A method and system for energy regulation of new energy vehicles

Abstract

This invention discloses a method and system for energy regulation of new energy vehicles, relating to the field of energy regulation technology. The method includes: deploying edge computing units in a preset area of ​​the new energy vehicle to construct a distributed edge computing network and obtain energy data for each preset area; preprocessing the energy data and, in conjunction with vehicle operating conditions, performing local energy regulation to obtain a local energy regulation strategy; based on the local energy regulation strategy, and according to the state space and action space of the new energy vehicle, performing reinforcement learning and iterative optimization to obtain a global energy regulation strategy with the goal of maximizing energy utilization; and adjusting the parameters of the distributed edge computing network and the global energy regulation unit through energy output and energy feedback to correct energy deviations. This invention can solve the technical problem in the prior art where energy is supplemented by a single and independent energy source, lacking cross-module data fusion capabilities, which makes it difficult to globally optimize the energy allocation strategy.

Description

Technical Field

[0001] This invention relates to the field of energy regulation technology, specifically to a method and system for energy regulation of new energy vehicles. Background Technology

[0002] In recent years, the global energy crisis and environmental issues have driven the rapid development of the new energy vehicle industry. Pure electric and hybrid vehicles, with their zero / low emission characteristics, have become the core direction for low-carbon transformation in the transportation sector. However, range anxiety and energy efficiency bottlenecks remain pain points for the industry. Data shows that the current mainstream new energy vehicle range cannot meet the actual needs of users, and the energy efficiency of new energy vehicles degrades significantly under extreme conditions (such as low temperature and high speed).

[0003] However, the energy management technology of new energy vehicles uses a single and independent energy source to replenish energy, lacking cross-module data fusion capabilities, which makes it difficult to optimize energy distribution strategies globally. For example, during regenerative braking, the recovery intensity is not dynamically adjusted based on the battery's state of charge (SOC), often resulting in battery overcharging or low recovery efficiency. Summary of the Invention

[0004] In view of the shortcomings of the prior art, the purpose of this invention is to provide a method and system for energy regulation of new energy vehicles, which aims to solve the technical problem that the existing technology relies on a single and independent energy source to supplement energy, lacks cross-module data fusion capabilities, and makes it difficult to optimize the energy allocation strategy globally.

[0005] This invention provides a method for regulating the energy of new energy vehicles, the method comprising:

[0006] Edge computing units are deployed in the designated areas of new energy vehicles, and a distributed edge computing network is constructed by combining the various edge computing units to obtain energy data for each designated area.

[0007] The energy data is preprocessed and combined with the vehicle operating conditions, then input to the local energy control unit for local energy control to obtain a local energy control strategy.

[0008] Based on the local energy regulation strategy, and taking the state space and action space of the new energy vehicle as the goal of maximizing energy utilization, the global energy regulation strategy is obtained through reinforcement learning by the global energy regulation unit and iterative optimization.

[0009] The global energy regulation strategy is distributed to each energy execution unit. Through energy output and energy feedback, the parameters of the distributed edge computing network and the global energy regulation unit are adjusted to correct energy deviations.

[0010] Compared with the prior art, the beneficial effects of the present invention are as follows: By deploying edge computing units in a preset area of ​​the new energy vehicle to collect energy data from different preset areas, and then achieving rapid response and global optimization through the collaboration of local energy control units and global energy control units, energy utilization is maximized and energy utilization efficiency is significantly improved. Furthermore, closed-loop control is formed through energy output and energy feedback, and execution errors are eliminated through energy deviation correction to ensure energy control accuracy. This solves the technical problem that the energy allocation strategy is difficult to optimize globally due to the common reliance on a single and independent energy source for energy replenishment and the lack of cross-module data fusion capabilities.

[0011] According to one aspect of the above technical solution, the steps of deploying edge computing units in a predetermined area of ​​the new energy vehicle, constructing a distributed edge computing network by combining the various edge computing units, and obtaining energy data for each predetermined area specifically include:

[0012] Edge computing units are deployed in a predetermined area of ​​the new energy vehicle, the predetermined area including a solar photovoltaic panel area, a braking system area, a thermoelectric conversion area, and an environmental energy capture area;

[0013] By combining various edge computing units to construct a distributed edge computing network, energy data for each preset area can be obtained.

[0014] Among them, energy data is acquired for each edge computing unit at a first preset time interval;

[0015] When the energy data fluctuations of each edge computing unit reach the fluctuation threshold, energy data is acquired every second preset time interval, triggering an anomaly detection algorithm to detect anomalies, and the data is prioritized for uploading to the local energy control unit.

[0016] The first preset time is greater than the second preset time.

[0017] According to one aspect of the above technical solution, the step of preprocessing the energy data specifically includes:

[0018] The energy data is filtered for outliers using the 3σ principle to determine whether the energy data in the preset area is dynamic data;

[0019] If so, the interactive window normalization algorithm is used for preprocessing;

[0020] If not, then a linear normalization algorithm is used for preprocessing.

[0021] According to one aspect of the above technical solution, the local energy control method of the local energy control unit includes:

[0022] When the energy data acquired by the braking system area exceeds the braking energy threshold, the local energy regulation unit preferentially stores the energy data in the battery. The braking energy threshold is calculated as follows:

[0023] ,

[0024] in, This represents the maximum energy that a battery can store. The braking energy threshold, It is the percentage coefficient, and It is related to the current state of battery charge;

[0025] When the photovoltaic power of the solar photovoltaic panel area is greater than the battery charging power threshold, the local energy control unit will supply the energy data of the solar photovoltaic panel area to the drive motor.

[0026] Under initial operating conditions, when the battery state of charge is less than the preset state of charge and the photovoltaic power is greater than the preset power, the local energy control unit will prioritize the activation of the energy data-assisted drive in the solar photovoltaic panel area.

[0027] During deceleration, when the braking intensity is greater than the preset braking intensity, the local energy control unit will preferentially store the energy data of the braking system area into the battery.

[0028] According to one aspect of the above technical solution, based on the local energy regulation strategy, and taking the maximization of energy utilization as the objective, the steps of deriving the global energy regulation strategy through reinforcement learning by a global energy regulation unit and iterative optimization, specifically include:

[0029] The state space includes vehicle operating condition information, energy states of various preset areas, and environmental parameters; the action space is the allocation ratio of each preset area.

[0030] Among these methods, based on the vehicle's basic information, a decision tree algorithm is used to determine the current vehicle operating condition.

[0031] Based on the current allocation ratio, the allocation ratio is predicted using the Long Short-Term Memory Network (LSTM).

[0032] Based on the local energy regulation strategy, according to the current state space and predicted action space of the new energy vehicle, with the goal of maximizing energy utilization, the global energy regulation unit performs reinforcement learning, and then the Q-learning algorithm is used to iteratively optimize and obtain the global energy regulation strategy, thus obtaining the optimal allocation ratio.

[0033] According to one aspect of the above technical solution, the function constructed with the goal of maximizing energy utilization is:

[0034] ,

[0035] in, To maximize energy efficiency, For energy efficiency, For the overall energy efficiency of the vehicle, The balance of the battery's state of charge in a vehicle. For energy loss costs, , , These are the weighting coefficients. For the first Discount factor of time, =0, 1, ..., .

[0036] According to one aspect of the above technical solution, the step of distributing the global energy regulation strategy to each energy execution unit and adjusting the parameters of the distributed edge computing network and the global energy regulation unit through energy output and energy feedback to correct energy deviation specifically includes:

[0037] The global energy control strategy is distributed to each energy execution unit, and the deviation type is determined through energy output and energy feedback. The deviation type includes system deviation and instantaneous deviation.

[0038] When the deviation type is a system deviation, it is uploaded to the vehicle cloud, and new parameters are generated through big data analysis to replace the parameters of the distributed edge computing network and the global energy control unit.

[0039] When the deviation type is instantaneous deviation, the current deviation is obtained, and the parameters of the distributed edge computing network and the global energy control unit are dynamically adjusted according to the PID algorithm.

[0040] According to one aspect of the above technical solution, the calculation formula of the PID algorithm is as follows:

[0041] ,

[0042] in, For correction amount, For the current deviation, This is the proportionality coefficient. The integral coefficient is... The differential coefficients are...

[0043] The correction amount is used to adjust the parameters of the distributed edge computing network and the global energy control unit.

[0044] The proportional coefficient, integral coefficient, and derivative coefficient are automatically adjusted according to the vehicle's operating conditions.

[0045] According to one aspect of the above technical solution, the method further includes:

[0046] Based on the vehicle's operating conditions, set the energy priority weights for each preset area;

[0047] Based on the energy priority weights and the energy conversion efficiency and power of each preset region, a local energy regulation strategy is determined, calculated as follows:

[0048] ,

[0049] in, For the first The energy allocation ratio of each preset area For the first Energy priority weights for each preset region For the first Energy conversion efficiency of a preset area For the first Power of each preset area This represents the total number of preset regions.

[0050] Another aspect of the present invention is to provide an energy regulation system for new energy vehicles, used to implement the above-mentioned energy regulation method for new energy vehicles, the system comprising:

[0051] The computing network construction module is used to deploy edge computing units in a preset area of ​​new energy vehicles, and to build a distributed edge computing network by combining the various edge computing units to obtain energy data for each preset area.

[0052] The local control module is used to preprocess the energy data and, in conjunction with the vehicle operating conditions, input it to the local energy control unit for local energy control to obtain a local energy control strategy.

[0053] The global control module is used to derive the global energy control strategy based on the local energy control strategy, according to the state space and action space of the new energy vehicle, with the goal of maximizing energy utilization, through reinforcement learning by the global energy control unit and iterative optimization.

[0054] The deviation correction module is used to distribute the global energy regulation strategy to each energy execution unit and adjust the parameters of the distributed edge computing network and the global energy regulation unit through energy output and energy feedback to correct the energy deviation. Attached Figure Description

[0055] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0056] Figure 1 This is a flowchart of the energy regulation method for new energy vehicles in Embodiment 1 of the present invention;

[0057] Figure 2 This is a structural block diagram of the new energy vehicle energy regulation system in Embodiment 2 of the present invention;

[0058] Explanation of symbols in the attached drawings:

[0059] The system includes a network construction module 100, a local control module 200, a global control module 300, and a deviation correction module 400. Detailed Implementation

[0060] To make the objectives, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Several embodiments of the present invention are shown in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that the disclosure of the present invention will be more thorough and complete.

[0061] Example 1

[0062] Please see Figure 1 The image shows a method for regulating the energy of a new energy vehicle according to the first embodiment of the present invention, the method comprising steps S10-S13:

[0063] Step S10: Deploy edge computing units in the preset area of ​​the new energy vehicle, and build a distributed edge computing network by combining the various edge computing units to obtain energy data for each preset area;

[0064] Among them, edge computing units are deployed in a preset area of ​​the new energy vehicle, the preset area including a solar photovoltaic panel area, a braking system area, a thermoelectric conversion area, and an environmental energy capture area;

[0065] By combining various edge computing units to construct a distributed edge computing network, energy data for each preset area can be obtained.

[0066] Among them, energy data is acquired for each edge computing unit at a first preset time interval;

[0067] When the energy data fluctuations of each edge computing unit reach the fluctuation threshold, energy data is acquired every second preset time interval, triggering an anomaly detection algorithm to detect anomalies, and the data is prioritized for uploading to the local energy control unit.

[0068] The first preset time is greater than the second preset time.

[0069] As an example, not a limitation, multiple types of fusion sensors are deployed at key locations in new energy vehicles (such as the roof, body, and braking system) to detect multi-dimensional energy data in real time, including solar photovoltaic output, regenerative braking energy, thermal energy conversion efficiency, and environmental energy.

[0070] For example, the edge computing unit in the solar photovoltaic panel area calculates the maximum power point tracking parameters of the solar photovoltaic panel to obtain the maximum photovoltaic power generation efficiency. The first preset time is 100ms, and data is collected from each edge computing unit every 100ms. When the fluctuation of the energy data of each edge computing unit reaches the fluctuation threshold, such as when the fluctuation exceeds 20%, energy data is collected every second preset time, which can be 50ms, and an anomaly detection algorithm is triggered to detect anomalies.

[0071] Furthermore, an anomaly detection algorithm using isolated forest (or adaptive threshold algorithm) is used to identify anomalies in the high-frequency energy data acquired at each second preset time interval, and the anomaly type (such as sudden rise / fall, continuous over-limit) and confidence level are output.

[0072] If an anomaly is detected, the edge computing unit immediately uploads it to the local energy control unit. The local energy control unit generates coarse control instructions (such as storing the photovoltaic energy in the battery or limiting the output when the photovoltaic energy exceeds the threshold), with a response time of ≤10ms, and simultaneously uploads the abnormal event and preliminary processing results to the global control unit.

[0073] As an example, not a limitation, when energy data fluctuations in the solar photovoltaic panel area trigger an anomaly, if the battery state of charge is less than 80%, the energy will be stored in the battery first. The duty cycle of the charging module will be adjusted through the PWM signal to limit the charging power increase to ≤10% / 50ms. When energy data fluctuations in the braking system area trigger an anomaly, the abnormal energy will be temporarily stored in the supercapacitor to prevent the battery from overcharging.

[0074] Step S11: The energy data is preprocessed and combined with the vehicle operating conditions, then input to the local energy control unit for local energy control to obtain a local energy control strategy.

[0075] The preprocessing step for the energy data specifically includes:

[0076] The energy data is filtered for outliers using the 3σ principle to determine whether the energy data in the preset area is dynamic data;

[0077] If so, the interactive window normalization algorithm is used for preprocessing;

[0078] If not, then a linear normalization algorithm is used for preprocessing.

[0079] Furthermore, the local energy regulation method of the local energy regulation unit includes:

[0080] When the energy data acquired by the braking system area exceeds the braking energy threshold, the local energy regulation unit preferentially stores the energy data in the battery. The braking energy threshold is calculated as follows:

[0081] ,

[0082] in, This represents the maximum energy that a battery can store. The braking energy threshold, It is the percentage coefficient, and It is related to the current state of battery charge;

[0083] When the photovoltaic power of the solar photovoltaic panel area is greater than the battery charging power threshold, the local energy control unit will supply the energy data of the solar photovoltaic panel area to the drive motor.

[0084] Under initial operating conditions, when the battery state of charge is less than the preset state of charge and the photovoltaic power is greater than the preset power, the local energy control unit will prioritize the activation of the energy data-assisted drive in the solar photovoltaic panel area.

[0085] Example, not limitation: The preset state of charge is 30% of the saturated battery state of charge. The preset power is 10kW.

[0086] During deceleration, when the braking intensity is greater than the preset braking intensity, the local energy control unit will preferentially store the energy data of the braking system area into the battery.

[0087] Example, not limitation, with a preset braking intensity of 0.3g.

[0088] In addition, the method also includes:

[0089] Based on the vehicle's operating conditions, set the energy priority weights for each preset area;

[0090] Based on the energy priority weights and the energy conversion efficiency and power of each preset region, a local energy regulation strategy is determined, calculated as follows:

[0091] ,

[0092] in, For the first The energy allocation ratio of each preset area For the first Energy priority weights for each preset region For the first Energy conversion efficiency of a preset area For the first Power of each preset area This represents the total number of preset regions.

[0093] For example, under cruise conditions, if the energy data collection in the solar photovoltaic panel area is stable and the energy priority weight is high, the energy data in the solar photovoltaic panel area will be used first to drive the new energy vehicle.

[0094] During deceleration, the energy priority of the braking system area is high, and the energy data of the braking system area is stored in the battery or provided to the new energy vehicle for other working consumption.

[0095] When energy data from the braking system is stored in the battery, the charging current is limited using the following formula:

[0096] ,

[0097] in, This is the charging current. The maximum charging current for the battery. Energy data for the braking system area. Battery voltage, This refers to the time during deceleration.

[0098] It should be noted that energy data from distributed edge computing networks can directly enter local energy control units, enabling rapid response. In sudden vehicle operating conditions (such as rapid acceleration and braking), this allows for effective and rapid decision-making, preventing energy waste caused by "decision delays" in the fine-tuning of energy control units. Furthermore, the initial adjustments made by local energy control units serve as "input conditions" for the global energy control unit, assisting it in improving the quality of input data and allowing the global energy control unit to converge its decisions more quickly.

[0099] Step S12: Based on the local energy regulation strategy, according to the state space and action space of the new energy vehicle, with the goal of maximizing energy utilization, the global energy regulation strategy is obtained by reinforcement learning through the global energy regulation unit and iterative optimization.

[0100] The state space includes vehicle operating condition information, energy states of various preset areas, and environmental parameters; the action space is the allocation ratio of each preset area.

[0101] Among these methods, based on the vehicle's basic information, a decision tree algorithm is used to determine the current vehicle operating condition.

[0102] Based on the current allocation ratio, the allocation ratio is predicted using the Long Short-Term Memory Network (LSTM).

[0103] Based on the local energy regulation strategy, according to the current state space and predicted action space of the new energy vehicle, with the goal of maximizing energy utilization, the global energy regulation unit performs reinforcement learning, and then iteratively optimizes the global energy regulation strategy through the Q-learning algorithm to obtain the optimal allocation ratio. The state space includes vehicle operating condition information, energy state of each preset area, and environmental parameters, and the action space is the allocation ratio of each preset area.

[0104] Furthermore, the function constructed with the goal of maximizing energy utilization is as follows:

[0105] ,

[0106] in, To maximize energy efficiency, For energy efficiency, For the overall energy efficiency of the vehicle, The balance of the battery's state of charge in a vehicle. For energy loss costs, , , These are the weighting coefficients. For the first Discount factor of time, =0, 1, ..., .

[0107] It should be noted that the global energy control unit combines state space and action space to enable the energy of each preset area to form effective coordination, thereby maximizing energy utilization and significantly improving energy efficiency.

[0108] Step S13: The global energy control strategy is distributed to each energy execution unit. The parameters of the distributed edge computing network and the global energy control unit are adjusted through energy output and energy feedback to correct the energy deviation.

[0109] Specifically, the global energy control strategy is distributed to each energy execution unit, and the deviation type is determined through energy output and energy feedback. The deviation type includes system deviation and instantaneous deviation.

[0110] When the deviation type is a system deviation, it is uploaded to the vehicle cloud, and new parameters are generated through big data analysis to replace the parameters of the distributed edge computing network and the global energy control unit.

[0111] When the deviation type is instantaneous deviation, the current deviation is obtained, and the parameters of the distributed edge computing network and the global energy control unit are dynamically adjusted according to the PID algorithm.

[0112] The calculation formula for the PID algorithm is as follows:

[0113] ,

[0114] in, For correction amount, For the current deviation, This is the proportionality coefficient. The integral coefficient is... The differential coefficients are...

[0115] The correction amount is used to adjust the parameters of the distributed edge computing network and the global energy control unit.

[0116] The proportional coefficient, integral coefficient, and derivative coefficient are automatically adjusted according to the vehicle's operating conditions.

[0117] It should be noted that a closed-loop control is formed through energy output and energy feedback, and then energy deviation correction is performed to eliminate execution errors and ensure the accuracy of energy regulation.

[0118] Compared with existing technologies, the new energy vehicle energy regulation method provided in this embodiment has the following advantages: By deploying edge computing units in a preset area of ​​the new energy vehicle to collect energy data from different preset areas, and then achieving rapid response and global optimization through local energy regulation units and global energy regulation units, energy utilization is maximized and energy efficiency is significantly improved. Furthermore, closed-loop control is formed through energy output and energy feedback, and execution errors are eliminated through energy deviation correction to ensure energy regulation accuracy. This solves the technical problem that the energy allocation strategy is difficult to optimize globally due to the common reliance on single and independent energy sources for energy replenishment and the lack of cross-module data fusion capabilities.

[0119] Example 2

[0120] Please see Figure 2 The image shows a second embodiment of the present invention providing an energy regulation system for a new energy vehicle, the system comprising:

[0121] The computing network construction module 100 is used to deploy edge computing units in a preset area of ​​the new energy vehicle, and to build a distributed edge computing network by combining the various edge computing units to obtain energy data of each preset area.

[0122] The local control module 200 is used to preprocess the energy data and, in conjunction with the vehicle operating conditions, input it to the local energy control unit for local energy control to obtain a local energy control strategy.

[0123] The global control module 300 is used to perform reinforcement learning and iterative optimization to obtain a global energy control strategy based on the local energy control strategy, according to the state space and action space of the new energy vehicle, with the goal of maximizing energy utilization.

[0124] The deviation correction module 400 is used to distribute the global energy control strategy to each energy execution unit and adjust the parameters of the distributed edge computing network and the global energy control unit through energy output and energy feedback to correct the energy deviation.

[0125] Compared to existing technologies, the new energy vehicle energy regulation system provided in this embodiment has the following advantages: By deploying edge computing units in a preset area of ​​the new energy vehicle through a computing network construction module, energy data from different preset areas is collected. Then, through the local energy regulation units of the local regulation module and the global energy regulation unit of the global regulation module, rapid response and global optimization are achieved in synergy, maximizing energy utilization and significantly improving energy efficiency. Furthermore, the energy output and energy feedback of the deviation correction module form a closed-loop control, and the execution error is eliminated through energy deviation correction, ensuring the accuracy of energy regulation. This solves the technical problem that the energy allocation strategy is difficult to optimize globally because there is a common problem of relying on a single and independent energy source for energy replenishment and a lack of cross-module data fusion capabilities.

[0126] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0127] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0128] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for energy regulation in new energy vehicles, characterized in that, The method includes: Edge computing units are deployed in a predetermined area of ​​the new energy vehicle. This predetermined area includes a solar photovoltaic panel area, a braking system area, a thermoelectric conversion area, and an environmental energy capture area. A distributed edge computing network is constructed by combining various edge computing units to obtain energy data for each preset area. Specifically, energy data is acquired for each edge computing unit at first preset time intervals. When the fluctuation of energy data of each edge computing unit reaches the fluctuation threshold, energy data is acquired every second preset time interval, an anomaly detection algorithm is triggered to detect anomalies, and the data is uploaded to the local energy control unit first. The first preset time interval is longer than the second preset time interval. The energy data is preprocessed and, in conjunction with vehicle operating conditions, input into a local energy control unit for local energy regulation, resulting in a local energy regulation strategy. The local energy regulation method of the local energy control unit includes: When the energy data acquired by the braking system area exceeds the braking energy threshold, the local energy regulation unit preferentially stores the energy data in the battery. The braking energy threshold is calculated as follows: ， in, This represents the maximum energy that a battery can store. The braking energy threshold, It is the percentage coefficient, and It depends on the current state of battery charge. When the photovoltaic power of the solar photovoltaic panel area exceeds the battery charging power threshold, the local energy control unit supplies the energy data of the solar photovoltaic panel area to the drive motor. Under initial operating conditions, when the battery state of charge is less than the preset state of charge and the photovoltaic power is greater than the preset power, the local energy control unit will prioritize activating the energy data-assisted drive of the solar photovoltaic panel area. During deceleration, when the braking intensity is greater than the preset braking intensity, the local energy control unit will preferentially store the energy data of the braking system area into the battery. Based on the aforementioned local energy regulation strategy, and taking into account the state space and action space of the new energy vehicle, with the goal of maximizing energy utilization, a global energy regulation strategy is derived through reinforcement learning via a global energy regulation unit and iterative optimization, including: The state space includes vehicle operating condition information, energy states of various preset areas, and environmental parameters; the action space is the allocation ratio of each preset area. Among these methods, based on the vehicle's basic information, a decision tree algorithm is used to determine the current vehicle operating condition. Based on the current allocation ratio, the allocation ratio is predicted using a long short-term memory network after a third preset time. Based on the local energy regulation strategy, according to the current state space and predicted action space of the new energy vehicle, with the goal of maximizing energy utilization, the global energy regulation unit performs reinforcement learning, and then the Q-learning algorithm iteratively optimizes to obtain the global energy regulation strategy and the optimal allocation ratio. The global energy regulation strategy is distributed to each energy execution unit. Through energy output and energy feedback, the parameters of the distributed edge computing network and the global energy regulation unit are adjusted to correct energy deviations.

2. The energy regulation method for new energy vehicles according to claim 1, characterized in that, The steps for preprocessing the energy data specifically include: The energy data is filtered for outliers using the 3σ principle to determine whether the energy data in the preset area is dynamic data; If so, then the sliding window normalization algorithm is used for preprocessing; If not, then a linear normalization algorithm is used for preprocessing.

3. The energy regulation method for new energy vehicles according to claim 1, characterized in that, The function constructed with the goal of maximizing energy utilization is as follows: ， in, To maximize energy utilization, For energy efficiency, For the overall energy efficiency of the vehicle, The balance of the battery's state of charge in a vehicle. For energy loss costs, , , These are the weighting coefficients. For the first Discount factor of time, =0, 1, ..., .

4. The energy regulation method for new energy vehicles according to claim 1, characterized in that, The steps of distributing the global energy regulation strategy to each energy execution unit and adjusting the parameters of the distributed edge computing network and the global energy regulation unit through energy output and energy feedback to correct energy deviations specifically include: The global energy control strategy is distributed to each energy execution unit, and the deviation type is determined through energy output and energy feedback. The deviation type includes system deviation and instantaneous deviation. When the deviation type is a system deviation, it is uploaded to the vehicle cloud, and new parameters are generated through big data analysis to replace the parameters of the distributed edge computing network and the global energy control unit. When the deviation type is instantaneous deviation, the current deviation is obtained, and the parameters of the distributed edge computing network and the global energy control unit are dynamically adjusted according to the PID algorithm.

5. The energy regulation method for new energy vehicles according to claim 4, characterized in that, The calculation formula for the PID algorithm is as follows: ， in, For correction amount, For the current deviation, This is the proportionality coefficient. The integral coefficient is... These are the differential coefficients. The correction amount is used to adjust the parameters of the distributed edge computing network and the global energy control unit. The proportional coefficient, integral coefficient, and derivative coefficient are automatically adjusted according to the vehicle's operating conditions.

6. The energy regulation method for new energy vehicles according to claim 2, characterized in that, The method further includes: Based on the vehicle's operating conditions, set the energy priority weights for each preset area; Based on the energy priority weights and the energy conversion efficiency and power of each preset region, a local energy regulation strategy is determined, calculated as follows: ， in, For the first The energy allocation ratio of each preset area For the first Energy priority weights for each preset region For the first Energy conversion efficiency of a preset area For the first Power of each preset area This represents the total number of preset regions.

7. An energy regulation system for new energy vehicles, characterized in that, The system is used to implement the energy regulation method for new energy vehicles according to any one of claims 1 to 6, the system comprising: A computing network construction module is used to deploy edge computing units in a predetermined area of ​​the new energy vehicle. This predetermined area includes a solar photovoltaic panel area, a braking system area, a thermoelectric conversion area, and an environmental energy capture area. A distributed edge computing network is constructed by combining various edge computing units to obtain energy data for each preset area. Specifically, energy data is acquired for each edge computing unit at first preset time intervals. When the fluctuation of energy data of each edge computing unit reaches the fluctuation threshold, energy data is acquired every second preset time interval, an anomaly detection algorithm is triggered to detect anomalies, and the data is uploaded to the local energy control unit first. The first preset time interval is longer than the second preset time interval. A local control module is used to preprocess the energy data and, in conjunction with vehicle operating conditions, input it to a local energy control unit for local energy control, thereby obtaining a local energy control strategy. The local energy control method of the local energy control unit includes: When the energy data acquired by the braking system area exceeds the braking energy threshold, the local energy regulation unit preferentially stores the energy data in the battery. The braking energy threshold is calculated as follows: ， in, This represents the maximum energy that a battery can store. The braking energy threshold, It is the percentage coefficient, and It depends on the current state of battery charge. When the photovoltaic power of the solar photovoltaic panel area exceeds the battery charging power threshold, the local energy control unit supplies the energy data of the solar photovoltaic panel area to the drive motor. Under initial operating conditions, when the battery state of charge is less than the preset state of charge and the photovoltaic power is greater than the preset power, the local energy control unit will prioritize activating the energy data-assisted drive of the solar photovoltaic panel area. During deceleration, when the braking intensity is greater than the preset braking intensity, the local energy control unit will preferentially store the energy data of the braking system area into the battery. A global control module, based on the local energy control strategy and according to the state space and action space of the new energy vehicle, aims to maximize energy utilization. It uses a global energy control unit to perform reinforcement learning and iterative optimization to derive a global energy control strategy, including: The state space includes vehicle operating condition information, energy states of various preset areas, and environmental parameters; the action space is the allocation ratio of each preset area. Among these methods, based on the vehicle's basic information, a decision tree algorithm is used to determine the current vehicle operating condition. Based on the current allocation ratio, the allocation ratio is predicted using a long short-term memory network after a third preset time. Based on the local energy regulation strategy, according to the current state space and predicted action space of the new energy vehicle, with the goal of maximizing energy utilization, the global energy regulation unit performs reinforcement learning, and then the Q-learning algorithm iteratively optimizes to obtain the global energy regulation strategy and the optimal allocation ratio. The deviation correction module is used to distribute the global energy regulation strategy to each energy execution unit and adjust the parameters of the distributed edge computing network and the global energy regulation unit through energy output and energy feedback to correct the energy deviation.

Images