Battery charging strategy intelligent optimization method and system

Abstract

The invention relates to the technical field of data processing, and provides an intelligent optimization method and system for a battery charging strategy, and the method comprises the steps: obtaining the historical charging and discharging data, environment parameter data and battery health state data of a battery, generating a battery operation data set, constructing a multi-dimensional charging characteristic spectrum, carrying out the weighted analysis of different charging characteristics, and carrying out the calculation of a multi-dimensional charging characteristic spectrum. Performing dynamic adjustment on the charging current and the charging voltage of the battery based on the charging strategy to obtain charging data; and updating the charging characteristic spectrum according to the charging data, and readjusting the charging strategy based on the updated charging characteristic spectrum to obtain an optimized charging strategy. Through a multi-iteration charging strategy, charging current and charging voltage are dynamically adjusted through real-time data in a complex and changeable charging environment, and the problem that the optimization of the charging efficiency and the service life of the battery is insufficient due to the fact that self-adaptive adjustment cannot be achieved in the face of the complex and changeable charging environment and different health states of the battery is solved.

Description

Technical Field

[0001] This application relates to the technical field of data processing, and in particular to an intelligent optimization method and system for battery charging strategies. Background Art

[0002] With the popularization of electronic devices and the rapid development of new energy technologies, batteries, as important carriers for electrical energy storage, are widely used in fields such as electric vehicles, portable electronic devices, and energy storage systems. How to improve the charging efficiency of batteries, extend battery life, and ensure their safety has become the focus of current battery technology research.

[0003] In related technical means, the optimization of battery charging strategies is carried out based on methods with fixed charging modes. These methods charge the battery with preset charging current and voltage values. During the charging process, the battery management system monitors the charging state of the battery and adjusts the charging parameters when necessary to avoid overcharging and overheating. This optimization strategy with a fixed charging mode can achieve efficient charging and protection of the battery.

[0004] For the above technical solutions, although the safe and efficient charging of the battery can be achieved in most cases through a fixed charging mode, in the face of complex and variable charging environments and different battery health states, there is a problem of inability to adaptively adjust, resulting in insufficient optimization of charging efficiency and battery life. Summary of the Invention

[0005] In order to improve the problem that in the face of complex and variable charging environments and different battery health states, there is an inability to adaptively adjust, resulting in insufficient optimization of charging efficiency and battery life, this application provides an intelligent optimization method and system for battery charging strategies.

[0006] The present invention provides an intelligent optimization method for battery charging strategies, including: obtaining historical charge and discharge data, environmental parameter data, and battery health state data of the battery, generating a battery operation data set based on the historical charge and discharge data, the environmental parameter data, and the battery health state data; constructing a multi-dimensional charging feature map through the battery operation data set, and based on the multi-dimensional charging feature map, using an adaptive weight dynamic calculation method to perform weighted analysis on different charging features to obtain a charging feature optimization weight set; constructing a charging mode optimization sequence based on the charging feature optimization weight set, inputting the charging mode optimization sequence into a preset deep reinforcement learning model to obtain a charging strategy, dynamically adjusting the charging current and charging voltage of the battery based on the charging strategy to obtain charging data; updating the charging feature map according to the charging data, and readjusting the charging strategy based on the updated charging feature map to obtain an optimized charging strategy.

[0007] As a preferred solution, the step of obtaining the historical charge and discharge data, environmental parameter data, and battery health status data of the battery, and generating a battery operation data set based on the historical charge and discharge data, the environmental parameter data, and the battery health status data includes: obtaining the historical charge and discharge data, environmental parameter data, and battery health status data of the battery through a data acquisition device, and generating a preliminary battery data set according to the historical charge and discharge data, the environmental parameter data, and the battery health status data; performing denoising processing on the preliminary battery data set to obtain a cleaned battery data set, and extracting key characteristic values such as the charge and discharge rate, temperature fluctuation, and number of cycles of the battery based on the cleaned battery data set to obtain a characteristic data set; inputting the characteristic data set into a preset battery health assessment model for evaluation to obtain a battery health index, selecting valid data that meet the standards through the battery health index, and merging them to form a battery operation data set.

[0008] As a preferred solution, the step of inputting the characteristic data set into a preset battery health assessment model for evaluation to obtain a battery health index, selecting valid data that meet the standards through the battery health index, and merging them to form a battery operation data set includes: performing principal component analysis on the characteristic data set, extracting the internal impedance, temperature change rate, charge and discharge rate, and capacity attenuation rate of the characteristic data set as the main components to obtain a battery health assessment characteristic set; inputting the battery health assessment characteristic set into a battery health assessment model, and using a battery degradation model to simulate the battery aging trend of the battery health assessment characteristic set to generate battery aging trend data; calculating the remaining available life of the current battery by using the battery aging trend data and historical battery health data to obtain battery life prediction data; grading the health status of the battery based on the battery life prediction data to obtain a battery health index, and screening out data that meet the preset health standards according to the battery health index to obtain a battery operation data set.

[0009] As a preferred solution, the step of constructing a multi-dimensional charging characteristic map based on the battery operation data set, and using an adaptive weight dynamic calculation method to perform weighted analysis on different charging characteristics to obtain a charging characteristic optimization weight set includes: extracting characteristics from the battery operation data set to obtain multi-dimensional charging characteristics, performing normalization processing on the multi-dimensional charging characteristics to obtain a charging characteristic data set, constructing a correlation matrix between charging characteristics based on the charging characteristic data set, analyzing the mutual influence relationship between each charging characteristic to obtain a charging characteristic correlation data set; performing weighted analysis on different charging characteristics based on the charging characteristic correlation data set through an adaptive weighting method to obtain the weight coefficient of each charging characteristic, and constructing a charging characteristic optimization weight set based on the weight coefficients of all the charging characteristics.

[0010] As a preferred solution, the step of constructing a charging mode optimization sequence based on the charging feature optimization weight set, inputting the charging mode optimization sequence into a preset deep reinforcement learning model to obtain a charging strategy, and dynamically adjusting the charging current and charging voltage of the battery based on the charging strategy to obtain charging data includes: evaluating the importance of each charging feature by using the charging feature optimization weight set to obtain charging feature evaluation data, and respectively optimizing different scenarios of the charging mode based on the charging feature evaluation data to obtain multiple charging mode candidate solutions; analyzing the effects of the charging mode candidate solutions, evaluating the charging efficiency and battery life under each charging mode to obtain a charging mode optimization sequence; inputting the charging mode optimization sequence into a preset deep reinforcement learning model, and in the deep reinforcement learning model, adaptively adjusting the charging mode based on the charging mode optimization sequence through a reinforcement learning algorithm to obtain a charging strategy, and dynamically adjusting the charging current and charging voltage of the battery based on the charging strategy to obtain charging data.

[0011] As a preferred solution, the step of updating the charging feature map according to the charging data and readjusting the charging strategy based on the updated charging feature map to obtain an optimized charging strategy includes: comparing and analyzing the charging data with historical charging data to identify abnormal fluctuations during the charging process to obtain a charging anomaly data set, and adjusting the charging features in a preset battery performance database according to the charging anomaly data set to obtain a battery performance data set; updating the charging feature map by using the battery performance data set to obtain an updated charging feature map, and optimizing the charging mode of the charging mode optimization sequence based on the updated charging feature map to obtain an optimized charging mode optimization sequence; inputting the optimized charging mode optimization sequence into the deep reinforcement learning model to obtain an updated charging strategy; charging the battery according to the updated charging strategy, and real-time monitoring key feature parameters during the charging process to obtain an optimized charging strategy.

[0012] As a preferred solution, the step of comparing the charging data with historical charging data, identifying abnormal fluctuations during the charging process, obtaining a set of charging abnormal data, and adjusting the charging characteristics in the preset battery performance database according to the set of charging abnormal data to obtain a set of battery performance data includes: By comparing and analyzing the charging data with historical charging data, calculating the current charging voltage deviation, the abnormal change rate of temperature rise, and the mutation of battery internal resistance, a set of charging abnormal parameters is obtained; Using an anomaly detection algorithm to identify abnormal charging states based on the set of charging abnormal parameters, generating a set of charging abnormal data, classifying the set of charging abnormal data, distinguishing normal fluctuations, minor anomalies, and severe anomalies, and assigning different anomaly labels to the normal fluctuations, the minor anomalies, and the severe anomalies; Analyzing the impact of the abnormal charging state on battery life and performance according to the anomaly labels, calculating the life attenuation rate caused by abnormal charging, obtaining charging abnormal impact assessment data, and correcting the charging characteristics in the preset battery performance database with the charging abnormal impact assessment data to obtain a set of battery performance data.

[0013] The present application also provides a battery charging strategy intelligent optimization system, including: an acquisition module, configured to acquire historical charge and discharge data, environmental parameter data, and battery health status data of the battery, and generate a set of battery operation data based on the historical charge and discharge data, the environmental parameter data, and the battery health status data; an analysis module, configured to construct a multi-dimensional charging characteristic map through the set of battery operation data, and based on the multi-dimensional charging characteristic map, use an adaptive weight dynamic calculation method to perform weighted analysis on different charging characteristics to obtain a set of optimized charging characteristic weights; an adjustment module, configured to construct an optimized charging mode sequence based on the set of optimized charging characteristic weights, input the optimized charging mode sequence into a preset deep reinforcement learning model to obtain a charging strategy, and dynamically adjust the charging current and charging voltage of the battery based on the charging strategy to obtain charging data; an optimization module, configured to update the charging characteristic map according to the charging data, and re-adjust the charging strategy based on the updated charging characteristic map to obtain an optimized charging strategy.

[0014] Compared with the prior art, the present application has the following beneficial effects: Adaptive adjustment. By obtaining the historical charge and discharge data, environmental parameter data, and battery health status data of the battery, a battery operation data set is generated. Then, a multi-dimensional charging feature map is constructed using these data. Based on the feature map, a weighted analysis is performed using an adaptive weight dynamic calculation method. After obtaining the optimized weight set of charging features, an optimized charging mode sequence is constructed and input into a deep reinforcement learning model for optimization to obtain a charging strategy. Through multiple iterations, it is possible to dynamically adjust the charging current and charging voltage in real time in a complex and changing charging environment, optimize the charging efficiency and extend the battery life, significantly improve the performance and safety of the battery, and solve the problem that it is impossible to adaptively adjust in the face of a complex and changing charging environment and different battery health statuses, resulting in insufficient optimization of charging efficiency and battery life. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0016] The structures, ratios, sizes, etc. shown in the drawings of this specification are only used to cooperate with the content disclosed in the specification for those who are familiar with this technology to understand and read, and are not used to limit the limiting conditions under which the present invention can be implemented. Therefore, they do not have a substantial technical meaning. Any modification of the structure, change in the proportional relationship, or adjustment of the size, without affecting the effects that the present invention can produce and the purposes that can be achieved, should still fall within the scope that can be covered by the technical content disclosed in the present invention.

[0017] Figure 1 is a schematic flow chart of the intelligent optimization method for the battery charging strategy provided by the embodiment of the present invention; Figure 2 is a schematic block diagram of the structure of the intelligent optimization system for the battery charging strategy provided by the embodiment of the present invention.

[0018] Description of the reference numerals: 10. Intelligent optimization system for battery charging strategy; 11. Acquisition module; 12. Analysis module; 13. Adjustment module; 14. Optimization module. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0020] The flowchart shown in the accompanying drawings is only an example illustration, and does not necessarily include all the content and operations / steps, nor does it necessarily need to be executed in the described order. For example, some operations / steps can also be decomposed, combined, or partially merged, so the actual execution order may change according to the actual situation.

[0021] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit this application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an", and "the" are intended to include the plural forms.

[0022] It should be further understood that the term "and / or" used in this application specification and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0023] The technical solutions of the present invention will be further described below in conjunction with the accompanying drawings and through specific embodiments.

[0024] Embodiment 1: As Figure 1 shown, this application provides an intelligent optimization method for battery charging strategies, including steps S100 to S400.

[0025] Step S100, obtain the historical charge and discharge data, environmental parameter data, and battery health status data of the battery, and generate a battery operation data set based on the historical charge and discharge data, environmental parameter data, and battery health status data.

[0026] In this step, the battery management system collects the historical charge and discharge data, environmental parameter data (such as temperature, humidity), and battery health status data (such as battery capacity, internal resistance, etc.) of the battery from multiple sources. Specifically, sensors and data recording devices are used to monitor the working state of the battery in real time and store the data in a database.

[0027] For example, the environmental temperature data collected by the temperature sensor, the charge and discharge current and voltage data collected by the voltage and current sensors, and the battery health status data collected by the battery management system together constitute the battery operation data set.

[0028] Step S200: Construct a multi-dimensional charging feature map based on the battery operation data set. Based on the multi-dimensional charging feature map, use the adaptive weight dynamic calculation method to perform weighted analysis on different charging features to obtain a set of optimized charging feature weights.

[0029] In this step, first use the data analysis algorithm to process the battery operation data set and extract key charging feature indicators (such as charging current, charging voltage, temperature change rate, etc.). Specifically, use multi-dimensional data analysis technology to construct a feature map containing multiple charging feature dimensions.

[0030] For example, through statistical analysis software, classify the data in the battery operation data set according to different feature dimensions, draw a multi-dimensional charging feature map, and then perform weighted analysis using the adaptive weight dynamic calculation method based on the data distribution in the feature map to obtain a set of optimized charging feature weights.

[0031] Step S300: Construct an optimized charging mode sequence based on the set of optimized charging feature weights, input the optimized charging mode sequence into a preset deep reinforcement learning model to obtain a charging strategy, and dynamically adjust the charging current and charging voltage of the battery based on the charging strategy to obtain charging data.

[0032] In this step, according to the set of optimized charging feature weights, generate multiple charging modes, sort them through an optimization algorithm, and select the best optimized charging mode sequence. Specifically, input the optimized charging mode sequence into a preset deep reinforcement learning model for training to obtain a charging strategy.

[0033] For example, use the Q-learning algorithm in the deep reinforcement learning model to optimize the charging strategies under different charging modes to obtain the best charging strategy based on the current battery operation state, and then dynamically adjust the charging current and charging voltage according to the charging strategy and record the charging data in real time.

[0034] Step S400: Update the charging feature map according to the charging data, and readjust the charging strategy based on the updated charging feature map to obtain an optimized charging strategy.

[0035] In this step, monitor the battery state during the charging process in real time, and update the multi-dimensional charging feature map according to the latest charging data. Specifically, through an iterative optimization algorithm, add the latest charging data to the battery operation data set, reconstruct the charging feature map, and readjust the charging strategy based on the updated feature map.

[0036] For example, during the actual charging process, it is found that the weights of some charging characteristics change. Through the updated charging characteristic map, the optimized weight set of charging characteristics is recalculated to obtain a more accurate charging strategy, thereby improving the charging effect.

[0037] In this embodiment, the historical charge and discharge data, environmental parameter data, and battery health status data of the battery are obtained, and a battery operation data set is generated based on these data. Then, a multi-dimensional charging characteristic map is constructed through the battery operation data set. Based on the multi-dimensional charging characteristic map, the adaptive weight dynamic calculation method is used to perform weighted analysis on different charging characteristics to obtain the optimized weight set of charging characteristics. Next, an optimized charging mode sequence is constructed based on the optimized weight set of charging characteristics and input into a preset deep reinforcement learning model to obtain a charging strategy. Finally, based on the charging strategy, the charging current and charging voltage of the battery are dynamically adjusted to obtain charging data, and the charging characteristic map is updated according to these charging data, and the charging strategy is readjusted to obtain an optimized charging strategy. The intelligence and adaptability of the charging strategy are realized, the charging efficiency is improved, and the battery life is extended. According to the historical charge and discharge data, environmental parameter data, and battery health status data of the battery, the charging current and charging voltage are adjusted in real time to ensure that charging can still be carried out efficiently and safely in a complex and changeable charging environment, and the use performance and life of the battery are optimized to the greatest extent, improving the problem that when facing a complex and changeable charging environment and different battery health statuses, it is impossible to adaptively adjust, resulting in insufficient optimization of charging efficiency and battery life.

[0038] Embodiment 2: In step S100, the historical charge and discharge data, environmental parameter data, and battery health status data of the battery are obtained through a data acquisition device, and a preliminary battery data set is generated according to the historical charge and discharge data, environmental parameter data, and battery health status data.

[0039] By integrating a variety of sensors (such as voltage sensors, temperature sensors, current sensors, etc.) into the battery management system, the historical charge and discharge data, environmental parameter data (such as temperature, humidity, etc.), and battery health status data (such as battery capacity, internal resistance, etc.) of the battery are monitored in real time. Specifically, using a data recording device, these data are regularly uploaded to a central database to form a preliminary battery data set.

[0040] For example, the battery voltage changes recorded by the voltage sensor, the charge and discharge current changes recorded by the current sensor, the environmental temperature changes recorded by the temperature sensor, and the battery health status data recorded by the battery management system constitute the preliminary battery data set.

[0041] Denoise the preliminary battery data set to obtain a cleaned battery data set. Based on the cleaned battery data set, extract the key characteristic values of the charging and discharging rate, temperature fluctuation, and number of cycles of the battery to obtain a characteristic data set.

[0042] By applying data preprocessing algorithms, filter and remove the noise data in the preliminary battery data set. Specifically, adopt data processing techniques such as mean filtering and interpolation method to clean and correct the data to ensure the accuracy and integrity of the data.

[0043] For example, use the mean filtering technique to eliminate the spike values in the battery voltage data, and use the interpolation method to fill in the missing values in the temperature data, so as to obtain a more accurate cleaned battery data set. Based on these cleaned data, extract the key characteristic values such as the charging and discharging rate, temperature fluctuation, and number of cycles to form a characteristic data set.

[0044] Input the characteristic data set into a preset battery health assessment model for evaluation to obtain a battery health index. Select the valid data that meet the standards through the battery health index and merge them to form a battery operation data set.

[0045] By constructing a battery health assessment model, use the characteristic data set to evaluate the health status of the battery. Specifically, apply machine learning algorithms, input the characteristic data into the model, and calculate the battery health index.

[0046] For example, use the support vector machine (SVM) algorithm to perform classification and regression analysis on the characteristics such as the charging and discharging rate, temperature fluctuation, and number of cycles of the battery to obtain the battery health index. By setting the battery health index threshold, screen out the valid data that meet the standards, and merge these valid data to form a battery operation data set.

[0047] Among them, the steps of inputting the characteristic data set into a preset battery health assessment model for evaluation to obtain a battery health index, selecting the valid data that meet the standards through the battery health index, and merging them to form a battery operation data set include: performing principal component analysis on the characteristic data set, and extracting the internal impedance, temperature change rate, charge and discharge rate, and capacity attenuation rate of the characteristic data set as the main components to obtain a battery health assessment characteristic set.

[0048] By using the principal component analysis (PCA) method, perform dimensionality reduction processing on the data in the characteristic data set. Specifically, extract the internal impedance, temperature change rate, charge and discharge rate, and capacity attenuation rate as the main components to form a battery health assessment characteristic set.

[0049] For example, by using the PCA method, redundant information in the original feature data set is removed, and only the main components that have a significant impact on battery health assessment are retained, thereby reducing the data dimension and improving the data processing efficiency.

[0050] The battery health assessment feature set is input into the battery health assessment model, and the battery degradation model is used to simulate the battery aging trend of the battery health assessment feature set to generate battery aging trend data.

[0051] By constructing a battery degradation model, the aging trend of the battery health assessment feature set is predicted. Specifically, the Bayesian regression algorithm is used to simulate the aging trend of the battery under different usage conditions to generate battery aging trend data.

[0052] For example, by using the Bayesian regression algorithm, based on features such as internal impedance and temperature change rate, the battery capacity decay rate is modeled to predict the capacity decay of the battery under different cycle numbers, and battery aging trend data is obtained.

[0053] The remaining available life of the current battery is calculated using the battery aging trend data and historical battery health data to obtain battery life prediction data.

[0054] By comparing and analyzing the battery aging trend data with historical battery health data, the remaining available life of the current battery is calculated. Specifically, the remaining life prediction algorithm is applied to estimate the remaining life of the battery.

[0055] For example, using the accelerated life model (ALM), combining the battery aging trend data and historical battery health data, the remaining available life of the battery within a certain future period is predicted to obtain battery life prediction data.

[0056] Based on the battery life prediction data, the battery is classified according to its health status to obtain a battery health index, and data that meets the preset health standard is screened according to the battery health index to obtain a battery operation data set.

[0057] By setting the battery health status classification standard, the battery life prediction data is converted into a battery health index. Specifically, according to the health status classification standard, the battery is divided into different health levels.

[0058] For example, the battery life prediction data is divided into four levels: "healthy", "sub-healthy", "mild decay", and "severe decay" according to the health status classification standard, and data that meets the preset health standard is screened according to these levels to form a battery operation data set.

[0059] In step S200, feature extraction is performed on the battery operation data set to obtain multi-dimensional charging features. The multi-dimensional charging features are normalized to obtain a charging feature data set. Based on the charging feature data set, a correlation matrix between charging features is constructed to analyze the mutual influence relationship between each charging feature, and a charging feature correlation data set is obtained.

[0060] Through feature engineering processing of the battery operation data set, multi-dimensional charging features such as charging current, charging voltage, and temperature change are extracted. Specifically, using a normalization algorithm, each dimension feature is normalized to obtain a charging feature data set.

[0061] For example, using the min-max normalization method, the data values of features such as charging current, charging voltage, and temperature change are converted to the range of 0 to 1 to ensure comparability between different features. Based on the normalized charging feature data set, a correlation matrix between charging features is constructed to analyze the mutual influence relationship between each charging feature, and a charging feature correlation data set is formed.

[0062] Through an adaptive weighting method, weighted analysis is performed on different charging features based on the charging feature correlation data set to obtain the weight coefficient of each charging feature, and a charging feature optimization weight set is constructed based on the weight coefficients of all charging features.

[0063] By applying an adaptive weighting algorithm and using the charging feature correlation data set, weighted analysis is performed on different charging features. Specifically, according to the correlation between charging features, the weight coefficient of each charging feature is dynamically adjusted.

[0064] For example, a genetic algorithm is used to perform weighted analysis on charging features, calculate the importance of each feature during the charging process, and obtain the weight coefficient of the charging feature. Based on the weight coefficients of all charging features, a charging feature optimization weight set is constructed.

[0065] In step S300, the importance of each charging feature is evaluated using the charging feature optimization weight set to obtain charging feature evaluation data. Based on the charging feature evaluation data, different scenarios of the charging mode are optimized respectively to obtain multiple charging mode candidate solutions.

[0066] Through the charging feature optimization weight set, the importance of each charging feature is evaluated to form charging feature evaluation data. Specifically, for different charging scenarios, the evaluation data is used to optimize the charging mode to generate multiple charging mode candidate solutions.

[0067] For example, using the charging feature evaluation data, different charging modes suitable for high-temperature environments, low-temperature environments, and standard environments are designed, and the corresponding charging mode candidate solutions are generated respectively.

[0068] Perform an effectiveness analysis on the charging mode candidate solutions, evaluate the charging efficiency and battery life under each charging mode, and obtain the optimized charging mode sequence.

[0069] By performing an effectiveness analysis on multiple charging mode candidate solutions, evaluate the charging efficiency and battery life of each solution under different charging conditions. Specifically, use a simulation tool to conduct simulation tests on each candidate solution.

[0070] For example, use Matlab to simulate the charging mode candidate solutions in a high-temperature environment, evaluate their charging efficiency and battery life under high-temperature conditions, finally screen out the optimal charging mode, and form the optimized charging mode sequence.

[0071] Input the optimized charging mode sequence into a preset deep reinforcement learning model. In the deep reinforcement learning model, through the reinforcement learning algorithm, adaptively adjust the charging mode based on the optimized charging mode sequence to obtain a charging strategy, and dynamically adjust the charging current and charging voltage of the battery based on the charging strategy to obtain charging data.

[0072] By constructing a deep reinforcement learning model, adaptively adjust the optimized charging mode sequence. Specifically, use the reinforcement learning algorithm to train and optimize the performance of the charging mode under different charging conditions.

[0073] For example, use the Deep Q-Network (DQN) algorithm to train and adjust the performance of each charging mode in different environments based on the optimized charging mode sequence, and finally obtain a charging strategy that can adapt to multiple charging environments. Based on the charging strategy, dynamically adjust the charging current and charging voltage of the battery, and record and analyze the charging data in real time to improve the charging efficiency and battery life.

[0074] In step S400, compare and analyze the charging data with the historical charging data, identify abnormal fluctuations during the charging process, obtain the charging anomaly data set, and adjust the charging characteristics in the preset battery performance database according to the charging anomaly data set to obtain the battery performance data set.

[0075] By comparing and analyzing the charging data and the historical charging data, identify abnormal fluctuations during the charging process. Specifically, use data mining techniques to analyze the changes in charging current, voltage, and temperature to identify abnormal charging states.

[0076] For example, by setting the threshold ranges of charging current, voltage, and temperature, discover abnormal data beyond the thresholds, and classify these data as the charging anomaly data set. According to the charging anomaly data set, adjust the charging characteristics in the battery performance database to ensure the effectiveness and safety of the charging strategy.

[0077] Update the charging characteristic map using the battery performance data set to obtain the updated charging characteristic map, and optimize the charging mode optimization sequence based on the updated charging characteristic map to obtain the optimized charging mode optimization sequence.

[0078] Update the charging characteristic map through the battery performance data set. Specifically, use the data update algorithm to incorporate the latest charging data into the characteristic map and reconstruct the charging characteristic map.

[0079] For example, through the incremental update algorithm, fuse the new charging data with the historical data to update the multi-dimensional charging characteristic map, thereby obtaining a more accurate and real-time charging characteristic map. Based on the updated charging characteristic map, optimize the charging mode optimization sequence to generate the optimized charging mode optimization sequence.

[0080] Input the optimized charging mode optimization sequence into the deep reinforcement learning model to obtain the updated charging strategy.

[0081] Further train and optimize by inputting the optimized charging mode optimization sequence into the deep reinforcement learning model. Specifically, use the reinforcement learning algorithm to adjust and optimize the charging strategy to obtain a more accurate updated charging strategy.

[0082] For example, use the Policy Gradient algorithm to train the optimized charging mode optimization sequence to improve the adaptability of the model and finally obtain a better charging strategy.

[0083] Charge the battery according to the updated charging strategy and monitor the key characteristic parameters during the charging process in real time to obtain the optimized charging strategy.

[0084] Charge the battery by executing the updated charging strategy. Specifically, monitor the key characteristic parameters such as voltage, current, and temperature during the charging process in real time to ensure the safety and efficiency of the charging process.

[0085] For example, use the charging management system to collect the voltage, current, and temperature data during the charging process in real time, compare these data with the preset safety range, and ensure that all parameters during the charging process are always within the safety range, thereby implementing the optimized charging strategy.

[0086] Among them, the steps of comparing the charging data with the historical charging data, analyzing to identify abnormal fluctuations during the charging process, obtaining a set of charging anomaly data, and adjusting the charging characteristics in the preset battery performance database according to the set of charging anomaly data to obtain a set of battery performance data include: By comparing and analyzing the charging data with the historical charging data, calculating the current charging voltage deviation, the abnormal temperature rise change rate, and the sudden change of the battery internal resistance to obtain a set of charging anomaly parameters.

[0087] By comparing the charging data in detail with the historical charging data, abnormal fluctuations during the charging process are identified. Specifically, the current charging voltage deviation, the abnormal temperature rise change rate, and the sudden change of the battery internal resistance are calculated to obtain a set of charging anomaly parameters.

[0088] For example, using time series analysis technology, comparing the current charging voltage and historical charging voltage data to calculate the voltage deviation; calculating the abnormal temperature rise change rate through the temperature sensor data; identifying the sudden change of the internal resistance through the internal resistance measurement data, and finally obtaining a set of charging anomaly parameters.

[0089] Using an anomaly detection algorithm to identify abnormal charging states based on the set of charging anomaly parameters, generating a set of charging anomaly data, classifying the set of charging anomaly data, distinguishing normal fluctuations, minor anomalies, and severe anomalies, and assigning different anomaly labels to normal fluctuations, minor anomalies, and severe anomalies.

[0090] By applying an anomaly detection algorithm, based on the set of charging anomaly parameters, identifying and classifying abnormal charging states. Specifically, the abnormal charging states are divided into normal fluctuations, minor anomalies, and severe anomalies, and different anomaly labels are assigned.

[0091] For example, adopting the support vector machine (SVM) algorithm to classify the set of charging anomaly parameters, distinguishing normal fluctuations, minor anomalies, and severe anomalies, and respectively labeling them with "normal", "minor", and "severe" labels.

[0092] Analyze the impact of abnormal charging states on battery life and performance according to the anomaly labels, calculate the life attenuation rate caused by abnormal charging to obtain charging anomaly impact assessment data, and correct the charging characteristics in the preset battery performance database with the charging anomaly impact assessment data to obtain a set of battery performance data.

[0093] By analyzing the set of charging anomaly data in detail, evaluate the impact of abnormal charging states on battery life and performance. Specifically, calculate the battery life attenuation rate caused by abnormal charging to obtain charging anomaly impact assessment data.

[0094] For example, using the multiple linear regression algorithm, model and analyze the relationship between the abnormal charging state and the battery life attenuation, calculate the life attenuation rate under different abnormal labels, and based on these data, correct the charging characteristics in the preset battery performance database, and finally form a battery performance data set.

[0095] In this embodiment, by using a data acquisition device to obtain the historical charge and discharge data, environmental parameter data, and battery health status data of the battery, a preliminary battery data set is generated. Further, through a data preprocessing algorithm for denoising, key characteristic values such as the charge and discharge rate, temperature fluctuation, and number of cycles of the battery are extracted to form a characteristic data set. Then, the characteristic data set is input into a preset battery health assessment model to calculate the battery health index, and based on this index, valid data that meets the standards is selected and merged to form a battery operation data set. Next, the principal component analysis method is used to extract the main components in the characteristic data set, and the battery aging trend is simulated through a battery degradation model to generate battery aging trend data, and these data are used to predict the remaining available life of the battery, and finally battery life prediction data is obtained. According to the battery life prediction data, the battery is graded according to its health status, and data that meets the preset health standards is screened out to form a battery operation data set. In the charging strategy optimization part, multi-dimensional charging characteristics in the battery operation data set are extracted through feature engineering and normalized to construct a charging characteristic correlation matrix. The adaptive weighting method is used for weighted analysis to obtain a charging characteristic optimization weight set. Then, based on this weight set, the charging characteristics are evaluated, and multiple charging mode candidate schemes are generated. Through effect analysis and simulation evaluation, the best charging mode optimization sequence is screened out and input into a deep reinforcement learning model to obtain the final charging strategy. This strategy can dynamically adjust the battery charging current and charging voltage, monitor and identify abnormal fluctuations in real time during the charging process, update the charging characteristic map, and further optimize the charging mode. It realizes the intelligent and adaptive optimization of the battery charging strategy, greatly improving the battery charging efficiency and life.

[0096] Embodiment 3: As Figure 2 shown, the present application also provides a battery charging strategy intelligent optimization system 10, including an acquisition module 11, an analysis module 12, an adjustment module 13, and an optimization module 14.

[0097] The acquisition module 11 is mainly used to obtain the historical charge and discharge data, environmental parameter data, and battery health status data of the battery, and generate a battery operation data set based on the historical charge and discharge data, environmental parameter data, and battery health status data.

[0098] The analysis module 12 is mainly used to construct a multi-dimensional charging feature map through the battery operation data set. Based on the multi-dimensional charging feature map, an adaptive weight dynamic calculation method is used to perform weighted analysis on different charging features to obtain a charging feature optimization weight set.

[0099] The adjustment module 13 is mainly used to construct an optimized charging mode sequence based on the charging feature optimization weight set, input the optimized charging mode sequence into a preset deep reinforcement learning model to obtain a charging strategy, and dynamically adjust the charging current and charging voltage of the battery based on the charging strategy to obtain charging data.

[0100] The optimization module 14 is mainly used to update the charging feature map according to the charging data, and readjust the charging strategy based on the updated charging feature map to obtain an optimized charging strategy.

[0101] In this embodiment, by integrating the acquisition module 11 into the battery management system, the historical charge and discharge data, environmental parameter data, and battery health status data of the battery are collected in real time to generate a battery operation data set. The acquisition module 11 uses a variety of sensors and data recording devices to ensure the accuracy and integrity of the data. The analysis module 12 constructs a charging feature map through multi-dimensional analysis of the battery operation data set, and performs weighted analysis on different charging features using an adaptive weight dynamic calculation method to obtain a charging feature optimization weight set. The adjustment module 13 generates an optimized charging mode sequence according to the charging feature optimization weight set and inputs it into the deep reinforcement learning model to obtain an optimized charging strategy. During the charging process, the adjustment module 13 dynamically adjusts the charging current and charging voltage of the battery based on the charging strategy and records the charging data in real time. The optimization module 14 compares and analyzes the latest charging data with the historical data to identify abnormal fluctuations during the charging process, and adjusts the charging features in the battery performance database based on the abnormal data set to update the charging feature map. The optimization module 14 further uses the updated charging feature map to optimize the optimized charging mode sequence, generates an optimized optimized charging mode sequence, and inputs it into the deep reinforcement learning model for retraining to obtain an updated charging strategy. Finally, through the above method, the battery charging strategy intelligent optimization system 10 realizes the intelligent and adaptive optimization of the charging strategy, improves the charging efficiency, extends the battery life, and significantly improves the charging safety and reliability of the battery in a complex and changeable environment.

[0102] It should be noted that those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described device and each module can refer to the corresponding processes in the foregoing embodiments of the battery charging strategy intelligent optimization method, and will not be elaborated herein.

[0103] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. An intelligent optimization method for battery charging strategies, characterized in that Including: Obtain the historical charge-discharge data, environmental parameter data, and battery health status data of the battery, and generate a battery operation data set based on the historical charge-discharge data, the environmental parameter data, and the battery health status data; Construct a multi-dimensional charging feature map through the battery operation data set, and based on the multi-dimensional charging feature map, use an adaptive weight dynamic calculation method to perform weighted analysis on different charging features to obtain a charging feature optimization weight set; Construct a charging mode optimization sequence based on the charging feature optimization weight set, input the charging mode optimization sequence into a preset deep reinforcement learning model to obtain a charging strategy, and dynamically adjust the charging current and charging voltage of the battery based on the charging strategy to obtain charging data; Update the charging feature map according to the charging data, and re-adjust the charging strategy based on the updated charging feature map to obtain an optimized charging strategy.

2. The intelligent optimization method for battery charging strategy according to claim 1, wherein The step of obtaining the historical charge-discharge data, environmental parameter data, and battery health status data of the battery, and generating a battery operation data set based on the historical charge-discharge data, the environmental parameter data, and the battery health status data includes: Obtain the historical charge-discharge data, environmental parameter data, and battery health status data of the battery through a data acquisition device, and generate a preliminary battery data set according to the historical charge-discharge data, environmental parameter data, and battery health status data; Perform denoising processing on the preliminary battery data set to obtain a cleaned battery data set, and extract key feature values of the charge-discharge rate, temperature fluctuation, and number of cycles of the battery based on the cleaned battery data set to obtain a feature data set; Input the feature data set into a preset battery health assessment model for evaluation to obtain a battery health index, select valid data that meet the standards through the battery health index, and merge them to form a battery operation data set.

3. The intelligent optimization method for battery charging strategy according to claim 2, wherein, The step of inputting the feature data set into a preset battery health assessment model for evaluation to obtain a battery health index, selecting valid data that meet the standards through the battery health index, and merging them to form a battery operation data set includes: Perform principal component analysis on the feature data set, extract the internal impedance, temperature change rate, charge-discharge rate, and capacity attenuation rate of the feature data set as the main components to obtain a battery health assessment feature set; Input the battery health assessment feature set into a battery health assessment model, and use a battery degradation model to simulate the battery aging trend of the battery health assessment feature set to generate battery aging trend data; Calculate the remaining available life of the current battery using the battery aging trend data and historical battery health data to obtain battery life prediction data; Classify the health status of the battery based on the battery life prediction data to obtain a battery health index, and screen out data that meet the preset health standards according to the battery health index to obtain a battery operation data set.

4. The intelligent optimization method for battery charging strategy according to claim 1, wherein The step of constructing a multi-dimensional charging feature map through the battery operation data set, and based on the multi-dimensional charging feature map, using an adaptive weight dynamic calculation method to perform weighted analysis on different charging features to obtain a charging feature optimization weight set includes: Performing feature extraction on the battery operation data set to obtain multi-dimensional charging features, performing normalization processing on the multi-dimensional charging features to obtain a charging feature data set, constructing a correlation matrix between charging features based on the charging feature data set, and analyzing the mutual influence relationship between each charging feature to obtain a charging feature correlation data set; Performing weighted analysis on different charging features based on the charging feature correlation data set through an adaptive weighting method to obtain the weight coefficient of each charging feature, and constructing a charging feature optimization weight set based on the weight coefficients of all the charging features.

5. The intelligent optimization method for battery charging strategy according to claim 1, characterized in that, The step of constructing an optimized charging mode sequence based on the charging feature optimization weight set, inputting the optimized charging mode sequence into a preset deep reinforcement learning model to obtain a charging strategy, and dynamically adjusting the charging current and charging voltage of the battery based on the charging strategy to obtain charging data includes: Evaluating the importance of each charging feature by using the charging feature optimization weight set to obtain charging feature evaluation data, and respectively optimizing different scenarios of the charging mode based on the charging feature evaluation data to obtain multiple charging mode candidate solutions; Performing effect analysis on the charging mode candidate solutions, evaluating the charging efficiency and battery life under each charging mode to obtain an optimized charging mode sequence; Inputting the optimized charging mode sequence into a preset deep reinforcement learning model. In the deep reinforcement learning model, adaptively adjusting the charging mode based on the optimized charging mode sequence through a reinforcement learning algorithm to obtain a charging strategy, and dynamically adjusting the charging current and charging voltage of the battery based on the charging strategy to obtain charging data.

6. The intelligent optimization method for battery charging strategy according to claim 1, wherein The step of updating the charging feature map according to the charging data, and readjusting the charging strategy based on the updated charging feature map to obtain an optimized charging strategy includes: Comparing and analyzing the charging data with historical charging data, identifying abnormal fluctuations during the charging process to obtain a charging anomaly data set, and adjusting the charging features in a preset battery performance database according to the charging anomaly data set to obtain a battery performance data set; Updating the charging feature map by using the battery performance data set to obtain an updated charging feature map, and optimizing the charging mode of the optimized charging mode sequence based on the updated charging feature map to obtain an optimized optimized charging mode sequence; Inputting the optimized optimized charging mode sequence into the deep reinforcement learning model to obtain an updated charging strategy; Charging the battery according to the updated charging strategy, and real-time monitoring the key feature parameters during the charging process to obtain an optimized charging strategy.

7. The intelligent optimization method for battery charging strategy according to claim 6, characterized in that The step of comparing and analyzing the charging data with historical charging data, identifying abnormal fluctuations during the charging process, obtaining a set of charging anomaly data, and adjusting the charging characteristics in a preset battery performance database according to the set of charging anomaly data to obtain a set of battery performance data includes: By comparing and analyzing the charging data with historical charging data, calculating the current charging voltage deviation, the abnormal change rate of temperature rise, and the mutation of battery internal resistance, a set of charging anomaly parameters is obtained; Using an anomaly detection algorithm to identify abnormal charging states based on the set of charging anomaly parameters, generating a set of charging anomaly data, classifying the set of charging anomaly data, distinguishing normal fluctuations, minor anomalies, and serious anomalies, and assigning different anomaly labels to the normal fluctuations, the minor anomalies, and the serious anomalies; Analyzing the impact of abnormal charging states on battery life and performance according to the anomaly labels, calculating the life attenuation rate caused by abnormal charging, obtaining charging anomaly impact evaluation data, and correcting the charging characteristics in a preset battery performance database with the charging anomaly impact evaluation data to obtain a set of battery performance data.

8. An intelligent optimization system for battery charging strategies, characterized in that, It includes: An acquisition module for acquiring historical charge and discharge data, environmental parameter data, and battery health status data of the battery, and generating a set of battery operation data based on the historical charge and discharge data, the environmental parameter data, and the battery health status data; An analysis module for constructing a multi-dimensional charging characteristic map through the set of battery operation data, and based on the multi-dimensional charging characteristic map, using an adaptive weight dynamic calculation method to perform weighted analysis on different charging characteristics to obtain a set of optimized charging characteristic weights; An adjustment module for constructing an optimized charging mode sequence based on the set of optimized charging characteristic weights, inputting the optimized charging mode sequence into a preset deep reinforcement learning model to obtain a charging strategy, and dynamically adjusting the charging current and charging voltage of the battery based on the charging strategy to obtain charging data; An optimization module for updating the charging characteristic map according to the charging data, and readjusting the charging strategy based on the updated charging characteristic map to obtain an optimized charging strategy.

Images