Intelligent battery replacement control method and system

Abstract

The present invention discloses an intelligent battery swap control method and system, which uses an intelligent scheduling algorithm to generate a first charging strategy, performs correlation analysis on historical vehicle flow data and historical load data to optimize the target segment division, optimizes the first charging strategy according to the optimization target segment division result, obtains a second charging strategy, and then obtains the corresponding future state data; uses the battery swap control potential boundary and the robot arm collaborative operation capability boundary under different vehicle flow data determined by the simulation method as constraints, constructs an integrated control model of battery swap scheduling and robot arm motion; uses future state data and reservation data as optimization conditions, performs optimization processing on the integrated control model, and generates a target battery swap scheduling strategy and a target robot arm motion control strategy. The method provided by the present application can cope with variable vehicle flow, improve battery swap efficiency, improve the robot arm collaborative operation capability, and ensure the reliability of battery swap operations.

Description

Technical Field

[0001] The present invention relates to the technical field of battery swap station dispatching management, and in particular to an intelligent battery swap control method and system. Background Art

[0002] An electric vehicle battery swap station is a centralized energy storage and charging station that provides more convenient and efficient charging and swapping services for electric vehicles through battery swapping. In recent years, with the rapid expansion of the electric vehicle market, the scale of electric vehicle battery swap stations has also grown rapidly, covering various scenarios such as cities, rural areas, and highways.

[0003] The regulation capabilities of battery swap stations face multiple technical contradictions and challenges. First, the capacity and charging rate of the battery pack are key factors affecting the regulation potential. The larger the capacity of the battery pack, the more power can be regulated, but it also means higher cost investment and space occupancy; and although the increase in charging rate is conducive to quickly replenishing reserve power, it may aggravate battery loss and attenuation. Secondly, changes in traffic flow at the battery swap station also bring uncertainty to regulation, and fluctuations in traffic flow make it difficult to accurately estimate power demand. Finally, although the introduction of the robotic arm has improved the efficiency of battery swapping, factors such as its movement speed, running trajectory, and grasping accuracy also bring new variables to regulation.

[0004] It can be seen that how to solve the problem that battery swap scheduling at battery swap stations cannot cope with the existence of multiple variables and cannot take into account multiple requirements such as benefits, efficiency and reliability has become a technical problem that technical personnel in this field need to solve urgently. Summary of the invention

[0005] The present invention provides an intelligent battery swap control method and system to solve the technical problems that battery swap scheduling at battery swap stations cannot cope with the existence of multiple variables and cannot take into account both benefits and efficiency. It can cope with the variables brought by the vehicle flow and robotic arms at battery swap stations and take into account multiple requirements such as benefits, efficiency and reliability.

[0006] In a first aspect, the present invention provides an intelligent battery replacement control method, the method comprising:

[0007] An intelligent scheduling algorithm is used to analyze the acquired battery pack status data, charging pile status data and reservation data of the target time period in the battery swap station to generate a first charging strategy;

[0008] Perform correlation analysis on the historical vehicle flow data and historical load data of the battery swap station within the preset time period length, and divide the preset time period length into optimization target segments according to the correlation analysis results. The optimization target segment division results are: the optimization target period with the shortest vehicle waiting time, the optimization target period with the maximum remaining power, and the optimization target period with the minimum remaining power;

[0009] Optimizing the first charging strategy according to the optimization target segment division result to obtain a second charging strategy, and acquiring future state data of the battery pack after the second charging strategy is executed;

[0010] Based on the preset battery swap scheduling strategy and the preset robot arm motion control strategy, simulate the working state of the robot arm under different preset traffic flow data, and determine the battery swap control potential boundary and the robot arm collaborative operation capability boundary under different traffic flow data;

[0011] Taking the battery swapping regulation potential boundary and the robotic arm collaborative operation capability boundary as constraints, a battery swapping scheduling and robotic arm motion integrated control model is constructed;

[0012] In the actual scheduling and control process, the future state data and the reservation data are used as optimization conditions to optimize the integrated control model of battery swap scheduling and robotic arm motion to generate a target battery swap scheduling strategy and a target robotic arm motion control strategy.

[0013] Preferably, the intelligent scheduling algorithm is used to analyze the acquired battery pack status data, charging pile status data and reservation data of the target time period in the battery swap station to generate a first charging strategy, including:

[0014] Based on the acquired battery pack status data, charging pile status data and reservation data of the target time period, an intelligent scheduling algorithm is used to build a battery pack and charging pile matching matrix;

[0015] Taking the optimal charging efficiency as the optimization goal, the matching matrix between the battery pack and the charging pile is optimized and solved to obtain the first charging strategy.

[0016] Preferably, the performing of correlation analysis on the historical flow data and historical load data of the battery swap station within the preset time period length, and dividing the preset time period length into optimized target segments according to the correlation analysis result, includes:

[0017] Calculating the charging demand under the historical vehicle flow data according to the historical vehicle flow data;

[0018] According to the historical load data, the preset time period length is divided into a number of high-load time periods and a number of low-load time periods;

[0019] A random forest regression algorithm is used to perform correlation analysis on the charging demand and the historical load data to establish a nonlinear mathematical model;

[0020] Calculating a load impact coefficient according to the nonlinear mathematical model, wherein the load impact coefficient is defined as a ratio of the charging demand to the historical load data at the same time;

[0021] Calculating the average value of the load influence coefficient in each of the high-load time periods and each of the low-load time periods respectively;

[0022] For the high-load time period, if the mean is greater than or equal to a first threshold, the corresponding high-load time period is divided into a target time period for optimizing the shortest vehicle waiting time; if the mean is less than the first threshold, the corresponding high-load time period is divided into a target time period for optimizing the minimum remaining power;

[0023] For the low-load time period, if the mean is greater than or equal to the second threshold, the corresponding low-load time period is divided into an optimization target period with the shortest vehicle waiting time; if the mean is less than the second threshold, the low-load time period is divided into an optimization target period with the maximum remaining power.

[0024] Preferably, the step of optimizing the first charging strategy of the target time period according to the optimization target segment division result to obtain the second charging strategy includes:

[0025] Based on the optimization target segment division result, determine the time period to which the target time period belongs, and if the determination result is that it belongs to the optimization target period with the shortest vehicle waiting time, optimize the first charging strategy with the shortest vehicle waiting time;

[0026] If the judgment result is that it belongs to the maximum remaining power optimization target period, optimizing the first charging strategy with the maximum remaining power of the battery pack;

[0027] If the judgment result is that it belongs to the optimization target period of minimum remaining power, optimizing the first charging strategy with the minimum remaining power of the battery pack;

[0028] The optimization results are integrated to obtain the second charging strategy.

[0029] Preferably, based on the preset battery swap scheduling strategy and the preset robot arm motion control strategy, simulating the working state of the robot arm under different preset traffic flow data, and determining the battery swap control potential boundary and the robot arm collaborative operation capability boundary under different traffic flow data, includes:

[0030] According to the preset robot arm motion control strategy, the robot arm attribute parameter data and the robot arm operation data are set, wherein the robot arm attribute parameter data at least includes: joint degrees of freedom, material strength and sensor accuracy, and the robot arm operation data at least includes: motion speed, motion trajectory and workload;

[0031] Extract the start timestamp and end timestamp of each battery replacement operation within the preset time period from the battery replacement operation record, and calculate the time series of a single battery replacement operation according to the start timestamp and the end timestamp;

[0032] Calculate the product of the single battery replacement operation time series and the robot arm grasping success rate, and use the product as the effective battery replacement time series;

[0033] Based on the effective battery replacement time series, a queuing system model is used to model the preset different vehicle flow data, and the vehicle arrival interval and battery replacement service time are obtained through distribution fitting;

[0034] The working state of the robotic arm and the simulated battery swap efficiency corresponding to the preset battery swap scheduling strategy and the preset robotic arm motion control strategy under the vehicle arrival interval and the battery swap service time are simulated by a Monte Carlo simulation method;

[0035] The battery swap efficiency prediction model constructed by using the preset different vehicle flow data, the preset battery swap scheduling strategy, the preset robot arm motion control strategy and the simulated battery swap efficiency is trained;

[0036] A genetic algorithm is used to search the trained battery swap efficiency prediction model to obtain an optimal battery swap scheduling strategy and an optimal robotic arm motion control strategy that maximize the simulated battery swap efficiency;

[0037] The optimal battery swap scheduling strategy, the optimal robotic arm motion control strategy, and the corresponding working states and simulated battery swap efficiencies are analyzed to obtain the battery swap regulation potential boundaries and robotic arm collaborative operation capability boundaries under different vehicle flow data.

[0038] Preferably, the battery swapping control potential boundary and the robot arm collaborative operation capability boundary are used as constraints to construct an integrated control model of battery swapping scheduling and robot arm motion, including:

[0039] Constructing a constraint equation with the battery swapping regulation potential boundary and the robotic arm collaborative operation capability boundary as constraint conditions;

[0040] Taking the battery pack status data, the workload and the vehicle flow data as decision variables, a mixed integer linear programming method is used to construct an objective function;

[0041] An integrated control model of battery swapping scheduling and robotic arm motion is constructed based on the constraint equations and the objective function.

[0042] Preferably, the future state data and the reservation data are used as optimization conditions to optimize the integrated control model of battery swap scheduling and robotic arm motion to generate a target battery swap scheduling strategy and a target robotic arm motion control strategy, including:

[0043] A Kalman filter is used to fuse the future state data and the reservation data into a unified data structure to obtain a state matrix of the battery swap station;

[0044] Input the state matrix of the battery swap station into the integrated control model of battery swap scheduling and robotic arm motion, and solve it using a model predictive control algorithm to generate multiple sets of candidate solutions, wherein the candidate solutions include a battery swap scheduling strategy and a robotic arm motion control strategy;

[0045] The battery swap efficiency and the robot arm utilization rate are converted into a single evaluation index by using a weighted sum method, each candidate solution is evaluated and ranked according to the single evaluation index, and a target candidate solution is selected according to the evaluation and ranking results;

[0046] Based on a deep reinforcement learning algorithm, the target candidate solution is used as an action space, the battery swap station state matrix is ​​used as a state space, and the single evaluation index is used as a reward function to construct a decision model;

[0047] The decision model is interactively learned with the simulation environment to generate a target battery swap scheduling strategy and a target robotic arm motion control strategy.

[0048] Preferably, the method further comprises:

[0049] According to the target robot arm motion control strategy, multiple collision-free motion trajectory solutions are generated, and the multiple collision-free motion trajectory solutions are optimized and solved to obtain the optimal robot arm motion path;

[0050] Based on the target battery replacement scheduling strategy and the optimal robotic arm motion path, the vehicle and battery are dynamically matched to generate a battery replacement task execution sequence.

[0051] Preferably, the constraint also includes the remaining battery life;

[0052] The method for obtaining the remaining service life of the battery includes:

[0053] Acquire battery pack charge and discharge history data, wherein the battery pack charge and discharge history data includes: charge and discharge cycle number, capacity attenuation, charging rate and battery temperature;

[0054] Training a battery health status prediction model constructed using a random forest algorithm based on the battery pack charge and discharge history data;

[0055] The collected real-time data of battery pack charge and discharge is input into the trained battery health status prediction model to predict the current health status of the battery pack and obtain the remaining service life of the battery.

[0056] In a second aspect, the present invention further provides an intelligent battery swap control system to implement the intelligent battery swap control method described above, the system comprising: a charging strategy generation unit, an optimization target segment division unit, a charging strategy optimization unit, a boundary determination unit, an integrated control model construction unit and a scheduling control unit;

[0057] The charging strategy generating unit is used to analyze the acquired battery pack status data, charging pile status data and reservation data of the target time period in the battery swap station by using an intelligent scheduling algorithm to generate a first charging strategy;

[0058] The optimization target segment division unit is used to perform correlation analysis on the historical vehicle flow data and historical load data of the battery swap station within the preset time period length, and divide the preset time period length into optimization target segments according to the correlation analysis result, and the optimization target segment division result is: the optimization target period with the shortest vehicle waiting time, the optimization target period with the maximum remaining power, and the optimization target period with the minimum remaining power;

[0059] The charging strategy optimization unit is used to optimize the first charging strategy according to the optimization target segment division result to obtain a second charging strategy, and obtain future state data of the battery pack after the second charging strategy is executed;

[0060] The boundary determination unit is used to simulate the working state of the robotic arm under different preset traffic flow data based on the preset battery swap scheduling strategy and the preset robotic arm motion control strategy, and determine the battery swap regulation potential boundary and the robotic arm collaborative operation capability boundary under different traffic flow data;

[0061] The integrated control model construction unit is used to construct an integrated control model of battery swap scheduling and robotic arm movement based on the battery swap regulation potential boundary and the robotic arm collaborative operation capability boundary as constraints;

[0062] The scheduling control unit is used to use the future state data and the reservation data as optimization conditions in the actual scheduling control process, optimize the integrated control model of battery swap scheduling and robotic arm motion, and generate a target battery swap scheduling strategy and a target robotic arm motion control strategy.

[0063] The present invention provides an intelligent battery replacement control method and system. Compared with the prior art, the beneficial effects of the embodiments of the present invention are at least one of the following:

[0064] (1) Improve the stability and safety of power operation in battery swap stations, optimize the division of target segments, and the setting of the first threshold and the second threshold, set different priorities for different time periods, and set corresponding optimization goals, thereby improving user experience while reducing the impact on the power system;

[0065] (2) The use of simulation methods can reduce the deviation caused by errors in the actual control process, obtain more accurate battery swap control potential boundaries and robotic arm collaborative operation capability boundaries, and improve the accuracy and practicality of the prediction of the constructed battery swap scheduling and robotic arm motion integrated control model.

[0066] (3) Fully consider the impact of traffic flow and robotic arms, and while improving the efficiency of battery replacement, enhance the ability of the robotic arms to work in coordination and ensure the reliability of battery replacement operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 It is a schematic diagram of the steps of an intelligent battery replacement control method provided by a preferred embodiment of the present invention;

[0068] Figure 2 It is a schematic diagram of the steps of a method for generating a first charging strategy provided by a preferred embodiment of the present invention;

[0069] Figure 3 It is a schematic diagram of the steps of a method for optimizing target segment division provided by a preferred embodiment of the present invention;

[0070] Figure 4 It is a schematic diagram of the steps of a method for optimizing charging strategy and acquiring future state data provided by a preferred embodiment of the present invention;

[0071] Figure 5 It is a schematic diagram of the steps of a method for obtaining the potential boundary of battery swap scheduling and the collaborative operation capability boundary of a robotic arm provided by a preferred embodiment of the present invention;

[0072] Figure 6 It is a schematic diagram of the method steps for constructing an integrated control model of battery swapping scheduling and mechanical arm motion provided by a preferred embodiment of the present invention;

[0073] Figure 7 It is a schematic diagram of the steps of a method for actual scheduling control provided by a preferred embodiment of the present invention;

[0074] Figure 8 It is a schematic diagram of the steps of a method for generating a battery replacement task execution sequence provided by a preferred embodiment of the present invention;

[0075] Fig. 9 It is a schematic diagram of the steps of a method for predicting the remaining service life of a battery provided by a preferred embodiment of the present invention;

[0076] Fig.10 It is a structural schematic diagram of an intelligent battery replacement control system provided by a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0077] The following is a detailed explanation of the implementation mode of the present invention in conjunction with the accompanying drawings. The embodiments are provided only for illustrative purposes and cannot be understood as limiting the present invention. The accompanying drawings are only for reference and illustration purposes and do not constitute a limitation on the scope of patent protection of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In the description of the present invention, the terms "first", "second", "third", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first", "second", "third", etc. may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, the meaning of "multiple" is two or more.

[0078] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be a connection between the two elements. The terms "vertical", "horizontal", "left", "right", "upper", "lower" and similar expressions used herein are only for illustrative purposes, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0079] In the description of the present invention, it should be noted that, unless otherwise defined, all technical and scientific terms used in the present invention have the same meanings as those commonly understood by those skilled in the art. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood by specific circumstances.

[0080] See also Figure 1 In an embodiment of the present invention, a smart battery replacement control method is provided, the method comprising:

[0081] S1. Use an intelligent scheduling algorithm to analyze the acquired battery pack status data, charging pile status data and reservation data of the target time period in the battery swap station to generate a first charging strategy.

[0082] A battery management system is deployed in the battery swap station to collect the status data of each battery pack in real time through the RS485 serial port or CAN bus. The battery management system monitors the status data of the battery packs in the battery swap station in real time through serial communication or CAN bus. The battery pack status data includes but is not limited to: battery status, current battery capacity, charging power, charging current, charging voltage, charging rate, charging time and expected charging time. The status data of each battery pack in the battery swap station is received from the battery management system according to the preset data collection time period. The preset data collection time period can be set according to demand, such as 5s, 10s, 10min, 30min, 1h or 2h. The battery management system stores the above data in the local SQLite database for quick access and analysis.

[0083] Among them, the battery status includes charging and waiting to be charged. For the battery pack being charged, the current battery capacity, charging power, charging current, charging voltage and charging rate are in dynamic change, and dynamic adjustment is required during the calculation of the estimated charging time data. For the battery pack to be charged, the charging power and charging rate are limited to the rated charging power and rated charging rate of the battery pack, so as to limit the charging power and charging rate of the future battery pack charging behavior.

[0084] After obtaining the current battery capacity, charging power, charging current and charging voltage, the theoretical value of the battery pack's charge per unit time is calculated based on the obtained charging current and charging power. Further, the estimated charging time of the battery pack is calculated based on the current battery capacity and the theoretical value of the charge per unit time.

[0085] For a battery pack being charged, determine whether the actual value of the battery pack's current charge per unit time reaches the theoretical value of the charge per unit time. If so, maintain the current charging state; if not, adjust the actual value of the battery pack's charge per unit time to the theoretical value of the charge per unit time.

[0086] Specifically, the battery model and ID number are obtained through the battery management system. The battery model includes at least: rated charging voltage, rated charging current, rated capacity, size, weight, chemical composition and charging and discharging performance. The theoretical value of the charging capacity per unit time is calculated based on the charging current. The calculation formula is:

[0087] Q=I×t

[0088] Where: Q represents the theoretical value of charging capacity per unit time, usually in ampere-hours (Ah) or milliampere-hours (mAh); I represents the rated charging current, in amperes (A) or milliamperes (mA); t represents the charging time, in hours (h).

[0089] The actual value of the charging current of the battery pack being charged is collected through the battery management system, and the actual value of the charging current is used instead of the rated charging current to calculate the actual value of the charge per unit time. It is determined whether the actual value of the battery pack's current charge per unit time reaches the theoretical value of the charge per unit time. If so, it indicates that the current charging current is appropriate and the current charging state is maintained; if not, it indicates that the current actual value of the charging current is too large or too small, and the actual value of the charging current is reduced or increased to reduce or increase the actual value of the charge per unit time to the theoretical value of the charge per unit time.

[0090] For the battery pack to be charged, the theoretical value of the charging amount per unit time is directly used as a parameter for future charging to set the charging current.

[0091] Analyze each battery pack separately and calculate the estimated charging time from the current battery capacity to the rated capacity. The specific calculation method is: calculate the difference between the rated capacity and the current battery capacity, and divide the difference by the theoretical value of the charge per unit time to obtain the theoretical charging time. The theoretical charging time is corrected based on factors such as battery temperature and charging efficiency. The theoretical charging time is multiplied by a coefficient of 1.1-1.2 as the correction value to obtain more accurate estimated charging time data. Based on the calculated estimated charging time data for each battery pack, a data list containing the battery ID and the estimated charging time is generated to facilitate subsequent analysis and management.

[0092] For example, the estimated charging time required for each battery pack from empty to full charge is calculated. Empty power is defined as 5% of the battery capacity and full charge is defined as 95%. The calculation formula is theoretical charging time = (full power - current power) / theoretical value of charging per unit time. Assume that a battery swap station has 50 battery packs and reads the data of the battery management system every 5 seconds, including parameters such as battery capacity, current remaining power, charging power, charging current and charging voltage. The battery management system stores this data in a local SQLite database for quick access and analysis. Extract the information of each battery from the database and calculate the theoretical value of charging per unit time and the theoretical charging time. Take a battery as an example, with a capacity of 100kWh, a current remaining power of 75kWh, and a theoretical value of charging per unit time of 50kW, then the theoretical charging time is (100kWh-75kWh) / 50kW=0.5 hours. Considering the efficiency loss in the actual charging process, the theoretical time is multiplied by a coefficient of 1.15 to correct it, and the estimated charging time is 0.575 hours. Repeat this process for 50 batteries to generate a data list containing battery IDs and expected charging times. During the charging process, the system checks the actual value of the charge per unit time every minute. If the actual value of the charge per unit time of a battery is 40kW, which is 80% lower than the expected speed of 50kW, the actual value of the charge per unit time will be increased to 50kW. Conversely, if the actual value of the charge per unit time reaches 65kW, which exceeds 120% of the expected speed, the actual value of the charge per unit time will be reduced to 50kW, but not exceeding the maximum charging power allowed by the battery, which is 60kW.

[0093] The charging pile status data mainly includes the current status of the charging pile, input voltage, input current, output voltage, output current and rated charging power. The current status of the charging pile includes: online, offline, charging and fault, etc.

[0094] The battery swap reservation system of the battery swap station is a convenient energy replenishment service system designed for electric vehicles. The system allows users to make reservations for battery swap services in advance through specific applications or platforms, thereby saving time and improving battery swap efficiency. The battery swap station can only provide battery swap services for electric vehicles if it has sufficient reserve battery packs, so the battery swap station generally strictly follows the reservation information for battery swap services. The number of reserve battery packs that can be used for battery swaps in the battery swap station will also be displayed in the battery swap reservation system. Therefore, the data of the battery swap reservation system is highly consistent with the flow data of real vehicles arriving at the battery swap station. For the setting of the target time period for reservation data collection, the current time is generally used as the starting point, and a time period of a preset length in the future is selected. The smaller the target time period, the higher the accuracy of the generated scheduling strategy and control instructions, but the scheduling strategy and control instructions change more frequently. The larger the target time period, the lower the accuracy of the generated scheduling strategy and control instructions, but the scheduling strategy and control instructions are relatively stable.

[0095] Intelligent scheduling algorithm is an algorithm that optimizes the allocation of tasks and resources to improve the overall system performance. Intelligent scheduling algorithm is based on artificial intelligence and machine learning theory. By training historical data, it can predict the demand and resource allocation of future tasks. Common types of algorithms include heuristic scheduling, meta-heuristic scheduling and neural network scheduling.

[0096] In a preferred embodiment of the present invention, an intelligent scheduling algorithm is used to analyze the acquired battery pack status data, charging pile status data and reservation data of the target time period in the battery swap station to generate a first charging strategy, such as Figure 2 As shown, the following steps are included:

[0097] S101. Based on the acquired battery pack status data, charging pile status data and reservation data of the target time period, an intelligent scheduling algorithm is used to construct a battery pack and charging pile matching matrix.

[0098] S102, taking the optimal charging efficiency as the optimization goal, optimizing and solving the matching matrix between the battery pack and the charging pile to obtain a first charging strategy.

[0099] An intelligent scheduling algorithm is used to analyze and calculate the battery pack status data, charging pile status data and reservation data of the target time period, and a battery pack and charging pile matching matrix is ​​constructed. The optimal charging efficiency of the battery pack is taken as the optimization goal, and the first charging strategy is obtained to achieve efficient charging.

[0100] S2. Perform a correlation analysis on the historical vehicle flow data and historical load data of the battery swap station within the preset time period length, and divide the preset time period length into optimized target segments according to the correlation analysis results. The optimized target segment division results are: the shortest vehicle waiting time optimization target period, the maximum remaining power optimization target period, and the minimum remaining power optimization target period.

[0101] The historical traffic flow data and historical load data of vehicles arriving at the battery swap station within a preset time period are obtained from the battery swap station management system database. For the setting of the preset time period length, a time period with typical periodic characteristics is generally selected, such as one day or one week. Generally, the number of vehicles arriving at the battery swap station is small in the morning and large in the afternoon and evening, and small on weekdays and large on weekends. The historical load data also has corresponding periodic fluctuations within a day or a week.

[0102] The historical traffic flow data of vehicles arriving at the battery swap station can be collected through the traffic flow monitoring system, which uses various sensors installed on the road, such as laser vehicle detectors and video cameras, to capture vehicle information in real time. These sensors work based on different principles, such as laser pulse ranging and image recognition, and can realize real-time detection, classification, counting, and speed measurement of vehicles.

[0103] The collection of historical load data is mainly based on the measurement principle of physical sensors. Various sensors installed in the power system of the power exchange station, such as current transformers, voltage transformers, power factor meters and frequency meters, are used to convert current, voltage and other parameters in the power system of the power exchange station into electrical signals. These electrical signals are then transmitted through data collectors and eventually reach data centers or monitoring systems for processing and storage.

[0104] In a preferred embodiment of the present invention, a correlation analysis is performed on the historical vehicle flow data and historical load data of the battery swap station within a preset time period length, and the preset time period length is divided into optimized target segments according to the correlation analysis result, such as Figure 3 As shown, the following steps are included:

[0105] S201. Calculate charging demand under the historical vehicle flow data according to the historical vehicle flow data.

[0106] S202. Divide the preset time period length into a plurality of high-load time periods and a plurality of low-load time periods according to the historical load data.

[0107] S203: using a random forest regression algorithm to perform a correlation analysis on the charging demand and the historical load data, and establishing a nonlinear mathematical model.

[0108] S204. Calculate a load impact coefficient according to the nonlinear mathematical model, where the load impact coefficient is defined as the ratio of the charging demand to the historical load data at the same moment.

[0109] S205. Calculate the average value of the load influence coefficient in each of the high-load time periods and each of the low-load time periods respectively.

[0110] S206. For the high-load time period, if the mean is greater than or equal to a first threshold, the corresponding high-load time period is divided into an optimized target time period with the shortest vehicle waiting time; if the mean is less than the first threshold, the corresponding high-load time period is divided into an optimized target time period with the minimum remaining power.

[0111] S207. For the low-load time period, if the mean is greater than or equal to a second threshold, the corresponding low-load time period is divided into an optimization target time period with the shortest vehicle waiting time; if the mean is less than the second threshold, the low-load time period is divided into an optimization target time period with the maximum remaining power.

[0112] In a preferred embodiment of the present invention, based on the relationship between the charging demand of the battery swap station and the historical load data, the typical preset time cycle length is divided into optimization target time periods to obtain different optimization target time periods for optimizing the first charging strategy.

[0113] Use Python's statsmodels library to perform time series analysis and apply the ARIMA model to process historical traffic data. Set the preset time period length to 24 hours, and count the number of vehicles every hour, such as 15 vehicles detected from 8:00 to 9:00, 20 vehicles detected from 9:00 to 10:00, and so on.

[0114] According to the historical traffic flow data, the charging demand under the historical traffic flow data is calculated. Specifically, the real-time video stream of the monitoring camera of the battery swap station is obtained, and the video frames in the real-time video stream are subjected to image denoising and histogram equalization to enhance the contrast. Further, the YOLOv5 algorithm is used to locate and count the vehicles in each frame of the image, and the features of the vehicles are extracted for vehicle classification. For example, the ResNet50 network is used to classify the extracted vehicle features, and the vehicles are divided into categories such as sedans, SUVs and trucks, and the number of vehicles of each type entering and leaving the station within the preset time period is counted. For vehicles of different categories, the power demand for full charge is different. The historical power demand for full charge of vehicles of different categories is averaged to obtain the average power demand for full charge of vehicles of different categories for the calculation of charging demand. For example, if the flow corresponding to sedans is 20 vehicles, and the average power demand for full charge of its vehicles is Q, then the charging demand of sedans is 20Q. The same method is used to calculate the charging demand corresponding to SUVs and trucks, and finally the sum of the charging demand of all vehicles is calculated to obtain the charging demand under the historical traffic flow data.

[0115] Furthermore, the preset time period length is evenly divided according to the preset time division length to obtain multiple time periods. According to the historical load data, the first mean of the historical load data in each time period is calculated. If the calculated first mean is greater than the load threshold, the time period is divided into a high-load time period. If the calculated first mean is less than or equal to the load threshold, the time period is divided into a low-load time period, and finally a number of high-load time periods and a number of low-load time periods are obtained.

[0116] Furthermore, the random forest regression algorithm is used to analyze the correlation between charging demand and historical load data and establish a nonlinear mathematical model.

[0117] Random Forest Regression is a regression algorithm based on ensemble learning. It improves the accuracy and robustness of the model by constructing multiple decision trees and integrating the prediction results of these trees.

[0118] The load impact coefficient is calculated based on the established nonlinear mathematical model. The load impact coefficient is defined as the ratio of the charging demand at the same time to the historical load data. The larger the ratio, the greater the impact of the charging action of the battery swap station on the power system of the battery swap station.

[0119] Furthermore, the second mean of the load influence coefficient in each high-load time period and each low-load time period is calculated. A first threshold is set for the high-load time period, and a second threshold is set for the low-load time period. If the second mean corresponding to the high-load time period is greater than or equal to the first threshold, indicating that the load of the power system of the battery swap station is high and the charging demand of the battery swap station is also high during the high-load time period, the shortest waiting time of the vehicle is prioritized, and the corresponding high-load time period is divided into the shortest vehicle waiting time optimization target period; if the second mean corresponding to the high-load time period is less than the first threshold, indicating that the load of the power system of the battery swap station is high during the high-load time period, but the charging demand of the battery swap station is relatively low, the stability of the power system is prioritized, and the minimum remaining power of the battery swap station is taken as the optimization target, and the corresponding high-load time period is divided into the minimum remaining power optimization target period.

[0120] If the second mean value corresponding to the low-load time period is greater than or equal to the second threshold value, indicating that the load of the power system of the battery swap station is low during the low-load time period, but the charging demand of the battery swap station is high, the waiting time of the vehicle is prioritized, and the corresponding low-load time period is divided into the optimization target period with the shortest vehicle waiting time; if the second mean value corresponding to the low-load time period is less than the second threshold value, indicating that the load of the power system of the battery swap station is low during the low-load time period, and the charging demand of the battery swap station is also low, the remaining power of the battery swap station is prioritized, and the low-load time period is divided into the optimization target period with the maximum remaining power.

[0121] The magnitude of the first threshold and the second threshold reflects the impact of the charging action of the battery swap station on the stability of the power system. For the first threshold and the second threshold, the first threshold is set to be less than the second threshold. Because the load of the power system of the battery swap station is relatively high during the high-load time period, the protection of the stability of the power system should be improved. During the low-load time period, the remaining power reserve in the battery swap station and the waiting time of the vehicle should be prioritized, and the protection of the stability of the power system should be reduced. Therefore, the division of the target segments and the setting of the first threshold and the second threshold are optimized in the present invention. Different priorities are set for different time periods to set corresponding optimization goals. While improving the user experience, the impact on the power system is reduced, and the stability and safety of the power operation of the battery swap station are improved.

[0122] S3. Optimize the first charging strategy according to the optimization target segment division result to obtain a second charging strategy, and obtain future state data of the battery pack after the second charging strategy is executed.

[0123] The first charging strategy performs charging control with the optimal charging efficiency as the optimization target. However, it is difficult to fully balance the needs of the power grid and the battery swap station with only the optimal charging efficiency. Therefore, in the preferred embodiment of the present invention, the first charging strategy is optimized based on the optimization target segment division result. Under the premise of optimal charging efficiency, the needs and dynamic changes of the power grid and the battery swap station are fully considered to obtain a more reasonable and efficient charging strategy. In the preferred embodiment of the present invention, the first charging strategy of the target time period is optimized according to the optimization target segment division result to obtain a second charging strategy, such as Figure 4 As shown, the following steps are included:

[0124] S301. Determine the time period to which the target time period belongs based on the optimization target segment division result. If the determination result is that the target time period belongs to the optimization target time period with the shortest vehicle waiting time, optimize the first charging strategy with the shortest vehicle waiting time.

[0125] S302: If the judgment result is that the target time period for optimizing the maximum remaining power is reached, the first charging strategy is optimized based on the maximum remaining power of the battery pack.

[0126] S303: If the judgment result is that the time period belongs to the optimization target period with the minimum remaining power, the first charging strategy is optimized with the remaining power of the battery pack being minimized.

[0127] S304: Fusion the optimization results to obtain a second charging strategy.

[0128] In a preferred embodiment of the present invention, the time period to which the target time period belongs is judged based on the optimization target segment division result. For example, 08:00-12:00 is the optimization target period with the maximum remaining power, 12:00-15:00 is the optimization target period with the minimum remaining power, and the target time period is 09:00-11:00, which is in the optimization target period with the maximum remaining power. Then, the first charging strategy is optimized with the maximum remaining power of the battery group to obtain the second charging strategy. If the target time period is 11:00-13:00, then in the 11:00-12:00 time period, the first charging strategy is optimized with the maximum remaining power of the battery group; in the 12:00-13:00 time period, the first charging strategy is optimized with the minimum remaining power of the battery group, and the optimization results of the 11:00-12:00 time period and the 12:00-13:00 time period are merged to obtain the second charging strategy of the target time period.

[0129] After obtaining the second charging strategy, the charging action of the battery pack is performed according to the second charging strategy to obtain the future state data of the battery pack in the target time period, and the future state data of the battery pack includes: battery state, battery capacity, charging power, charging rate and expected charging time.

[0130] S4. Based on the preset battery swap scheduling strategy and the preset robotic arm motion control strategy, simulate the working state of the robotic arm under the preset different traffic flow data, and determine the battery swap control potential boundary and the robotic arm collaborative operation capability boundary under the different traffic flow data.

[0131] The joint freedom, material strength and sensor accuracy of the robot arm, as well as the movement speed, movement trajectory and workload have a great impact on the battery replacement efficiency. At the same time, different traffic flow data also have a great impact on the battery replacement efficiency. Therefore, in this application, based on the preset battery replacement scheduling strategy and the preset robot arm motion control strategy, the working state of the robot arm under different traffic flow data is simulated to determine the battery replacement control potential boundary and the robot arm collaborative operation capability boundary under different traffic flow data, such as Figure 5 As shown, the following steps are included:

[0132] S401. Set the robot arm attribute parameter data and the robot arm operation data according to the preset robot arm motion control strategy, wherein the robot arm attribute parameter data at least includes: joint degrees of freedom, material strength and sensor accuracy, and the robot arm operation data at least includes: motion speed, motion trajectory and workload; grasping accuracy, and battery replacement operation record.

[0133] S402. Extract the start timestamp and end timestamp of each battery replacement operation within the preset time period length from the battery replacement operation record, and calculate the time series of a single battery replacement operation based on the start timestamp and the end timestamp.

[0134] S403. Calculate the product of the single battery replacement operation time series and the robot arm grasping success rate, and use the product as the effective battery replacement time series.

[0135] S404. Based on the effective battery replacement time series, a queuing system model is used to model the preset different vehicle flow data, and the vehicle arrival interval and battery replacement service time are obtained through distribution fitting.

[0136] S405. Simulate the working state of the robotic arm and the simulated battery swapping efficiency corresponding to the preset battery swapping scheduling strategy and the preset robotic arm motion control strategy under the vehicle arrival interval and the battery swapping service time through the Monte Carlo simulation method.

[0137] S406. Using the preset different vehicle flow data, the preset battery swap scheduling strategy, the preset robotic arm motion control strategy and the simulated battery swap efficiency, train the battery swap efficiency prediction model constructed using the random forest algorithm.

[0138] S407. Use a genetic algorithm to search the trained battery swap efficiency prediction model to obtain an optimal battery swap scheduling strategy and an optimal robotic arm motion control strategy that maximize the simulated battery swap efficiency.

[0139] S408. Analyze the optimal battery swap scheduling strategy and the optimal robotic arm motion control strategy, as well as the corresponding working states and simulated battery swap efficiency, to obtain the battery swap control potential boundary and the robotic arm collaborative operation capability boundary under different vehicle flow data.

[0140] According to the preset robot arm motion control strategy and the robot arm design specification, the robot arm attribute parameter data and the robot arm operation data are set. The robot arm attribute parameter data at least include: joint degrees of freedom, material strength and sensor accuracy. The robot arm operation data at least include: motion speed, motion trajectory and workload.

[0141] Furthermore, the start timestamp and end timestamp of each battery replacement operation are extracted from the battery replacement operation record, and the time difference between the start timestamp and the end timestamp is calculated to obtain the time consumed by a single battery replacement operation, which is the single battery replacement operation time.

[0142] The robot arm grasping success rate within a preset time period is obtained from the robot arm monitoring system, and the product of the single battery replacement operation time series and the robot arm grasping success rate within the preset time period is calculated, and the product is used as the effective battery replacement time series.

[0143] For example, a battery swap operation starts at 10:15:30 and ends at 10:19:45, and the calculated time for a single battery swap is 4 minutes and 15 seconds. All single battery swap operation times within the preset time period length are used to form a single battery swap operation time sequence within the preset time period length.

[0144] Multiply the time series of a single battery swap operation within the preset time period by the obtained robot grab success rate, and use the product as the effective battery swap time series. For example, if the robot grab success rate is 95% obtained from the monitoring system of the robot controller, the effective battery swap time is calculated to be 4 minutes and 15 seconds * 95% = 4 minutes and 02 seconds.

[0145] Based on the effective battery replacement time series, the queuing system model is used to model the preset different traffic flow data, and the vehicle arrival interval and battery replacement service time are obtained by distribution fitting. In a preferred embodiment of the present invention, the G / G / c queuing theory model is used to model the rules of the preset different traffic flow data, and the vehicle arrival interval and battery replacement service time are fitted by general distribution. Furthermore, the Monte Carlo simulation method is used to simulate the working state and simulated battery replacement efficiency of the robotic arm corresponding to the preset battery replacement scheduling strategy and the preset robotic arm motion control strategy under different vehicle arrival intervals and battery replacement service times. The preset battery scheduling strategy includes "shortest charging time priority" and "highest remaining power priority", and the preset robotic arm motion control strategy includes "shortest path planning" and "load balancing". Multiple simulation experiments are carried out for each strategy combination, and the working state and simulated battery replacement efficiency of the robotic arm are recorded.

[0146] The battery replacement efficiency is the ratio of the number of battery replacements that a battery replacement station can successfully complete to the rated number within a specific time, or the ratio of the speed at which the battery replacement station can successfully complete to the rated speed.

[0147] Furthermore, principal component analysis is used to perform feature analysis on the simulation results to reduce the feature dimension, and then the random forest algorithm is used to analyze the processed data. The preset different traffic flow data, preset battery swap scheduling strategy and preset robot arm motion control strategy are used as input features, and the simulated battery swap efficiency is used as the output label. The battery swap efficiency prediction model constructed using the random forest algorithm is trained to predict the battery swap efficiency of the battery swap station under different traffic flow data. Based on the trained battery swap efficiency prediction model, the genetic algorithm is used to search the trained battery swap efficiency prediction model under the preset different traffic flow data to obtain the optimal strategy combination under different traffic flow data. The optimal strategy combination is the optimal battery swap scheduling strategy and the optimal robot arm motion control strategy that maximize the simulated battery swap efficiency. The optimal battery swap scheduling strategy, the optimal robot arm motion control strategy and the corresponding simulated battery swap efficiency are analyzed to obtain the battery swap regulation potential boundary and the robot arm collaborative operation capability boundary under different traffic flow data.

[0148] For example, in the G / G / c queuing model, c is set to 3 service windows. Through the analysis of historical data, it is found that the vehicle arrival interval conforms to the Erlang distribution, with parameters k=2, λ=10 vehicles / hour, and the battery replacement service time conforms to the log-normal distribution, μ=1.1, σ=0.2. The Monte Carlo simulation is run 10,000 times to simulate the change of traffic flow from 10 vehicles / hour to 100 vehicles / hour. The battery scheduling strategy includes "shortest charging time priority" and "highest remaining power priority" with weights of 0.6 and 0.4 respectively, and the robot motion control strategy includes "shortest path planning" and "load balancing" with weights of 0.7 and 0.3 respectively. The principal component analysis reduces the original 14 features to 6 principal components. The random forest model contains 100 decision trees, the maximum depth of each tree is 10, and the minimum number of leaf node samples is 5. The genetic algorithm sets the population size to 100, the number of iterations to 1000, the crossover rate to 0.8, and the mutation rate to 0.1. According to the preset traffic flow data, different traffic flow intervals are set. In the first traffic flow interval, the optimal strategy combination is "highest remaining power priority" and "load balancing", the battery swap efficiency reaches 95%, and the robot arm operates 60 times per hour; in the second traffic flow interval, the optimal strategy is "shortest charging time priority" and "shortest path planning", the battery swap efficiency is 85%, and the robot arm operates 50 times per hour. Analyze the above information to obtain the battery swap regulation potential boundary and the robot arm collaborative operation capability boundary. From the above information, it can be seen that in the first traffic flow interval, the battery swap regulation potential boundary is 95%, and the robot arm collaborative operation capability boundary is 50 operations per hour. In the second traffic flow interval, the battery swap regulation potential boundary is 85%, and the robot arm collaborative operation capability boundary is 60 operations per hour, which are stored in the database of the battery swap station.

[0149] In a preferred embodiment of the present invention, a simulation method is used to simulate the working state of the robotic arm under different vehicle flow data, and then the battery battery swapping regulation potential boundary and the robotic arm collaborative operation capability boundary under different vehicle flow data are obtained, eliminating the error caused by the unreasonable use of historical data to predict the actual battery swapping scheduling strategy and the robotic arm control strategy. The simulation method can reduce the deviation caused by errors in the actual control process, and the obtained battery battery swapping regulation potential boundary and the robotic arm collaborative operation capability boundary are more accurate, which improves the rigor of constraint setting when constructing an integrated control model of battery swapping scheduling and robotic arm motion, and the accuracy of model prediction.

[0150] S5. Taking the battery swapping regulation potential boundary and the robotic arm collaborative operation capability boundary as constraints, construct an integrated control model of battery swapping scheduling and robotic arm movement.

[0151] Taking the battery swapping control potential boundary and the robot arm collaborative operation capability boundary as constraint conditions, an integrated control model of battery swapping scheduling and robot arm motion is constructed, such as Figure 6 As shown, the following steps are included:

[0152] S501. Construct a constraint equation with the battery swapping control potential boundary and the robotic arm collaborative working capability boundary as constraint conditions.

[0153] S502: Taking the battery pack status data, the workload and the flow data as decision variables, a mixed integer linear programming method is used to construct an objective function.

[0154] S503. Construct an integrated control model of battery swapping scheduling and robotic arm motion based on the constraint equations and the objective function.

[0155] Obtain the battery swap control potential boundary and the robotic arm collaborative operation capability boundary stored in the database, fuse these data into a unified data structure through the Kalman filter, construct the battery swap station state constraint matrix, and construct the constraint equation.

[0156] The battery pack status data, workload and traffic flow data in the robot arm parameters are used as decision variables, and the battery swap efficiency and robot arm utilization are used as optimization targets. The mixed integer linear programming method is used to construct the objective function. Based on the constraint equations and the objective function, an integrated control model for battery swap scheduling and robot arm motion is constructed. Constructing a control model based on the objective function is a conventional method of mixed integer linear programming, which will not be repeated here.

[0157] Mixed integer linear programming is a special type of linear programming problem, which is characterized by the fact that the decision variables contain both continuous variables and integer variables. It is widely used in production planning and scheduling, logistics and supply chain management, resource allocation problems and network design.

[0158] S6. Using the future state data and the reservation data as optimization conditions, the integrated control model of battery swap scheduling and robotic arm motion is optimized to generate a target battery swap scheduling strategy and a target robotic arm motion control strategy.

[0159] In a preferred embodiment of the present invention, Figure 7 As shown, step S6 includes the following steps:

[0160] S601. Use a Kalman filter to merge the future state data and the reservation data into a unified data structure to obtain a state matrix of the battery swap station.

[0161] S602. Input the battery swap station state matrix into the integrated control model of battery swap scheduling and robotic arm motion, and use a model predictive control algorithm to solve it to generate multiple groups of candidate solutions, wherein the candidate solutions include a battery swap scheduling strategy and a robotic arm motion control strategy.

[0162] S603. Use the weighted sum method to convert the battery exchange efficiency and the robot arm utilization rate into a single evaluation index, evaluate and rank each candidate solution according to the single evaluation index, and select the target candidate solution according to the evaluation and ranking results.

[0163] S604. Based on a deep reinforcement learning algorithm, the target candidate solution is used as an action space, the battery swap station state matrix is ​​used as a state space, and the single evaluation index is used as a reward function to construct a decision model.

[0164] S605. Interactive learning is performed on the decision model and the simulation environment to generate a target battery swap scheduling strategy and a target robotic arm motion control strategy.

[0165] In a preferred embodiment of the present invention, a Kalman filter is used to integrate the future state data and reservation data of the battery pack, as well as the real-time workload of the robotic arm, into a unified data structure to obtain a state matrix of the battery swap station. The state matrix of the battery swap station is input into the integrated control model of battery swap scheduling and robotic arm motion, and the model predictive control algorithm is used to solve it, predict the state in the future period of time, and generate multiple sets of candidate solutions, which include battery swap scheduling strategies and robotic arm motion control strategies. The weighted sum method is used to convert the battery swap efficiency and robotic arm utilization rate into a single evaluation index, and each candidate solution is evaluated and ranked according to the single evaluation index, and the target candidate solution is selected according to the evaluation and ranking results. A decision model is constructed based on the DeepQ-Network algorithm, and the evaluated target candidate solution is used as the action space, the battery swap station state matrix is ​​used as the state space, and the single evaluation index is used as the reward function. The decision model is interactively learned with the simulation environment, and the optimal battery swap scheduling strategy and robotic arm motion control strategy are generated in real time as the target battery swap scheduling strategy and target robotic arm motion control strategy.

[0166] In a specific embodiment of the present invention, during the operation of the battery swap station, the battery regulation potential boundary is extracted from the real-time database through the reservation data, which is 90%-95%, and the robot arm operation capacity boundary is 50-60 operations per hour. The future state data of the battery pack includes the battery status, battery capacity, charging power, charging rate and expected charging time of 50 battery packs, as well as the robot arm workload of 30kg, and the reservation data is 50 units / hour. The Kalman filter fuses these data at a frequency of 0.1 seconds to construct a state matrix of the battery swap station containing 50 batteries and 3 robot arms. The objective function is the weighted sum of maximizing the battery swap efficiency and the utilization rate of the robot arm, and the single operation time of the robot arm does not exceed 3 minutes. The model predictive control algorithm uses 5 minutes as the prediction cycle to generate 10 groups of candidate solutions. Each group of candidate solutions contains a scheduling plan for the target time period, such as battery 1 is replaced after 10 minutes, and robot arm 2 performs pick-and-place operations. In the single evaluation index calculated by the weighted sum method, the weight of the battery swap efficiency is 0.6 and the weight of the robot arm utilization is 0.4. The DeepQ-Network algorithm uses a three-layer neural network, with the number of nodes in the input layer equal to the state matrix dimension, 100 nodes in the hidden layer, and 10 actions in the output layer. After 1,000 iterations of training, the ε value of the ε-greedy strategy decayed from 0.9 to 0.1, and the target battery swap scheduling strategy and target robotic arm motion control strategy generated in the end can increase the battery swap efficiency to 95% and the robotic arm utilization rate to 85%.

[0167] Further, in a preferred embodiment of the present invention, Figure 8 As shown, the intelligent battery replacement control method also includes the following steps:

[0168] S01. Generate multiple collision-free motion trajectory solutions according to the target robot arm motion control strategy, optimize and solve the multiple collision-free motion trajectory solutions to obtain an optimal robot arm motion path.

[0169] S02. Based on the target battery replacement scheduling strategy and the optimal robotic arm motion path, dynamically match the vehicle and battery to generate a battery replacement task execution sequence.

[0170] In a preferred embodiment of the present invention, according to the target robot arm motion control strategy, multiple collision-free motion trajectory solutions are generated, and the collision-free motion trajectory solutions are optimized based on the particle swarm optimization algorithm to obtain the optimal robot arm motion path. Among them, the objective function of the particle swarm optimization algorithm includes the vehicle waiting time weight and the robot arm energy consumption weight; if the vehicle waiting time weight is greater than the robot arm energy consumption weight, the motion path that minimizes the vehicle waiting time is given priority; if the robot arm energy consumption weight is greater than the vehicle waiting time weight, the motion path that minimizes the robot arm energy consumption is given priority.

[0171] In a preferred specific embodiment of the present invention, according to the target robot motion control strategy, the robot motion path is planned using a rapidly expanding random tree algorithm, a collision detection module is added to the algorithm, and obstacles and other robots in the robot workspace are considered to generate multiple feasible collision-free motion trajectory solutions. The generated motion trajectory solution is optimized based on the particle swarm optimization algorithm, the vehicle waiting time weight is set to 0.6, the robot energy consumption weight is set to 0.4, a weighted sum objective function is constructed, and the path length, execution time and energy consumption are comprehensively considered to select the optimal trajectory, and finally generate the battery replacement task execution sequence.

[0172] Furthermore, in a preferred embodiment of the present invention, the constraint condition also includes the remaining battery life, such as Fig. 9 As shown, the method for obtaining the remaining battery life includes the following steps:

[0173] S001. Acquire battery pack charge and discharge history data, wherein the battery pack charge and discharge history data includes: number of charge and discharge cycles, capacity attenuation, charging rate, and battery temperature.

[0174] S002. Training a battery health status prediction model constructed using a random forest algorithm based on the battery pack charge and discharge history data.

[0175] S003. Input the collected real-time data of battery pack charge and discharge into the trained battery health status prediction model to predict the current health status of the battery pack and obtain the remaining service life of the battery.

[0176] The number of charge and discharge cycles, capacity decay data and charging rate records of each battery pack are extracted from the battery management system database. At the same time, the temperature data of the battery is obtained from the environmental monitoring system, and the random forest algorithm is used to build a battery health status prediction model. The battery health status prediction model is trained based on the historical data of the number of charge and discharge cycles, capacity decay, charging rate and battery temperature for health monitoring and remaining service life prediction of the battery pack. The real-time data of the number of charge and discharge cycles, capacity decay, charging rate and battery temperature of the collected battery pack are input into the trained battery health status prediction model to output the battery health status score and the remaining service life of the battery. The remaining service life obtained is used as another constraint condition for constructing an integrated control model for battery swap scheduling and robotic arm motion.

[0177] In a preferred specific embodiment of the present invention, a battery has completed 500 charge and discharge cycles, the current capacity is 85% of the initial capacity, and the average charging rate is 1C. At the same time, the environmental monitoring system records the ambient temperature of the battery every 5 minutes, such as 25°C. These data are used as input features of the random forest algorithm to build a model containing 100 decision trees. Through cross-validation, the battery health status score is 82 points, and the remaining service life is predicted to be 300 days.

[0178] In a preferred embodiment of the present invention, an intelligent scheduling algorithm is used to analyze the battery pack status data, charging pile status data and reservation data of the target time period in the battery swap station to generate a first charging strategy; a correlation analysis is performed on the historical vehicle flow data and historical load data of the battery swap station within the preset time period length, and the preset time period length is divided into optimized target segments according to the correlation analysis results. The optimized target segment division results are: the shortest vehicle waiting time optimization target period, the maximum remaining power optimization target period and the minimum remaining power optimization target period; the first charging strategy is optimized according to the optimized target segment division results to obtain the second charging strategy, and the first charging strategy is obtained. The future state data of the battery pack after the execution of the second charging strategy; based on the preset battery swap scheduling strategy and the preset robot arm motion control strategy, simulate the working state of the robot arm under different preset traffic flow data, and determine the battery swap control potential boundary and the robot arm collaborative operation capability boundary under different traffic flow data; use the battery swap control potential boundary and the robot arm collaborative operation capability boundary as constraints to construct an integrated control model of battery swap scheduling and robot arm motion; in the actual scheduling and control process, use the future state data and reservation data as optimization conditions to optimize the integrated control model of battery swap scheduling and robot arm motion, and generate the target battery swap scheduling strategy and target robot arm motion control strategy. The intelligent battery swap control method provided in this application can cope with the variability of traffic flow, improve the efficiency of battery swapping, and improve the ability of the robot arm to work collaboratively, thereby ensuring the reliability of battery swapping operations.

[0179] Accordingly, if Fig.10 As shown, based on an intelligent battery swapping control method, an embodiment of the present invention further provides an intelligent battery swapping control system to implement the intelligent battery swapping control method disclosed in an embodiment of the present invention, including: a charging strategy generating unit 1, an optimization target segment dividing unit 2, a charging strategy optimizing unit 3, a boundary determining unit 4, an integrated control model building unit 5 and a scheduling control unit 6;

[0180] The charging strategy generating unit 1 is used to analyze the acquired battery pack status data, charging pile status data and reservation data of the target time period in the battery swap station by using an intelligent scheduling algorithm to generate a first charging strategy.

[0181] The optimization target segment division unit 2 is used to perform correlation analysis on the historical vehicle flow data and historical load data of the battery swap station within a preset time period length, and divide the preset time period length into optimization target segments according to the correlation analysis results. The optimization target segment division results are: the optimization target period with the shortest vehicle waiting time, the optimization target period with the maximum remaining power, and the optimization target period with the minimum remaining power.

[0182] The charging strategy optimization unit 3 is used to optimize the first charging strategy according to the optimization target segment division result to obtain a second charging strategy, and obtain future state data of the battery pack after the second charging strategy is executed.

[0183] The boundary determination unit 4 is used to simulate the working state of the robotic arm under preset different vehicle flow data based on the preset battery replacement scheduling strategy and the preset robotic arm motion control strategy, and determine the battery replacement control potential boundary and the robotic arm collaborative operation capability boundary under different vehicle flow data.

[0184] The integrated control model construction unit 5 is used to construct an integrated control model of battery replacement scheduling and robotic arm movement based on the battery replacement regulation potential boundary and the robotic arm collaborative operation capability boundary as constraints.

[0185] The scheduling control unit 6 is used to use the future state data and the reservation data as optimization conditions in the actual scheduling control process, optimize the integrated control model of battery swap scheduling and robotic arm motion, and generate a target battery swap scheduling strategy and a target robotic arm motion control strategy.

[0186] For the specific definition of an intelligent battery exchange control system, please refer to the above-mentioned definition of an intelligent battery exchange control method, which will not be repeated here. A person of ordinary skill in the art can appreciate that the various modules and steps described in conjunction with the embodiments disclosed in the present invention can be implemented in hardware, software, or a combination of both. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0187] An intelligent battery swap control method and system provided in this embodiment is aimed at the technical problem that the battery swap scheduling of battery swap stations cannot cope with the existence of multiple variables and cannot take into account multiple requirements such as benefits, efficiency and reliability. The present invention uses an intelligent scheduling algorithm to analyze the battery pack status data, charging pile status data and reservation data of the target time period obtained in the battery swap station to generate a first charging strategy; a correlation analysis is performed on the historical traffic flow data and historical load data of the battery swap station within the preset time period length, and the preset time period length is divided into optimized target segments according to the correlation analysis results. The optimized target segment division results are: the shortest vehicle waiting time optimization target period, the maximum remaining power optimization target period and the minimum remaining power optimization target period; the first charging strategy is optimized according to the optimization target segment division results to obtain a second charging strategy, and the second charging strategy is obtained. The future state data of the battery pack after the strategy is executed; based on the preset battery swap scheduling strategy and the preset robot arm motion control strategy, simulate the working state of the robot arm under different preset traffic flow data, and determine the battery swap regulation potential boundary and the robot arm collaborative operation capability boundary under different traffic flow data; use the battery swap regulation potential boundary and the robot arm collaborative operation capability boundary as constraints to construct an integrated control model of battery swap scheduling and robot arm motion; in the actual scheduling and control process, use the future state data and reservation data as optimization conditions to optimize the integrated control model of battery swap scheduling and robot arm motion, and generate the target battery swap scheduling strategy and target robot arm motion control strategy. The intelligent battery swap control method provided in this application can cope with the variability of traffic flow, improve the efficiency of battery swapping, and improve the ability of the robot arm to work collaboratively, thereby ensuring the reliability of battery swapping operations.

[0188] Each embodiment in this specification is described in a progressive manner, and the same or similar parts of each embodiment can be directly referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. It should be noted that the technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above-mentioned embodiments are not 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.

[0189] The above-mentioned embodiments only express several preferred implementation modes of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this technical field, several improvements and substitutions can be made without departing from the technical principle of the present invention, and these improvements and substitutions should also be regarded as the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be based on the protection scope of the claims.

Claims

1. An intelligent battery replacement control method, characterized in that: The method comprises: An intelligent scheduling algorithm is used to analyze the acquired battery pack status data, charging pile status data and reservation data of the target time period in the battery swap station to generate a first charging strategy; Perform correlation analysis on the historical vehicle flow data and historical load data of the battery swap station within the preset time period length, and divide the preset time period length into optimization target segments according to the correlation analysis results. The optimization target segment division results are: the optimization target period with the shortest vehicle waiting time, the optimization target period with the maximum remaining power, and the optimization target period with the minimum remaining power; Optimizing the first charging strategy according to the optimization target segment division result to obtain a second charging strategy, and acquiring future state data of the battery pack after the second charging strategy is executed; Based on the preset battery swap scheduling strategy and the preset robotic arm motion control strategy, simulate the working state of the robotic arm under different preset traffic flow data, and determine the battery swap control potential boundary and the robotic arm collaborative operation capability boundary under different traffic flow data; Taking the battery swapping regulation potential boundary and the robotic arm collaborative operation capability boundary as constraints, a battery swapping scheduling and robotic arm motion integrated control model is constructed; In the actual dispatching and control process, the future state data and the reservation data are used as optimization conditions to optimize the integrated control model of battery swapping dispatching and robotic arm motion, and generate a target battery swapping dispatching strategy and a target robotic arm motion control strategy; The step of optimizing the first charging strategy of the target time period according to the optimization target segment division result to obtain a second charging strategy includes: Based on the optimization target segment division result, determine the time period to which the target time period belongs, and if the determination result is that it belongs to the optimization target period with the shortest vehicle waiting time, optimize the first charging strategy with the shortest vehicle waiting time; If the judgment result is that it belongs to the maximum remaining power optimization target period, optimizing the first charging strategy with the maximum remaining power of the battery pack; If the judgment result is that it belongs to the optimization target period of minimum remaining power, optimizing the first charging strategy with the minimum remaining power of the battery pack; The optimization results are integrated to obtain a second charging strategy; The determining of the battery swapping control potential boundary and the robot arm collaborative operation capability boundary under different vehicle flow data includes: A Monte Carlo simulation method is used to simulate the working state and simulated battery swap efficiency of the robotic arm corresponding to the preset battery swap scheduling strategy and the preset robotic arm motion control strategy under the preset different vehicle flow data; The battery swap efficiency prediction model constructed by using the preset different vehicle flow data, the preset battery swap scheduling strategy, the preset robot arm motion control strategy and the simulated battery swap efficiency is trained; A genetic algorithm is used to search the trained battery swap efficiency prediction model to obtain an optimal battery swap scheduling strategy and an optimal robotic arm motion control strategy that maximize the simulated battery swap efficiency; The optimal battery swap scheduling strategy, the optimal robotic arm motion control strategy, and the corresponding working states and simulated battery swap efficiencies are analyzed to obtain the battery swap regulation potential boundaries and robotic arm collaborative operation capability boundaries under different vehicle flow data.

2. The intelligent battery replacement control method according to claim 1, characterized in that: The intelligent scheduling algorithm is used to analyze the acquired battery pack status data, charging pile status data and reservation data of the target time period in the battery swap station to generate a first charging strategy, including: Based on the acquired battery pack status data, charging pile status data and reservation data of the target time period, an intelligent scheduling algorithm is used to build a battery pack and charging pile matching matrix; Taking the optimal charging efficiency as the optimization goal, the matching matrix between the battery pack and the charging pile is optimized and solved to obtain the first charging strategy.

3. The intelligent battery replacement control method according to claim 1, characterized in that: The performing of correlation analysis on the historical vehicle flow data and historical load data of the battery swap station within the preset time period, and dividing the preset time period into optimized target segments according to the correlation analysis result, includes: Calculating the charging demand under the historical vehicle flow data according to the historical vehicle flow data; According to the historical load data, the preset time period length is divided into a number of high-load time periods and a number of low-load time periods; A random forest regression algorithm is used to perform correlation analysis on the charging demand and the historical load data to establish a nonlinear mathematical model; Calculating a load impact coefficient according to the nonlinear mathematical model, wherein the load impact coefficient is defined as a ratio of the charging demand to the historical load data at the same time; Calculating the average value of the load influence coefficient in each of the high-load time periods and each of the low-load time periods respectively; For the high-load time period, if the mean is greater than or equal to a first threshold, the corresponding high-load time period is divided into a target time period for optimizing the shortest vehicle waiting time; if the mean is less than the first threshold, the corresponding high-load time period is divided into a target time period for optimizing the minimum remaining power; For the low-load time period, if the mean is greater than or equal to the second threshold, the corresponding low-load time period is divided into an optimization target period with the shortest vehicle waiting time; if the mean is less than the second threshold, the low-load time period is divided into an optimization target period with the maximum remaining power.

4. The intelligent battery replacement control method according to claim 1, characterized in that: The Monte Carlo simulation method is used to simulate the working state of the mechanical arm and the simulated battery swap efficiency corresponding to the preset battery swap scheduling strategy and the preset mechanical arm motion control strategy under the preset different vehicle flow data, including: According to the preset robot arm motion control strategy, the robot arm attribute parameter data and the robot arm operation data are set, wherein the robot arm attribute parameter data at least includes: joint freedom, material strength and sensor accuracy, and the robot arm operation data at least includes: motion speed, motion trajectory and workload; Extract the start timestamp and end timestamp of each battery replacement operation within the preset time period from the battery replacement operation record, and calculate the time series of a single battery replacement operation according to the start timestamp and the end timestamp; Calculate the product of the single battery replacement operation time series and the robot arm grasping success rate, and use the product as the effective battery replacement time series; Based on the effective battery replacement time series, a queuing system model is used to model the preset different vehicle flow data, and the vehicle arrival interval and battery replacement service time are obtained through distribution fitting; The Monte Carlo simulation method is used to simulate the working state of the robotic arm and the simulated battery swap efficiency corresponding to the preset battery swap scheduling strategy and the preset robotic arm motion control strategy under the vehicle arrival interval and the battery swap service time.

5. The intelligent battery replacement control method according to claim 4, characterized in that: The battery swapping control potential boundary and the robot arm collaborative operation capability boundary are used as constraints to construct an integrated control model of battery swapping scheduling and robot arm movement, including: Constructing a constraint equation with the battery swapping regulation potential boundary and the robotic arm collaborative operation capability boundary as constraint conditions; Taking the battery pack status data, the workload and the vehicle flow data as decision variables, a mixed integer linear programming method is used to construct an objective function; An integrated control model of battery swapping scheduling and robotic arm motion is constructed based on the constraint equations and the objective function.

6. The intelligent battery replacement control method according to claim 1, characterized in that: The method uses the future state data and the reservation data as optimization conditions, performs optimization processing on the integrated control model of battery swap scheduling and robotic arm motion, and generates a target battery swap scheduling strategy and a target robotic arm motion control strategy, including: A Kalman filter is used to fuse the future state data and the reservation data into a unified data structure to obtain a state matrix of the battery swap station; Input the state matrix of the battery swap station into the integrated control model of battery swap scheduling and robotic arm motion, and solve it using a model predictive control algorithm to generate multiple sets of candidate solutions, wherein the candidate solutions include a battery swap scheduling strategy and a robotic arm motion control strategy; The battery swap efficiency and the robot arm utilization rate are converted into a single evaluation index by using a weighted sum method, each candidate solution is evaluated and ranked according to the single evaluation index, and a target candidate solution is selected according to the evaluation and ranking results; Based on a deep reinforcement learning algorithm, the target candidate solution is used as an action space, the battery swap station state matrix is ​​used as a state space, and the single evaluation index is used as a reward function to construct a decision model; The decision model is interactively learned with the simulation environment to generate a target battery swap scheduling strategy and a target robotic arm motion control strategy.

7. The intelligent battery replacement control method according to claim 1, characterized in that: The method further comprises: According to the target robot arm motion control strategy, multiple collision-free motion trajectory solutions are generated, and the multiple collision-free motion trajectory solutions are optimized and solved to obtain the optimal robot arm motion path; Based on the target battery replacement scheduling strategy and the optimal robotic arm motion path, the vehicle and battery are dynamically matched to generate a battery replacement task execution sequence.

8. The intelligent battery replacement control method according to claim 1, characterized in that: The constraints also include remaining battery life; The method for obtaining the remaining service life of the battery includes: Acquire battery pack charge and discharge history data, wherein the battery pack charge and discharge history data includes: charge and discharge cycle number, capacity attenuation, charging rate and battery temperature; Training a battery health status prediction model constructed using a random forest algorithm based on the battery pack charge and discharge history data; The collected real-time data of battery pack charge and discharge is input into the trained battery health status prediction model to predict the current health status of the battery pack and obtain the remaining service life of the battery.

9. An intelligent battery swap control system, implementing the intelligent battery swap control method according to any one of claims 1 to 8, characterized in that: The system comprises: a charging strategy generating unit, an optimization target segment dividing unit, a charging strategy optimizing unit, a boundary determining unit, an integrated control model building unit and a dispatching control unit; The charging strategy generating unit is used to analyze the acquired battery pack status data, charging pile status data and reservation data of the target time period in the battery swap station by using an intelligent scheduling algorithm to generate a first charging strategy; The optimization target segment division unit is used to perform correlation analysis on the historical vehicle flow data and historical load data of the battery swap station within the preset time period length, and divide the preset time period length into optimization target segments according to the correlation analysis result, and the optimization target segment division result is: the optimization target period with the shortest vehicle waiting time, the optimization target period with the maximum remaining power, and the optimization target period with the minimum remaining power; The charging strategy optimization unit is used to optimize the first charging strategy according to the optimization target segment division result to obtain a second charging strategy, and obtain future state data of the battery pack after the second charging strategy is executed; The boundary determination unit is used to simulate the working state of the robotic arm under different preset traffic flow data based on the preset battery swap scheduling strategy and the preset robotic arm motion control strategy, and determine the battery swap regulation potential boundary and the robotic arm collaborative operation capability boundary under different traffic flow data; The integrated control model construction unit is used to construct an integrated control model of battery swap scheduling and robotic arm movement based on the battery swap regulation potential boundary and the robotic arm collaborative operation capability boundary as constraints; The scheduling control unit is used to use the future state data and the reservation data as optimization conditions in the actual scheduling control process, optimize the integrated control model of battery swap scheduling and robotic arm motion, and generate a target battery swap scheduling strategy and a target robotic arm motion control strategy; The step of optimizing the first charging strategy of the target time period according to the optimization target segment division result to obtain a second charging strategy includes: Based on the optimization target segment division result, determine the time period to which the target time period belongs, and if the determination result is that it belongs to the optimization target period with the shortest vehicle waiting time, optimize the first charging strategy with the shortest vehicle waiting time; If the judgment result is that it belongs to the maximum remaining power optimization target period, optimizing the first charging strategy with the maximum remaining power of the battery pack; If the judgment result is that it belongs to the optimization target period of minimum remaining power, optimizing the first charging strategy with the minimum remaining power of the battery pack; The optimization results are integrated to obtain a second charging strategy; The determining of the battery swapping control potential boundary and the robot arm collaborative operation capability boundary under different vehicle flow data includes: A Monte Carlo simulation method is used to simulate the working state and simulated battery swap efficiency of the robotic arm corresponding to the preset battery swap scheduling strategy and the preset robotic arm motion control strategy under the preset different vehicle flow data; The battery exchange efficiency prediction model constructed by using the preset different vehicle flow data, the preset battery exchange scheduling strategy, the preset robot arm motion control strategy and the simulated battery exchange efficiency is trained; A genetic algorithm is used to search the trained battery swap efficiency prediction model to obtain an optimal battery swap scheduling strategy and an optimal robotic arm motion control strategy that maximize the simulated battery swap efficiency; The optimal battery swap scheduling strategy, the optimal robotic arm motion control strategy, and the corresponding working states and simulated battery swap efficiencies are analyzed to obtain the battery swap regulation potential boundaries and robotic arm collaborative operation capability boundaries under different vehicle flow data.

Images