Railway vehicle power optimization method based on energy routing control

Abstract

The invention discloses a railway vehicle power optimization method based on energy routing control, which comprises the following steps: acquiring a velocity vector and an acceleration parameter from a channel allocation state, fusing the parameters through a Kalman filtering algorithm, determining an energy demand curve in a future time period, and obtaining a predicted power demand sequence; a reference value of a hierarchical scheduling frame is obtained from the power flow distribution scheme, instantaneous power waves are processed through a multi-time-scale hierarchical architecture, a control instruction sequence from the millisecond level to the minute level is determined, and a coordinated short-term scheduling result is obtained; and acquiring position information data from the long-term energy network topology, fusing the data through a real-time synchronization mechanism, determining a final energy routing control state, and obtaining a closed-loop feedback cycle of vehicle power optimization.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of rail vehicle power optimization, and particularly relates to a rail vehicle power optimization method based on energy routing control. BACKGROUND

[0002] Rail vehicle power optimization is an important research direction in the field of modern transportation, which is directly related to transportation efficiency, energy utilization rate and operation cost control. With the acceleration of urbanization and the continuous expansion of rail transportation network, how to achieve efficient energy distribution in a complex operating environment has become a key issue in the industry development. Optimizing power distribution not only reduces energy consumption, but also improves the stability of vehicle operation and the comfort of passengers, so it occupies a core position in technological innovation.

[0003] However, the existing methods often appear to be inadequate in dealing with dynamic operating environment, especially when facing changes in vehicle position and fluctuations in energy demand, lack of real-time and adaptability. Many schemes cannot effectively handle the instantaneous demand changes of vehicles in different sections, and it is also difficult to coordinate energy distribution in multiple time scales, resulting in frequent waste of resources or power shortage. This limitation makes the rail transportation system often face the challenge of uneven energy distribution during peak hours or complex sections, and new technological breakthroughs are needed to solve these problems.

[0004] Moreover, in this field, the core technical difficulties are concentrated in the precise acquisition of vehicle position and the dynamic matching of energy distribution network. Since the position information of the vehicle needs to reach a very high precision in high-speed operation, otherwise it is difficult to accurately determine the track section it is in, and it is also difficult to achieve effective connection with the energy supply point. The deeper problem is that even if the position information is accurate, how to dynamically adjust the energy distribution path according to the real-time state and future running trend of the vehicle is also a complex problem. For example, when a train quickly passes through a track branching point, if the system cannot identify its position and switch to the appropriate energy supply channel in a very short time, it may cause power interruption, affecting the stability and safety of train operation.

[0005] Therefore, how to accurately grasp the position of the vehicle in the scenario of high-speed movement, and how to adjust the energy distribution path in real time according to the position change and running state, has become a key problem in the field of rail vehicle power optimization. SUMMARY

[0006] The present application proposes a rail vehicle power optimization method based on energy routing control to solve the above-mentioned problems existing in the prior art.

[0007] To achieve the above-mentioned purpose, the present application provides a rail vehicle power optimization method based on energy routing control, comprising the following steps: acquire vehicle centimeter-level positioning data, and establish a dynamic mapping relationship with the energy routing node to obtain a power supply channel configuration corresponding to the current section; determine an energy routing node switching point that the vehicle is about to approach according to the dynamic mapping relationship, and if the distance of the switching point is less than a preset threshold, update the channel allocation state; based on the updated channel allocation state, predict an energy demand curve in a future period to generate a power demand sequence; for the power demand sequence, calculate allocation strategy parameters using a particle swarm optimization algorithm, and if a multi-time scale requirement is met, generate an energy routing allocation instruction to obtain an optimized power flow distribution scheme; based on the reference value of the hierarchical scheduling architecture extracted from the power flow distribution scheme, obtain a short-term scheduling result; according to the short-term scheduling result, obtain hourly-level reference information in combination with passenger flow prediction data, and if the matching degree of the hourly-level reference information with the all-day configuration is higher than a preset threshold, update the all-day reference configuration scheme to form a long-term energy network topology; based on the position information and related data in the long-term energy network topology, determine a final energy routing control state to realize a closed-loop feedback cycle of vehicle power optimization.

[0008] Optionally, obtaining the power supply channel configuration corresponding to the current section comprises: extract the centimeter-level positioning information of the vehicle from real-time data by combining the Beidou signal and inertial navigation to obtain preliminary vehicle position coordinates; according to the preliminary vehicle position coordinates, calculate the distance relationship between the vehicle and each node in combination with the pre-established energy routing node distribution data to determine the nearest energy routing node; if the determined nearest energy routing node is inconsistent with the node corresponding data of the current section, update the dynamic relationship through real-time data to obtain the latest node corresponding information; for the latest node corresponding information, query a preset power supply strategy database to determine the power supply scheme applicable to the current section to obtain corresponding power allocation parameters; generate specific channel configuration instructions by combining the power allocation parameters with the channel configuration rules of the current section to determine the final power supply channel configuration.

[0009] Optionally, updating the channel allocation state comprises: analyze the relative position relationship between the current vehicle and the energy routing node by combining the vehicle position information with the dynamic mapping relationship to determine the node switching point to be involved; for the determined node switching point, calculate the real-time distance between the vehicle and the switching point, and if the distance is less than a preset threshold, trigger the subsequent power supply adjustment process to obtain a switching demand signal; According to the switching demand signal, state data of the current power supply is acquired, the feasibility of channel switching is analyzed, and a corresponding seamless switching instruction draft is generated; The seamless switching instruction draft is adopted, the distribution information of the energy routing node is combined, a suitable channel switching scheme is matched, and the final instruction content is determined; Through the final instruction content, a control signal of channel switching is sent to the power supply system, real-time update of the distribution state is completed, and an updated channel distribution state is obtained; According to the updated channel distribution state, the distance change between the vehicle position and the energy routing node is continuously monitored, and if a new switching demand is detected, the above process is repeatedly executed to maintain the dynamic adaptation of channel distribution.

[0010] Optionally, the power demand sequence is generated, including: The speed vector and acceleration parameter are extracted from the channel distribution state, the parameters are cleaned by using a pre-established screening rule, and a sorted parameter set is obtained; According to the sorted parameter set, the speed vector and acceleration parameter are processed by using a fusion algorithm to determine a corrected parameter combination; Through the corrected parameter combination, if the parameter exceeds the threshold range, boundary adjustment is performed, and an adjusted parameter mapping is obtained by combining a pre-set threshold range; For the adjusted parameter mapping, a time series tool is used to segment the mapping to obtain a predicted power demand sequence.

[0011] Optionally, the optimized power flow distribution scheme is obtained, including: The power demand sequence data is obtained, the sequence is segmented by using a pre-established data processing module, a plurality of time period sub-sequences are divided according to the time scale, and the segmented demand data is obtained; For the segmented demand data, the particle swarm optimization algorithm is used to iteratively calculate the distribution strategy parameters of each time period to determine the preliminary parameter configuration in each time period; According to the preliminary parameter configuration, it is judged whether the preset time scale constraint condition is met, if the parameters of a time period do not meet the constraint requirement, the parameters of the time period are recalculated to obtain an adjusted parameter set; Through the adjusted parameter set, a corresponding energy routing distribution instruction is generated, a corresponding routing path is matched for the demand data of each time period, and the specific content of the distribution instruction is determined; According to the distribution instruction, the initial scheme of power flow distribution is calculated, the power flow distribution is simulated by using a pre-set simulation tool, and the distribution result data is obtained; For the distribution result data, it is judged whether there is a local imbalance phenomenon, if there is, the allocation instruction of the related time period is fine-tuned to obtain a final optimized power flow distribution scheme.

[0012] Optionally, the short-term scheduling result includes: Through the power flow distribution data source, initial power grid operation state information is obtained, a multi-level scheduling framework data is decomposed for hierarchical scheduling requirements, and a preliminary hierarchical scheduling benchmark is obtained; According to the hierarchical scheduling benchmark, combined with the division of time scales, a hierarchical architecture model is constructed, real-time collection is performed for the dynamic change of instantaneous power, and the fluctuation characteristics of power data are determined; Through the fluctuation characteristics of the power data, fluctuation analysis processing is performed, and if the fluctuation value exceeds the threshold range, a corresponding adjustment signal is generated, and the scheduling node that needs to be intervened is determined; According to the intervention requirement of the scheduling node, a millisecond-level control instruction is generated, the rapid change of instantaneous power is responded to, and a high-precision instant control sequence is obtained; Through the instant control sequence, the control instruction generation is extended to the minute level, combined with the periodic requirement of short-term scheduling, the instruction combination across time scales is determined; The result of the instruction combination is obtained, the global data of the power flow distribution and the local data of the hierarchical scheduling are integrated according to the requirement of the coordination result, and the final short-term scheduling coordination scheme is obtained.

[0013] Optionally, the long-term energy network topology includes: According to the short-term scheduling result, the hourly passenger flow prediction data is obtained from the passenger flow prediction system, the time period is divided according to the prediction data, the peak and valley information in each time period is extracted, and the hourly benchmark information is determined; According to the hourly benchmark information, the resource allocation data in the all-day configuration scheme is obtained, the corresponding relationship between the benchmark information and the resource allocation data in the time period is analyzed, and the matching degree between the two is calculated; According to the comparison of the matching degree and the preset threshold value, if the matching degree is higher than the preset threshold value, the configuration scheme updating process is triggered, and the updated all-day configuration scheme data is obtained; Through the updated all-day configuration scheme data, combined with the long-term network planning requirement, the network resource distribution information is extracted, and the preliminary topology structure of the long-term energy network is determined; According to the preliminary topology structure, the energy allocation data of each node in the network is obtained, the energy flow path between nodes is analyzed, and it is judged whether there is an uneven resource allocation, if there is an uneven resource allocation, the path allocation ratio is adjusted, and the optimized topology structure is obtained; For the optimized topology, a support vector machine model is used to predict and analyze the energy flow path, obtain the energy demand distribution in the future time period, and determine the final long-term energy network topology.

[0014] Optionally, the closed-loop feedback loop for optimizing vehicle power includes: By obtaining topological data from long-term energy networks, analyzing their distribution characteristics, determining the connection relationships of key nodes, and obtaining preliminary network layout information; Based on the preliminary network layout information, the corresponding location information is extracted, and the relevant dynamically updated data is obtained in combination with the real-time synchronization mechanism. The data integrity is judged. If the data missing exceeds the preset threshold, the supplementary collection process is triggered to obtain the complete dataset. For the complete dataset, an integration mechanism is used to classify the data and separate the part that is directly related to routing control. If the separated data does not conform to the preset standard, the classification rules are readjusted to determine the data subset suitable for routing control. From the data subset of routing control, analyze the real-time changes in control state. If an abnormal state is detected, correct it through pre-established mapping rules to obtain stable control state parameters. By obtaining stable control state parameters and combining them with the dynamic power requirements of the vehicle, the support vector machine algorithm is used to optimize the power allocation, determine whether the allocation result meets the requirements of the closed-loop process, and obtain an optimized power allocation scheme. By optimizing the power allocation scheme, updating the feedback data in the closed-loop process, and making real-time adjustments to the deviation between the feedback data and the power optimization target, the final energy network operation strategy is determined.

[0015] Compared with the prior art, the present invention has the following advantages and technical effects: This invention discloses a centimeter-level vehicle positioning and energy routing optimization control method based on the fusion of BeiDou and track-side inertial navigation. Addressing the interconnected challenges of high-precision vehicle positioning, dynamic energy allocation, and multi-timescale power scheduling in rail transit, the method acquires real-time vehicle location data and establishes a dynamic mapping relationship with energy routing nodes to achieve seamless switching of power supply channels. Simultaneously, it utilizes a Kalman filter algorithm to predict energy demand curves and combines this with a particle swarm optimization algorithm to adjust allocation strategy parameters, ensuring optimized power flow distribution across multiple time scales. Finally, a hierarchical architecture handles instantaneous power fluctuations, integrating short-term and long-term scheduling results to form a closed-loop feedback loop. This invention effectively solves the challenge of balancing vehicle positioning accuracy and energy allocation efficiency, significantly improving the energy utilization and operational stability of rail transit systems, and providing innovative technical support for green energy conservation in the field of intelligent transportation. Attached Figure Description

[0016] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart of a method according to an embodiment of the present invention. Detailed Implementation

[0017] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0018] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0019] Example 1 like Figure 1 As shown, this embodiment provides a power optimization method for rail vehicles based on energy routing control, including the following steps: Acquire centimeter-level positioning data of vehicles and establish a dynamic mapping relationship with energy routing nodes to obtain the power supply channel configuration corresponding to the current segment; The energy routing node switching point that the vehicle is about to approach is determined based on the dynamic mapping relationship. If the distance to the switching point is less than a preset threshold, the channel allocation status is updated. Based on the updated channel allocation status, predict the energy demand curve for future periods and generate a power demand sequence. For the power demand sequence, the particle swarm optimization algorithm is used to calculate the allocation strategy parameters. If the multi-time scale requirements are met, an energy routing allocation instruction is generated to obtain an optimized power flow distribution scheme. Based on the power flow distribution scheme, extract the hierarchical scheduling architecture reference value to obtain short-term scheduling results; Based on the short-term scheduling results, hourly baseline information is obtained by combining passenger flow forecast data. If the matching degree between the baseline information and the daily configuration is higher than a preset threshold, the daily baseline configuration scheme is updated to form a long-term energy network topology. Based on the location information and related data in the long-term energy network topology, the final energy routing control state is determined, realizing a closed-loop feedback loop for vehicle power optimization.

[0020] Specifically, the following steps are included: S101. By fusing BeiDou with track-side inertial navigation, centimeter-level vehicle positioning data is obtained from real-time signals, and a dynamic mapping relationship with energy routing nodes is established based on the data to obtain the power supply channel configuration corresponding to the current section.

[0021] Specifically, by combining BeiDou signals and inertial navigation, centimeter-level positioning information of the vehicle is extracted from real-time data to obtain preliminary vehicle position coordinates. Based on the preliminary vehicle position coordinates and pre-established energy routing node distribution data, the distance relationship between the vehicle and each node is calculated to determine the nearest energy routing node. If the determined nearest energy routing node is inconsistent with the node correspondence data of the current segment, the dynamic relationship is updated through real-time data to obtain the latest node correspondence information. For the latest node correspondence information, a preset power supply strategy database is queried to determine the applicable power supply scheme for the current segment and obtain the corresponding power allocation parameters. Based on the power allocation parameters and the channel configuration rules of the current segment, specific channel configuration instructions are generated to determine the final power supply channel configuration.

[0022] For example, in scenarios involving vehicle dynamic positioning and energy supply, the combination of BeiDou signals and inertial navigation is crucial for achieving centimeter-level positioning. BeiDou signals provide high-precision satellite positioning data, while inertial navigation uses accelerometers and gyroscopes to calculate the vehicle's displacement and orientation changes in real time. This combination compensates for the shortcomings of a single system; for instance, BeiDou signals may be lost in tunnels, but inertial navigation can temporarily maintain positioning accuracy. Assuming a vehicle is traveling on urban rail, the BeiDou system updates its coordinates in real time to 116.3 degrees east longitude and 39.9 degrees north latitude, while inertial navigation corrects for errors using the vehicle's acceleration and angular velocity, ultimately obtaining centimeter-level accurate coordinate data. This high-precision positioning lays the foundation for subsequent energy routing node matching.

[0023] Specifically, based on the initial location coordinates, the system calculates the distance between the vehicle and each node using pre-established energy routing node distribution data. Assuming the current vehicle's coordinates are 116.3 degrees East longitude and 39.9 degrees North latitude, and there are three nearby nodes located at coordinates A, B, and C, the system calculates the geometric distance and finds that node B is only 50 meters away from the vehicle, while the other nodes are more than 200 meters away. Therefore, node B is determined to be the closest node. This distance calculation ensures the efficiency of energy supply and reduces transmission losses.

[0024] For example, if the nearest node B is inconsistent with the preset node A for the current segment, the system will update the dynamic relationship through real-time data. Suppose a vehicle enters a new road segment, and node B is identified as the new corresponding node, the system will immediately update the node information for that segment in the database to ensure that subsequent power allocation is based on the latest data. This dynamic update mechanism can adapt to node changes caused by vehicle movement, improving the system's flexibility and response speed.

[0025] Specifically, based on the latest node information, the system queries the power supply strategy database to determine the appropriate solution for the current segment. Assuming node B is located on a highway, the database pre-determines that this segment requires high power supply to support fast charging, with a power allocation parameter of 80 kW output power, instead of the 50 kW required for ordinary road segments. This strategy matching ensures that power supply matches actual demand, avoiding resource waste.

[0026] For example, when generating channel configuration instructions, the system combines power allocation parameters and segment channel configuration rules to determine the final solution. Suppose the current segment rules require priority to dual-channel transmission to improve stability, the system will generate instructions to evenly distribute 80 kW of power across the two channels, 40 kW each. This configuration satisfies the power requirements while improving transmission reliability through multi-channel design.

[0027] Specifically, each step of the above method is closely linked, forming a complete technology chain from positioning and node matching to power allocation and channel configuration. High-precision positioning ensures accurate node selection, dynamic updates adapt to environmental changes, and strategy matching and channel configuration optimize resource utilization, ultimately achieving efficient and stable energy supply. This systematic design significantly improves the intelligence level of vehicle energy management, providing technical support for the future development of intelligent transportation.

[0028] S102. Based on the dynamic mapping relationship, determine the energy routing node switching point that is about to approach from the vehicle location information. If the distance of the switching point is less than a preset threshold, trigger the seamless switching command of the power supply channel to obtain the updated channel allocation status.

[0029] Specifically, by analyzing vehicle location information and dynamic mapping relationships, the relative positional relationship between the current vehicle and energy routing nodes is determined, identifying the node switching points that will be involved. For each identified switching point, the real-time distance between the vehicle and the switching point is calculated. If this distance is less than a preset threshold, a subsequent power supply adjustment process is triggered, generating a switching demand signal. Based on the switching demand signal, the current power supply status data is obtained, the feasibility of channel switching is analyzed, and a corresponding seamless instruction draft is generated. Using the seamless instruction draft and the distribution information of energy routing nodes, a suitable channel switching scheme is matched to determine the final instruction content. Based on the final instruction content, a channel switching control signal is sent to the power supply system to complete the real-time update of the allocation status, obtaining the updated channel allocation status. Based on the updated channel allocation status, the distance changes between the vehicle location and the energy routing nodes are continuously monitored. If a new switching demand is detected, the above process is repeated to maintain dynamic adaptation of channel allocation.

[0030] For example, in the interaction scenario between vehicles and energy routing nodes, obtaining vehicle location information is fundamental to the entire process. The vehicle updates its coordinate data in real time using a high-precision positioning system. Assuming the vehicle is currently located in a specific section of a track, its coordinates show it is approximately 200 meters away from an energy routing node ahead. Based on this location information, the system combines it with a pre-stored node distribution map to analyze the relative position of the vehicle to each node and determine whether the node ahead is the upcoming switching point. This method provides accurate spatial reference for subsequent decision-making.

[0031] For example, determining the node switching point can be based on the real-time distance between the vehicle and the node to decide whether a switch is needed. Assuming the system sets a distance threshold of 100 meters, when the distance between the vehicle and the node ahead gradually decreases from 200 meters to 90 meters, the system will trigger a switching request signal. This signal is generated based on continuous monitoring of distance changes, ensuring a timely response as the vehicle approaches the node, thus allowing sufficient time for power supply adjustments.

[0032] For example, when analyzing the feasibility of channel switching, the system considers the current power supply status data. Assuming the current channel's load rate is 70%, while the target channel's load rate is only 30%, the system will assess the feasibility of switching and generate a draft seamless instruction. This draft comprehensively considers channel load balancing and stability to ensure that the switching process does not affect the vehicle's energy supply.

[0033] For example, in the process of matching seamless instruction drafts, the system selects the optimal channel switching scheme based on the distribution information of energy routing nodes. Assuming there are three channels within the coverage area of ​​the forward node, the system will prioritize the channel with the lowest load and the most stable signal as the target channel and generate the final instruction content. This process effectively improves the smoothness of the switching process.

[0034] For example, when sending a channel switching control signal, the system issues a command to the power supply system via the communication interface. Assuming the command is to switch the vehicle from channel A to channel B, after the switch is completed, the system updates the allocation status in real time, recording the new channel information as channel B. This real-time update mechanism ensures that the system always has the latest allocation status.

[0035] For example, continuously monitoring changes in vehicle position and distance to nodes is key to dynamic adaptation. Suppose a vehicle switches to lane B and continues moving. If the system detects that the distance to the next node is only 150 meters and the distance is gradually decreasing, it will trigger a switching request signal again, repeating the above process. This continuous monitoring method maintains a dynamic match between lane allocation and vehicle position.

[0036] For example, establishing dynamic mapping relationships plays a crucial role throughout the entire process. Assuming the vehicle travels through different sections, the system updates the mapping data with nodes in real time based on location information, ensuring that each switch accurately corresponds to the appropriate node and channel. This dynamic mapping significantly improves the system's adaptability and response speed, providing stable and reliable energy support for the vehicle.

[0037] S103. Obtain the velocity vector and acceleration parameters from the channel allocation state, fuse the parameters using the Kalman filter algorithm, determine the energy demand curve for future periods, and obtain the predicted power demand sequence.

[0038] Specifically, velocity vectors and acceleration parameters are extracted from the channel allocation status. These parameters are then cleaned using pre-established filtering rules to obtain a refined parameter set. Based on this refined parameter set, a fusion algorithm is applied to process the velocity vectors and acceleration parameters to determine a corrected parameter combination. Using this corrected parameter combination and a preset threshold range, if any parameter exceeds the threshold range, boundary adjustments are performed to obtain an adjusted parameter mapping. Finally, time series analysis is used to segment the adjusted parameter mapping to obtain a predicted power demand sequence.

[0039] For example, in the field of vehicle energy management, extracting velocity vectors and acceleration parameters from channel allocation status is a crucial step. Velocity vectors typically reflect the vehicle's direction and speed of movement at a given moment, while acceleration parameters reflect the trend of velocity changes. Assuming a rail vehicle is traveling on a track, the system collects real-time velocity vectors of 30 km / h in the direction of true north, and acceleration parameters with a rate of change of 2 meters per second. Through pre-established filtering rules, such as removing outliers or noisy data, unreasonable parameter values ​​can be filtered out, resulting in a refined set of parameters, such as a velocity vector stable between 28 and 32 km / h.

[0040] Specifically, when processing the processed parameters using a fusion algorithm, the velocity vector and acceleration parameters can be combined with historical data for weighted analysis to determine the corrected parameter combination. For example, if the system detects significant fluctuations in the current acceleration parameter, and considers the data trend over the past 5 minutes, the corrected acceleration value is adjusted to 1.5 meters per second. This correction helps to more accurately reflect the vehicle's actual operating state. Next, considering a preset threshold range, such as acceleration not exceeding 3 meters per second, if the corrected parameters exceed this range, boundary adjustments are made. If, after a correction, the acceleration reaches 3.2 meters per second, the system will adjust it to 3 meters per second, forming an adjusted parameter mapping.

[0041] For example, when segmenting the adjusted parameter mapping, time series tools can divide the parameter mapping into multiple segments, each lasting 5 minutes, to predict future power demand sequences. Suppose that vehicle speed vectors show a continuous upward trend within a certain time period, the system predicts that power demand will increase by approximately 20% in the next 10 minutes, thus generating a corresponding demand sequence. This prediction can provide a reference for the allocation of energy routing nodes, ensuring a stable energy supply for vehicles on different road segments.

[0042] Specifically, each of the above steps is closely aligned with the core requirements of vehicle energy management. From parameter extraction and cleaning, the aim is to obtain a reliable data foundation; through fusion algorithms and correction processing, the accuracy of the parameters can be improved; boundary adjustments avoid the impact of abnormal parameters on the system; and the final power demand sequence prediction provides forward-looking guidance for subsequent channel allocation. This progressive approach not only improves the reliability of data processing but also lays the foundation for dynamic adaptation of energy supply, effectively ensuring the system's responsiveness, especially in scenarios with frequent vehicle speed changes or complex road conditions.

[0043] S104. For the power demand sequence, the particle swarm optimization algorithm is used to calculate the allocation strategy adjustment parameters, and it is determined whether the parameters meet the multi-time scale requirements. If they do, the adjusted energy routing allocation instruction is generated to obtain the optimized power flow distribution scheme.

[0044] Specifically, the process includes: Step 1: Acquire power demand sequence data, segment the sequence using a pre-established data processing module, and divide it into multiple time-scale subsequences to obtain segmented demand data. Step 2: For the segmented demand data, use the particle swarm optimization algorithm to iteratively calculate the allocation strategy parameters for each time-scale to determine the initial parameter configuration for each time-scale. Step 3: Based on the initial parameter configuration, determine whether it meets the preset time-scale constraints. If the parameters for a certain time-scale do not meet the constraints, recalculate the parameters for that time-scale to obtain an adjusted parameter set. Step 4: Generate corresponding energy routing allocation instructions using the adjusted parameter set, match the corresponding routing path to the demand data for each time-scale, and determine the specific content of the allocation instructions. Step 5: Calculate the initial power flow distribution scheme based on the allocation instructions, simulate the power flow distribution using a preset simulation tool, and obtain distribution result data. Step 6: For the distribution result data, determine whether there are local imbalances. If so, fine-tune the allocation instructions for the relevant time-scales to obtain the final optimized power flow distribution scheme.

[0045] For example, when processing power demand sequence data, the raw data can first be segmented using a data processing module. Assuming a system collects power demand data continuously for 24 hours daily, this data can be divided into six sub-sequences, each lasting four hours. This segmentation method facilitates targeted analysis of demand fluctuations across different time periods.

[0046] It should be noted that the scale of the segmentation can be adjusted according to actual business needs. For example, during peak periods, a more granular time division may be needed to capture short-term fluctuations.

[0047] For example, when using particle swarm optimization (PSO) for parameter iteration on segmented demand data, we can assume that multiple allocation strategy parameters need to be optimized within a given time period. Suppose the initial parameter configuration for a certain time period is a power allocation ratio of 60% and 40%. After algorithm iteration, this might be adjusted to 55% and 45% to better reflect actual needs. This iterative process can gradually approach the optimal solution, especially effective in cases of multi-parameter coupling.

[0048] For example, when determining whether the initial parameter configuration meets the time-scale constraints, a constraint range can be set, such as the power allocation ratio during a certain time period must not be lower than 50%. If the calculated result is 48%, it needs to be readjusted, perhaps by increasing the resource input during that time period, eventually adjusting it to 52%. This approach ensures the rationality of the parameter configuration and avoids resource waste caused by parameter misalignment.

[0049] For example, when generating energy routing and allocation instructions, specific routing paths can be matched for each time period based on the adjusted parameter set. If demand is high during a certain time period, the system will prioritize paths with higher transmission efficiency to ensure timely energy delivery. This matching process effectively improves resource utilization while reducing transmission losses.

[0050] For example, when calculating and simulating the initial power flow distribution scheme, simulation tools can be used to model the power flow distribution over a certain period. Assuming a path has a load rate of 80%, simulation can reveal potential overload risks, and the system will issue an early warning. This simulation process helps identify potential problems and provides a basis for subsequent optimization.

[0051] For example, when identifying local imbalances in power flow distribution data, if the power flow distribution over a certain period shows that a node has a load rate as high as 90%, while adjacent nodes have a load rate of only 30%, then the distribution instructions need to be fine-tuned. This might involve transferring some of the load to the low-load node, ultimately achieving load balancing. This fine-tuning mechanism can effectively avoid local overload and improve the overall stability of the system.

[0052] It should be noted that the above-mentioned steps are closely interconnected, forming a complete closed-loop process from data analysis to the final optimization solution. Detailed processing at each stage provides a reliable foundation for subsequent steps. Especially in the fields of power demand forecasting and energy allocation, this systematic approach can significantly improve resource allocation efficiency and ensure the stability of system operation.

[0053] S105. Obtain the reference value of the hierarchical scheduling frame from the power flow distribution scheme, process the instantaneous power wave through the multi-time-scale hierarchical architecture, determine the control command sequence from millisecond to minute level, and obtain the coordinated short-term scheduling result.

[0054] Specifically, initial power grid operating status information is obtained through power flow distribution data sources. To meet the requirements of hierarchical scheduling, multi-level scheduling framework data is decomposed to obtain a preliminary hierarchical scheduling benchmark. Based on the hierarchical scheduling benchmark and time scale division, a hierarchical architecture model is constructed. Real-time data acquisition is performed to collect dynamic changes in instantaneous power and determine the fluctuation characteristics of power data. Fluctuation analysis is performed based on these characteristics, using a preset threshold range. If the fluctuation value exceeds the threshold range, a corresponding adjustment signal is generated to identify the scheduling nodes requiring intervention. Based on the intervention requirements of the scheduling nodes, millisecond-level control commands are generated to respond to rapid changes in instantaneous power, resulting in a high-precision real-time control sequence. This real-time control sequence is extended to minute-level control command generation. Combining the periodic requirements of short-term scheduling, cross-time scale command combinations are determined. The results of the command combinations are obtained, and based on the coordination requirements, global power flow distribution data and local hierarchical scheduling data are integrated to obtain the final short-term scheduling coordination scheme.

[0055] For example, in acquiring power grid operation status information, basic parameters such as voltage and current of the power grid can be collected in real time through power flow distribution data sources to construct a dynamic database that comprehensively reflects the power grid's operating status. Suppose that the instantaneous power data of a certain regional power grid shows 500 MW during peak hours and drops to 200 MW during off-peak hours; this fluctuation directly reflects the instability of the power grid's operating status. Preliminary analysis of this data can provide data support for subsequent hierarchical dispatching.

[0056] For example, to decompose the demand for tiered dispatching, the power grid can be divided into three levels: provincial, municipal, and regional, each corresponding to different dispatching authority and scope. At the provincial level, the dispatching framework may focus on cross-regional power balance, while at the municipal level, it focuses on the allocation of local load. Suppose a municipal power grid needs to import an additional 100 megawatts of power during peak hours; this demand will be fed back upwards through the tiered dispatching benchmark to ensure the rationality of overall resource allocation.

[0057] For example, when constructing a hierarchical architecture model, the time scale can be divided into three levels: milliseconds, minutes, and hours, enabling real-time responses to instantaneous power changes. Suppose a millisecond-level data acquisition point detects a sudden power surge to 550 megawatts; the system will immediately analyze its fluctuation characteristics to determine if it is an anomaly. This hierarchical model can effectively handle dynamic changes at different time scales.

[0058] For example, in analyzing power fluctuation characteristics, a threshold range can be preset, such as power fluctuations not exceeding ±10% being considered normal. If the power fluctuation reaches 15% within a certain time period, the system will generate an adjustment signal to identify the scheduling nodes requiring intervention. This analytical approach helps to quickly locate problem areas.

[0059] For example, when generating millisecond-level control commands, if a node experiences a sudden power surge, the system will issue a command within one millisecond to adjust the operating status of the relevant equipment. This rapid response mechanism can effectively prevent local grid overload.

[0060] For example, when extending to minute-level control commands, periodic command combinations can be generated by combining short-term scheduling needs. Suppose that the load in a certain area is expected to increase by 20% in the next 30 minutes, the system will allocate resources in advance to form command combinations across time scales, ensuring the continuity of power supply.

[0061] For example, when integrating global power flow distribution data with hierarchical scheduling local data, provincial scheduling instructions can be matched with municipal-level execution data based on the requirements of the coordination results. Suppose a provincial instruction requires a 50 MW load reduction; the municipal-level dispatcher will refine the allocation plan based on local conditions, ultimately forming a short-term scheduling coordination plan. This approach enables effective coordination between global and local data.

[0062] S106. Based on the short-term scheduling results, obtain hourly baseline information from the passenger flow forecast data, determine the matching degree between the baseline information and the all-day configuration, and if the matching degree is higher than a preset threshold, update the all-day baseline configuration scheme to obtain the integrated long-term energy network topology.

[0063] Specifically, the process includes: Step 1: Based on the short-term scheduling results, obtain hourly passenger flow forecast data from the passenger flow forecasting system. Divide the forecast data into time periods, extract peak and trough traffic information for each time period, and determine hourly baseline information. Step 2: Based on the hourly baseline information, obtain resource allocation data from the all-day configuration plan. Analyze the correspondence between the baseline information and the resource allocation data over time periods and calculate the degree of matching between them. Step 3: Compare the degree of matching with a preset threshold. If the degree of matching is higher than the preset threshold, trigger the configuration plan update process and obtain the updated all-day configuration plan data. Step 4: Using the updated all-day configuration plan data and combining it with long-term network planning requirements, extract network resource distribution information and determine the preliminary topology of the long-term energy network. Step 5: Based on the preliminary topology, obtain energy allocation data for each node in the network. Analyze the energy flow paths between nodes to determine if there is uneven resource allocation. If uneven allocation exists, adjust the path allocation ratio to obtain an optimized topology. Step Six: For the optimized topology, a support vector machine model is used to predict and analyze the energy flow paths, obtain the energy demand distribution in the future time period, and determine the final long-term energy network topology.

[0064] For example, when combining short-term scheduling results with passenger flow forecast data, one can start with hourly forecast data to analyze the peak passenger flow during the morning rush hour in a certain power grid area. Assuming a predicted passenger flow of 5,000 people per hour during a certain time period and 1,000 people per hour during off-peak periods, this time period division clarifies the need for a higher resource allocation benchmark during the morning rush hour. This benchmark information provides a basic reference for subsequent resource allocation, ensuring that the scheduling plan aligns with actual needs.

[0065] In one possible implementation, by analyzing the matching degree between hourly baseline information and the all-day configuration plan, the capacity configuration for the morning peak period in the resource allocation data can be set to a capacity of 6,000 passengers per hour. By comparing the baseline information and the configuration data, a high degree of matching is found, but some redundancy still exists. Assuming the preset threshold is 90% and the actual matching degree is 85%, the plan does not need to be updated immediately, but this data can be recorded for subsequent optimization to ensure efficient resource utilization.

[0066] For example, regarding the configuration update process when the matching degree exceeds a threshold, we can envision a day when the matching degree reaches 95%, triggering the update process. The system will then adjust the morning peak capacity configuration to 5,500 passengers per hour based on the latest passenger flow forecast data, reducing resource waste. This dynamic adjustment helps improve the adaptability of the scheduling scheme and ensures operational efficiency.

[0067] In one possible implementation, when extracting resource distribution information in conjunction with long-term network planning requirements, the energy demand distribution in different regions of the power grid can be analyzed. Assuming the core area has a daily demand of 1 million kilowatt-hours and the peripheral area has 300,000 kilowatt-hours, the initial topology will prioritize energy supply to the core area. This design helps to form a rational network layout, laying the foundation for long-term planning.

[0068] For example, analyzing and adjusting energy flow paths between nodes can reveal that the energy transmission ratio from the core area to the edge area is too high on a certain path, causing some nodes to be overloaded. After adjustment, the transmission ratio is reduced from 70% to 50%, and backup paths are added to share the traffic, ensuring balanced resource allocation. This optimization method can improve network stability.

[0069] In one possible implementation, when using a support vector machine (SVM) model to predict energy demand distribution, the energy demand trends of each node over the next 24 hours can be analyzed based on historical data. For example, if the prediction shows that a node's demand will increase by 20% during the evening peak period, the final topology will allocate more resources to it in advance. This predictive analysis helps improve the network's responsiveness to future demands and ensures long-term operational reliability. Through the above multi-faceted analysis and examples, a logically rigorous implementation path has been formed for each stage, from short-term scheduling to long-term planning, ensuring the feasibility and adaptability of the solution. Furthermore, through dynamic adjustment and predictive analysis, the accuracy of resource allocation and the stability of network operation are significantly improved.

[0070] S107. Obtain location information data from the long-term energy network topology, fuse the data through a real-time synchronization mechanism, determine the final energy routing control state, and obtain a closed-loop feedback loop for vehicle power optimization.

[0071] Specifically, by acquiring topology data from the long-term energy network, analyzing its distribution characteristics, and determining the connection relationships of key nodes, preliminary network layout information is obtained. Based on the preliminary network layout information, corresponding location information is extracted, and dynamically updated relevant data is acquired using a real-time synchronization mechanism. Data integrity is assessed; if data loss exceeds a preset threshold, a supplementary acquisition process is triggered to obtain a complete dataset. For the complete dataset, an integration mechanism is used to classify the data, separating the parts directly related to routing control. If the separated data does not conform to preset standards, the classification rules are readjusted to determine the data subset suitable for routing control. From the routing control data subset, real-time changes in the control state are analyzed. If an anomaly is detected, it is corrected using pre-established mapping rules to obtain stable control state parameters. With stable control state parameters obtained, combined with the dynamic power requirements of vehicles, a support vector machine algorithm is used to optimize power allocation, determining whether the allocation result meets the requirements of the closed-loop process, and obtaining an optimized power allocation scheme. Using the optimized power allocation scheme, the feedback data in the closed-loop process is updated, and real-time adjustments are made based on the deviation between the feedback data and the power optimization target to determine the final energy network operation strategy.

[0072] For example, when acquiring topology data and analyzing distribution characteristics from long-term energy networks, one can begin by examining the overall network layout to identify which nodes act as hubs in energy transmission. Suppose that in an urban rail transit energy network, the central station transmits an average of 5000 kWh of energy per day, while the surrounding stations transmit only 1000 kWh. This distribution characteristic indicates that the central station is a critical node. Analyzing its connectivity reveals that the central station and surrounding stations form a star-shaped connection structure, thus providing a preliminary understanding of the network layout. This approach helps to clarify the main energy flow paths, providing a foundation for subsequent optimization.

[0073] For example, when extracting location information and acquiring dynamic data using a real-time synchronization mechanism, energy consumption data can be collected in real time by sensors deployed at various sites. Assuming a site updates data every 5 minutes during peak hours, if three consecutive data gaps are detected, exceeding a preset threshold of two, a supplementary data collection process is triggered, supplementing the data using backup data sources or historical data. This method ensures data integrity and provides reliable support for subsequent analysis.

[0074] For example, when classifying and separating routing control-related data from a complete dataset, the data can be divided into two categories: energy allocation and routing control. If, in the separated routing control data, the latency of a certain path exceeds the standard value of 10 milliseconds, the classification rules need to be adjusted to re-select the data subset with lower latency. This adjustment helps ensure the accuracy of routing control.

[0075] For example, when analyzing real-time changes in control status and correcting anomalies, if the control status of a certain node shows energy output fluctuations exceeding 20%, the output value can be corrected to a stable range using a preset mapping rule. This correction mechanism can effectively avoid sudden problems during system operation.

[0076] For example, in power allocation optimization calculations, combined with dynamic vehicle demand, the power demand during peak hours can be analyzed using the support vector machine algorithm. Assuming a certain route has a peak demand of 3000 kWh, the optimized allocation scheme will distribute power evenly across all nodes, meeting the closed-loop process requirements. This optimization method can improve energy utilization efficiency.

[0077] For example, when updating closed-loop process feedback data and adjusting deviations, if the feedback data shows that the power allocation of a certain node deviates from the target by 15%, the allocation ratio will be adjusted in real time to within the target range. This dynamic adjustment mechanism ensures the stability of the energy network operation strategy, thereby improving overall operating efficiency.

[0078] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A power optimization method for rail vehicles based on energy routing control, characterized in that, Includes the following steps: Acquire centimeter-level positioning data of vehicles and establish a dynamic mapping relationship with energy routing nodes to obtain the power supply channel configuration corresponding to the current segment; The energy routing node switching point that the vehicle is about to approach is determined based on the dynamic mapping relationship. If the distance to the switching point is less than a preset threshold, the channel allocation status is updated. Based on the updated channel allocation status, predict the energy demand curve for future periods and generate a power demand sequence. For the power demand sequence, the particle swarm optimization algorithm is used to calculate the allocation strategy parameters. If the multi-time scale requirements are met, an energy routing allocation instruction is generated to obtain an optimized power flow distribution scheme. Based on the power flow distribution scheme, extract the hierarchical scheduling architecture reference value to obtain short-term scheduling results; Based on the short-term scheduling results, hourly baseline information is obtained by combining passenger flow forecast data. If the matching degree between the baseline information and the daily configuration is higher than a preset threshold, the daily baseline configuration scheme is updated to form a long-term energy network topology. Based on the location information and related data in the long-term energy network topology, the final energy routing control state is determined, realizing a closed-loop feedback loop for vehicle power optimization.

2. The method according to claim 1, characterized in that, The power supply channel configuration corresponding to the current segment includes: By combining BeiDou signals and inertial navigation, centimeter-level positioning information of the vehicle is extracted from real-time data to obtain preliminary vehicle position coordinates. Based on the preliminary vehicle location coordinates and combined with the pre-established energy routing node distribution data, the distance relationship between the vehicle and each node is calculated, and the nearest energy routing node is determined. If the identified nearest energy routing node does not match the node corresponding data of the current segment, the latest node corresponding information is obtained by updating the dynamic relationship through real-time data. Based on the latest node information, the preset power supply strategy database is queried to determine the applicable power supply scheme for the current segment and obtain the corresponding power allocation parameters. Based on the power allocation parameters and the channel configuration rules of the current segment, specific channel configuration instructions are generated to determine the final power supply channel configuration.

3. The method according to claim 1, characterized in that, Update channel allocation status includes: By analyzing the relative positional relationship between the current vehicle and the energy routing node using the vehicle location information and the dynamic mapping relationship, the node switching point that will be involved is determined. For a given node switching point, calculate the real-time distance between the vehicle and the switching point. If the distance is less than a preset threshold, trigger the subsequent power supply adjustment process to obtain a switching demand signal. Based on the switching demand signal, obtain the current power supply status data, analyze the feasibility of channel switching, and generate a corresponding seamless switching instruction draft. Using the aforementioned seamless switching instruction draft, combined with the distribution information of energy routing nodes, a suitable channel switching scheme is matched to determine the final instruction content; Based on the final instruction content, a channel switching control signal is sent to the power supply system to complete the real-time update of the allocation status and obtain the updated channel allocation status. Based on the updated channel allocation status, the distance between the vehicle location and the energy routing node is continuously monitored. If a new switching requirement is detected, the above process is executed repeatedly to maintain dynamic adaptation of channel allocation.

4. The method according to claim 1, characterized in that, The generated power demand sequence includes: The velocity vector and acceleration parameters are extracted from the channel allocation state, and the parameters are cleaned using pre-established filtering rules to obtain a cleaned parameter set. Based on the sorted parameter set, the velocity vector and acceleration parameters are processed using a fusion algorithm to determine the corrected parameter combination; By combining the corrected parameter combination with a preset threshold range, if the parameter exceeds the threshold range, boundary adjustment is performed to obtain the adjusted parameter mapping. For the adjusted parameter mapping, a time series tool is used to segment the mapping to obtain the predicted power demand sequence.

5. The method according to claim 1, characterized in that, The optimized power flow distribution schemes include: The power demand sequence data is obtained, and the sequence is segmented using a pre-established data processing module. The sequence is divided into multiple time periods according to the time scale to obtain the segmented demand data. For the segmented demand data, the particle swarm optimization algorithm is used to iteratively calculate the allocation strategy parameters for each time period to determine the initial parameter configuration for each time period. Based on the initial parameter configuration, it is determined whether the preset time scale constraints are met. If the parameters for a certain time period do not meet the constraints, the parameters for that time period are recalculated to obtain the adjusted parameter set. The adjusted parameter set is used to generate corresponding energy routing allocation instructions, match the corresponding routing path for the demand data of each time period, and determine the specific content of the allocation instructions. According to the allocation instructions, calculate the initial scheme of power flow distribution, use a preset simulation tool to simulate the power flow distribution, and obtain distribution result data; Based on the distribution results data, it is determined whether there is a local imbalance. If so, the allocation instructions for the relevant time period are fine-tuned to obtain the final optimized power flow distribution scheme.

6. The method according to claim 1, characterized in that, Obtaining short-term scheduling results includes: By using power flow distribution data sources, initial power grid operation status information is obtained. Based on the hierarchical scheduling requirements, multi-level scheduling framework data is decomposed to obtain a preliminary hierarchical scheduling benchmark. Based on the hierarchical scheduling benchmark and the division of time scales, a hierarchical architecture model is constructed to collect data in real time on the dynamic changes of instantaneous power and determine the fluctuation characteristics of power data. By analyzing the fluctuation characteristics of the power data, fluctuation analysis is performed. If a preset threshold range is used, and the fluctuation value exceeds the threshold range, a corresponding adjustment signal is generated to determine the scheduling node that needs intervention. Based on the intervention requirements of the scheduling node, millisecond-level control commands are generated to respond to rapid changes in instantaneous power and obtain a high-precision real-time control sequence. Through the aforementioned real-time control sequence, control command generation is extended to the minute level, and combined with the periodic requirements of short-term scheduling, instruction combinations across time scales are determined. The results of the command combination are obtained, and the global data of power flow distribution and the local data of hierarchical scheduling are integrated according to the requirements of the coordination results to obtain the final short-term scheduling coordination scheme.

7. The method according to claim 1, characterized in that, The formation of long-term energy network topology includes: Based on the short-term scheduling results, hourly passenger flow forecast data is obtained from the passenger flow forecast system. The forecast data is divided into time periods, and the peak and trough information of traffic flow in each time period is extracted to determine the hourly baseline information. Based on the hourly baseline information, obtain the resource allocation data in the all-day configuration plan, analyze the correspondence between the baseline information and the resource allocation data in the time period, and calculate the degree of matching between the two. If the matching degree is higher than the preset threshold, the configuration scheme update process is triggered to obtain the updated configuration scheme data for the whole day. By using the updated all-day configuration scheme data and combining it with long-term network planning requirements, network resource distribution information is extracted to determine the preliminary topology of the long-term energy network. Based on the preliminary topology, energy distribution data of each node in the network is obtained, energy flow paths between nodes are analyzed, and it is determined whether there is uneven resource distribution. If uneven distribution exists, the path distribution ratio is adjusted to obtain an optimized topology. For the optimized topology, a support vector machine model is used to predict and analyze the energy flow path, obtain the energy demand distribution in the future time period, and determine the final long-term energy network topology.

8. The method according to claim 1, characterized in that, The closed-loop feedback loop for optimizing vehicle power includes: By obtaining topological data from long-term energy networks, analyzing their distribution characteristics, determining the connection relationships of key nodes, and obtaining preliminary network layout information; Based on the preliminary network layout information, the corresponding location information is extracted, and the relevant dynamically updated data is obtained in combination with the real-time synchronization mechanism. The data integrity is judged. If the data missing exceeds the preset threshold, the supplementary collection process is triggered to obtain the complete dataset. For the complete dataset, an integration mechanism is used to classify the data and separate the part that is directly related to routing control. If the separated data does not conform to the preset standard, the classification rules are readjusted to determine the data subset suitable for routing control. From the data subset of routing control, analyze the real-time changes in control state. If an abnormal state is detected, correct it through pre-established mapping rules to obtain stable control state parameters. By obtaining stable control state parameters and combining them with the dynamic power requirements of the vehicle, the support vector machine algorithm is used to optimize the power allocation, determine whether the allocation result meets the requirements of the closed-loop process, and obtain an optimized power allocation scheme. By optimizing the power allocation scheme, updating the feedback data in the closed-loop process, and making real-time adjustments to the deviation between the feedback data and the power optimization target, the final energy network operation strategy is determined.

Images