A train timetable replanning method and system considering grid-side power matching

Abstract

This invention belongs to the field of train delay recovery technology, specifically disclosing a train timetable replanning method and system considering grid-side power matching. The method includes: extending the stopping time of subsequent trains based on train delay information and determining the delay impact range; obtaining the power location information of trains within the delay impact range and performing AC / DC power flow calculations to obtain the original power of the main substation; calculating the allocated power of the main substation according to the allocation coefficient table; performing power matching between the original power and the allocated power, using the sum of the power matching values ​​and the safe following penalty value between the train and the preceding train as the reward function, minimizing the total incoming energy of the main substation as the optimization objective, and using a reinforcement learning algorithm to adjust train running time and stopping time to update the timetable. This invention solves the problems of existing technologies, such as a single optimization objective, limitation to global planning, failure to directly consider the power consumption of the substation, and potential neglect of power loss during transmission, leading to calculation errors.

Description

Technical Field

[0001] This invention belongs to the field of train delay recovery technology, specifically relating to a train timetable replanning method and system that considers grid-side power matching. Background Technology

[0002] In train traction systems, although the overhead contact line voltage is high, the train's traction power is extremely large, resulting in a very high current flowing through the line. This leads to significant power losses during transmission, even with low line resistance. Furthermore, during periods of high train density, the total traction power and power losses are substantial. Because current technologies typically simplify the traction power supply network as an ideal source capable of unconditionally meeting any power demand, they neglect the inherent energy dissipation processes. This disregard for power losses during transmission causes fundamental deviations in optimization models when pursuing energy-saving goals.

[0003] In terms of solution strategies, delay events typically have spatiotemporal limitations; their propagation follows certain physical laws, and they only have a substantial impact on subsequent trains within a certain time and space range behind the fault point. Replanning trains that are far away and unaffected is a huge waste of resources. Furthermore, once global planning is initiated, the entire complex calculation must be completed, lacking the agility to handle dynamic changes.

[0004] To address the aforementioned issues, one existing technology studies how to establish a mathematical model based on route timetables and passenger flow data after train delays occur, adjusting train arrival and departure times and travel times to minimize the impact of delays on the entire train fleet. Simultaneously, another method takes minimizing the change in substation energy consumption during the adjustment process as the optimization objective, considering the constraints of passenger flow factors on station dwell time and the absorption of regenerative braking energy, ensuring energy conservation while addressing delays. However, existing technologies still have the following problems: 1) The calculation may be incorrect because it does not directly take into account the power consumption of the substation and may ignore the power loss during transmission.

[0005] 2) Single optimization objective. Existing technologies mostly aim to minimize total delay time or energy consumption, which cannot effectively achieve efficient and economical operation of the traction power supply system during delayed recovery.

[0006] 3) Limited to global planning. Existing technologies replan all trains when performing delay recovery, which may lead to low computational efficiency, wasted computing resources, or secondary disturbances. Summary of the Invention

[0007] The purpose of this invention is to address the problems of existing technologies, such as a single optimization objective, limitation to global planning, failure to directly consider the power consumption of substations, potential neglect of power loss during transmission, and resulting calculation errors. This invention proposes a train timetable replanning method and system that considers grid-side power matching, and plans within the scope of delay impact based on the minimum safe following interval, thus avoiding global planning.

[0008] The technical solution of the present invention is as follows: Firstly, a train timetable replanning method considering grid-side power matching, comprising the following steps: Acquire train delay information, real-time train status information, train timetable, and constraint information; real-time train status information includes a power allocation coefficient table for each traction substation; constraint information includes the minimum safe tracking interval. Based on train delay information, real-time train status information, and train schedules, the stopping time of following trains is extended to meet the minimum safe tracking interval. Determine whether the next train meets the minimum safe following interval with the train following it. If it does, determine the scope of the delay. If it does not, define the next train as the new following train and extend the stopping time of the new following train to meet the minimum safe following interval. The power location information of trains within the delay impact area is obtained to perform AC / DC power flow calculations, and the operating status of other trains is generated based on the train timetable. Based on the operating status of the other trains, AC / DC power flow calculations are performed on the other trains to obtain the original power of the main substation; Based on the allocation coefficient table, calculate the power that needs to be adjusted for trains to be allocated to each traction substation within the power supply section, and then add the power of the traction substation to the corresponding main substation to obtain the allocated power of the main substation. The original power and allocated power of the main substation are matched to obtain the power matching value. The sum of the power matching values ​​and the safety tracking penalty value between the train and the preceding train are used as the reward function. The optimization objective is to minimize the total incoming line energy of the main substation. The reinforcement learning algorithm is used to adjust the train running time and stopping time and update the timetable. Determine whether the next train meets the minimum safe departure interval after the timetable is updated. If it does, generate the adjusted timetable for trains within the affected range. If not, recalculate the original power and allocated power of the main substation and iteratively adjust the timetable for the next affected train.

[0009] Preferably, train delay information includes the delayed train number, the location of the delay, the current delay time, and a preliminary estimate of the time required to restore normal operation. The real-time train status information also includes the original timetables of delayed trains and trains following them, as well as the precise locations of delayed trains and trains following them. The constraint information also includes stop time constraints, interval running time constraints, and catch-up point constraints.

[0010] As a preferred option, the minimum safe tracking interval is:

[0011] in, This indicates the time when the current vehicle arrives at position s. This indicates the time when the preceding vehicle arrives at position s. This indicates the minimum safe tracking interval.

[0012] As a preferred method, AC / DC power flow calculation specifically includes the following steps: Based on the current power and location information of the train, construct the node admittance matrix of the DC system and initialize the node voltage; The node current is calculated based on the train power and traction substation status. The node voltage is iteratively updated using the DC system node admittance matrix until the voltage error meets the preset convergence accuracy. The working state of the traction substation and the node admittance matrix of the DC system are corrected based on the updated node voltage. If the traction substation state does not converge, the node voltage is iteratively updated until the voltage error meets the preset convergence accuracy, and the final node voltage and the node admittance matrix of the DC system are obtained. Based on the final node voltage and DC system node admittance matrix, the branch currents on the grid side and rail side are calculated, and the traction substation current is calculated using KCL law. The voltage, current and DC power of the traction substation are output to complete the DC power flow solution. The DC power of the traction substation is converted to the AC power, AC power flow is calculated, input information under the current simulation step size is obtained, initial power values ​​are set, AC system power equations are written, and Jacobian matrix is ​​constructed. The AC power deviation is calculated based on the Jacobian matrix. If the AC power deviation does not meet the preset convergence accuracy, the AC system power equation is corrected and the AC power deviation is recalculated. If the convergence accuracy requirement is met, the node voltage and current of the AC system are output and the power of the main substation is calculated.

[0013] As a preferred option, the constraints of the reinforcement learning algorithm include catch-up point constraints, interval running time adjustment constraints, and stop time adjustment constraints. The catch-up point constraint is:

[0014] in, This represents the arrival time of the nth train. This represents the original scheduled arrival time of train n. Indicates the time deviation threshold for on-time arrival; The interval running time adjustment constraint is:

[0015] in, This represents the lower bound of the running time for the m-th interval. This represents the upper bound of the running time of the m-th interval. This represents the travel time of the nth train in the m-th interval. Represents the set of non-negative integers; The constraints for adjusting stop time are:

[0016] in, Indicates the first Minimum stopping time at each station, Indicates the first Maximum stopping time at each station This indicates that the nth train is on the [missing information]. Stop time at each station.

[0017] As a preferred approach, the state space of reinforcement learning algorithms... for:

[0018] in, Indicates the current station of the train. Indicates the current delay time; Action space of reinforcement learning algorithms for:

[0019] in, This indicates the amount of time adjustment for the interval. This indicates the amount of time the station stops are adjusted.

[0020] As a preferred approach, the objective function of the reinforcement learning algorithm is:

[0021] in, This indicates the total energy required for the train to travel on the line. Indicates the total number of trains. Let x represent the total simulation runtime, and let x represent the set of adjustments made to all delayed trains. Indicates in Time of the first The incoming power of each main substation; The specific formula for the reward function of reinforcement learning algorithms is as follows:

[0022]

[0023] in, Indicates the main substation The power matching value, Indicates the main substation The allocated power value, This indicates that the main substation does not include the AC / DC power flow calculations obtained from the decision-making train. The power value, This indicates the time when the current vehicle arrives at position s. This indicates the time when the preceding vehicle arrives at position s. Indicates the minimum safe tracking interval. This represents the weighting coefficient of the power matching reward item. This represents the function indicating the safe tracking interval. This represents the original power value of main substation i. If the current state is the last decision stage, a catch-up point reward value is set. ,in, This represents the train's original planned arrival time at the terminal station. This represents the function indicating on-time arrival at the destination station. This indicates the actual arrival time of the train at the final station. This indicates the time error threshold for arrival at the final destination. The state transition function of the reinforcement learning algorithm is:

[0024] in, Represents the next state space. Indicates the time of delay in the next state space. This represents the state transition probability, which is either 1 or 0. This represents the state transition function.

[0025] The beneficial effects of this invention are: This invention considers power matching on the grid side and an adaptive adjustment strategy to achieve safe, efficient, and energy-saving delay recovery. First, this invention uses the power matching degree of the main substation as the optimization target, taking into account the traction power, auxiliary power, and regenerative braking power of the trains. It avoids ignoring power losses during transmission, greatly improving the accuracy of optimization. Second, this invention uses a local adaptive adjustment strategy to replan the timetables only for trains directly affected by delays, avoiding unnecessary waste of computational resources, reducing computational costs, and enhancing the system's dynamic response capability. The method proposed in this invention can effectively curb the impact of delays while achieving energy saving and safety.

[0026] Secondly, a train timetable replanning system considering grid-side power matching includes: The first module is used to acquire train delay information, real-time train status information, train timetable and constraint information; the real-time train status information includes the power allocation coefficient table for each traction station; the constraint information includes the minimum safe tracking interval. The second module is used to extend the stopping time of the following train based on train delay information, real-time train status information and train timetable to meet the minimum safe tracking interval. The third module is used to determine whether the next train meets the minimum safe following interval with the train behind it. If it does, the scope of the delay is determined; if it does not, the next train is defined as the new following train, and the stopping time of the new following train is extended to meet the minimum safe following interval. The fourth module is used to obtain the power location information of trains within the delay impact range, perform AC / DC power flow calculations, and generate the operating status of other trains based on the train timetable. The fifth module is used to perform AC / DC power flow calculations on the remaining trains based on their operating status, and to obtain the original power of the main substation. The sixth module is used to calculate the power that needs to be adjusted for trains to be allocated to each traction substation within the power supply section according to the allocation coefficient table, and to add the power of the traction substation to the corresponding main substation to obtain the allocated power of the main substation. The seventh module is used to match the original power and the allocated power of the main substation to obtain the power matching value. The sum of the power matching values ​​and the safety tracking penalty value between the train and the preceding train are used as the reward function. The optimization objective is to minimize the total incoming line energy of the main substation. The reinforcement learning algorithm is used to adjust the train running time and stopping time and update the timetable. The eighth module is used to determine whether the next train meets the minimum safe departure interval after the timetable is updated. If it does, it generates the adjusted timetable for trains within the affected range. If it does not, it recalculates the original power and allocated power of the main substation and iteratively adjusts the timetable for the next affected train.

[0027] Thirdly, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method as described in the first aspect.

[0028] Fourthly, a non-transitory computer-readable storage medium is provided that stores computer instructions for causing a computer to perform the method as described in the first aspect. Attached Figure Description

[0029] Figure 1 The diagram shows a flowchart of a train timetable replanning method that considers grid-side power matching.

[0030] Figure 2 The diagram shown is a schematic of the equivalent circuit model. Detailed Implementation

[0031] Exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be understood that the embodiments shown and described in the drawings are merely exemplary and are intended to illustrate the principles and spirit of the invention, and are not intended to limit the scope of the invention.

[0032] Example 1: like Figure 1 As shown, a train timetable replanning method considering grid-side power matching includes the following steps: S1. Obtain train delay information, real-time train status information, train timetable, and constraint information; real-time train status information includes a power allocation coefficient table for each traction station; constraint information includes the minimum safe tracking interval. S2. Based on train delay information, real-time train status information, and train timetable, extend the stopping time of the following train to meet the minimum safe tracking interval; S3. Determine whether the next train meets the minimum safe following interval with the train behind it. If it does, determine the scope of the delay. If it does not, define the next train as the new following train and extend the stopping time of the new following train to meet the minimum safe following interval. Specifically, taking a train with 5 trains as an example, based on the delay information of the first train, the real-time status information of the train, and the train timetable, the stopping time of the second train is extended to meet the minimum safe following interval; it is determined whether the third train meets the minimum safe following interval with the fourth train. If it does, the scope of the delay impact is determined; if it does not, the third train is defined as a new following train, and the stopping time of the third train is extended to meet the minimum safe following interval.

[0033] S4. Obtain the power location information of trains within the delay impact range, perform AC / DC power flow calculations, and generate the operating status of other trains based on the train timetable; S5. Based on the operating status of the other trains, perform AC / DC power flow calculations on the other trains to obtain the original power of the main substation; S6. Based on the allocation coefficient table, calculate the power that needs to be adjusted for trains to be allocated to each traction substation within the power supply section, and add the power of the traction substation to the corresponding main substation to obtain the allocated power of the main substation; S7. Match the original power and allocated power of the main substation to obtain the power matching value. Use the sum of the power matching values ​​and the safety tracking penalty value between the train and the preceding train as the reward function. With minimizing the total incoming line energy of the main substation as the optimization objective, use reinforcement learning algorithm to adjust the train running time and stopping time and update the timetable. S8. Determine whether the next train meets the minimum safe departure interval after updating the timetable. If it does, generate the adjusted timetable for trains within the affected range. If it does not, recalculate the original power and allocated power of the main substation and iteratively adjust the timetable for the next affected train.

[0034] In this embodiment, the train delay information includes the delayed train number, the location of the delay, the current delay time, and a preliminary estimate of the time required to restore normal operation. The real-time train status information also includes the original timetables of delayed trains and trains following them, as well as the precise locations of delayed trains and trains following them. The constraint information also includes stop time constraints, interval running time constraints, and catch-up point constraints.

[0035] In this embodiment, the power allocation coefficient table is obtained as follows: The traction substation is equivalent to a voltage source with a series resistor, and the train is equivalent to a power load. After simplifying the line impedance model, the equivalent circuit model is obtained as follows: Figure 2 As shown. Based on the equivalent circuit model, we obtain:

[0036]

[0037] In the formula, Indicates the current matching the train. These represent the current on the left and the current on the right, respectively. This indicates the proportion of the matched train's location within the power supply zone. The equivalent resistance of the interval is... These represent the equivalent resistance on the left and the equivalent resistance on the right, respectively. These represent the left and right allocation coefficients of the train, respectively.

[0038] In this embodiment, the urban rail train group is subject to a safe following interval constraint during operation; the safe following interval between trains cannot exceed the minimum safe following interval.

[0039] In the formula, Let be the time when the current vehicle arrives at position s. Let be the time when the preceding vehicle arrives at position s. This is the minimum safe tracking interval.

[0040] In this embodiment, the power location information of trains within the delay impact range includes the algebraic sum of positive traction power, positive auxiliary power, and negative regenerative braking power.

[0041] In this embodiment, the AC / DC power flow calculation specifically includes the following steps: Step S01: Based on the power and location information of the train at the current moment, construct the DC system node admittance matrix and initialize the node voltage; Step S02: Calculate the node current based on the train power and traction substation status, and iteratively update the node voltage using the DC system node admittance matrix until the voltage error meets the preset convergence accuracy. Step S03: Correct the working state of the traction substation and the node admittance matrix of the DC system according to the updated node voltage. If the traction substation state does not converge, continue to iterate and update the node voltage until the voltage error meets the preset convergence accuracy, and obtain the final node voltage and the node admittance matrix of the DC system. Step S04: Based on the final node voltage and DC system node admittance matrix, calculate the branch currents on the grid side and rail side, and calculate the traction substation current using KCL law. Output the voltage, current and DC side power of the traction substation to complete the DC power flow calculation. Step S05: Convert the DC power of the traction substation to the AC power, perform AC power flow calculation, obtain the input information under the current simulation step size, set the initial power value, write the AC system power equation, and construct the Jacobian matrix; Specifically, for any node i, its injected power is expressed by voltage and current as follows:

[0042] Among them, S i P represents the complex power of the node. i Q represents active power. i This indicates reactive power.

[0043] According to Kirchhoff's Current Law (KCL), the relationship between the node injection current and the connected branch current is obtained, and expressed by the admittance matrix Y as follows:

[0044] Substituting the current equation into the power equation:

[0045] Among them, V i This represents the voltage phasor of node i.

[0046] Expressing voltage and admittance in polar coordinates, the final power flow equations are derived:

[0047] in, This represents the voltage phase difference between two points. This represents the real part of the admittance, i.e., the conductance; This represents the imaginary part of admittance, i.e., susceptance.

[0048] The AC power flow calculation is performed using the Newton-Raphson method. The power deviation at the PQ node is:

[0049] Among them, P i and Q i These are the active power and reactive power injected by the PQ node, respectively. This represents the voltage phase angle at node i. Let represent the voltage phase angle at node j. Based on the above derivation, the corrected equation for AC power flow calculation can be obtained as follows:

[0050] Step S06: Calculate the AC power deviation based on the Jacobian matrix. If the AC power deviation does not meet the preset convergence accuracy, then correct the AC system power equation and recalculate the AC power deviation. If the convergence accuracy requirement is met, then output the node voltage and current measured by AC and calculate the power of the main substation.

[0051] In this embodiment, a reinforcement learning algorithm is used to adjust train running time and station stopping time. The constraints when updating the timetable include: Catch-up point constraint: If a train arrives on time after receiving interference a certain number of stations and subsequently resumes operation according to the original timetable, the catch-up point constraint is denoted as:

[0052] In the formula, Let n be the arrival time of train n. Let be the original planned arrival time of train n. The threshold for time deviation of arrival on time.

[0053] Section running time adjustment constraints: Considering train characteristics and line speed limits, the time cannot be lower than the minimum section running time. At the same time, to quickly restore the timetable from disturbances, excessively long running times should be avoided. Therefore, upper and lower bounds are set for section running time adjustment:

[0054] In the formula, This is the lower bound of the running time for the m-th interval. This is the upper bound of the running time for the m-th interval. Let be the travel time of the nth train in the mth section.

[0055] Stop time adjustment constraints: To meet passenger demand and disturbance recovery requirements, upper and lower bounds are set for stop time adjustment:

[0056] In the formula, Let m be the minimum stopping time at the m-th station. Let m be the maximum dwell time at the m-th station. Let be the stopping time of the nth train at the mth station.

[0057] The state space of the reinforcement learning algorithm is:

[0058] In the formula, Indicates the current station of the train. Indicates the current delay time (in seconds).

[0059] The action space of the reinforcement learning algorithm is:

[0060] In the formula, This is the adjustment amount for the interval running time. This is the amount of time to adjust the stop time.

[0061] The objective function of the reinforcement learning algorithm is to minimize the total energy consumption of the traction substation. In the timetable replanning model, the optimization objective is to minimize the total incoming line energy of the main substation. The decision variables are the interval travel time and stop time of the train during its operation between several stations after a delay, and the corresponding formula is:

[0062] in, This indicates the total energy required for the train to travel on the line. Indicates the total number of trains. Let x represent the total simulation runtime, and let x represent the set of adjustments made to all delayed trains. Indicates in Time of the first The incoming power of each main substation; The reward function of the reinforcement learning algorithm is as follows: after the action is executed, the reward value is calculated by summing the power matching values ​​of the train during the time period from departure from the current station to departure from the next station, and the safe following penalty value with the preceding train. The corresponding formula is: The specific formula for the reward function of reinforcement learning algorithms is as follows:

[0063]

[0064] Furthermore, if the current state is the final decision-making stage, there will be an additional catch-up point bonus:

[0065] If the catch-up point constraint is satisfied at the current moment after the decision, the reward value is positive; otherwise, the reward value is negative.

[0066] in, Indicates the main substation The power matching value, Indicates the main substation The allocated power value, This indicates that the main substation does not include the AC / DC power flow calculations obtained from the decision-making train. The power value, This indicates the time when the current vehicle arrives at position s. This indicates the time when the preceding vehicle arrives at position s. Indicates the minimum safe tracking interval. This represents the weighting coefficient of the power matching reward item. This represents the function indicating the safe tracking interval. This represents the original power value of main substation i. If the current state is the last decision stage, a catch-up point reward value is set. ,in, This represents the train's original planned arrival time at the terminal station. This represents the function indicating on-time arrival at the destination station. This indicates the actual arrival time of the train at the final station. This represents the time error threshold for arrival at the final destination.

[0067] Each action triggers a transition to the next state. The state transition function of the reinforcement learning algorithm is:

[0068] in, Represents the next state space. Indicates the time of delay in the next state space. This represents the state transition probability, which is either 1 or 0. Represents the state transition function. This is the adjustment amount for the interval running time. This is the amount of time to adjust the stop time.

[0069] Example 2: Based on Embodiment 1, this embodiment of the invention provides a train timetable replanning system considering grid-side power matching, which can be used to implement the train timetable replanning method considering grid-side power matching as described in the foregoing embodiments. The system includes: The first module is used to acquire train delay information, real-time train status information, train timetable and constraint information; the real-time train status information includes the power allocation coefficient table for each traction station; the constraint information includes the minimum safe tracking interval. The second module is used to extend the stopping time of the following train based on train delay information, real-time train status information and train timetable to meet the minimum safe tracking interval. The third module is used to determine whether the next train meets the minimum safe following interval with the train behind it. If it does, the scope of the delay is determined; if it does not, the next train is defined as the new following train, and the stopping time of the new following train is extended to meet the minimum safe following interval. The fourth module is used to obtain the power location information of trains within the delay impact range, perform AC / DC power flow calculations, and generate the operating status of other trains based on the train timetable. The fifth module is used to perform AC / DC power flow calculations on the remaining trains based on their operating status, and to obtain the original power of the main substation. The sixth module is used to calculate the power that needs to be adjusted for trains to be allocated to each traction substation within the power supply section according to the allocation coefficient table, and to add the power of the traction substation to the corresponding main substation to obtain the allocated power of the main substation. The seventh module is used to match the original power and the allocated power of the main substation to obtain the power matching value. The sum of the power matching values ​​and the safety tracking penalty value between the train and the preceding train are used as the reward function. The optimization objective is to minimize the total incoming line energy of the main substation. The reinforcement learning algorithm is used to adjust the train running time and stopping time and update the timetable. The eighth module is used to determine whether the next train meets the minimum safe departure interval after the timetable is updated. If it does, it generates the adjusted timetable for trains within the affected range. If it does not, it recalculates the original power and allocated power of the main substation and iteratively adjusts the timetable for the next affected train.

[0070] According to embodiments of the present invention, the present invention also provides an electronic device, a readable storage medium, and a computer program product.

[0071] In an exemplary embodiment, the electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the train timetable replanning method considering grid-side power matching as described in Embodiment 1 above.

[0072] In an exemplary embodiment, the readable storage medium may be a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the train timetable replanning method considering grid-side power matching as described in Embodiment 1 above.

[0073] In an exemplary embodiment, the computer program product includes a computer program that, when executed by a processor, implements the train timetable replanning method considering grid-side power matching as described in Embodiment 1 above.

[0074] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0075] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0076] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0077] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0078] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0079] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A train timetable replanning method considering grid-side power matching, characterized in that, Includes the following steps: Acquire train delay information, real-time train status information, train timetable, and constraint information; real-time train status information includes a power allocation coefficient table for each traction substation; constraint information includes the minimum safe tracking interval. Based on train delay information, real-time train status information, and train schedules, the stopping time of following trains is extended to meet the minimum safe tracking interval. Determine whether the next train meets the minimum safe following interval with the train following it. If it does, determine the scope of the delay. If it does not, define the next train as the new following train and extend the stopping time of the new following train to meet the minimum safe following interval. The power location information of trains within the delay impact area is obtained to perform AC / DC power flow calculations, and the operating status of other trains is generated based on the train timetable. Based on the operating status of the other trains, AC / DC power flow calculations are performed on the other trains to obtain the original power of the main substation; Based on the allocation coefficient table, calculate the power that needs to be adjusted for trains to be allocated to each traction substation within the power supply section, and then add the power of the traction substation to the corresponding main substation to obtain the allocated power of the main substation. The original power and allocated power of the main substation are matched to obtain the power matching value. The sum of the power matching values ​​and the safety tracking penalty value between the train and the preceding train are used as the reward function. The optimization objective is to minimize the total incoming line energy of the main substation. The reinforcement learning algorithm is used to adjust the train running time and stopping time and update the timetable. Determine whether the following trains meet the minimum safe departure interval after the timetable is updated. If so, generate the adjusted timetable for trains within the affected area. If the conditions are not met, the original power and allocated power of the main substation are recalculated, and the timetable of the next affected train is iteratively adjusted.

2. The train timetable replanning method considering grid-side power matching according to claim 1, characterized in that, Train delay information includes the delayed train number, the location of the delay, the current delay time, and a preliminary estimate of the time required to restore normal operation. The real-time train status information also includes the original timetables of delayed trains and trains following them, as well as the precise locations of delayed trains and trains following them. The constraint information also includes stop time constraints, interval running time constraints, and catch-up point constraints.

3. The train timetable replanning method considering grid-side power matching according to claim 1, characterized in that, The minimum safe tracking interval is: in, This indicates the time when the current vehicle arrives at position s. This indicates the time when the preceding vehicle arrives at position s. This indicates the minimum safe tracking interval.

4. The train timetable replanning method considering grid-side power matching according to claim 1, characterized in that, AC / DC power flow calculation specifically includes the following steps: Based on the current power and location information of the train, construct the node admittance matrix of the DC system and initialize the node voltage; The node current is calculated based on the train power and traction substation status. The node voltage is iteratively updated using the DC system node admittance matrix until the voltage error meets the preset convergence accuracy. The working state of the traction substation and the node admittance matrix of the DC system are corrected based on the updated node voltage. If the traction substation state does not converge, the node voltage is iteratively updated until the voltage error meets the preset convergence accuracy, and the final node voltage and the node admittance matrix of the DC system are obtained. Based on the final node voltage and DC system node admittance matrix, the branch currents on the grid side and rail side are calculated, and the traction substation current is calculated using KCL law. The voltage, current and DC power of the traction substation are output to complete the DC power flow solution. The DC power of the traction substation is converted to the AC power, AC power flow is calculated, input information under the current simulation step size is obtained, initial power values ​​are set, AC system power equations are written, and Jacobian matrix is ​​constructed. The AC power deviation is calculated based on the Jacobian matrix. If the AC power deviation does not meet the preset convergence accuracy, the AC system power equation is corrected and the AC power deviation is recalculated. If the convergence accuracy requirement is met, the node voltage and current of the AC system are output and the power of the main substation is calculated.

5. The train timetable replanning method considering grid-side power matching according to claim 1, characterized in that, The constraints of reinforcement learning algorithms include catch-up point constraints, interval running time adjustment constraints, and stop time adjustment constraints. The catch-up point constraint is: in, This represents the arrival time of the nth train. This represents the original scheduled arrival time of train n. Indicates the time deviation threshold for on-time arrival; The interval running time adjustment constraint is: in, This represents the lower bound of the running time for the m-th interval. This represents the upper bound of the running time of the m-th interval. This represents the travel time of the nth train in the m-th interval. Represents the set of non-negative integers; The constraints for adjusting stop time are: in, Indicates the first Minimum stopping time at each station, Indicates the first Maximum stopping time at each station This indicates that the nth train is on the [missing information]. Stop time at each station.

6. The train timetable replanning method considering grid-side power matching according to claim 1, characterized in that, State space of reinforcement learning algorithms for: in, Indicates the current station of the train. Indicates the current delay time; Action space of reinforcement learning algorithms for: in, This indicates the amount of time adjustment for the interval. This indicates the amount of time the station stops are adjusted.

7. The train timetable replanning method considering grid-side power matching according to claim 6, characterized in that, The objective function of the reinforcement learning algorithm is: in, This indicates the total energy required for the train to travel on the line. Indicates the total number of trains. Let x represent the total simulation runtime, and let x represent the set of adjustments made to all delayed trains. Indicates in Time of the first The incoming power of each main substation; The specific formula for the reward function of reinforcement learning algorithms is as follows: in, Indicates the main substation The power matching value, Indicates the main substation The allocated power value, This indicates that the main substation does not include the AC / DC power flow calculations obtained from the decision-making train. The power value, This indicates the time when the current vehicle arrives at position s. This indicates the time when the preceding vehicle arrives at position s. Indicates the minimum safe tracking interval. This represents the weighting coefficient of the power matching reward item. This represents the function indicating the safe tracking interval. This represents the original power value of main substation i. If the current state is the last decision stage, a catch-up point reward value is set. ,in, This represents the train's original planned arrival time at the terminal station. This represents the function indicating on-time arrival at the destination station. This indicates the actual arrival time of the train at the final station. This indicates the time error threshold for arrival at the final destination. The state transition function of the reinforcement learning algorithm is: in, Represents the next state space. Indicates the time of delay in the next state space. This represents the state transition probability, which is either 1 or 0. This represents the state transition function.

8. A train timetable replanning system considering grid-side power matching, characterized in that, include: The first module is used to acquire train delay information, real-time train status information, train timetable, and constraint information. Real-time train status information includes a power allocation coefficient table for each traction substation; constraint information includes the minimum safe tracking interval. The second module is used to extend the stopping time of the following train based on train delay information, real-time train status information and train timetable to meet the minimum safe tracking interval. The third module is used to determine whether the next train meets the minimum safe following interval with the train behind it. If it does, the scope of the delay is determined; if it does not, the next train is defined as the new following train, and the stopping time of the new following train is extended to meet the minimum safe following interval. The fourth module is used to obtain the power location information of trains within the delay impact range, perform AC / DC power flow calculations, and generate the operating status of other trains based on the train timetable. The fifth module is used to perform AC / DC power flow calculations on the remaining trains based on their operating status, and to obtain the original power of the main substation. The sixth module is used to calculate the power that needs to be adjusted for trains to be allocated to each traction substation within the power supply section according to the allocation coefficient table, and to add the power of the traction substation to the corresponding main substation to obtain the allocated power of the main substation. The seventh module is used to match the original power and the allocated power of the main substation to obtain the power matching value. The sum of the power matching values ​​and the safety tracking penalty value between the train and the preceding train are used as the reward function. The optimization objective is to minimize the total incoming line energy of the main substation. The reinforcement learning algorithm is used to adjust the train running time and stopping time and update the timetable. The eighth module is used to determine whether the following trains meet the minimum safe departure interval after the timetable is updated. If so, it generates the adjusted timetable for trains within the affected area. If the conditions are not met, the original power and allocated power of the main substation are recalculated, and the timetable of the next affected train is iteratively adjusted.

9. An electronic device, characterized in that, include: At least one processor; And a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to cause the at least one processor to perform the method according to any one of claims 1-7.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-7.

Images