New energy heavy truck battery replacement system and method based on intelligent management

Abstract

The invention discloses a new energy heavy truck battery replacement system and method based on intelligent management. The method comprises the following steps that S1, the system is initialized; s2, generating a battery state curve by using a deep recurrent neural network, detecting an abnormal battery and marking the abnormal battery as a non-distributable battery; s3, extracting candidate sites to generate a preliminary candidate path set; s4, optimizing a path by using a simulated annealing tree search method, processing dynamic information by embedding a decision tree search algorithm, and screening out an optimal power conversion path; s5, full-automatic disassembly and assembly of the batteries are completed, and the optimal batteries are distributed; s6, generating a battery health state curve and long-term charge and discharge characteristic simulation data by using the generative adversarial network, and dynamically adjusting a battery use priority and a distribution strategy; s7, dynamically planning a charging task, and using new energy to complete the charging task; and S8, establishing a battery health file, and performing health state detection on the battery. According to the invention, through deep learning and simulated annealing tree search methods and the like, intelligent management of new energy heavy truck battery replacement is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of new energy vehicles, and in particular to a new energy heavy-duty truck battery replacement system and method based on intelligent management. Background Art

[0002] In recent years, with the increasing awareness of environmental protection and the rapid development of new energy technologies, new energy heavy-duty trucks have been widely used as a green means of transportation. New energy heavy-duty trucks usually use batteries as the main power source. Their environmental protection and low carbon emission characteristics make them an ideal substitute for traditional fuel heavy-duty trucks. However, the large-scale popularization of new energy heavy-duty trucks still faces many challenges, among which the efficiency and intelligence of the battery replacement process and the battery management capabilities directly affect the operating efficiency and economy of the vehicle.

[0003] The existing battery replacement technologies for new energy heavy trucks mainly include two modes: manual battery replacement and automated battery replacement. In the manual battery replacement mode, the battery replacement process takes a long time and requires a lot of manual participation, which not only increases operating costs, but also poses certain safety hazards. Although the automated battery replacement mode achieves partial automation through equipment such as robotic arms, the degree of intelligence in key links such as path planning, battery replacement station selection, and battery allocation is low, and it often cannot fully adapt to complex operating environments. In addition, most existing battery replacement systems rely on a fixed battery replacement station network and a simple battery scheduling algorithm, which cannot dynamically respond to the real-time needs and traffic conditions of vehicles, resulting in low resource utilization.

[0004] In terms of battery management, the existing technology mainly adopts a simple battery status monitoring method, which evaluates the remaining capacity and health status of the battery by recording basic parameters such as the battery's voltage and current. This method has obvious deficiencies in processing multi-dimensional data and predicting the long-term performance degradation trend of the battery. At the same time, the traditional battery management system lacks intelligent abnormality detection and maintenance mechanisms, making it difficult to identify the potential failure risks of the battery in a timely manner, which can easily lead to performance degradation and even safety problems during battery operation. In addition, in the existing technology, the battery allocation strategy is mostly based on static rules, and it is impossible to dynamically adjust according to the actual needs, load conditions, and cruising range of heavy trucks, resulting in low battery utilization efficiency, affecting the economy and reliability of operations.

[0005] In the field of path planning, existing battery swapping systems usually use a simple optimization algorithm based on the shortest distance for path selection. This algorithm ignores dynamic factors such as the real-time load of the battery swapping station, queue conditions, and traffic congestion, resulting in the lack of adaptability of the planned path, which often cannot meet the actual operational needs of heavy trucks. At the same time, the existing technology lacks in-depth optimization of the path planning algorithm, and it is easy to fall into the local optimal solution during the path selection process, and it is impossible to effectively provide a global optimal battery swapping path for heavy trucks. In addition, in the process of path planning, the existing technology usually does not make full use of real-time data and fails to achieve a rapid response to complex traffic environments, which further reduces the efficiency and flexibility of the battery swapping system.

[0006] In terms of charging task control, existing technologies usually use fixed plans to allocate charging tasks, and lack the ability to dynamically control the volatility of renewable energy power generation and the load status of the power grid. Renewable energy power generation (such as photovoltaics and wind power) has obvious intermittent and volatile characteristics, and the existing charging system fails to make full use of real-time renewable energy power generation data when allocating charging power, resulting in low utilization of clean energy. In addition, during peak load periods of the power grid, existing technologies lack intelligent load control methods, which can easily cause power grid overload and increased operating costs. At the same time, the existing charging system has insufficient research on the relationship between battery health status and charging priority, and has failed to design personalized charging strategies for different battery states, thus affecting the health management effect and service life of the battery.

[0007] In general, the existing new energy heavy-duty truck battery replacement technology and system have obvious deficiencies in the following aspects: First, in terms of battery status prediction and management, the existing technology lacks the ability to dynamically process multi-dimensional data and cannot accurately predict the battery health status and performance degradation trend; second, in terms of path planning, the existing technology has insufficient optimization capabilities and lacks efficient processing of complex traffic environments and real-time data; third, in terms of charging task control, the existing technology fails to make full use of new energy power generation data and power grid status, resulting in low charging efficiency and clean energy utilization; fourth, in terms of battery allocation and usage priority adjustment, the existing technology fails to achieve dynamic optimization and is difficult to meet actual operational needs.

[0008] Therefore, how to provide a new energy heavy-duty truck battery replacement system and method based on intelligent management is an issue that technical personnel in this field urgently need to solve. Summary of the invention

[0009] One purpose of the present invention is to propose a new energy heavy-duty truck battery replacement system and method based on intelligent management. The present invention integrates advanced algorithms and technologies such as deep learning, generative adversarial networks, and simulated annealing tree search, and realizes the intelligent management of the new energy heavy-duty truck battery replacement system through multi-module collaboration. Specifically, it includes dynamic prediction and anomaly detection of battery status, real-time planning of optimal battery replacement path, automated battery replacement execution, dynamic regulation of charging tasks, and battery life cycle management. The system has the characteristics of efficient, safe, and precise operation, while significantly improving the utilization efficiency of new energy and the service life of the battery, solving the problems of low intelligence, insufficient efficiency, and waste of resources in the prior art, and providing reliable technical support for the widespread application of new energy heavy-duty trucks.

[0010] The battery replacement method for a new energy heavy truck based on intelligent management according to an embodiment of the present invention comprises the following steps: S1. Initialize the system. The battery swap station and the heavy truck establish communication through the vehicle terminal and the Internet of Things platform. The battery swap station uploads battery inventory data, and the heavy truck uploads operating parameters. S2. Based on the battery inventory data and operating parameters, a deep recurrent neural network is used to train and predict the battery operating status, generate a battery status curve, and the battery status curve includes the battery remaining capacity, battery health status and energy decay trend, and detect abnormal batteries and mark them as unallocatable; S3. Based on the battery status curve and marking information, extract candidate sites from all battery swap stations in the area to generate a preliminary candidate path set; S4. Based on the preliminary candidate path set, the simulated annealing tree search method is used to optimize the path, the temperature decay strategy is combined to gradually eliminate the suboptimal path, and the decision tree search algorithm is embedded to process traffic congestion and site load dynamic information to screen out the optimal battery replacement path; S5. Based on the optimal battery swapping path, after the heavy truck arrives at the battery swapping station, the battery swapping robot arm completes the fully automatic removal and installation of the battery, and allocates the most suitable battery according to the battery status curve and allocation rules; S6. Based on the battery operation data after battery replacement, the generative adversarial network is used to generate the battery health status curve and long-term charge and discharge characteristic simulation data, and dynamically adjust the battery usage priority and allocation strategy; S7. Based on the battery inventory status and new energy power generation data of the battery swap station, dynamically plan the charging task, optimize the charging power allocation in combination with the electricity price, and use new energy to complete the charging task; S8. Based on the battery health status curve and the completion of the charging task, a battery health profile is established, and the health status of the replaced battery is tested. When the health status is detected to be lower than the set threshold, retirement or maintenance is arranged.

[0011] Optionally, the S2 specifically includes: S21, extracting multi-dimensional characteristic parameters of the battery from the battery inventory data uploaded by the battery swap station, wherein the multi-dimensional characteristic parameters include voltage , Current ,temperature 、Internal resistance , Remaining capacity and charge and discharge cycle times ; Extract the current battery load requirement from the running parameters uploaded by the heavy truck , Estimated cruising range and usage time ; S22. Construct the input data set of the deep recurrent neural network and define the time series data: ; in, represents a time series, represents the time step, represents the total number of time steps; S23, using a deep recurrent neural network to Perform training, the training goal is to fit the battery state curve : ; in, Indicates the remaining battery capacity. Indicates the battery health status. Indicates the energy decay trend; S24, inputting the real-time time series data into the trained deep recurrent neural network model for prediction to generate a real-time battery status curve; S25, perform abnormal detection based on the battery status curve, when the battery health status Less than threshold When the battery temperature Exceeding safety threshold , or internal resistance Higher than the upper limit of internal resistance When the energy decay trend is When the display shows a nonlinear rapid decline, an early warning is triggered, marking the battery as requiring priority maintenance; S26. Synchronously storing the battery status curve and abnormal marking result to the cloud.

[0012] Optionally, the S3 specifically includes: S31, according to the remaining battery capacity in the battery status curve , Battery health status and energy decay trend , combined with the tag information to filter out batteries that are in good health and not marked as unallocated; based on the remaining battery capacity and current load demand , determine the range of heavy trucks As a filter condition; S32. Extract site information from the database of all battery swap stations in the region, including the geographical location of the site , battery inventory, real-time load, number of queued vehicles and traffic congestion coefficient; S33, calculate the current position of the heavy truck To each battery swap station Traffic weighted distance : ; in, Indicates the battery swap station site The real-time traffic congestion coefficient of the road section, Indicates arrival at the power station site The path complexity is calculated by the number of turns and the number of traffic lights. and represents the weight parameter; S34, based on traffic weighted distance and driving range , filter the valid battery swap station site set that meets the conditions: ; in, represents the set of valid battery swap stations, Indicates the battery swap station site Battery inventory; S35. Collect effective battery swap stations Priority scoring is performed by comprehensively considering the distance to the battery swap station, battery inventory, real-time load and queue status through a nonlinear function: ; in, Indicates the battery swap station site The priority score of , and represents the adjustment coefficient, Indicates the battery inventory. Indicates real-time load, represents a nonlinear penalty for increasing the number of vehicles in the queue; S36. Score by priority The results with a score higher than the threshold are selected The battery swap stations are selected as candidate sites, a preliminary candidate path set is generated, and the geographical location, battery inventory, real-time load and queuing data information of the candidate sites are stored in the database.

[0013] Optionally, the S4 specifically includes: S41, extracting path information from the preliminary candidate path set, including the path starting point and end point, path length, traffic congestion coefficient, and site load and battery inventory data, as initial input for path optimization; S42, initializing the simulated annealing algorithm, generating an initial path set and an initial temperature, and setting a temperature attenuation coefficient and a termination temperature; the initial path set is selected based on the candidate path in a random perturbation manner as an optimization starting point; S43. Define path objective function Comprehensively evaluate the advantages and disadvantages of the path. The objective function takes into account the dynamic effects of path length, traffic congestion, site load and battery inventory: ; in, represents the path length, represents the smoothing factor, represents the real-time traffic congestion coefficient, Indicates the battery inventory. Indicates real-time load, Indicates the site load threshold, , and represents the weight parameter; S44. During the simulated annealing optimization process, randomly perturb the current path set Generate a new set of paths , calculate the objective function and ; if , directly accept the new path; if , with probability Receive new paths; S45, after each iteration, the temperature is updated using a temperature decay strategy until the temperature is lower than the set termination temperature, and the simulated annealing iteration process is stopped; S46. Based on the path set selected by the simulated annealing algorithm, a decision tree search algorithm is nested to adjust the path selection according to the real-time traffic congestion coefficient and site load dynamic information, and the path that meets the optimal objective function value is selected as the optimal battery replacement path.

[0014] Optionally, the S6 specifically includes: S61, collecting battery operation data after battery replacement, including battery remaining capacity, battery health status, internal resistance, temperature, and long-term charge and discharge characteristics, and constructing a data set; S62, defining a generator network and a discriminator network of a generative adversarial network, wherein the generator network accepts random noise as input data and outputs simulated battery data; the discriminator network accepts real data and simulated data, distinguishes whether the input data is real, and improves the quality of generated data through adversarial training; S63, train the generative adversarial network, and use the real battery operation dataset and the dataset generated by the generator for multiple rounds of adversarial optimization; S64, generating simulation data through the trained generator network, wherein the simulation data includes a battery health status curve and long-term charge and discharge characteristic simulation data, and enriching the data set in combination with real data; S65, based on the generated simulation data and real data, dynamically adjust the battery usage priority and define the optimization objective function For balancing battery life and allocation efficiency: ; in, Indicates time The remaining battery capacity, Indicates charging time. Indicates the change in health status, Indicates the time interval, Indicates the current load requirement of the battery. Indicates the allocation capability. , and represents the weight parameter; S66. Based on the result of the optimization objective function, dynamically assign battery priorities, assign batteries with higher health status to vehicles with greater endurance requirements, and assign batteries with lower health status to short-distance vehicles or arrange battery health maintenance.

[0015] Optionally, the S7 specifically includes: S71, obtaining battery inventory status, wherein the battery inventory status includes the current number of available batteries, battery health status, remaining capacity, charging status, and a list of batteries to be charged, analyzing the usage priority of the inventory batteries, and classifying and marking the batteries according to their health status and remaining capacity; S72, collecting real-time power generation data from the new energy power generation system, the real-time power generation data including current power generation power, power generation type and power generation fluctuation characteristics; combining the real-time data of the power grid to obtain the current power grid load situation and electricity price fluctuation information, and analyzing the available charging power of the power grid and the off-peak electricity price period; S73. Analyze the charging demand of the battery swap station based on the battery inventory status and new energy power generation data, and dynamically generate a charging task list; select batteries with low remaining capacity and good health status to enter the charging queue, and adjust the charging task order according to the fluctuation characteristics of new energy power generation; S74. According to the charging task list and the renewable energy power generation data, a dynamic charging plan is formulated to use the renewable energy power generation to charge the battery; when the renewable energy power is insufficient, the charging mode is dynamically switched in combination with the grid load status and the electricity price fluctuation, and the grid low load and low electricity price period is selected to replenish the battery; S75. Real-time monitoring of new energy power generation fluctuations and battery charging status, and dynamic allocation of charging power; allocating higher charging power to high-priority batteries, and dynamically adjusting the battery charging power according to real-time power generation and load demand; S76. After the charging task is completed, the battery health status, remaining capacity and charging record are updated, and the data is uploaded to the battery swap station database to analyze the changes in battery performance during the charging process, provide a decision-making basis for charging task planning, and evaluate the execution effect of the charging plan in combination with new energy power generation data.

[0016] The new energy heavy truck battery replacement system based on intelligent management according to an embodiment of the present invention includes the following modules: Data acquisition module, used to collect battery operation data and system operation parameters of battery swap stations and heavy trucks; The battery status prediction module is used to generate battery status curves using a deep recurrent neural network, detect abnormal batteries and mark them as unallocatable; The path planning module is used to extract candidate sites from all battery swap stations in the area, generate a preliminary set of candidate paths, optimize the paths using the simulated annealing tree search method, and select the optimal battery swap path in combination with real-time traffic and load information; Automatic battery replacement module, used to control the battery replacement robot arm to complete the fully automatic removal and installation of the battery under the guidance of the optimal battery replacement path, and allocate the most suitable battery according to the battery status and allocation rules; A battery management module, which uses a generative adversarial network to generate battery simulation data, dynamically adjust battery usage priorities and allocation strategies, and optimize battery life management; The charging task control module is used to dynamically plan charging tasks based on battery inventory status and new energy power generation data, and optimize charging power allocation based on electricity prices and grid load conditions; The health profile management module is used to establish a battery health profile, update the battery health status based on the charging completion status, and arrange retirement or maintenance for batteries below the set threshold.

[0017] The beneficial effects of the present invention are: First of all, the present invention realizes multi-dimensional dynamic prediction of battery operating status by combining the data acquisition module with the deep recurrent neural network. The generated battery status curve includes the remaining battery capacity, battery health status and energy decay trend, which can accurately reflect the real-time operating status and long-term performance changes of the battery. At the same time, the abnormality detection function can automatically identify and mark abnormal conditions of the battery, eliminate potential faulty batteries in time, and ensure the safety and stability of the battery replacement system. This method overcomes the defects of the prior art that the battery status evaluation is single and the prediction accuracy is low.

[0018] Secondly, the present invention realizes the global optimization of the preliminary candidate path set by combining the simulated annealing tree search method with the temperature attenuation strategy in the path planning module. The embedding of the decision tree algorithm enables the path planning process to dynamically process real-time information such as traffic congestion and site load, and provides a battery replacement path selection solution that is more adaptable to complex environments. This path planning method effectively solves the problem that path selection in the prior art is limited to local optimality, significantly shortens the battery replacement time and driving distance of heavy trucks, and improves the operational efficiency of the battery replacement system. In addition, through the seamless connection between path optimization and automatic battery replacement modules, the battery replacement robot arm can quickly complete the precise disassembly and installation of the battery according to the optimized path, and at the same time allocates the most suitable battery in combination with the battery status curve and allocation rules, further improving the automation level of the system.

[0019] In addition, the charging task control module of the present invention fully combines the battery inventory status, new energy power generation data and grid load conditions to dynamically plan charging tasks and optimize charging power distribution. In the case of strong volatility in new energy power generation, the system can give priority to the use of real-time clean energy to complete charging tasks, thereby greatly improving the utilization rate of new energy; when power generation is insufficient, charging is carried out in conjunction with low electricity price periods, effectively reducing charging costs. At the same time, the system adjusts the charging strategy according to the battery health status and charging priority, avoiding waste of resources and the risk of overcharging the battery. Through this module, the charging process is more efficient and flexible, and can adapt to diverse operational needs.

[0020] Finally, through the health record management module, the present invention has established a complete battery life cycle management system. The system can update the battery file in real time according to the battery health status curve and the completion of the charging task, and arrange maintenance or retirement in time for batteries whose health status is lower than the set threshold. This function not only ensures the safety of battery use, but also extends the overall service life of the battery, improves resource utilization efficiency, and reduces system operating costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 This is an overall flow chart of the new energy heavy truck battery replacement method based on intelligent management proposed by the present invention; Figure 2 This is a schematic diagram of the structure of the new energy heavy-duty truck battery replacement system based on intelligent management proposed in the present invention. DETAILED DESCRIPTION

[0022] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0023] refer to Figure 1 The battery replacement method for new energy heavy trucks based on intelligent management includes the following steps: S1. Initialize the system. The battery swap station and the heavy truck establish communication through the vehicle terminal and the Internet of Things platform. The battery swap station uploads battery inventory data, and the heavy truck uploads operating parameters. S2. Based on the battery inventory data and operating parameters, a deep recurrent neural network is used to train and predict the battery operating status, generate a battery status curve, and the battery status curve includes the battery remaining capacity, battery health status and energy decay trend, and detect abnormal batteries and mark them as unallocatable; S3. Based on the battery status curve and marking information, extract candidate sites from all battery swap stations in the area to generate a preliminary candidate path set; S4. Based on the preliminary candidate path set, the simulated annealing tree search method is used to optimize the path, the temperature decay strategy is combined to gradually eliminate the suboptimal path, and the decision tree search algorithm is embedded to process traffic congestion and site load dynamic information to screen out the optimal battery replacement path; S5. Based on the optimal battery swapping path, after the heavy truck arrives at the battery swapping station, the battery swapping robot arm completes the fully automatic removal and installation of the battery, and allocates the most suitable battery according to the battery status curve and allocation rules; S6. Based on the battery operation data after battery replacement, the generative adversarial network is used to generate the battery health status curve and long-term charge and discharge characteristic simulation data, and dynamically adjust the battery usage priority and allocation strategy; S7. Based on the battery inventory status and new energy power generation data of the battery swap station, dynamically plan the charging task, optimize the charging power allocation in combination with the electricity price, and use new energy to complete the charging task; S8. Based on the battery health status curve and the completion of the charging task, a battery health profile is established, and the health status of the replaced battery is tested. When the health status is detected to be lower than the set threshold, retirement or maintenance is arranged.

[0024] In this implementation, S2 specifically includes: S21, extracting multi-dimensional characteristic parameters of the battery from the battery inventory data uploaded by the battery swap station, wherein the multi-dimensional characteristic parameters include voltage , Current ,temperature 、Internal resistance , Remaining capacity and charge and discharge cycle times ; Extract the current battery load requirement from the running parameters uploaded by the heavy truck , Estimated cruising range and usage time ; S22. Construct the input data set of the deep recurrent neural network and define the time series data: ; in, represents a time series, represents the time step, represents the total number of time steps; S23, using a deep recurrent neural network to Perform training, the training goal is to fit the battery state curve : ; in, Indicates the remaining battery capacity. Indicates the battery health status. Indicates the energy decay trend; S24, inputting the real-time time series data into the trained deep recurrent neural network model for prediction to generate a real-time battery status curve; S25, perform abnormal detection based on the battery status curve, when the battery health status Less than threshold When the battery temperature Exceeding safety threshold , or internal resistance Higher than the upper limit of internal resistance When the energy decay trend is When the display shows a nonlinear rapid decline, an early warning is triggered, marking the battery as requiring priority maintenance; S26. Synchronously storing the battery status curve and abnormal marking result to the cloud.

[0025] In this implementation, S3 specifically includes: S31, according to the remaining battery capacity in the battery status curve , Battery health status and energy decay trend , combined with the tag information to filter out batteries that are in good health and not marked as unallocated; based on the remaining battery capacity and current load demand , determine the range of heavy trucks As a filter condition; S32. Extract site information from the database of all battery swap stations in the region, including the geographical location of the site , battery inventory, real-time load, number of queued vehicles and traffic congestion coefficient; S33, calculate the current position of the heavy truck To each battery swap station Traffic weighted distance : ; in, Indicates the battery swap station site The real-time traffic congestion coefficient of the road section, Indicates arrival at the power station site The path complexity is calculated by the number of turns and the number of traffic lights. and represents the weight parameter; S34, based on traffic weighted distance and driving range , filter the valid battery swap station site set that meets the conditions: ; in, represents the set of valid battery swap stations, Indicates the battery swap station site Battery inventory; S35. Collect effective battery swap stations Priority scoring is performed by comprehensively considering the distance to the battery swap station, battery inventory, real-time load and queue status through a nonlinear function: ; in, Indicates the battery swap station site The priority score of , and represents the adjustment coefficient, Indicates the battery inventory. Indicates real-time load, represents a nonlinear penalty for increasing the number of vehicles in the queue; S36. Score by priority The results with a score higher than the threshold are selected The battery swap stations are selected as candidate sites, a preliminary candidate path set is generated, and the geographical location, battery inventory, real-time load and queuing data information of the candidate sites are stored in the database.

[0026] In this implementation manner, the S4 specifically includes: S41, extracting path information from the preliminary candidate path set, including the path starting point and end point, path length, traffic congestion coefficient, and site load and battery inventory data, as initial input for path optimization; S42, initializing the simulated annealing algorithm, generating an initial path set and an initial temperature, and setting a temperature attenuation coefficient and a termination temperature; the initial path set is selected based on the candidate path in a random perturbation manner as an optimization starting point; S43. Define path objective function Comprehensively evaluate the advantages and disadvantages of the path. The objective function takes into account the dynamic effects of path length, traffic congestion, site load and battery inventory: ; in, represents the path length, represents the smoothing factor, represents the real-time traffic congestion coefficient, Indicates the battery inventory. Indicates real-time load, Indicates the site load threshold, , and represents the weight parameter; S44. During the simulated annealing optimization process, randomly perturb the current path set Generate a new set of paths , calculate the objective function and ; if , directly accept the new path; if , with probability Receive new paths; S45, after each iteration, the temperature is updated using a temperature decay strategy until the temperature is lower than the set termination temperature, and the simulated annealing iteration process is stopped; S46. Based on the path set selected by the simulated annealing algorithm, a decision tree search algorithm is nested to adjust the path selection according to the real-time traffic congestion coefficient and site load dynamic information, and the path that meets the optimal objective function value is selected as the optimal battery replacement path.

[0027] In this implementation manner, S6 specifically includes: S61, collecting battery operation data after battery replacement, including battery remaining capacity, battery health status, internal resistance, temperature, and long-term charge and discharge characteristics, and constructing a data set; S62, defining a generator network and a discriminator network of a generative adversarial network, wherein the generator network accepts random noise as input data and outputs simulated battery data; the discriminator network accepts real data and simulated data, distinguishes whether the input data is real, and improves the quality of generated data through adversarial training; S63, train the generative adversarial network, and use the real battery operation dataset and the dataset generated by the generator for multiple rounds of adversarial optimization; S64, generating simulation data through the trained generator network, wherein the simulation data includes a battery health status curve and long-term charge and discharge characteristic simulation data, and enriching the data set in combination with real data; S65, based on the generated simulation data and real data, dynamically adjust the battery usage priority and define the optimization objective function For balancing battery life and allocation efficiency: ; in, Indicates time The remaining battery capacity, Indicates charging time. Indicates the change in health status, Indicates the time interval, Indicates the current load requirement of the battery. Indicates the allocation capability. , and represents the weight parameter; S66. Based on the result of the optimization objective function, dynamically assign battery priorities, assign batteries with higher health status to vehicles with greater endurance requirements, and assign batteries with lower health status to short-distance vehicles or arrange battery health maintenance.

[0028] In this implementation manner, the S7 specifically includes: S71, obtaining battery inventory status, wherein the battery inventory status includes the current number of available batteries, battery health status, remaining capacity, charging status, and a list of batteries to be charged, analyzing the usage priority of the inventory batteries, and classifying and marking the batteries according to their health status and remaining capacity; S72, collecting real-time power generation data from the new energy power generation system, the real-time power generation data including current power generation power, power generation type and power generation fluctuation characteristics; combining the real-time data of the power grid to obtain the current power grid load situation and electricity price fluctuation information, and analyzing the available charging power of the power grid and the off-peak electricity price period; S73. Analyze the charging demand of the battery swap station based on the battery inventory status and new energy power generation data, and dynamically generate a charging task list; select batteries with low remaining capacity and good health status to enter the charging queue, and adjust the charging task order according to the fluctuation characteristics of new energy power generation; S74. According to the charging task list and the renewable energy power generation data, a dynamic charging plan is formulated to use the renewable energy power generation to charge the battery; when the renewable energy power is insufficient, the charging mode is dynamically switched in combination with the grid load status and the electricity price fluctuation, and the grid low load and low electricity price period is selected to replenish the battery; S75. Real-time monitoring of new energy power generation fluctuations and battery charging status, and dynamic allocation of charging power; allocating higher charging power to high-priority batteries, and dynamically adjusting the battery charging power according to real-time power generation and load demand; S76. After the charging task is completed, the battery health status, remaining capacity and charging record are updated, and the data is uploaded to the battery swap station database to analyze the changes in battery performance during the charging process, provide a decision-making basis for charging task planning, and evaluate the execution effect of the charging plan in combination with new energy power generation data.

[0029] refer to Figure 2 The new energy heavy truck battery replacement system based on intelligent management includes the following modules: Data acquisition module, used to collect battery operation data and system operation parameters of battery swap stations and heavy trucks; The battery status prediction module is used to generate battery status curves using a deep recurrent neural network, detect abnormal batteries and mark them as unallocatable; The path planning module is used to extract candidate sites from all battery swap stations in the area, generate a preliminary set of candidate paths, optimize the paths using the simulated annealing tree search method, and select the optimal battery swap path in combination with real-time traffic and load information; Automatic battery replacement module, used to control the battery replacement robot arm to complete the fully automatic removal and installation of the battery under the guidance of the optimal battery replacement path, and allocate the most suitable battery according to the battery status and allocation rules; A battery management module, which uses a generative adversarial network to generate battery simulation data, dynamically adjust battery usage priorities and allocation strategies, and optimize battery life management; The charging task control module is used to dynamically plan charging tasks based on battery inventory status and new energy power generation data, and optimize charging power allocation based on electricity prices and grid load conditions; The health profile management module is used to establish a battery health profile, update the battery health status based on the charging completion status, and arrange retirement or maintenance for batteries below the set threshold.

[0030] Embodiment 1: In order to verify the feasibility of the present invention in implementation, the present invention is applied in a logistics park to support the efficient operation of large-scale new energy heavy trucks. The logistics park is located in the center of a busy city, with five battery swap stations around it. The average daily transportation task volume of new energy heavy trucks is as high as 800 times. In the traditional battery swap mode, due to inaccurate battery status assessment, unscientific battery swap path planning, low efficiency of the battery swap process, and insufficient utilization of new energy power generation, the system operation efficiency is low and battery resources are seriously wasted. In order to solve these problems, the intelligent battery swap system of the present invention has been fully deployed and applied.

[0031] In this embodiment, before the start of each day's mission, new energy heavy trucks connect to the system through the on-board terminal and upload the current battery operating status, including remaining power, health status, vehicle load and other parameters. The battery swap station uploads the current battery inventory, charging status and site load in real time. The system uses a deep recurrent neural network model to analyze the battery status, generate a battery health status curve, energy decay trend, and mark abnormal batteries. On a particular day, the system analysis found that a total of 25 batteries in five battery swap stations were in a high internal resistance state and could not be distributed normally. This discovery avoided possible battery swap failures.

[0032] At the same time, the system generates the optimal battery swap route based on the day's transportation task volume and traffic data. A heavy truck needs to depart from the logistics park and go to a distribution point 20 kilometers away. Its current battery remaining power is 45%, and the estimated range is 70 kilometers. After comprehensively analyzing the distance to the surrounding battery swap stations, the number of batteries in stock, and the queue situation, the system recommends that it go to the nearest battery swap station with a planned route length of 12 kilometers. During the journey, the system monitors traffic data in real time and updates the route, successfully avoiding 3 kilometers of traffic congestion.

[0033] At the battery swap station, the heavy truck completed the battery replacement through an automated robotic arm. The whole process took 5 minutes, while it would have taken 15 minutes in the traditional mode. The replaced battery was in good health, with 90% remaining power, which could support the vehicle to complete the transportation task of the day. The replaced old battery was marked as requiring maintenance by the system, and then entered the health record management module to arrange for repair or retirement.

[0034] In the charging task control, the system gives priority to using the photovoltaic power generation system in the logistics park to charge the batteries to be charged. On that day, the total photovoltaic power generation was 500 kWh. The system used the power generation fluctuation data to adjust the charging task, using 72% of clean energy for battery charging, and the rest of the power was supplemented by the low-peak electricity price at night. Through this control strategy, the average charging cost on that day decreased by 18% and the utilization rate of clean energy increased by 25%.

[0035] After 5 consecutive days of system operation data analysis, we obtained specific operational efficiency improvements. The following table shows the specific experimental comparison data.

[0036] Table 1 Comparison of the battery exchange efficiency of the traditional mode and the battery exchange efficiency of the system of the present invention

[0037] Table 2 Path optimization effect table of the traditional mode and the system of the present invention

[0038] Table 3 Comparison of new energy utilization efficiency between traditional mode and the system of the present invention

[0039] It can be seen from the above data table 1 that after adopting the intelligent battery replacement system, the average battery replacement time of the vehicle is significantly reduced. From an average of 15 minutes in the traditional mode to about 5 minutes in the intelligent system, the battery replacement efficiency has increased by about 66.67%. For example, on November 1, 2024, the total time to complete 820 battery replacement tasks was reduced by approximately 8,200 minutes (about 136.7 hours). The time saved can be directly converted into vehicle transportation time, improving transportation efficiency. This result is due to the full-process automated operation of the automatic battery replacement module and the time optimization capability of the optimal path planning, which reduces the cost of manual intervention and waiting time for logistics companies.

[0040] The above data table 2 shows that the intelligent system comprehensively considers real-time traffic and station load information in route planning, so that the total mileage of the vehicle is effectively optimized. For example, the total mileage on November 1, 2024 was reduced from 18,400 kilometers in the traditional mode to 16,100 kilometers, a mileage reduction ratio of up to 12.5%. This optimization not only reduces the operating cost of the vehicle, but also reduces unnecessary energy consumption and carbon emissions. Overall, the effect of route optimization is also reflected in the reduction of traffic congestion delay time by 22%, further improving logistics efficiency.

[0041] As can be seen from the above data table 3, the intelligent system has greatly improved the utilization efficiency of new energy. In the traditional mode, the utilization rate of new energy power generation is only about 48%, while the intelligent system increases the utilization rate to 72%-74% by dynamically adjusting the charging task priority and power allocation. For example, on November 3, 2024, the total power generation was 510 kWh, of which 74% was used for battery charging, an increase of 25% over the traditional utilization rate. This improvement not only reduces dependence on the power grid, but also reduces operating costs and improves the sustainability of the system. At the same time, the dynamic nature of charging task allocation avoids overload problems during peak hours of the power grid, providing a strong guarantee for the stable operation of the power grid.

[0042] Through a comprehensive analysis of battery exchange efficiency, path optimization and new energy utilization, it can be seen that the intelligent battery exchange system has achieved remarkable results in many aspects. First, the improvement of battery exchange efficiency directly saves time costs and provides higher turnover efficiency for the operation of logistics companies. Secondly, the optimization of path planning reduces unnecessary mileage and time losses, saves fuel costs for enterprises and reduces carbon emissions. Finally, the efficient utilization and intelligent regulation of new energy power generation not only improves the efficiency of energy use, but also promotes the widespread application of green energy in the transportation industry, with good economic and social benefits. These data fully demonstrate the significant advantages of the present invention in practical applications, and provide reliable technical support for the large-scale promotion of new energy heavy trucks.

[0043] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A new energy heavy truck battery replacement method based on intelligent management, characterized in that: The steps include: S1. Initialize the system. The battery swap station and the heavy truck establish communication through the vehicle terminal and the Internet of Things platform. The battery swap station uploads battery inventory data, and the heavy truck uploads operating parameters. S2. Based on the battery inventory data and operating parameters, a deep recurrent neural network is used to train and predict the battery operating status, generate a battery status curve, and the battery status curve includes the battery remaining capacity, battery health status and energy decay trend, and detect abnormal batteries and mark them as unallocatable; S3. Based on the battery status curve and marking information, extract candidate sites from all battery swap stations in the area to generate a preliminary candidate path set; S4. Based on the preliminary candidate path set, the simulated annealing tree search method is used to optimize the path, the temperature decay strategy is combined to gradually eliminate the suboptimal path, and the decision tree search algorithm is embedded to process traffic congestion and site load dynamic information to screen out the optimal battery replacement path; S5. Based on the optimal battery swapping path, after the heavy truck arrives at the battery swapping station, the battery swapping robot arm completes the fully automatic removal and installation of the battery, and allocates the most suitable battery according to the battery status curve and allocation rules; S6. Based on the battery operation data after battery replacement, the generative adversarial network is used to generate the battery health status curve and long-term charge and discharge characteristic simulation data, and dynamically adjust the battery usage priority and allocation strategy; S7. Based on the battery inventory status and new energy power generation data of the battery swap station, dynamically plan the charging task, optimize the charging power allocation in combination with the electricity price, and use new energy to complete the charging task; S8. Based on the battery health status curve and the completion of the charging task, a battery health profile is established, and the health status of the replaced battery is tested. When the health status is detected to be lower than the set threshold, retirement or maintenance is arranged.

2. The method for replacing a new energy heavy truck battery based on intelligent management according to claim 1 is characterized in that: The S2 specifically includes: S21, extracting multi-dimensional characteristic parameters of the battery from the battery inventory data uploaded by the battery swap station, wherein the multi-dimensional characteristic parameters include voltage , Current ,temperature 、Internal resistance , Remaining capacity and charge and discharge cycle times ; Extract the current battery load requirement from the running parameters uploaded by the heavy truck , Estimated cruising range and usage time ; S22. Construct the input data set of the deep recurrent neural network and define the time series data: ； in, represents a time series, represents the time step, represents the total number of time steps; S23, using a deep recurrent neural network to Perform training, the training goal is to fit the battery state curve : ； in, Indicates the remaining battery capacity. Indicates the battery health status. Indicates the energy decay trend; S24, inputting the real-time time series data into the trained deep recurrent neural network model for prediction to generate a real-time battery status curve; S25, perform abnormal detection based on the battery status curve, when the battery health status Less than threshold When the battery temperature Exceeding safety threshold , or internal resistance Higher than the upper limit of internal resistance When the energy decay trend is When the display shows a nonlinear rapid decline, an early warning is triggered, marking the battery as requiring priority maintenance; S26. Synchronously storing the battery status curve and abnormal marking result to the cloud.

3. The method for replacing a new energy heavy truck battery based on intelligent management according to claim 1 is characterized in that: The S3 specifically includes: S31, according to the remaining battery capacity in the battery status curve , Battery health status and energy decay trend , combined with the tag information to filter out batteries that are in good health and not marked as unallocated; based on the remaining battery capacity and current load demand , determine the cruising range of heavy trucks As a filter condition; S32. Extract site information from the database of all battery swap stations in the region, including the geographical location of the site , battery inventory, real-time load, number of queued vehicles and traffic congestion coefficient; S33, calculate the current position of the heavy truck To each battery swap station Traffic weighted distance : ； in, Indicates the battery swap station site The real-time traffic congestion coefficient of the road section, Indicates arrival at the power station site The path complexity is calculated by the number of turns and the number of traffic lights. and represents the weight parameter; S34, based on traffic weighted distance and driving range , filter the valid battery swap station site set that meets the conditions: ； in, represents the set of valid battery swap stations, Indicates the battery swap station site Battery inventory; S35. Collect effective battery swap stations Priority scoring is performed by comprehensively considering the distance to the battery swap station, battery inventory, real-time load and queue status through a nonlinear function: ； in, Indicates the battery swap station site The priority score of , and represents the adjustment coefficient, Indicates the battery inventory. Indicates real-time load, represents a nonlinear penalty for increasing the number of vehicles in the queue; S36. Score by priority The results with a score higher than the threshold are selected The battery swap stations are selected as candidate sites, a preliminary candidate path set is generated, and the geographical location, battery inventory, real-time load and queuing data information of the candidate sites are stored in the database.

4. The method for replacing a new energy heavy truck battery based on intelligent management according to claim 1 is characterized in that: The S4 specifically includes: S41, extracting path information from the preliminary candidate path set, including the path starting point and end point, path length, traffic congestion coefficient, and site load and battery inventory data, as initial input for path optimization; S42, initializing the simulated annealing algorithm, generating an initial path set and an initial temperature, and setting a temperature attenuation coefficient and a termination temperature; the initial path set is selected based on the candidate path in a random perturbation manner as an optimization starting point; S43. Define path objective function Comprehensively evaluate the advantages and disadvantages of the path. The objective function takes into account the dynamic effects of path length, traffic congestion, site load and battery inventory: ； in, represents the path length, represents the smoothing factor, represents the real-time traffic congestion coefficient, Indicates the battery inventory. Indicates real-time load, Indicates the site load threshold, , and represents the weight parameter; S44. During the simulated annealing optimization process, randomly perturb the current path set Generate a new set of paths , calculate the objective function and ; if , directly accept the new path; if , with probability Receive new paths; S45, after each iteration, the temperature is updated using a temperature decay strategy until the temperature is lower than the set termination temperature, and the simulated annealing iteration process is stopped; S46. Based on the path set selected by the simulated annealing algorithm, a decision tree search algorithm is nested to adjust the path selection according to the real-time traffic congestion coefficient and site load dynamic information, and the path that meets the optimal objective function value is selected as the optimal battery replacement path.

5. The method for replacing a new energy heavy truck battery based on intelligent management according to claim 1 is characterized in that: The S6 specifically includes: S61, collecting battery operation data after battery replacement, including battery remaining capacity, battery health status, internal resistance, temperature, and long-term charge and discharge characteristics, and constructing a data set; S62, defining a generator network and a discriminator network of a generative adversarial network, wherein the generator network accepts random noise as input data and outputs simulated battery data; the discriminator network accepts real data and simulated data, distinguishes whether the input data is real, and improves the quality of generated data through adversarial training; S63, train the generative adversarial network, and use the real battery operation dataset and the dataset generated by the generator for multiple rounds of adversarial optimization; S64, generating simulation data through the trained generator network, wherein the simulation data includes a battery health status curve and long-term charge and discharge characteristic simulation data, and enriching the data set in combination with real data; S65, based on the generated simulation data and real data, dynamically adjust the battery usage priority and define the optimization objective function For balancing battery life and allocation efficiency: ； in, Indicates time The remaining battery capacity, Indicates charging time. Indicates the change in health status, Indicates the time interval, Indicates the current load requirement of the battery. Indicates the allocation capability. , and represents the weight parameter; S66. Based on the result of the optimization objective function, dynamically assign battery priorities, assign batteries with higher health status to vehicles with greater endurance requirements, and assign batteries with lower health status to short-distance vehicles or arrange battery health maintenance.

6. The method for replacing a new energy heavy truck battery based on intelligent management according to claim 1 is characterized in that: The S7 specifically includes: S71, obtaining battery inventory status, wherein the battery inventory status includes the current number of available batteries, battery health status, remaining capacity, charging status, and a list of batteries currently to be charged, analyzing the usage priority of the inventory batteries, and classifying and marking the batteries according to their health status and remaining capacity; S72, collecting real-time power generation data from the new energy power generation system, the real-time power generation data including current power generation power, power generation type and power generation fluctuation characteristics; combining the real-time power grid data to obtain the current power grid load situation and electricity price fluctuation information, and analyzing the available charging power of the power grid and the off-peak electricity price period; S73. Analyze the charging demand of the battery swap station based on the battery inventory status and new energy power generation data, and dynamically generate a charging task list; select batteries with low remaining capacity and good health status to enter the charging queue, and adjust the charging task order according to the fluctuation characteristics of new energy power generation; S74. According to the charging task list and the renewable energy power generation data, a dynamic charging plan is formulated to use the renewable energy power generation to charge the battery; when the renewable energy power is insufficient, the charging mode is dynamically switched in combination with the grid load status and the electricity price fluctuation, and the grid low load and low electricity price period is selected to replenish the battery; S75, real-time monitoring of new energy power generation fluctuations and battery charging status, and dynamic allocation of charging power; allocate higher charging power to high-priority batteries, and dynamically adjust the battery charging power according to real-time power generation and load demand; S76. After the charging task is completed, the battery health status, remaining capacity and charging record are updated, and the data is uploaded to the battery swap station database to analyze the changes in battery performance during the charging process, provide a decision-making basis for charging task planning, and evaluate the execution effect of the charging plan in combination with new energy power generation data.

7. A new energy heavy truck battery replacement system based on intelligent management, applying the new energy heavy truck battery replacement method based on intelligent management as described in any one of claims 1 to 6, characterized in that: Includes the following modules: Data acquisition module, used to collect battery operation data and system operation parameters of battery swap stations and heavy trucks; The battery status prediction module is used to generate battery status curves using a deep recurrent neural network, detect abnormal batteries and mark them as unallocatable; The path planning module is used to extract candidate sites from all battery swap stations in the area, generate a preliminary set of candidate paths, optimize the paths using the simulated annealing tree search method, and select the optimal battery swap path in combination with real-time traffic and load information; Automatic battery replacement module, used to control the battery replacement robot arm to complete the fully automatic removal and installation of the battery under the guidance of the optimal battery replacement path, and allocate the most suitable battery according to the battery status and allocation rules; A battery management module, which uses a generative adversarial network to generate battery simulation data, dynamically adjust battery usage priorities and allocation strategies, and optimize battery life management; The charging task control module is used to dynamically plan charging tasks based on battery inventory status and new energy power generation data, and optimize charging power allocation based on electricity prices and grid load conditions; The health profile management module is used to establish a battery health profile, update the battery health status based on the charging completion status, and arrange retirement or maintenance for batteries below the set threshold.

Images