Energy management method, device and system for connecting new energy to traction power supply system

Abstract

The application discloses a kind of energy management method, device and system of new energy access traction power supply system, belong to traction power supply system field, for reducing train operation energy consumption and improving the utilization of new energy generation, solve the problem that train operation energy consumption is high and new energy generation utilization is low;Through target control sequence, traction energy consumption minimization and regenerative braking energy recovery maximization can be realized simultaneously, equivalent to reduce train traction power consumption while making full use of regenerative braking energy, so that train overall operation energy consumption can be reduced;At the same time, according to new energy output prediction data to determine the upper limit of traction power and take it as the constraint condition of dynamic optimization model, so that the available power supply capacity of train traction power can adapt to new energy generation system, when new energy output is high, make full use of new energy generation, when new energy output is low, avoid traction power supply system power imbalance, improve new energy generation utilization and the power balance of traction power supply system.

Description

Technical Field

[0001] This invention relates to the field of traction power supply systems, and in particular to an energy management method, apparatus, and system for integrating new energy sources into a traction power supply system. Background Technology

[0002] In scenarios where new energy sources are integrated into the traction power supply system, optimizing train operation energy consumption is one of the core tasks of the energy management system. Train operation energy consumption is related to many factors, and there is a lack of mature energy management methods for integrating new energy sources into the traction power supply system. In current technologies, train operation energy consumption is high, the utilization rate of new energy power generation is low, and the power balance of the traction power supply system is poor.

[0003] Therefore, how to provide a solution to the above-mentioned technical problems is a problem that needs to be solved by those skilled in the art. Summary of the Invention

[0004] The purpose of this invention is to provide an energy management method, device, and system for integrating new energy sources into the traction power supply system. Since the target control sequence can simultaneously minimize traction energy consumption and maximize regenerative braking energy recovery, it is equivalent to fully utilizing regenerative braking energy while reducing train traction power consumption, thereby reducing the overall energy consumption of the train operation. Furthermore, by determining the upper limit of traction power based on new energy output prediction data and using it as a constraint condition for the dynamic optimization model, the train traction power can adapt to the available power supply capacity of the new energy power generation system. This allows for full utilization of new energy power generation when new energy output is high and avoids power imbalance in the traction power supply system when new energy output is low, thereby improving the utilization rate of new energy power generation and the power balance of the traction power supply system.

[0005] To address the aforementioned technical problems, this invention provides an energy management method for integrating new energy sources into a traction power supply system, applicable to the energy management system of a traction power supply system, comprising: The train's real-time operating parameters are obtained, including at least one of the following: train position, train speed, train acceleration, traction and braking status, and the gradient of the track ahead. Based on the new energy power output forecast data, the upper limit of the train's traction power in the future forecast time domain is determined; the new energy power output forecast data includes the forecast power generation of the new energy power generation system in the future preset time period. Based on a preset dynamic optimization model of train traction and braking sequence, with the objectives of minimizing traction energy consumption and maximizing regenerative braking energy recovery, a target control sequence that satisfies the constraints is solved according to the real-time operating parameters of the train. The target control sequence includes the traction power and braking power of each control step in the future prediction time domain; the constraints include the upper limit of the traction power. The traction and braking power of the train are controlled according to the target control sequence.

[0006] On the other hand, before solving for the target control sequence that satisfies the constraints, the energy management method for integrating new energy sources into the traction power supply system further includes: Obtain train power flow scheduling instructions, which include regenerative braking energy utilization priority, target zonal load level, and constraint time window; Based on the priority of regenerative braking energy utilization, the upper limit of braking power in the constraint conditions is adjusted within the constraint time window. The priority of regenerative braking energy utilization and the upper limit of braking power in the constraint conditions are positively correlated. Based on the target zone load level, within the constraint time window, the upper limit of traction power in the constraint conditions of the target zone is adjusted, and the target zone load level and the upper limit of traction power in the constraint conditions are positively correlated.

[0007] On the other hand, controlling the traction power and braking power of the train according to the target control sequence includes: An energy-saving coasting strategy is generated based on the target control sequence. The energy-saving coasting strategy includes the coasting start point mileage, the coasting duration, and the coasting exit condition. An adaptive velocity curve is generated based on the target control sequence, the adaptive velocity curve including the target velocity sequence and the timestamp sequence; The energy-saving coasting strategy and the adaptive speed curve are sent to the train for execution.

[0008] On the other hand, generating the energy-saving lazy driving strategy based on the target control sequence includes: When the absolute value of the train's current acceleration is less than the preset acceleration threshold and the gradient of the track ahead is downhill or flat, the conditions for coasting are met, and the train enters coasting mode. Calculate the optimal coasting time based on the predicted gradient of the route ahead and the constraints of on-time operation. When the train speed drops to the lower speed limit or reaches the speed limit point or signal point ahead, it exits the coasting mode. Among them, speed limit points are specific locations on the line where trains are required to reduce their operating speed; signal points are locations where pass signals are installed.

[0009] On the other hand, generating the adaptive velocity curve based on the target control sequence includes: Substitute the traction power and braking power of each control step in the target control sequence into the train dynamics model, and obtain the speed trajectory in the time domain based on the integration of the current train speed. The velocity trajectory in the time domain is converted into a velocity-distance curve; The speed-distance curve is compared point by point with the line speed limit curve, and the smaller value is taken to obtain the speed curve after the speed limit constraint. The acceleration of the velocity curve after the speed limit constraint is smoothed to obtain an adaptive velocity curve.

[0010] On the other hand, the preset dynamic optimization model of the train traction and braking sequence is: a dynamic optimization model of the train traction and braking sequence constructed based on model predictive control. The dynamic optimization model of the train traction and braking sequence includes a train dynamics model; The state vector of the train dynamics model is: ; in, x This represents the state vector of the train dynamics model. s For the train's position, v For train speed, E The remaining kinetic energy, T This is the transpose of the matrix.

[0011] On the other hand, the process of acquiring the new energy output prediction data includes: Collect real-time meteorological data, wherein the real-time meteorological data includes at least one of irradiance, wind speed and temperature; The real-time meteorological data is input into a pre-trained new energy output prediction model to determine the new energy output prediction data; The new energy output prediction model is a prediction model constructed based on a Bayesian optimized long short-term memory network. The Bayesian optimization is used to automatically tune the hyperparameters of the long short-term memory network.

[0012] On the other hand, the energy management method for connecting new energy sources to the traction power supply system also includes: Based on reinforcement learning algorithms, power dispatch instructions are determined according to the state of charge of the energy storage system, the predicted output data of the new energy source, and the predicted train load data. The power dispatch instructions include at least one of the energy storage charging and discharging power and the proportion of new energy source grid connection.

[0013] On the other hand, the dynamic optimization model based on the preset train traction and braking sequence aims to minimize traction energy consumption and maximize regenerative braking energy recovery. It solves for the target control sequence that satisfies the constraints based on the real-time train operating parameters, and employs a rolling optimization method using model predictive control, including: In any control cycle, a state prediction in the prediction time domain is constructed based on the current train state, and the optimal control sequence that satisfies the constraints is solved in the prediction time domain. The optimal control sequence includes the traction power and braking power of each control step in the prediction time domain. Execute the traction power and braking power corresponding to the first control step in the optimal control sequence.

[0014] On the other hand, the constraints include energy recovery constraints and power constraints; The energy recovery constraint is: ; The power constraint is: 0≤ P traction ≤ P tractionmax ; 0≤ P brake ≤ P brakemax ; in, η Energy recovery efficiency for regenerative braking energy E potential For potential recyclable braking energy, E recovered For the actual braking energy recovered, P traction For traction power, P tractionmax This is the upper limit of traction power. P brake For braking power, P brakemax This is the upper limit of braking power.

[0015] On the other hand, determining the upper limit of the train's traction power in the future forecast time domain based on the new energy output forecast data includes: Based on the new energy output forecast data, the predicted power generation of the new energy power generation system in the future forecast time domain is determined; The sum of the predicted power generation of the new energy power generation system in the future prediction time domain, the available discharge power of the energy storage system in the future prediction time domain, and the available power supply of the power grid in the future prediction time domain is used as the estimated value of the available power supply in the future prediction time domain. Obtain the estimated value of other load power of the traction power supply system in the future prediction time domain, wherein the other load power is the load power of the traction power supply system other than the train traction load; The upper limit of traction power in the constraints is determined based on the difference between the estimated available power supply and the estimated power of other loads.

[0016] On the other hand, after controlling the traction power and braking power of the train according to the target control sequence, the energy management method for connecting new energy sources to the traction power supply system further includes: Based on the target control sequence and the train dynamics model, the predicted energy consumption in the prediction time domain is determined; Obtain the actual energy consumption of the train; Based on the deviation between the actual energy consumption and the predicted energy consumption, the preset adjustable parameters in the dynamic optimization model are adjusted; the preset adjustable parameters include train mass and drag coefficient.

[0017] On the other hand, adjusting the preset adjustable parameters in the dynamic optimization model based on the deviation between the actual energy consumption and the predicted energy consumption includes: When the deviation between the actual energy consumption and the predicted energy consumption is greater than a preset deviation threshold, power and speed data over the past N control cycles are acquired. The power and speed data includes the traction power, braking power, and corresponding actual speed for each control step. Using the traction power and braking power of each control step in the power and speed data as inputs and the actual speed corresponding to the control step as the observation, the estimated values ​​of the preset adjustable parameters in the dynamic optimization model are determined through the rolling time-domain estimation model. The values ​​of the preset adjustable parameters in the dynamic optimization model are updated to the estimated values.

[0018] To address the aforementioned technical problems, this invention also provides an energy management device for integrating new energy sources into a traction power supply system, applied to the energy management system of the traction power supply system, comprising: The acquisition module is used to acquire real-time train operating parameters, which include at least one of the following: train position, train speed, train acceleration, traction and braking status, and the gradient of the track ahead. The determination module is used to determine the upper limit of the train's traction power in the future prediction time domain based on the new energy output prediction data; the new energy output prediction data includes the predicted power generation of the new energy power generation system in the future preset time period. The solution module is used to solve for the target control sequence that meets the constraints based on the preset dynamic optimization model of train traction and braking sequence, with the objectives of minimizing traction energy consumption and maximizing regenerative braking energy recovery. The target control sequence includes the traction power and braking power of each control step in the future prediction time domain; the constraints include the upper limit of the traction power. The execution module is used to control the traction power and braking power of the train according to the target control sequence.

[0019] To address the aforementioned technical problems, the present invention also provides an energy management system for integrating new energy sources into a traction power supply system, comprising: Memory, used to store computer programs; A processor is used to execute the computer program to implement the steps of the energy management method for accessing the traction power supply system with new energy sources as described above.

[0020] Beneficial Effects: This invention provides an energy management method for integrating new energy sources into a traction power supply system. Considering that traction energy consumption and regenerative braking energy recovery simultaneously affect train operation energy consumption, and by solving the control sequence through a dynamic optimization model, the method can minimize train traction energy consumption and maximize regenerative braking energy recovery. This invention first obtains real-time train operating parameters. Based on a preset dynamic optimization model of the train traction and braking sequence, with the objectives of minimizing traction energy consumption and maximizing regenerative braking energy recovery, it solves for a target control sequence that satisfies the constraints based on the real-time train operating parameters. This target control sequence includes the traction power and braking power for each control step in the future prediction time domain. Finally, based on this target control sequence... The traction and braking power of the train are controlled. Since the target control sequence can simultaneously minimize traction energy consumption and maximize regenerative braking energy recovery, it is equivalent to making full use of regenerative braking energy while reducing the train's traction power consumption, thereby reducing the overall operating energy consumption of the train. At the same time, since the upper limit of traction power is determined based on the predicted data of renewable energy output and used as a constraint condition for the dynamic optimization model, the train's traction power can adapt to the available power supply capacity of the renewable energy power generation system. When the renewable energy output is high, renewable energy power generation is fully utilized, and when the renewable energy output is low, the power imbalance of the traction power supply system is avoided, thereby improving the utilization rate of renewable energy power generation and the power balance of the traction power supply system.

[0021] The present invention also provides an energy management device and system for new energy access to traction power supply system, which has the same beneficial effects as the energy management method for new energy access to traction power supply system described above. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention, the relevant technologies and the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 A schematic diagram of the first process of the energy management method for the new energy access traction power supply system provided by the present invention; Figure 2 A schematic diagram of the second process of the energy management method for accessing a new energy traction power supply system provided by the present invention; Figure 3 A schematic diagram of the structure of the energy management device for the new energy access traction power supply system provided by the present invention; Figure 4A schematic diagram of the energy management system for the new energy access traction power supply system provided by the present invention. Detailed Implementation

[0024] The core of this invention is to provide an energy management method, device, and system for integrating new energy sources into the traction power supply system. Since the target control sequence can simultaneously minimize traction energy consumption and maximize regenerative braking energy recovery, it is equivalent to fully utilizing regenerative braking energy while reducing train traction power consumption, thereby reducing the overall energy consumption of the train. Furthermore, by determining the upper limit of traction power based on new energy output prediction data and using it as a constraint condition for the dynamic optimization model, the train traction power can adapt to the available power supply capacity of the new energy power generation system. This allows for full utilization of new energy power generation when new energy output is high and avoids power imbalance in the traction power supply system when new energy output is low, thereby improving the utilization rate of new energy power generation and the power balance of the traction power supply system.

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Please refer to Figure 1 , Figure 1 This is a first flowchart illustrating the energy management method for a new energy source integration traction power supply system provided by the present invention. This energy management method for a new energy source integration traction power supply system is applied to the energy management system of the traction power supply system and includes: S101: Obtain real-time train operation parameters, which include at least one of the following: train position, train speed, train acceleration, traction and braking status, and the gradient of the track ahead. Specifically, considering that the energy consumption of train operation in the traction power supply system is closely related to the spatiotemporal state of the train and the track conditions, the primary prerequisite for optimizing the control of the train's traction and braking power is to accurately obtain the real-time status information of the train during operation. Therefore, in this embodiment of the invention, the real-time operating parameters of the train are first obtained. In this embodiment of the invention, the real-time operating parameters of the train specifically include the following five categories of information: Train position: Used to determine the current power supply zone where the train is located and the track conditions ahead, it is a basic parameter for judging the train's operating environment; Train speed: Reflects the current speed of the train and is a core input for estimating traction / braking power demand and speed curve planning; Train acceleration: including traction acceleration and braking deceleration, reflects the rate of change of the train's motion state and is an important basis for determining whether the train is in the acceleration, deceleration or constant speed phase. Traction and braking status: This indicates the discrete operating state of the train, whether it is currently traction, braking, coasting, or stopped, and carries information on the current traction and braking power. The gradient of the track ahead reflects the relative elevation change of the section of track ahead of the train. It is a key track condition that determines the magnitude of the train's running resistance and thus affects traction / braking energy consumption.

[0027] Specifically, to ensure consistent data standards, the definitions, precision, and value ranges of the real-time train operation parameters in this embodiment of the invention are explained as follows: Train position: expressed as track mileage, unit m, accuracy ±10m, used to determine the power supply zone where the train is located and the track conditions ahead; Train speed: train operating speed, unit km / h, accuracy ±1km / h, range 0~350km / h; Train acceleration: traction / braking acceleration, unit m / s², accuracy ±0.01m / s², range -1.5~1.0m / s²; Traction and braking status: discrete status, taking one of traction, braking, coasting, and stopping, and carrying traction power (unit kW, accuracy ±50kW) and braking power (unit kW, accuracy ±50kW); Track gradient ahead: gradient value, unit ‰, accuracy ±1‰, range -30‰~+30‰, sampling interval 100m.

[0028] By acquiring the above five types of parameters in a coordinated manner, a complete and real-time data foundation can be provided for subsequent solving of the target control sequence based on the dynamic optimization model.

[0029] S102: Based on the new energy power output forecast data, determine the upper limit of the train's traction power in the future forecast time domain; the new energy power output forecast data includes the forecast power generation of the new energy power generation system in the future preset time period; Specifically, after obtaining the real-time operating parameters of the train, this embodiment of the invention determines the upper limit of the train's traction power in the future prediction time domain based on the new energy power output prediction data. The new energy power output prediction data includes the predicted power generation of the new energy power generation system within a preset future time period; the specific acquisition process can be found in the supplementary embodiment below. By pre-determining the upper limit of traction power, a constraint basis that conforms to the actual power supply capacity of the new energy power generation system can be provided for subsequent solving of the target control sequence based on the dynamic optimization model, enabling the optimization results of the train's traction power to adapt to changes in new energy power output.

[0030] S103: Based on the preset train traction and braking sequence, the dynamic optimization model aims to minimize traction energy consumption and maximize regenerative braking energy recovery. According to the real-time train operation parameters, the target control sequence that meets the constraints is solved. The target control sequence includes the traction power and braking power of each control step in the future prediction time domain. Specifically, after obtaining the real-time operating parameters of the train, this embodiment of the invention comprehensively plans the traction and braking actions of the train over a future period based on a preset dynamic optimization model of the train traction and braking sequence. This dynamic optimization model has dual optimization objectives: minimizing traction energy consumption and maximizing regenerative braking energy recovery. It seeks the optimal traction and braking scheme while considering the inherent contradiction between these two objectives. During the solution process, the model must simultaneously satisfy multiple constraints on train operation (such as speed constraints, acceleration constraints, power constraints, time constraints, and energy recovery constraints, as detailed in the supplementary embodiments below) to ensure that the obtained control sequence is engineering-feasible and within the safety and comfort range. The obtained target control sequence covers all control steps within a future prediction time domain. Each control step in the sequence corresponds to a set of decision values ​​for traction power and braking power, thus constituting a time-progressing traction and braking action sequence oriented towards energy saving and energy recovery.

[0031] S104: Control the traction and braking power of the train according to the target control sequence.

[0032] Specifically, after obtaining the target control sequence based on the dynamic optimization model, closed-loop control can be performed on the traction and braking outputs during actual train operation based on the traction and braking power given by each control step in the sequence, so that the train operates according to the optimized traction and braking rhythm. Since this target control sequence simultaneously minimizes traction energy consumption and maximizes regenerative braking energy recovery, it effectively reduces the overall operating energy consumption of the train while fully utilizing regenerative braking energy, thereby reducing the train's operating costs.

[0033] Specifically, the Energy Management System (EMS) can establish communication with the Automatic Train Operation (ATO), for example, by using IEC 61375-1 (the standard for train communication networks) or proprietary protocols, thereby sending the target control sequence to the ATO, which then controls the train's traction and braking power.

[0034] This invention provides an energy management method for integrating new energy sources into a traction power supply system. Considering that traction energy consumption and regenerative braking energy recovery simultaneously affect train operating energy consumption, and by solving the control sequence through a dynamic optimization model, the method can minimize train traction energy consumption and maximize regenerative braking energy recovery. This invention first obtains real-time train operating parameters. Based on a preset dynamic optimization model of the train traction and braking sequence, with the objectives of minimizing traction energy consumption and maximizing regenerative braking energy recovery, it solves for a target control sequence that satisfies the constraints based on the real-time train operating parameters. This target control sequence includes the traction power and braking power for each control step in the future prediction time domain. Finally, the method adjusts the train operation according to this target control sequence. The traction power and braking power of the train are controlled. Since the traction energy consumption can be minimized and the regenerative braking energy recovery can be maximized simultaneously through the target control sequence, it is equivalent to making full use of regenerative braking energy while reducing the train's traction power consumption, thereby reducing the overall operating energy consumption of the train. At the same time, since the upper limit of traction power is determined based on the predicted data of renewable energy output and used as a constraint condition for the dynamic optimization model, the train's traction power can adapt to the available power supply capacity of the renewable energy power generation system. When the renewable energy output is high, renewable energy power generation is fully utilized, and when the renewable energy output is low, the power imbalance of the traction power supply system is avoided, thereby improving the utilization rate of renewable energy power generation and the power balance of the traction power supply system.

[0035] Based on the above embodiments: As an optional embodiment, before solving for the target control sequence that satisfies the constraints, the energy management method for integrating new energy sources into the traction power supply system further includes: S201: Obtain train power flow scheduling instructions, which include regenerative braking energy utilization priority, target zonal load level, and constraint time window; Specifically, for a better explanation of the embodiments of the present invention, please refer to... Figure 2 , Figure 2 This is a second flowchart illustrating the energy management method for integrating new energy into the traction power supply system provided by the present invention. Considering that the load levels and new energy consumption demands of each power supply zone in the traction power supply system dynamically change with the operating mode, the energy management system needs to obtain train-level power flow scheduling instructions from the upper layer or regional coordination layer to transmit the energy scheduling intent of the entire network to the train traction and braking optimization stage. In this embodiment of the invention, the power flow scheduling instruction includes three key components: first, the priority of regenerative braking energy utilization, used to indicate whether regenerative braking energy should be prioritized or restricted within the current constraint time window; second, the target zone load level, used to indicate the current load tension of a specific power supply zone; and third, the constraint time window, defining the time interval for the above scheduling intent to take effect in minutes. By obtaining the above power flow scheduling instructions, the traction and braking optimization on the train side and the power flow scheduling on the system side can be coordinated.

[0036] S202: Based on the priority of regenerative braking energy utilization, adjust the upper limit of braking power in the constraint conditions within the constraint time window. The priority of regenerative braking energy utilization and the upper limit of braking power in the constraint conditions are positively correlated. Specifically, this embodiment of the invention establishes a positive correlation between the priority of regenerative braking energy utilization and the upper limit of braking power: when the priority of regenerative braking energy utilization is high, it indicates that the system has a high demand for recovering regenerative braking energy. In this case, the value of the upper limit of braking power in the constraint conditions is increased accordingly within the constraint time window, freeing up more space for the generation and recovery of regenerative braking energy; conversely, when the priority is low, the upper limit of braking power is decreased accordingly. Thus, the higher the priority of regenerative braking energy utilization, the larger the allowed upper limit of braking power, and the two change positively, thereby transforming the system-side recovery intention into a quantifiable adjustment at the model constraint level.

[0037] S203: Based on the target zone load level, within the constraint time window, adjust the upper limit of traction power in the constraint conditions of the target zone. The target zone load level and the upper limit of traction power in the constraint conditions are positively correlated.

[0038] Specifically, similarly, embodiments of the present invention establish a positive correlation between the target zone load level and the upper limit of traction power: within the constraint time window, the upper limit of traction power in the constraint conditions of the target zone is adjusted in the same direction according to the load level of the target zone; the higher the load level of the target zone, the higher the upper limit of traction power, and the two show a positive correlation. This adjustment ensures that the train traction flow collection capacity is consistent with the load level set for the target zone, thereby achieving refined control of the zone load and avoiding the impact of train traction flow collection on key zones.

[0039] It should be noted that the above-mentioned adjustments to the upper limits of braking power and traction power are effective within the constraint time window. After the window ends, the relevant constraints can be restored to the default benchmark, thus taking into account both the overall scheduling intention and the local flexibility of train operation.

[0040] Specifically, the details of the power flow scheduling instructions are as follows: Instruction content and format: The instruction content includes the priority of regenerative braking energy utilization, target zonal load level, and constraint time window (unit: minutes), encapsulated in JSON format, including instruction ID, timestamp, and target parameter list; Reception method: Subscribe to receive via IEC60870-5-104 protocol, with an update cycle of 10 seconds; Model embedding: Power flow scheduling commands are incorporated as constraints into the MPC (Model Predictive Control) optimization model. Specifically, if the command requires increased utilization of regenerative braking energy, the upper limit of braking power is increased. P brakemax The weighting; if the instruction requires reducing the load of a certain zone, then the traction power is reduced when the train is in that zone. P tractionmax And prioritize using lazy driving; The energy-saving coasting command specifically includes the coasting start point (mileage), coasting duration, and coasting exit conditions; the adaptive speed curve specifically includes the target speed sequence (unit km / h, accuracy ±1 km / h) and the timestamp sequence; both can be sent to the onboard ATP (Automatic Train Protection) or ATO equipment through the IEC 61375-1 train communication network, and the ATO will execute speed control.

[0041] As an optional embodiment, the traction power and braking power of the train are controlled according to the target control sequence, including: S301: Generate an energy-saving coasting strategy based on the target control sequence. The energy-saving coasting strategy includes the coasting start point mileage, coasting duration, and coasting exit conditions. Specifically, to further reduce energy consumption in sections where train traction work is not required, this embodiment of the invention generates an energy-saving coasting strategy based on the energy-saving intent inherent in the target control sequence. This strategy explicitly defines three elements: the coasting start point mileage, i.e., the starting point from which the train enters coasting; the coasting duration, i.e., the duration of maintaining the coasting state; and the coasting exit condition, i.e., the condition for terminating coasting and resuming controlled traction or braking. Through the energy-saving coasting strategy, coasting can replace some traction operations while meeting punctuality requirements, reducing unnecessary traction energy consumption.

[0042] S302: Generate an adaptive velocity curve based on the target control sequence. The adaptive velocity curve includes the target velocity sequence and the timestamp sequence. Specifically, in this embodiment of the invention, an adaptive speed curve is generated based on the target control sequence. The curve is expressed in the form of a target speed sequence and a timestamp sequence: the target speed sequence gives the operating speed that the train should reach at each time point in the future, and the timestamp sequence corresponds to the time corresponding to each speed value in the speed sequence. The two correspond one-to-one, forming a speed planning curve that adaptively adjusts with time and track conditions.

[0043] S303: Send the energy-saving coasting strategy and adaptive speed curve to the train for execution.

[0044] Specifically, after generating the energy-saving coasting strategy and adaptive speed curve, these are sent as control commands to the train's onboard equipment, such as ATP or ATO, for execution. The train then completes the actual operation control according to the issued coasting strategy and speed curve, thereby achieving a closed loop from optimization decision-making to on-site execution.

[0045] As an optional embodiment, generating an energy-saving lazy propagation strategy based on a target control sequence includes: S401: When the absolute value of the current acceleration of the train is less than the preset acceleration threshold and the gradient of the track ahead is downhill or flat, the conditions for coasting are met and the train enters coasting mode. Specifically, coasting cannot be entered at any time. This embodiment of the invention stipulates that entering coasting mode requires two conditions to be met simultaneously: First, the absolute value of the train's current acceleration is less than a preset acceleration threshold, indicating that the train is currently in a stable operating state with a near-uniform speed, and cutting traction will not cause a significant speed drop. Second, the gradient of the track ahead is downhill or flat, meaning the track has the conditions to maintain or even increase speed by gravity or inertia, and coasting will not cause an excessively rapid speed decrease. When both conditions are met simultaneously, the coasting condition is satisfied, and the train enters coasting mode. The preset acceleration threshold is used to define the boundary for sufficiently small acceleration, and its specific value can be flexibly set according to the train type and comfort requirements.

[0046] The preset acceleration threshold can be flexibly set, for example, it can be 0.1 m / s². 2 The embodiments of the present invention are not limited herein.

[0047] S402: Calculate the optimal coasting time based on the predicted gradient of the track ahead and the on-time running time constraint; Specifically, after entering coasting mode, it is necessary to determine how long the coasting should last. This embodiment of the invention calculates the optimal coasting duration that balances energy saving and punctuality based on the predicted gradient of the track ahead (i.e., the distribution and magnitude of downhill / flat slopes in each section ahead) and the time constraint for train punctuality (i.e., the remaining time allowed to reach the next station or signal point). This ensures that the train can still meet timetable requirements after the coasting period ends.

[0048] S403: When the train speed drops to the lower speed limit or reaches the speed limit point or signal point ahead, exit the coasting mode; Among them, speed limit points are specific locations on the line where trains are required to reduce their operating speed; signal points are locations where pass signals are installed.

[0049] Specifically, the coasting mode is exited under any of the following circumstances: first, the train speed drops to the preset lower speed limit, and continuing coasting would jeopardize operational efficiency or punctuality; second, the train reaches the speed limit point ahead, i.e., a specific location on the line where the train is required to reduce its operating speed; third, the train reaches a signal point, i.e., a location with a passing signal. When any of the above circumstances occur, the coasting mode is exited, and controlled traction or braking dominated by the target control sequence is restored.

[0050] Among them, speed limit points are specific locations on the track where trains are required to reduce their operating speed; signal points are locations where pass signals are installed. By clearly defining these two, the clarity and enforceability of the coasting exit criteria can be ensured.

[0051] As an optional embodiment, generating an adaptive velocity curve based on the target control sequence includes: S501: Substitute the traction power and braking power of each control step in the target control sequence into the train dynamics model, and obtain the speed trajectory in the time domain based on the integration of the current train speed. Specifically, the generation of the adaptive speed curve is based on the train dynamics model. In this embodiment of the invention, the traction power and braking power corresponding to each control step in the target control sequence are substituted into the train dynamics model as inputs, and the current speed of the train is used as the initial value to numerically integrate the dynamic equations, thereby recursively obtaining the speed change process of the train over a period of time in the future, that is, the speed trajectory in the time domain.

[0052] S502: Converts the velocity trajectory in the time domain into a velocity-distance curve; Specifically, since the focus of traction power supply and train operation is not only on when but also on where, this embodiment of the invention maps the speed trajectory in the time domain from time coordinates to spatial coordinates, converting it into a speed-distance curve, that is, giving the train's operating speed at each mileage position, which is convenient for point-by-point comparison with the line conditions.

[0053] S503: Compare the speed-distance curve with the line speed limit curve point by point and take the smaller value to obtain the speed curve after speed limit constraint; Specifically, to ensure driving safety, this embodiment of the invention compares the speed-distance curve with the existing speed limit curve of the line, i.e., the maximum operating speed allowed at each mileage position, and takes the smaller value of the two for each mileage point, thereby embedding the hard constraint of the line speed limit into the speed curve, and obtaining the speed curve after speed limit constraint, ensuring that the optimized speed does not exceed the line allowable value at any position.

[0054] S504: Smooth the acceleration of the velocity curve after speed limit constraint to obtain an adaptive velocity curve.

[0055] Specifically, the speed curve after speed limit constraint may experience acceleration jumps near the speed limit point, affecting ride comfort and equipment lifespan. Therefore, this embodiment of the invention smooths the acceleration of the curve, eliminates abrupt changes, and makes the acceleration transition continuously, ultimately obtaining an adaptive speed curve that takes into account energy saving, speed limit and comfort, and issues it for execution in the form of a target speed sequence and a timestamp sequence.

[0056] In generating the adaptive speed curve, in addition to smoothing the speed curve after the speed limit constraint, the jerk is also limited, for example, to ≤1m / s³, in order to further ensure ride comfort. As an optional embodiment, the preset dynamic optimization model of the train traction and braking sequence is: a dynamic optimization model of the train traction and braking sequence constructed based on model predictive control. The dynamic optimization model of train traction and braking sequence includes a train dynamics model; The state vector of the train dynamics model is: ; in, x This represents the state vector of the train dynamics model. s For the train's position, v For train speed, E The remaining kinetic energy, T This is the transpose of the matrix.

[0057] Specifically, the dynamic optimization model in this embodiment of the invention is constructed using the Model Predictive Control (MPC) method, which can predict the future operating state in the time domain based on the current state of the train and solve for the optimal traction and braking sequence. The core physical basis of this model is the train dynamics model, which describes the train's motion state using state vectors: s represents the real-time position of the train along the track (unit: m), v represents the real-time speed of the train (unit: m / s), and E represents the remaining kinetic energy of the train (unit: kJ). By collecting the above state vectors as initial values ​​in each control cycle, and combining them with model predictive control, subsequent control steps can be optimized, thereby achieving the synergistic goal of minimizing traction energy consumption and maximizing regenerative braking energy recovery.

[0058] As an optional embodiment, the process of acquiring new energy output forecast data includes: Collect real-time meteorological data, which includes at least one of irradiance, wind speed, and temperature. Real-time meteorological data is input into a pre-trained new energy output prediction model to determine the new energy output prediction data; Among them, the new energy output prediction model is a prediction model constructed based on a Bayesian optimized long short-term memory network. Bayesian optimization is used to automatically tune the hyperparameters of the long short-term memory network.

[0059] Specifically, in this embodiment of the invention, real-time meteorological data may include: irradiance (unit: W / m², sampling period: for example, 1 minute, value range: for example, 0~1200 W / m²), used to predict photovoltaic output; wind speed (unit: m / s, sampling period: for example, 1 minute, value range: for example, 0~30 m / s), used to predict wind power output; and temperature (unit: ℃, sampling period: for example, 5 minutes, value range: for example, -20~45℃), used to correct photovoltaic module efficiency. By collecting the above real-time meteorological data, real-time input features can be provided for the new energy output prediction model, thereby enabling the new energy output prediction data to reflect changes in meteorological conditions in a timely manner.

[0060] Specifically, the construction process of the new energy output prediction model includes: a data preprocessing stage, using historical meteorological data (irradiance, wind speed, temperature), time features (hours, days of the week, months), and historical output data over t time steps as input feature sequences to form an input matrix of dimension [t, n_features]. The input features are then normalized using Min-Max to map them to the [0,1] interval. A logarithmic transformation is applied to the output target to handle long-tailed distributions. Sliding window sampling is used, with a window length of, for example, T=24, and a prediction step size of, for example, H=1~4, corresponding to a prediction range of 15 minutes to 1 hour. The long short-term memory network architecture stage includes an input layer, at least two long short-term memory network layers (the number of hidden units and the dropout rate in each layer are hyperparameters to be optimized; the activation function can be, for example, tanh), and a fully connected layer. The number of neurons in the connection layer corresponds to the prediction step size. In the Bayesian optimization hyperparameter tuning phase, a search space is constructed for hyperparameters such as learning rate, batch size, number of hidden units in each long short-term memory network layer, and dropout rate. The objective function is to minimize the root mean square error (RMSE) of the validation set. This is achieved by initializing a Gaussian process surrogate model and iteratively selecting evaluation points using an expected improvement acquisition function, training the model, evaluating the RMSE of the validation set, and updating the surrogate model. This process is repeated until the preset maximum number of iterations is reached. Finally, the hyperparameter combination with the smallest RMSE is selected as the final configuration of the new energy power output prediction model. Through this Bayesian optimization process, the hyperparameters of the long short-term memory network can be automatically determined, thereby improving the prediction accuracy of new energy power output prediction data without the need for repeated manual parameter tuning.

[0061] Specifically, after training, the new energy power output prediction model can output the predicted value of new energy power output for a preset time period in the future (e.g., the next 15 minutes, 30 minutes, or 1 hour), and can also provide the corresponding confidence interval (e.g., 95% confidence level) to characterize the credibility of the prediction results, providing a data basis for determining the upper limit of train traction power based on the new energy power output prediction data.

[0062] As an optional embodiment, the energy management method for integrating new energy sources into the traction power supply system further includes: Based on reinforcement learning algorithms, power dispatch instructions are determined according to the state of charge of the energy storage system, the power output prediction data of new energy sources, and the train load prediction data. The power dispatch instructions include at least one of the energy storage charging and discharging power and the proportion of new energy grid connection.

[0063] Specifically, in this embodiment of the invention, the determination process of the above-mentioned power scheduling command can be modeled as a Markov decision process: the state space includes at least the state of charge of the energy storage system (e.g., a value range of 0.2~0.95), the predicted output data of new energy sources (predicted output values ​​for the current and several future control steps), and the predicted load data of the train (predicted load values ​​for the current and several future control steps); the action space includes at least the energy storage charging power, the energy storage discharging power, and the proportion of new energy grid connection; the state transition function is determined based on the physical model of the energy storage state of charge changing with charging and discharging and the system power balance relationship; the reward function is composed of a weighted average of economic benefits, operating cost penalties, and constraint violation penalties, wherein the economic benefits include peak-valley price arbitrage benefits and new energy consumption rewards, the operating cost penalties include energy storage charging and discharging loss costs, and the constraint violation penalties include negative rewards corresponding to state of charge exceeding limits, voltage exceeding limits, and power imbalance.

[0064] Specifically, in this embodiment of the invention, a deep Q-network can be used to implement the above reinforcement learning algorithm: the state vector is used as input, and after passing through at least two fully connected hidden layers (each layer can use the ReLU activation function, for example), the Q value corresponding to each action is output; during training, an experience replay buffer is established to store historical transition samples, and a main network and a target network are set up. The main network parameters are updated by exploring with an ε-greedy strategy and calculating the target Q value using sampled batches of samples. Every preset number of steps, the main network parameters are copied to the target network; during inference, the current energy storage state of charge, new energy output prediction data and train load prediction data are collected as the current state, input into the trained deep Q-network to calculate the Q value of each action, select the action that satisfies the constraints and has the largest Q value, and convert the action into a specific power scheduling instruction.

[0065] Specifically, power dispatch instructions can include second-level and minute-level instructions: second-level instructions are used to adjust the charging and discharging power of the energy storage converter in real time to respond to rapid power fluctuations; minute-level instructions are used to optimize the proportion of renewable energy connected to the grid and the energy storage dispatch plan based on updated predicted values; the content of these instructions can include energy storage charging power, energy storage discharging power, and reference values ​​for the output of renewable energy inverters. By generating these power dispatch instructions based on reinforcement learning algorithms, the power output of the energy storage system and the renewable energy power generation system can be dynamically coordinated according to the state of charge and predicted data, thereby further improving the utilization rate of renewable energy power generation and the power balance of the traction power supply system.

[0066] As an optional embodiment, the state equation of the train dynamics model is: ; ; ; ; ; ; in, F traction For traction force, F brake For braking force, F resistance For running resistance, P traction For traction power, P brake For braking power, P loss For power loss, A , B and C The drag coefficient, M For train quality, G It is the acceleration due to gravity. i θ For the gradient of the line, θ The slope angle.

[0067] Specifically, in this embodiment of the invention, the physical quantities in the state equation are defined and coupled in the following manner, so that the train dynamics model can be directly called, wherein... F traction Traction force (unit: N) is determined by traction power and speed, i.e., traction power divided by speed; when the train is in traction condition, this force is positive and is used to overcome resistance and accelerate the train. F brakeBraking force (unit: N) is determined by braking power and speed, i.e., braking power divided by speed; when the train is in braking condition, this force is positive and is used to consume the train's kinetic energy. F resistance The operating resistance (unit: N) consists of two parts: basic resistance and additional resistance. The basic resistance is composed of ( A + B · v + C · v ²)· M express, A , B , C These are the drag coefficients, corresponding to the components of mechanical friction, rolling resistance, and air resistance, respectively. M The train mass (in tons); additional resistance is caused by the gradient and is expressed as... G · M · i θ ,in G It is the acceleration due to gravity. i θ The gradient of the line is expressed in per mille (‰). θ This represents the gradient angle corresponding to the route mileage.

[0068] Specifically, the state equation dE / dt = P traction - P brake - P loss The evolution of the train's remaining kinetic energy is described: P traction Traction power (unit: kW, a decision variable). P brake Braking power (unit: kW, a decision variable) P loss The power loss is expressed in kW and is related to traction / braking efficiency. The above state equations accurately represent the evolution of the train's position, speed, and remaining kinetic energy with increasing traction / braking power, providing physical constraints for rolling optimization under model predictive control and making the solved target control sequence physically achievable.

[0069] As an optional implementation, the objective function aims to minimize traction energy consumption and maximize regenerative braking energy recovery, and is achieved through the following objective function: ; in, J The objective function value, E total Total traction energy consumption, E potential For potential recyclable braking energy,E recovered For the actual braking energy recovered, α , β The weighting coefficients are and satisfy the following conditions: α + β =1; Under default operating conditions α The value is greater than β The value of is determined so that minimizing traction energy consumption is prioritized over maximizing regenerative braking energy recovery.

[0070] Specifically, E total Total traction energy consumption (in kWh), and the target minimum value; E potential The potential recoverable braking energy (unit: kWh) is determined by the braking conditions. E recovered The target is to maximize the actual recovered braking energy (in kWh), which is equivalent to minimizing the difference between it and the potential recoverable amount. α , β The weighting coefficient can be taken as follows, under the default operating conditions of this embodiment of the invention: α =0.6、 β =0.4, and satisfies α + β =1; where α The value is greater than β The value of is chosen to prioritize minimizing traction energy consumption over maximizing regenerative braking energy recovery.

[0071] The objective function incorporates traction energy consumption and regenerative braking energy recovery into a single optimization index. By balancing the inverse relationship between the two through weighting coefficients, the obtained target control sequence can reduce traction energy consumption while recovering as much regenerative braking energy as possible, thus achieving the dual goals of train energy conservation and new energy consumption.

[0072] Of course, the objective function can take other forms besides this specific form, and this embodiment of the invention does not limit it here.

[0073] As an optional embodiment, after obtaining the train power flow scheduling instruction, the energy management method for integrating new energy sources into the traction power supply system further includes: The weighting coefficients are updated in real time based on the regenerative braking energy utilization priority and / or target zoning load level in the power flow dispatching instructions. α , β The value.

[0074] Specifically, the power flow scheduling command carries fields such as regenerative braking energy utilization priority, target partition load level, and constraint time window, encapsulated in JSON format, and received via IEC 60870-5-104 protocol subscription, with an update cycle of, for example, 10 seconds; the embodiments of the present invention adjust the weight coefficients in the objective function accordingly. α , β Adaptive updates are performed to make the optimization objective change dynamically with the scheduling intent of the system.

[0075] The update rules for the weighting coefficients can be flexibly set, and may include, for example: When the priority of regenerative braking energy utilization is increased, the weighting coefficient is increased. β The value of and the corresponding decrease α The value of [value] is chosen to emphasize the proportion of regenerative braking energy recovery in the optimization objective; when the load level of the target zone increases, the weighting coefficient is increased. α The value of and the corresponding decrease β The values ​​are selected to prioritize reducing traction energy consumption and alleviating load pressure on the target zone. The above weight updates take effect within the constraint time window; after the window ends, the relevant weights can be restored to the default baseline. α =0.6、 β =0.4), thus balancing the overall scheduling intent with the local flexibility of train operation.

[0076] It should be noted that the updated version α , β Then, it is substituted into the objective function and used to participate in the rolling optimization solution of the model predictive control in the next control cycle, thereby realizing energy management of vehicle-network collaboration.

[0077] As an optional implementation, based on a preset dynamic optimization model of train traction and braking sequence, with the objectives of minimizing traction energy consumption and maximizing regenerative braking energy recovery, the target control sequence that satisfies the constraints is solved according to the real-time train operating parameters. This is achieved using a rolling optimization method based on model predictive control, including: In any control cycle, a state prediction in the prediction time domain is constructed based on the current train state, and the optimal control sequence that satisfies the constraints is solved in the prediction time domain. The optimal control sequence includes the traction power and braking power of each control step in the prediction time domain. Execute the traction power and braking power corresponding to the first control step in the optimal control sequence.

[0078] Specifically, through the rolling optimization process, the dynamic optimization model can replan the control sequence for the future prediction time domain based on the latest collected train status in each control cycle, and execute the first step, thereby achieving continuous tracking and optimization of time-varying operating conditions, such as changes in line gradient and changes in power flow scheduling instructions, with limited computational cost.

[0079] In one specific embodiment, the rolling optimization of model predictive control includes: (1) At the current moment t Collect the train's state vector x ( t )=[ s ( t ), v ( t ), E ( t )] T As initial values ​​for optimization s ( t () represents the current position of the train. v ( t ( ) represents the train's speed at the current moment. E ( t The remaining kinetic energy of the train at the current moment is denoted as . (2) Constructing the prediction time domain N p In the embodiments of the present invention, it is possible to adopt N p =20, corresponding to the state prediction within a 20-second prediction window: x ( t + k | t )= f ( x ( t + k -1| t ), u ( t + k -1| t The initial state of the optimization problem x ( t | t )=x( t ), k For control step index, u To control the input vector, k =1,2,…, N p ,in u ( k )=[ P traction ( k ), P brake ( k )] T For control input; (3) Solving constrained optimization problems: , JGiven the objective function, the state equations and constraints such as velocity, acceleration, power, time, and energy recovery; (4) The optimal control sequence u is obtained by using a quadratic programming (QP) solver. (t),u (t+1),…,u (t+N p -1), This serves as an identifier for the optimal solution. (5) Execute the first control step u in the optimal control sequence. (t), which is then converted into traction / braking commands and sent to the train; (6) Wait for the control cycle, for example, Δtcycle=1 second, and return to step (1) to perform rolling iteration.

[0080] As an optional embodiment, the constraints include energy recovery constraints and power constraints; The energy recovery constraint is: ; The power constraint is: 0≤ P traction ≤ P tractionmax ; 0≤ P brake ≤ P brakemax ; in, η Energy recovery efficiency for regenerative braking energy E potential For potential recyclable braking energy, E recovered For the actual braking energy recovered, P traction For traction power, P tractionmax This is the upper limit of traction power. P brake For braking power, P brakemax This is the upper limit of braking power.

[0081] Specifically, η The energy recovery efficiency of regenerative braking energy is used to characterize the actual recoverable proportion of braking energy after feedback and grid connection. In the embodiments of this invention, it can be taken as... η ≈0.7~0.8; E potential For potential recyclable braking energy, E recoveredThe actual recovered braking energy; the energy recovery constraint limits the actual recovered braking energy to no more than the product of the potential recoverable amount and the recovery efficiency, thereby ensuring that the recovery target is achievable in engineering physics.

[0082] Specifically, power constraints limit the upper and lower limits of traction and braking power. The positive correlation adjustment of the upper limit of braking power and traction power according to the power flow scheduling command in the aforementioned embodiments corresponds to the dynamic correction of the upper limit of power constraints. Energy recovery constraints and power constraints together define the feasible region of the target control sequence, making the solved control sequence feasible within the range of safety, comfort, and equipment capabilities.

[0083] Specifically, in embodiments of the present invention, the constraints may also include other types, such as: (1) Velocity constraint: v min ≤v≤v max , where v min As the lower limit of speed, v max The upper and lower speed limits can be set by considering both line speed limits and signal speed limits. (2) Acceleration constraint: a brakemax ≤a≤a tractionmax a brakemax a is the upper limit of braking deceleration. tractionmax This is the upper limit of traction acceleration, in order to take into account passenger comfort; (3) Time constraint: t arrival -t departure =Δt target , t arrival t represents the actual time the train arrives at the next station. departure Δt represents the actual departure time of the train from the current station (or origin). target This refers to the planned travel time for a train from the current station to the next station, in order to meet on-time requirements; The aforementioned constraints collectively define the feasible region of the target control sequence.

[0084] As an optional implementation, the upper limit of the train's traction power in the future forecast time domain is determined based on the new energy output forecast data, including: S1101: Based on the new energy output forecast data, determine the predicted power generation of the new energy power generation system in the future forecast time domain; Specifically, the new energy output forecast data itself includes the predicted power generation value of the new energy power generation system at each time point or control step in the future forecast time domain. Based on this, the embodiments of the present invention directly determine the predicted power generation, which serves as one of the basic data for subsequent calculation of the estimated available power supply.

[0085] S1102: The sum of the predicted power generation of the new energy power generation system in the future prediction time domain, the available discharge power of the energy storage system in the future prediction time domain, and the available power supply of the power grid in the future prediction time domain is used as the estimated value of the available power supply in the future prediction time domain. Specifically, the available discharge power of the energy storage system in the future forecast time domain can be determined based on parameters such as the current state of charge, rated discharge power, and depth of discharge limit of the energy storage system; the available power supply of the power grid in the future forecast time domain can be determined based on the rated power supply capacity of the traction substation and the dispatch constraints on the grid side. This embodiment of the invention adds the aforementioned predicted power generation of new energy sources, available discharge power of energy storage, and available power supply of the power grid to obtain the estimated value of the available power supply of the traction power supply system in the future forecast time domain, thereby enabling a unified assessment of power supply capacity based on new energy output, energy storage status, and power grid supply capacity.

[0086] S1103: Obtain the estimated value of other load power of the traction power supply system in the future prediction time domain. Other load power refers to the load power in the traction power supply system other than the train traction load. Specifically, other load power may include the power consumption of non-traction loads such as signal systems, communication systems, and lighting systems in the traction power supply system. This part of the load power can be estimated based on historical power consumption data statistics or preset load curves, and is the part that needs to be deducted in advance when calculating the upper limit of traction power.

[0087] S1104: Determine the upper limit of traction power in the constraints based on the difference between the estimated available power supply and the estimated power of other loads.

[0088] Specifically, in this embodiment of the invention, the difference between the estimated available power supply and the estimated power supply of other loads is used as the upper limit of traction power in the constraint conditions, i.e., upper limit of train traction power = estimated available power supply - estimated power supply of other loads. The upper limit of traction power determined in this way can match the train's traction power demand with the actual available power supply capacity of the traction power supply system in the future predicted time domain. During periods of high renewable energy output, the train is allowed to obtain a higher upper limit of traction power to fully utilize renewable energy generation; during periods of low renewable energy output, the upper limit of traction power is correspondingly reduced to avoid power imbalance in the traction power supply system, thereby balancing train operating efficiency and renewable energy absorption.

[0089] As an optional embodiment, after controlling the traction power and braking power of the train according to the target control sequence, the energy management method for integrating new energy sources into the traction power supply system further includes: S1201: Determine the predicted energy consumption in the prediction time domain based on the target control sequence and the train dynamics model; Specifically, considering that the target control sequence solved by model predictive control itself contains the traction power and braking power of each future control step, the embodiments of the present invention can combine the train dynamics model to calculate the energy consumption evolution process of the train in the prediction time domain, thereby determining the predicted energy consumption in the prediction time domain as a theoretical benchmark for comparison with the actual energy consumption.

[0090] The predicted energy consumption can be derived from the MPC model.

[0091] S1202: Obtain the actual energy consumption of the train; Specifically, during the actual operation of the train according to the target control sequence, the onboard equipment collects and feeds back the train's real energy consumption data in real time, thereby obtaining the train's actual energy consumption.

[0092] Specifically, the onboard ATP or ATO equipment can periodically, for example, provide feedback on the actual energy consumption (unit kWh, accuracy ±0.1 kWh) and the actual regenerative braking energy recovery (unit kWh, accuracy ±0.1 kWh) per second. S1203: Adjust the preset adjustable parameters in the dynamic optimization model based on the deviation between actual energy consumption and predicted energy consumption; the preset adjustable parameters include train mass and drag coefficient.

[0093] Specifically, since there are inevitably differences between the actual operating environment of the train and the model assumptions, a deviation will occur between the actual energy consumption and the predicted energy consumption. This embodiment of the invention calculates the deviation and corrects the preset adjustable parameters in the dynamic optimization model accordingly. The preset adjustable parameters include at least the train mass and drag coefficient, with the drag coefficient being one of the coefficients in the train running resistance formula. By dynamically correcting the model parameters based on actual operating feedback, the dynamic model can be made to better reflect the real train, thereby improving the accuracy of subsequent target control sequence solutions and energy-saving effects.

[0094] As an optional implementation, adjusting the preset adjustable parameters in the dynamic optimization model based on the deviation between actual energy consumption and predicted energy consumption includes: S1301: When the deviation between actual energy consumption and predicted energy consumption is greater than the preset deviation threshold, acquire the power and speed data over the past N control cycles. The power and speed data include the traction power, braking power and corresponding actual speed of each control step. Specifically, to avoid frequent disturbances to model parameters by normal-level random fluctuations, this embodiment of the invention sets a preset deviation threshold as a trigger condition: the parameter correction process is only initiated when the deviation between actual energy consumption and predicted energy consumption exceeds this preset deviation threshold. After triggering, power and speed data from the past N control cycles are acquired. The data covers the traction power, braking power, and corresponding actual train speed for each control step, forming a data sample for parameter identification.

[0095] The preset deviation threshold can be set flexibly and independently, for example, it can be set as a preset percentage of the rated energy consumption, such as 5%, etc., but this embodiment of the invention does not limit it.

[0096] S1302: Using the traction power and braking power of each control step in the power and speed data as inputs and the actual speed corresponding to the control step as the observation, the estimated values ​​of the preset adjustable parameters in the dynamic optimization model are determined through the rolling time domain estimation model. Specifically, in this embodiment of the invention, the acquired power and speed data are organized as the input and output of the estimation problem: the traction power and braking power of each control step are used as the model input, and the actual train speed corresponding to the control step is used as the observation value. With the help of the Receding Horizon Estimation (RHE) model, the preset adjustable parameters in the dynamic optimization model are recursively estimated within the rolling time window, so as to obtain parameter estimates that can better interpret the actual observation data.

[0097] S1303: Update the values ​​of the preset adjustable parameters in the dynamic optimization model to the estimated values.

[0098] Specifically, after obtaining the estimated values ​​of the preset adjustable parameters, they are written back to the dynamic optimization model to replace the original parameter values. This allows the model to use the updated parameters for solving in the next round of rolling optimization, thus forming a continuous closed loop that continuously improves the model's fitting accuracy to real trains and the energy-saving benefits of energy management.

[0099] Please refer to Figure 3 , Figure 3 This is a schematic diagram of the structure of the energy management device for a new energy access traction power supply system provided by the present invention. The energy management device for the new energy access traction power supply system is applied to the energy management system of the traction power supply system and includes: The acquisition module 31 is used to acquire real-time train operating parameters, which include at least one of the following: train position, train speed, train acceleration, traction and braking status, and the gradient of the track ahead. The determination module 32 is used to determine the upper limit of the train's traction power in the future prediction time domain based on the new energy power output prediction data; the new energy power output prediction data includes the predicted power generation of the new energy power generation system in the future preset time period. The solver module 33 is used to solve the target control sequence that meets the constraints based on the dynamic optimization model of the preset train traction and braking sequence, with the goal of minimizing traction energy consumption and maximizing regenerative braking energy recovery. The target control sequence includes the traction power and braking power of each control step in the future prediction time domain; the constraints include the upper limit of traction power. The execution module 34 is used to control the traction power and braking power of the train according to the target control sequence.

[0100] For an introduction to the energy management device of the new energy access traction power supply system provided in the embodiments of the present invention, please refer to the aforementioned embodiments of the energy management method of the new energy access traction power supply system. The embodiments of the present invention will not be repeated here.

[0101] Please refer to Figure 4 , Figure 4 This is a schematic diagram of the energy management system of the new energy access traction power supply system provided by the present invention. The energy management system of the new energy access traction power supply system includes: Memory 41 is used to store computer programs; The processor 42 is used to execute computer programs to implement the steps of the energy management method for accessing the traction power supply system of new energy sources as described in the foregoing embodiments.

[0102] For an introduction to the energy management system of the new energy access traction power supply system provided in the embodiments of the present invention, please refer to the aforementioned embodiments of the energy management method of the new energy access traction power supply system. The embodiments of the present invention will not be repeated here.

[0103] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section. It should also be noted that in this specification, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising a..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0104] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An energy management method for integrating new energy sources into a traction power supply system, characterized in that, An energy management system applied to traction power supply systems includes: The train's real-time operating parameters are obtained, including at least one of the following: train position, train speed, train acceleration, traction and braking status, and the gradient of the track ahead. Based on the new energy power output forecast data, the upper limit of the train's traction power in the future forecast time domain is determined; the new energy power output forecast data includes the forecast power generation of the new energy power generation system in the future preset time period. Obtain train power flow scheduling instructions, which include regenerative braking energy utilization priority, target zonal load level, and constraint time window; Based on the priority of regenerative braking energy utilization, the upper limit of braking power in the constraint conditions is adjusted within the constraint time window. The priority of regenerative braking energy utilization and the upper limit of braking power in the constraint conditions are positively correlated. Based on the target zone load level, within the constraint time window, the upper limit of traction power in the constraint conditions of the target zone is adjusted, and the target zone load level and the upper limit of traction power in the constraint conditions are positively correlated. Based on a preset dynamic optimization model of train traction and braking sequence, with the objectives of minimizing traction energy consumption and maximizing regenerative braking energy recovery, a target control sequence that satisfies the constraints is solved according to the real-time operating parameters of the train. The target control sequence includes the traction power and braking power of each control step in the future prediction time domain; the constraints include the upper limit of the traction power. The traction power and braking power of the train are controlled according to the target control sequence. The determination of the upper limit of train traction power in the future forecast time domain based on new energy output forecast data includes: Based on the new energy output forecast data, the predicted power generation of the new energy power generation system in the future forecast time domain is determined; The sum of the predicted power generation of the new energy power generation system in the future prediction time domain, the available discharge power of the energy storage system in the future prediction time domain, and the available power supply of the power grid in the future prediction time domain is used as the estimated value of the available power supply in the future prediction time domain. Obtain the estimated value of other load power of the traction power supply system in the future prediction time domain, wherein the other load power is the load power of the traction power supply system other than the train traction load; The upper limit of traction power in the constraints is determined based on the difference between the estimated available power supply and the estimated power of other loads.

2. The energy management method for integrating new energy sources into the traction power supply system according to claim 1, characterized in that, The step of controlling the traction power and braking power of the train according to the target control sequence includes: An energy-saving coasting strategy is generated based on the target control sequence. The energy-saving coasting strategy includes the coasting start point mileage, the coasting duration, and the coasting exit condition. An adaptive velocity curve is generated based on the target control sequence, the adaptive velocity curve including the target velocity sequence and the timestamp sequence; The energy-saving coasting strategy and the adaptive speed curve are sent to the train for execution.

3. The energy management method for accessing a new energy source to a traction power supply system according to claim 2, characterized in that, The step of generating an energy-saving lazy propagation strategy based on the target control sequence includes: When the absolute value of the train's current acceleration is less than the preset acceleration threshold and the gradient of the track ahead is downhill or flat, the conditions for coasting are met, and the train enters coasting mode. Calculate the optimal coasting time based on the predicted gradient of the route ahead and the constraints of on-time operation. When the train speed drops to the lower speed limit or reaches the speed limit point or signal point ahead, it exits the coasting mode. Among them, speed limit points are locations on the line where trains are required to reduce their operating speed; signal points are locations where pass signals are installed.

4. The energy management method for new energy access to the traction power supply system according to claim 2, characterized in that, The step of generating the adaptive velocity curve based on the target control sequence includes: Substitute the traction power and braking power of each control step in the target control sequence into the train dynamics model, and obtain the speed trajectory in the time domain based on the integration of the current train speed. The velocity trajectory in the time domain is converted into a velocity-distance curve; The speed-distance curve is compared point by point with the line speed limit curve, and the smaller value is taken to obtain the speed curve after the speed limit constraint. The acceleration of the velocity curve after the speed limit constraint is smoothed to obtain an adaptive velocity curve.

5. The energy management method for new energy access to the traction power supply system according to claim 1, characterized in that, The preset dynamic optimization model for the train traction and braking sequence is: a dynamic optimization model for the train traction and braking sequence constructed based on model predictive control. The dynamic optimization model of the train traction and braking sequence includes a train dynamics model; The state vector of the train dynamics model is: ； in, x This represents the state vector of the train dynamics model. s For the train's position, v For train speed, E The remaining kinetic energy, T This is the transpose of the matrix.

6. The energy management method for new energy access to the traction power supply system according to claim 1, characterized in that, The process of acquiring the new energy output forecast data includes: Collect real-time meteorological data, wherein the real-time meteorological data includes at least one of irradiance, wind speed and temperature; The real-time meteorological data is input into a pre-trained new energy output prediction model to determine the new energy output prediction data; The new energy output prediction model is a prediction model constructed based on a Bayesian optimized long short-term memory network. The Bayesian optimization is used to automatically tune the hyperparameters of the long short-term memory network.

7. The energy management method for integrating new energy sources into the traction power supply system according to claim 1, characterized in that, The energy management method for integrating new energy sources into the traction power supply system also includes: Based on reinforcement learning algorithms, power dispatch instructions are determined according to the state of charge of the energy storage system, the predicted output data of the new energy source, and the predicted train load data. The power dispatch instructions include at least one of the energy storage charging and discharging power and the proportion of new energy source grid connection.

8. The energy management method for integrating new energy sources into the traction power supply system according to claim 1, characterized in that, The dynamic optimization model based on the preset train traction and braking sequence aims to minimize traction energy consumption and maximize regenerative braking energy recovery. It solves for the target control sequence that satisfies the constraints based on the real-time train operating parameters, and employs a rolling optimization method using model predictive control, including: In any control cycle, a state prediction in the prediction time domain is constructed based on the current train state, and the optimal control sequence that satisfies the constraints is solved in the prediction time domain. The optimal control sequence includes the traction power and braking power of each control step in the prediction time domain. Execute the traction power and braking power corresponding to the first control step in the optimal control sequence.

9. The energy management method for accessing a new energy source to a traction power supply system according to claim 1, characterized in that, The constraints include energy recovery constraints and power constraints; The energy recovery constraint is: ; The power constraint is: 0≤ P traction ≤ P tractionmax ； 0≤ P brake ≤ P brakemax ； in, η Energy recovery efficiency for regenerative braking energy E potential For potential recyclable braking energy, E recovered For the actual braking energy recovered, P traction For traction power, P tractionmax This is the upper limit of traction power. P brake For braking power, P brakemax This is the upper limit of braking power.

10. The energy management method for integrating new energy sources into a traction power supply system according to any one of claims 1 to 9, characterized in that, After controlling the traction power and braking power of the train according to the target control sequence, the energy management method for connecting new energy sources to the traction power supply system further includes: Based on the target control sequence and the train dynamics model, the predicted energy consumption in the prediction time domain is determined; Obtain the actual energy consumption of the train; Based on the deviation between the actual energy consumption and the predicted energy consumption, the preset adjustable parameters in the dynamic optimization model are adjusted; the preset adjustable parameters include train mass and drag coefficient.

11. The energy management method for new energy access to traction power supply system according to claim 10, characterized in that, The step of adjusting the preset adjustable parameters in the dynamic optimization model based on the deviation between the actual energy consumption and the predicted energy consumption includes: When the deviation between the actual energy consumption and the predicted energy consumption is greater than a preset deviation threshold, power and speed data over the past N control cycles are acquired. The power and speed data includes the traction power, braking power, and corresponding actual speed for each control step. Using the traction power and braking power of each control step in the power and speed data as inputs and the actual speed corresponding to the control step as the observation, the estimated values ​​of the preset adjustable parameters in the dynamic optimization model are determined through the rolling time-domain estimation model. The values ​​of the preset adjustable parameters in the dynamic optimization model are updated to the estimated values.

12. An energy management device for integrating new energy sources into a traction power supply system, characterized in that, An energy management system applied to traction power supply systems includes: The acquisition module is used to acquire real-time train operating parameters, which include at least one of the following: train position, train speed, train acceleration, traction and braking status, and the gradient of the track ahead. The determination module is used to determine the upper limit of the train's traction power in the future prediction time domain based on the new energy output prediction data; the new energy output prediction data includes the predicted power generation of the new energy power generation system in the future preset time period. The solution module is used to solve for the target control sequence that meets the constraints based on the preset dynamic optimization model of train traction and braking sequence, with the objectives of minimizing traction energy consumption and maximizing regenerative braking energy recovery. The target control sequence includes the traction power and braking power of each control step in the future prediction time domain; the constraints include the upper limit of the traction power. An execution module is used to control the traction power and braking power of the train according to the target control sequence; The energy management device for the new energy access traction power supply system also includes: Obtain train power flow scheduling instructions, which include regenerative braking energy utilization priority, target zonal load level, and constraint time window; Based on the priority of regenerative braking energy utilization, the upper limit of braking power in the constraint conditions is adjusted within the constraint time window. The priority of regenerative braking energy utilization and the upper limit of braking power in the constraint conditions are positively correlated. Based on the target zone load level, within the constraint time window, the upper limit of traction power in the constraint conditions of the target zone is adjusted, and the target zone load level and the upper limit of traction power in the constraint conditions are positively correlated. The determination of the upper limit of train traction power in the future forecast time domain based on new energy output forecast data includes: Based on the new energy output forecast data, the predicted power generation of the new energy power generation system in the future forecast time domain is determined; The sum of the predicted power generation of the new energy power generation system in the future prediction time domain, the available discharge power of the energy storage system in the future prediction time domain, and the available power supply of the power grid in the future prediction time domain is used as the estimated value of the available power supply in the future prediction time domain. Obtain the estimated value of other load power of the traction power supply system in the future prediction time domain, wherein the other load power is the load power of the traction power supply system other than the train traction load; The upper limit of traction power in the constraints is determined based on the difference between the estimated available power supply and the estimated power of other loads.

13. An energy management system for integrating new energy sources into a traction power supply system, characterized in that, include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the energy management method for accessing a new energy source to a traction power supply system as described in any one of claims 1 to 11.

Images