power supply system; and analyzing and/or optimizing the at least one virtual scene model by using the simulated electrical data, so as to monitor and predict the traction power supply system. The virtual scene model of the traction power supply system corresponding to the operating scene can be used to monitor the state of the traction power supply system in the current operating scene of the traction power supply system, simulate the situation of the traction power supply system in other operating scenes, and predict the state of the traction power supply system in other operating scenes, thereby providing operation and maintenance recommendations for operators of the rail transit.

Description

Method and device for monitoring and predicting traction power supply system of rail transit

technical field

The present disclosure relates to the technical field of rail transit, and more particularly, to a monitoring and prediction method, apparatus, computing device, computer-readable storage medium, and program product of a traction power supply system of rail transit.

Background technique

Due to the rapid development of urbanization and the gradual increase of urban population, urban rail transit (such as light rail, subway, intercity train, etc.) has gradually become a popular choice for urban residents due to its high efficiency, strong passenger carrying capacity and small impact on the environment. important means of transportation. At present, urban rail transit usually adopts electric-driven rail trains, which use electric energy as traction power and obtain electric energy from the outside of the train from the pantograph or collector shoes equipped with them. The pantograph or collector shoes are connected to the contact wires or contact rails erected along the track, and the electrical energy of the contact wires or contact rails comes from the traction substations built at certain distances along the rail transit. The capacity and station setting of the traction substation are related to many factors such as line design, train type, traffic density, train formation, and train speed. Therefore, the entire traction power supply system of rail transit includes multiple components such as trains, power supply networks, stations, and environments. With the gradual development of urban rail transit, it is necessary to operate and maintain the traction power supply system of rail transit.

The operation and maintenance of a traction power supply system generally requires three aspects to be considered. The first is safety, which is mainly reflected in: the rail potential should not be too high, otherwise it will bring personnel safety hazards, that is, personnel safety; the contact line potential must be within the safe potential range of train operation, that is, train safety; traction substation The load rate of the rectifier in the center cannot be too high, that is, the safety of the equipment. The second is energy consumption, that is, the total amount of electricity consumed by the entire rail transit line in a unit time (such as peak hours, a day and a night, or a year). For economical and environmental protection purposes, it is necessary to reduce energy consumption as much as possible. The third is the transportation capacity, that is, the total number of passengers that the entire rail transit line can transport in unit time (such as peak hours, one day and one year, or one year). These three aspects are usually mutually restrictive and therefore need to be balanced against each other.

At present, high-density distributed monitoring systems with intelligent edge devices are usually used to realize the operation and maintenance of traction power supply systems for rail transit. Specifically, in addition to the sensors that are usually arranged at the entrance and exit of the traction substation, contact lines, and parts of the track, additional sensors need to be arranged at other measurement locations of interest, so as to increase the sensor density to collect all the data. as much data as possible and analyze it to monitor the status of the traction power system.

SUMMARY OF THE INVENTION

The high-density distributed monitoring system in the prior art needs to add a large number of sensors, cables and power supplies, so the cost is high, and the harsh working environment will also affect the reliability of data communication. In addition, in reality, some measurement locations of interest are difficult to deploy sensors and cables due to practical reasons such as geographic environment, or require high costs. In addition, some factors that affect energy consumption, such as tunnel factor, track wear, etc., cannot be measured by sensors to measure the actual value, resulting in incomplete state monitoring of the traction power supply system of rail transit. More importantly, this system is only suitable for monitoring the current state of the traction power supply system of rail transit, and it is difficult to predict or simulate its operation in other operating scenarios (such as failure of a rectifier in a traction substation, train interval time, etc.). shortening, etc.).

A first embodiment of the present disclosure proposes a method for monitoring and predicting a traction power supply system of rail transit, including: determining at least one operation scenario of the traction power supply system of rail transit; simulated electrical data of the corresponding at least one virtual scene model of the traction power supply system; and analyzing and/or optimizing the at least one virtual scene model using the simulated electrical data to monitor and predict the traction power supply system.

In this embodiment, the virtual scene model of the traction power supply system corresponding to the operating scene can not only monitor the state of the traction power supply system in the current operating scene of the traction power supply system, but also simulate the situation in other operating scenarios of the traction power supply system, predicting Its status in other operating scenarios can provide operation and maintenance suggestions for rail transit operators, and find a balance in three aspects of safety, energy consumption and transportation capacity. In the embodiments of the present disclosure, there is no need to add additional sensors and wiring in the actual traction power supply system, which significantly reduces time and economic costs. In addition, by using the virtual scene model, some parameters that cannot or are inconvenient to measure in the traction power supply system can also be introduced, thus making the monitoring and prediction of the traction power supply system more comprehensive and accurate.

A second embodiment of the present disclosure proposes a monitoring and prediction device for a traction power supply system of rail transit, including: a scene determination unit configured to determine at least one operation scenario of the traction power supply system of rail transit; simulation data acquisition a unit configured to acquire simulated electrical data of the corresponding at least one virtual scene model of the traction power supply system; and an analysis and optimization unit configured to analyze the at least one virtual scene model using the simulated electrical data and/or optimization to monitor and forecast the traction power system.

A third embodiment of the present disclosure proposes a computing device comprising: a processor; and a memory for storing computer-executable instructions that, when executed, cause the processor to perform the first implementation method in the example.

A fourth embodiment of the present disclosure proposes a computer-readable storage medium having computer-executable instructions stored thereon for performing the method of the first embodiment.

A fifth embodiment of the present disclosure proposes a computer program product tangibly stored on a computer-readable storage medium and comprising computer-executable instructions that, when executed, cause at least one The processor executes the method of the first embodiment.

Description of drawings

The features, advantages and other aspects of various embodiments of the present disclosure will become more apparent when taken in conjunction with the accompanying drawings and with reference to the following detailed description, several embodiments of which are shown here by way of illustration and not limitation. , in the attached image:

1 shows a flowchart of a method for monitoring and predicting a traction power supply system of rail transit according to an embodiment of the present disclosure;

2 shows a schematic diagram of training a deep neural network model using deep reinforcement learning in the embodiment of FIG. 1;

3 shows a schematic block diagram of a monitoring and prediction system for a traction power supply system of rail transit according to an embodiment of the present disclosure;

Fig. 4 shows a schematic block diagram of the display interface of the client device in the embodiment of Fig. 3;

FIG. 5 shows a schematic block diagram of a monitoring and forecasting device for a traction power supply system of rail transit according to an embodiment of the present disclosure; and

FIG. 6 shows a schematic block diagram of a computing device for monitoring and forecasting a traction power supply system for rail transit in accordance with one embodiment of the present disclosure.

Detailed ways

Various exemplary embodiments of the present disclosure are described in detail below with reference to the accompanying drawings. Although the example methods, apparatuses described below include software and/or firmware executing on hardware among other components, it should be noted that these examples are merely illustrative and should not be regarded as limiting. For example, it is contemplated that any or all hardware, software and firmware components may be implemented exclusively in hardware, exclusively in software, or in any combination of hardware and software. Accordingly, while exemplary methods and apparatus have been described below, those skilled in the art will readily appreciate that the examples provided are not intended to limit the manner in which these methods and apparatus may be implemented.

Additionally, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of methods and systems in accordance with various embodiments of the present disclosure. It should be noted that the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may in fact be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It should also be noted that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented using dedicated hardware-based systems that perform the specified functions or operations , or can be implemented using a combination of dedicated hardware and computer instructions.

As used herein, the terms "including", "comprising" and similar terms are open-ended terms, ie, "including/including but not limited to," meaning that other content may also be included. The term "based on" is "based at least in part on." The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment" and so on.

FIG. 1 shows a monitoring and prediction method of a traction power supply system of rail transit according to an embodiment of the present disclosure. In this embodiment, the method of Figure 1 may be performed by a server in communication with the client device. The user of the client device can select one or more operating scenarios of the traction power supply system via a user interface (such as a user interface) in a web manner or through an application program, and implement the generation of a virtual scenario model for the one or more operating scenarios, Simulation, analysis and/or optimization. At the server, the corresponding function is implemented according to the user's selection of the running scenario. In another embodiment, the method of FIG. 1 may also be performed by a device that directly interacts with the user.

Referring to FIG. 1 , first, the method 100 begins with step 101 . In step 101, at least one operation scenario of the traction power supply system of the rail transit is determined. As mentioned above, the entire traction power supply system of rail transit includes multiple components such as trains, power supply networks, stations, and environments, and each component has its specific parameters or configurations. Some parameters or configurations are fixed when the rail transit line is built, such as maximum train acceleration, length, self-weight, maximum load, geographic information of each station and tunnel, number and location of traction substations, etc.; Configurations may change during operations and maintenance, such as train intervals, load factor, whether rectifiers in traction substations are working properly, etc. The operating scenario refers to the situation in which the traction power supply system operates under a set of parameters or configurations. For example, when other parameters or configurations do not change, the train interval of 90 seconds and 160 seconds are two different operating scenarios, and the 50% and 80% train passenger rates are also two different operating scenarios. Operational scenarios may be selected by a user (eg, a technician or administrator) via a user interface.

Next, in step 102, based on the at least one operating scenario, the simulated electrical data of the corresponding at least one virtual scenario model of the traction power supply system is acquired. When there are multiple operating scenarios determined in step 101, the simulated electrical data of the virtual scenario model corresponding to each of the multiple operating scenarios needs to be acquired respectively. In this embodiment, the virtual scene model created each time is stored in the database. As the virtual scene models in the database are continuously expanded and accumulated, the required virtual scene models can be searched from the database. When the virtual scene model corresponding to the determined running scene does not exist in the database, the virtual scene model needs to be generated. In addition, some virtual scene models have been simulated previously, and the simulated electrical data is saved in the database, so the simulated electrical data of these virtual scene models can be obtained directly from the database. However, when the simulated electrical data of the virtual scene model does not exist in the database, the simulation of the virtual scene model needs to be performed.

Finally, in step 103, at least one virtual scene model is analyzed and/or optimized using the simulated electrical data to monitor and predict the traction power supply system. Each virtual scenario model is capable of simulating the traction power system under a specific operating scenario. Therefore, when the operation scene corresponding to the virtual scene model is the current operation scene of the traction power supply system, the simulated electrical data can be used to monitor the current state of the traction power supply system; when the operation scene corresponding to the virtual scene model is the predicted operation of the traction power supply system In scenarios, simulated electrical data can be used to predict the future state of the traction power system. Predicted operating scenarios may include changing train interval and/or train occupancy rates, failure of a rectifier in a traction substation, etc., which are difficult to simulate in reality.

Using the virtual scene model of the traction power supply system corresponding to the operating scenario can not only monitor the state of the traction power supply system in the current operating scenario, but also simulate the situation in other operating scenarios of the traction power supply system, and predict its performance in other operating scenarios. Therefore, it can provide operation and maintenance suggestions for rail transit operators, and find a balance in the three aspects of safety, energy consumption and transportation capacity. In the embodiments of the present disclosure, there is no need to add additional sensors and wiring in the actual traction power supply system, which significantly reduces time and economic costs. In addition, by using the virtual scene model, some parameters that cannot or are inconvenient to measure in the traction power supply system can also be introduced, thus making the monitoring and prediction of the traction power supply system more comprehensive and accurate.

In one embodiment according to the present disclosure, step 102 further includes: generating at least one virtual scene model of the traction power supply system; and simulating each virtual scene model in the at least one virtual scene model to obtain each virtual scene model Simulated electrical data for the scene model. Generating the virtual scene model may be to directly establish the virtual scene model or to modify it on the basis of an associated or similar virtual scene model. Therefore, the determined operating scene can be compared with the operating scene corresponding to the virtual scene model saved in the database to determine whether there is an associated or similar virtual scene model in the database, and there is an associated or similar virtual scene in the database. Use this virtual scene model when modeling. For example, the determined running scene is a train interval of 90 seconds. When a virtual scene model with a train interval of 160 seconds and other parameters is saved in the database, the train interval of the virtual scene model is modified to 90 seconds. Can. In this way, the model generation time can be greatly shortened. However, when there is no associated or similar virtual scene model in the database, the virtual scene model corresponding to the determined running scene needs to be re-established.

In an embodiment according to the present disclosure, generating the at least one virtual scene model of the traction power supply system further includes: collecting raw data related to the at least one virtual scene model; performing data processing on the offline data and the online data as a model modeling data; and establishing at least one virtual scene model based on the modeling data. The raw data includes offline data and online data of the traction power supply system, and includes at least one of the following: power supply network parameters of the traction power supply system, train parameters, running route and geographic information, additional load parameters, and train scheduling information. Offline data includes data collected from various databases as well as data entered by a user via a user interface. The database may be, for example, a database for storing design data of the traction power supply system, a database for storing historical operating data of the traction power supply system, and the like. Online data includes data received from data acquisition equipment in the traction power supply system, such as actual voltage output values received from data acquisition equipment (eg, sensors) provided at the exit of the traction substation. Using the data received from the data acquisition equipment can bring the virtual scene model closer to the actual traction power supply system. The data input by the user may be data that cannot or is inconvenient to measure in the actual traction power supply system, such as expert experience value or theoretical calculation value. By collecting raw data including offline data and online data, the entire traction power supply system can be described more comprehensively and accurately through the virtual scene model, so as to more accurately monitor and predict the state of the traction power supply system.

As can be understood by those skilled in the art, the raw data includes all relevant data required to build a virtual scenario model for the traction power system. The power supply network includes components such as traction substations, contact lines and return rails. Therefore, power supply network parameters include but are not limited to rectifier parameters (such as short-circuit current, wire type, load loss, coupling factor, etc.), circuit breaker parameters (such as connection relationship, rated insulation voltage, rated impulse withstand voltage, etc.), and contact wires and return rail parameters (such as feed distance, wire type, wire impedance, inner diameter, outer diameter, resistivity, wear, temperature coefficient, joint type, feed point, etc.). Train parameters include but are not limited to maximum acceleration, train class, length, dead weight, rotating mass, maximum load, maximum speed, inverter parameters, motor parameters, etc. The running route and geographic information include, but are not limited to, running direction, station number and physical coordinates, marshalling arrangement, tunnel factor, route terrain information (such as gradient value), etc. Additional load parameters include, but are not limited to, vehicle-mounted equipment (such as ventilation and lighting equipment, display equipment) parameters, platform equipment (such as elevators, ventilation and lighting equipment, communication equipment) parameters, etc. The train scheduling information includes, but is not limited to, train interval time, stop time at each station, and the like. Those skilled in the art can understand that the above only lists part of the data required for establishing the virtual scene model for the traction power supply system, and they are only for the purpose of example and not limitation.

Since the raw data come from different data sources, they usually have different forms such as photos, tables, text, etc. Therefore, after collecting the raw data, it is necessary to convert these raw data with different formats into the target format, and perform processing such as data filtering as modeling data. These raw data can be processed using any data processing technique known in the art. Afterwards, at least one virtual scene model is established based on the modeling data. The established virtual scene model can be a plane model or a three-dimensional model. When the virtual scene model is a three-dimensional model, more accurate simulation results can be obtained because factors such as the influence of the air flow that the train is subjected to during running are considered in the model.

The simulation can be configured by the user via the user interface. For example, when three-dimensional and two-dimensional virtual scene models are established for a running scene at the same time, the user can select one or both of them for simulation. For another example, the user can select the simulated electrical data to be generated, such as the energy consumption of each traction substation. For another example, the simulated electrical data to be displayed may be selected by the user. In the process of simulation, according to the scheduling information of the train, the network topology of the virtual scene model at each moment is converted into an equivalent power supply model, and the simulated electrical data of the virtual scene model is obtained through power flow calculation and accumulation in time. The simulated electrical data includes, for example, the highest and lowest rail potentials as a function of distance, the highest and lowest contact line potentials as a function of distance, the current and voltage of each traction substation and the load rate of the rectifier, energy flow, and a virtual scene model in the simulation. Total energy consumption, total loss, etc. over time. It should be understood by those skilled in the art that some of the simulated electrical data listed above are for illustrative and non-limiting purposes only.

In an embodiment according to the present disclosure, step 103 further includes: for a single virtual scene model in the at least one virtual scene model, comparing its simulated electrical data with a preset threshold; Analysis of the scene model. In this embodiment, a single virtual scene model is analyzed. The preset thresholds may be industry standard data, data input by a user via a user interface, and/or actual data collected by a data collection device in the traction power supply system. The content and results of the analysis depend on the specific type of simulated electrical data and thresholds. For the convenience of description, the following lists several examples of comparing the simulated electrical data with a preset threshold, and obtaining analysis results for a single virtual scene model according to the comparison results.

Comparing the simulated electrical data with the actual data collected by the data acquisition equipment can determine whether the modeling of the virtual scene model is accurate. If the difference between the simulated electrical data and the actual data is large (for example, outside a certain threshold range), the analysis result that the virtual scene model is not accurate enough and the modeling data needs to be corrected can be obtained according to the comparison result; Otherwise, according to the comparison result, it is concluded that the virtual scene model is an accurate analysis result.

Comparing simulated electrical data with industry standard data or user input data enables state monitoring and prediction of traction power supply systems in corresponding operating scenarios. For example, when the simulated electrical data of interest include rail potential, contact line potential, and the load rate of the rectifiers in each traction substation, the preset thresholds may be user-entered expert experience values and/or industry standard values: such as rail The potential should not exceed 135V, the contact line potential should be between 1350V and 1800V, and the load rate of the rectifier in each traction substation should not exceed 80%. The rail potential, contact line potential and the load rate of the rectifier in each traction substation simulated by a single virtual scene model are compared with the above thresholds respectively. If any one of the simulated electrical data exceeds the threshold or is outside the threshold range, the analysis result of the traction power supply system in this operating scenario can be obtained according to the comparison result; otherwise, the traction power supply system can be obtained according to the comparison result Analysis results of the safe operation of the power supply system in this operating scenario. For example, in an operation scenario where a rectifier in a traction substation fails, the analysis result indicates whether there is a safety problem in the traction power supply system in this operation scenario. For another example, when the simulated electrical data concerned is the total energy consumption, the preset threshold may be an expert experience value input by the user and/or an industry standard value: for example, the target energy consumption is 110MWh. The total energy consumption simulated by a single virtual scene model is compared with the target energy consumption. If the total energy consumption exceeds the target energy consumption, the analysis result that the traction power supply system does not meet the energy consumption requirements can be obtained according to the comparison result; otherwise, according to the comparison result, the traction power supply system is in the operation Analysis results that can meet the energy consumption requirements when running in the scenario. For example, in an operation scenario with a shortened train interval, the analysis result indicates whether the total energy consumption of the traction power supply system in this operation scenario will exceed the maximum total energy consumption or the maximum planned total energy consumption that the traction substation can provide. It should be understood by those skilled in the art that the above description is for purposes of illustration and not limitation. By analyzing a single virtual scene model with simulated electrical data, it is possible to monitor or predict the state of the traction power supply system in the corresponding operating scene, thereby guiding users to make operation and maintenance decisions.

In an embodiment according to the present disclosure, step 103 further includes: for a plurality of virtual scene models in the at least one virtual scene model, according to a preset rule, analyze a plurality of virtual scene models based on the simulated electrical data of the plurality of virtual scene models Relationships between virtual scene models. In this embodiment, the relationship between a plurality of virtual scenes is analyzed. A plurality of virtual scene models may be related to a certain extent. For example, these virtual scene models differ only in train interval time and/or train occupancy rate, while other parameters are the same. The preset rules can be set according to different analysis goals. For the convenience of explanation, several examples of analyzing the relationship between the plurality of virtual scene models based on the simulated electrical data of the plurality of virtual scene models are listed below.

In some cases, it is desirable to know the effect of train interval time on the overall energy consumption of the traction power system. Comparing the total energy consumption of multiple virtual scene models that differ only in the train interval time (such as 90 seconds, 120 seconds, 160 seconds, and 180 seconds, etc.), the train interval time at which the total energy consumption will increase sharply can be determined. In some cases, it is desirable to know the effect of train occupancy on the overall energy consumption of the traction power system. Similarly, comparing the total energy consumption of multiple virtual scene models that differ only in the occupancy rate of trains (such as 50%, 60%, 70%, and 80%, etc.), it can be determined that the total energy consumption will increase sharply. Train occupancy rate. In other cases, the difference between the multiple virtual scene models may be that both the train interval time and the train occupancy rate are different.

In some cases, it is desirable to know the overall situation of different operating scenarios in terms of safety, transportation capacity and energy consumption, so as to find the best operating scenario. For each of the multiple virtual scene models to be analyzed, using the respective modeling data and simulated electrical results, calculate its comprehensive score S by the following formula:

S=f(f ₁ (U ₁ ,U ₂ ,R ₁₁ ,R ₁₂ ,R ₂₁ ,…),f ₂ (N ₁ ,P,N ₂ ),A) (1)

Among them, f ₁ (U ₁ , U ₂ , R) is used to calculate the safety. The higher the value, the higher the safety. U ₁ is the simulated rail potential, U ₂ is the simulated contact line potential, and R ₁₁ , R ₁₂ , R ₂₁ , etc. are the load rates of the rectifiers in each traction substation obtained by simulation; f ₂ (N ₁ , P, N ₂ ) is used to calculate the transportation volume, and the higher the value, the higher the transportation capacity. N ₁ is the number of people fully loaded in each vehicle, P is the load rate of the train, N ₂ is the total number of trains in the simulation time, which can be calculated by the train interval time; A represents the total energy consumption obtained by the simulation, and the higher the value, the total energy consumption higher. The higher the safety, the higher the transportation volume, and the lower the total energy consumption, the higher the comprehensive score S. The comprehensive scores S of multiple virtual scene models can be compared, so as to obtain a virtual scene model with the highest comprehensive score S. It should be understood by those skilled in the art that the above description is for purposes of illustration and not limitation. By using simulated electrical data to analyze multiple virtual scene models, it is possible to compare the status of the traction power supply system in different operating scenarios, thereby guiding users to make operation and maintenance decisions.

In an embodiment according to the present disclosure, step 103 further includes: using the trained deep neural network model to generate train running data for at least one virtual scene model, wherein the deep neural network model is trained by a deep reinforcement learning algorithm , and the simulated electrical data is used as the input of the reward function in the deep reinforcement learning algorithm to adjust the model parameters of the deep neural network model. In this embodiment, the train running data of the virtual scene model is optimized. Train operation data includes train driving patterns and train operation diagrams. The train driving mode includes a list of acceleration values for each train at different locations. The train operation diagram includes the train stop time, the number of trains, the interval time between trains, and the running direction and interval.

The following describes the process of training a deep neural network model through a deep reinforcement learning algorithm with reference to FIG. 2 . As shown in FIG. 2 , the input of the deep neural network model 201 is the initial state (eg, initial position and speed) of each train in the virtual scene model, and the output is a set of acceleration values of each train. In block 202, the position and velocity of each train at the next instant are calculated from the acceleration values of each train. The calculated position and velocity are fed back to the deep neural network model 201 and provided to the virtual scene model and reward function 205 . In block 203, the position and speed of each train in the virtual scene model are updated. In block 204, the virtual scene model is simulated based on the updated position and speed of each train, and the total energy consumption is obtained, which is provided to the reward function 205 as another input thereof. The reward function is set according to the four constraints of arrival speed, arrival time, station spacing and total energy consumption. When the train arrives at the target station at the preset time and the arrival speed is zero, the output of the reward function 205 represents a positive reward, otherwise it represents a negative reward. At the same time, the smaller the total energy consumption is, the output of the reward function 205 indicates that the positive reward is larger, otherwise it indicates that the positive reward is smaller. The reward function 205 provides its output to the deep neural network model 201 . The deep neural network model 201 adjusts its model parameters according to the output of the reward function 205 . Next, the deep neural network model 201 outputs the next set of acceleration values for each train according to the input positions and speeds of each train. In block 202, the position and velocity of each train at the next moment are calculated again from the acceleration values of each train. During the training process, the above process is performed cyclically and repeatedly until the deep neural network model 201 converges. After the deep neural network model 201 is trained, it can be used to generate a list of acceleration values of each train at different positions. From this list of acceleration values, train running data can be generated for the virtual scene model. In other embodiments, the list of acceleration values may also be used for automatic driving control of the train. By using the total simulated energy consumption as an input to the reward function and adjusting the model parameters of the deep neural network model, the required energy consumption can be minimized, resulting in significant cost savings.

FIG. 3 shows a schematic block diagram of a monitoring and prediction system for a traction power supply system of rail transit according to an embodiment of the present disclosure. As shown in FIG. 3, the system 300 includes an application program 31 installed on a client device and monitoring and forecasting software 32 installed on a server. The users of the rail transit operator (such as technicians or managers) use the application 31 to monitor and forecast the traction power supply system of the rail transit, so as to realize the operation and maintenance of the traction power supply system. The user's input is received or information is displayed to the user via the user interface of the client device. FIG. 4 shows a schematic block diagram of the display interface of the client device in the embodiment of FIG. 3 . As shown in FIG. 4 , five option buttons of modeling 401 , simulation 402 , analysis 403 , optimization 404 , and management 405 are shown on the display interface 400 . Some or all of the above option buttons may be displayed according to different personnel roles. For example, all the above-mentioned option buttons are displayed for technical personnel, while only three option buttons for analysis 403, optimization 404 and management 405 are displayed for management personnel. In FIG. 3 , the application 31 includes a user interaction module 310 and a result display module 311 . When the user selects an option button on the display interface 400, the user interaction module 310 prompts the user via the display interface 400 to determine the running scenario, and when selecting certain options, it is also necessary to input data (such as the expert experience value that needs to be input during modeling). , simulation configuration to be entered during simulation, expert experience value and/or industry standard value to be entered during analysis, etc.). The user interaction module 310 receives user input information and sends the user input information to the monitoring and prediction software 32 . The monitoring and prediction software 32 implements corresponding functions according to the received input information. The monitoring and prediction software 32 includes a scenario determination module 320 , a data query module 321 , a model generation module 322 , a model simulation module 323 , a model analysis module 324 and a model optimization module 325 .

The actions performed by the monitoring and prediction software 32 are described below through a specific application scenario. For a completed rail transit line, the user expects to know the train interval that achieves the optimal level in terms of safety, transportation capacity and energy consumption under the condition that other parameters remain unchanged. Therefore, it is necessary to simulate the operation of the traction power supply system under different train interval times and determine the optimal train interval time. First, the user selects the option button of analysis 403 via the display interface 400, and selects or inputs three operation scenarios of the traction power supply system of the rail transit line: the train interval is 90 seconds, 120 seconds and 160 seconds respectively. The scenario determination module 320 in the monitoring and prediction software 32 determines the above-mentioned operating scenario according to the received input information. The data query module 321 searches the database for whether there is a corresponding virtual scene model and its simulated electrical data according to the determined operation scene. When it does not exist, the model generation module 322 needs to generate a corresponding virtual scene model. As mentioned above, the virtual scene model corresponding to the above-mentioned three operating scenes can be directly established, or it can be modified on the basis of the associated or similar virtual scene model. If the virtual scene model is directly established, the model generation module 322 collects relevant original data from different data sources, processes them as modeling data, and establishes a virtual scene model based on the modeling data. After generating the models, the model simulation module 323 simulates the above three virtual scene models respectively, and obtains simulated electrical data such as rail potential, contact line potential, load rate of rectifiers in each traction substation, and total energy consumption. After that, the model analysis module 324 uses the modeling data of the three virtual scene models and the simulated electrical data to calculate the comprehensive scores S ₁ , S ₂ and S ₃ of the three virtual scene models according to the above formula (1). , and send the train interval (eg, 120 seconds) of the virtual scene model with the highest comprehensive score to the result display module 311 as the recommended interval. The result display module 311 displays the recommended interval time to the user via the display interface 400 . It should be understood by those skilled in the art that the above description is for purposes of example and not limitation, and that the monitoring and prediction system 300 may be used in many other application scenarios. In some other application scenarios, the model optimization module 325 generates train running data for a virtual scenario model corresponding to a specific running scenario through the trained deep neural network model. The deep neural network model is trained by a deep reinforcement learning algorithm as described earlier. During the training process, the train acceleration value output by the deep neural network model is used to update the train state in the virtual scene model and the input of the deep neural network model itself, and the total energy consumption obtained from the simulation of the virtual scene model is used as the reward function to adjust the model parameters of the deep neural network model.

In the above-mentioned embodiment, the virtual scene model of the traction power supply system corresponding to the operating scene can not only monitor the state of the traction power supply system in the current operating scene of the traction power supply system, but also simulate the situation in other operating scenarios of the traction power supply system, predicting Its status in other operating scenarios can provide operation and maintenance suggestions for rail transit operators, and find a balance in three aspects of safety, energy consumption and transportation capacity. In the embodiments of the present disclosure, there is no need to add additional sensors and wiring in the actual traction power supply system, which significantly reduces time and economic costs. In addition, by using the virtual scene model, some parameters that cannot or are inconvenient to measure in the traction power supply system can also be introduced, thus making the monitoring and prediction of the traction power supply system more comprehensive and accurate.

FIG. 5 shows a monitoring and forecasting device for a traction power supply system of rail transit according to an embodiment of the present disclosure. Each unit in FIG. 5 can be implemented by software, hardware (eg, integrated circuit, FPGA, etc.), or a combination of software and hardware. Referring to FIG. 5 , the apparatus 500 includes a scene determination unit 501 , a simulation data acquisition unit 502 and an analysis and optimization unit 503 . The scenario determination unit 501 is configured to determine at least one operation scenario of the traction power supply system of the rail transit. The simulation data acquisition unit 502 is configured to acquire simulated electrical data of the corresponding at least one virtual scene model of the traction power supply system. The analysis and optimization unit 503 is configured to analyze and/or optimize the at least one virtual scene model using the simulated electrical data for monitoring and prediction of the traction power supply system.

Optionally, in an embodiment according to the present disclosure, the simulation data acquisition unit 502 further includes a model generation unit and a model simulation unit (not shown in FIG. 5 ). The model generation unit is configured to generate at least one virtual scene model of the traction power supply system. The model simulation unit is configured to simulate each of the at least one virtual scene models to obtain simulated electrical data of each virtual scene model.

Optionally, in an embodiment according to the present disclosure, the model generation unit further includes a data collection unit, a data processing unit and a model establishment unit (not shown in FIG. 5 ). The data collection unit is configured to collect raw data related to the at least one virtual scene model. The data processing unit is configured to perform data processing on the raw data as modeling data. The model generation unit is configured to build at least one virtual scene model based on the modeling data.

Optionally, in an embodiment according to the present disclosure, the original data includes offline data and online data of the traction power supply system, and includes at least one of the following items: power supply network parameters of the traction power supply system, train parameters , operating route and geographic information, additional load parameters, and train scheduling information.

Optionally, in an embodiment according to the present disclosure, the analysis and optimization unit 503 is further configured to: for a single virtual scene model in the at least one virtual scene model, compare its simulated electrical data with a preset threshold ; and an analysis of a single virtual scene model based on the comparison results.

Optionally, in an embodiment according to the present disclosure, the analysis and optimization unit 503 is further configured to: for a plurality of virtual scene models in the at least one virtual scene model, according to a preset rule, based on the plurality of virtual scene models The simulated electrical data of the virtual scene model analyzes the relationship between the plurality of virtual scene models.

Optionally, in one embodiment in accordance with the present disclosure, the simulated electrical data includes at least one of the following: rail potential, contact line potential, load factor of rectifiers in each traction substation, and total energy consumption.

Optionally, in an embodiment according to the present disclosure, the analysis and optimization unit 503 is further configured to: use the trained deep neural network model to generate train running data for at least one virtual scene model, wherein the deep neural network The model is trained by the deep reinforcement learning algorithm, and the simulated electrical data is used as the input of the reward function in the deep reinforcement learning algorithm to adjust the model parameters of the deep neural network model.

6 shows a schematic block diagram of a computing device for monitoring and forecasting of traction power systems for rail transit, according to one embodiment of the present disclosure. As can be seen in FIG. 6 , a computing device 601 for operating and maintaining a traction power system for rail transit includes a central processing unit (CPU) 601 (eg, a processor) and a memory 602 coupled to the central processing unit (CPU) 601 . The memory 602 is used to store computer-executable instructions, which, when executed, cause a central processing unit (CPU) 601 to execute the methods in the above embodiments. A central processing unit (CPU) 601 and a memory 602 are connected to each other through a bus to which an input/output (I/O) interface is also connected. The computing device 601 may also include a number of components (not shown in FIG. 6 ) connected to the I/O interface, including but not limited to: an input unit, such as a keyboard, mouse, etc.; an output unit, such as various types of displays, speakers etc.; storage units, such as magnetic disks, optical discs, etc.; and communication units, such as network cards, modems, wireless communication transceivers, and the like. The communication unit allows the computing device 601 to exchange information/data with other devices through a computer network such as the Internet and/or various telecommunication networks.

Also, alternatively, the above-described method can be implemented by a computer-readable storage medium. A computer-readable storage medium carries computer-readable program instructions for carrying out various embodiments of the present disclosure. A computer-readable storage medium may be a tangible device that can hold and store instructions for use by the instruction execution device. The computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (non-exhaustive list) of computer readable storage media include: portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM) or flash memory), static random access memory (SRAM), portable compact disk read only memory (CD-ROM), digital versatile disk (DVD), memory sticks, floppy disks, mechanically coded devices, such as printers with instructions stored thereon Hole cards or raised structures in grooves, and any suitable combination of the above. Computer-readable storage media, as used herein, are not to be construed as transient signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (eg, light pulses through fiber optic cables), or through electrical wires transmitted electrical signals.

Accordingly, in another embodiment, the present disclosure proposes a computer-readable storage medium having computer-executable instructions stored thereon for performing the functions of the present disclosure. The method of various embodiments.

In another embodiment, the present disclosure proposes a computer program product tangibly stored on a computer-readable storage medium and comprising computer-executable instructions, which when executed At least one processor is caused to perform the methods of various embodiments of the present disclosure.

In general, the various example embodiments of the present disclosure may be implemented in hardware or special purpose circuits, software, firmware, logic, or any combination thereof. Certain aspects may be implemented in hardware, while other aspects may be implemented in firmware or software that may be executed by a controller, microprocessor or other computing device. While aspects of the embodiments of the present disclosure are illustrated or described as block diagrams, flowcharts, or using some other graphical representation, it is to be understood that the blocks, apparatus, systems, techniques, or methods described herein may be taken as non-limiting Examples of are implemented in hardware, software, firmware, special purpose circuits or logic, general purpose hardware or controllers or other computing devices, or some combination thereof.

Computer-readable program instructions or computer program products for executing various embodiments of the present disclosure can also be stored in the cloud, and when invoking is required, users can access the data stored in the cloud through the mobile Internet, fixed network or other network. The computer-readable program instructions of one embodiment of the present disclosure are executed, thereby implementing the technical solutions disclosed in accordance with various embodiments of the present disclosure.

Although embodiments of the present disclosure have been described with reference to several specific embodiments, it should be understood that embodiments of the present disclosure are not limited to the specific embodiments of the disclosure. The embodiments of the present disclosure are intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims. The scope of the claims is to be accorded the broadest interpretation so as to encompass all such modifications and equivalent structures and functions.

Claims

Monitoring and forecasting methods for traction power supply systems of rail transit, including:

determining at least one operation scenario of the traction power supply system of the rail transit;

Based on the at least one operating scenario, obtain simulated electrical data of the corresponding at least one virtual scenario model of the traction power supply system; and

The at least one virtual scenario model is analyzed and/or optimized using the simulated electrical data for monitoring and prediction of the traction power supply system.
The method according to claim 1, wherein acquiring the simulated electrical data of the corresponding at least one virtual scene model of the traction power supply system further comprises:

generating the at least one virtual scene model of the traction power system; and

Each virtual scene model in the at least one virtual scene model is simulated to obtain simulated electrical data of each virtual scene model.
The method of claim 2, wherein generating the at least one virtual scene model of the traction power system further comprises:

collecting raw data related to the at least one virtual scene model;

performing data processing on the raw data as modeling data; and

The at least one virtual scene model is established based on the modeling data.
The method according to claim 3, wherein the raw data includes offline data and online data of the traction power supply system, and includes at least one of the following: power supply network parameters of the traction power supply system, train parameters, operating route and geographic information, additional load parameters, and train scheduling information.
The method of claim 1, wherein analyzing and/or optimizing the at least one virtual scene model using the simulated electrical data further comprises:

for a single virtual scene model of the at least one virtual scene model, comparing its simulated electrical data to a preset threshold; and

The single virtual scene model is analyzed according to the comparison result.
The method of claim 1, wherein analyzing and/or optimizing the at least one virtual scene model using the simulated electrical data further comprises:

For a plurality of virtual scene models in the at least one virtual scene model, according to a preset rule, the relationship between the plurality of virtual scene models is analyzed based on the simulated electrical data of the plurality of virtual scene models.
6. The method of claim 5 or 6, wherein the simulated electrical data includes at least one of the following: rail potential, contact line potential, duty ratio of rectifiers in each traction substation, and total energy consumption.
The method of claim 1, wherein analyzing and/or optimizing the at least one virtual scene model using the simulated electrical data further comprises:

generating train operation data for the at least one virtual scene model using a trained deep neural network model, wherein the deep neural network model is trained by a deep reinforcement learning algorithm, and using the simulated electrical data as the depth The input of the reward function in the reinforcement learning algorithm is used to adjust the model parameters of the deep neural network model.
Monitoring and forecasting devices for traction power supply systems of rail transit, including:

a scenario determination unit configured to determine at least one operating scenario of the traction power supply system of the rail transit;

a simulation data acquisition unit configured to acquire simulated electrical data of the corresponding at least one virtual scene model of the traction power supply system; and

An analysis and optimization unit configured to analyze and/or optimize the at least one virtual scene model using the simulated electrical data to monitor and predict the traction power supply system.
The apparatus according to claim 9, wherein the simulation data acquisition unit further comprises:

a model generation unit configured to generate the at least one virtual scene model of the traction power system; and

A model simulation unit configured to simulate each of the at least one virtual scene models to obtain simulated electrical data of each of the virtual scene models.
The apparatus of claim 10, wherein the model generating unit further comprises:

a data collection unit configured to collect raw data related to the at least one virtual scene model;

a data processing unit configured to perform data processing on the raw data as modeling data; and

A model building unit configured to build the at least one virtual scene model based on the modeling data.
The apparatus of claim 11, wherein the raw data includes offline data and online data of the traction power supply system, and includes at least one of the following: power supply network parameters of the traction power supply system, train parameters, operating route and geographic information, additional load parameters, and train scheduling information.
The apparatus of claim 9, wherein the analysis optimization unit is further configured to:

for a single virtual scene model of the at least one virtual scene model, comparing its simulated electrical data to a preset threshold; and

The single virtual scene model is analyzed according to the comparison result.
The apparatus of claim 9, wherein the analysis optimization unit is further configured to:

For a plurality of virtual scene models in the at least one virtual scene model, according to a preset rule, the relationship between the plurality of virtual scene models is analyzed based on the simulated electrical data of the plurality of virtual scene models.
14. The apparatus of claim 13 or 14, wherein the simulated electrical data includes at least one of the following: rail potential, contact line potential, load factor of rectifiers in each traction substation, and total energy consumption.
The apparatus of claim 9, wherein the analysis optimization unit is further configured to:

generating train operation data for the at least one virtual scene model using a trained deep neural network model, wherein the deep neural network model is trained by a deep reinforcement learning algorithm, and using the simulated electrical data as the depth The input of the reward function in the reinforcement learning algorithm is used to adjust the model parameters of the deep neural network model.
Computing equipment, including:

processor; and

a memory for storing computer-executable instructions which, when executed, cause the processor to perform the method of any of claims 1-8.
A computer-readable storage medium having computer-executable instructions stored thereon for performing the method of any of claims 1-8.
A computer program product tangibly stored on a computer-readable storage medium and comprising computer-executable instructions which, when executed, cause at least one processor to perform the execution according to claims 1-8 The method of any of the above.