A cloud battery management system and method based on digital twin and blockchain technology
The cloud-based battery management system, utilizing digital twin and blockchain technologies, addresses the issue of insufficient computing and data storage in in-vehicle BMS, enabling accurate estimation and prediction of battery status and enhancing system safety and reliability.
Patent Information
- Application Number
- CN202011417811.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-12-07
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2040-12-07
AI Technical Summary
Existing vehicle-mounted BMS systems suffer from insufficient computing power and data storage when faced with a large number of battery cells and complex algorithms, making it difficult to achieve accurate battery state estimation and prediction, and aging prediction and strategy optimization are also challenging.
A cloud battery management system based on digital twin and blockchain technologies is adopted. It communicates with the battery system through the cloud platform, uses the Internet of Things to transmit data, and combines the JADE algorithm to estimate the battery state, and establishes a virtual battery model for accurate estimation and prediction.
It improves computing and data storage capabilities, enables accurate estimation and prediction of battery state, enhances the safety, reliability and performance of battery systems, and supports aging prediction and strategy optimization.
Smart Images

Figure CN112711892B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of battery management, and specifically relates to a cloud battery management system and method based on digital twin and blockchain technology. Background Art
[0002] BMS battery management system (BATTERY MANAGEMENT SYSTEM), commonly known as battery nanny or battery butler, is mainly used to intelligently manage and maintain each battery unit, prevent the battery from overcharging and over-discharging, extend the battery life, and monitor the battery status.
[0003] The BMS battery management system unit includes a BMS battery management system, a control module, a display module, a wireless communication module, electrical equipment, a battery pack for powering the electrical equipment, and a collection module for collecting battery information of the battery pack. The BMS battery management system is connected to the wireless communication module and the display module respectively through communication interfaces, the output end of the collection module is connected to the input end of the BMS battery management system, the output end of the BMS battery management system is connected to the input end of the control module, the control module is connected to the battery pack and the electrical equipment respectively, and the BMS battery management system is connected to the Server server through the wireless communication module.
[0004] The application of battery management system (BMS) in electric vehicles can be traced back to the management of nickel-metal hydride batteries in Toyota HEV models. Unlike the management of lithium batteries, nickel-metal hydride batteries have the characteristics of high consistency, good safety, and low single-cell voltage (1.0-1.7V), so the BMS of nickel-metal hydride batteries usually does not require balancing functions, does not need to control contactors, and does not need to collect voltage for each battery (6 batteries can be connected in series as a whole for voltage monitoring). Although the hardware functions of nickel-metal hydride battery BMS are relatively simple, due to the memory effect of nickel-metal hydride batteries and the complex relationship between voltage characteristics and SOC, the difficulty lies in how to estimate SOC and how to control and adjust the charge and discharge range to avoid rapid battery degradation. With the application of lithium battery technology, the power battery system has higher energy density, larger capacity, and longer operating time, which also puts forward new requirements for the functions of BMS. From the perspective of topological architecture, BMS is divided into two categories: centralized and distributed according to different project requirements.
[0005] However, as the number of battery cells and algorithm complexity increase, the onboard BMS faces problems in computing power and data storage when accurately estimating and predicting the battery status based on model-based algorithms. In addition, new functions such as aging prediction and strategy optimization based on historical operating data of the battery pack are difficult to implement on the device. Summary of the invention
[0006] In order to solve the problems in the prior art, the present invention proposes a cloud battery management system and method based on digital twin and blockchain technology, which can accurately estimate and predict the battery status, improve the computing power and data storage capacity, and can perform aging prediction and strategy optimization, thereby enhancing the safety, reliability and performance of the battery system.
[0007] In order to achieve the above objectives, the present invention provides a cloud battery management system based on digital twin and blockchain technology, characterized in that it includes a cloud platform, which is communicated with the battery system through the Internet of Things, and the cloud platform traces the relevant information of the battery system through the blockchain. The cloud platform includes a digital twin module and a battery status estimation module. The digital twin module establishes a virtual battery model consistent with the characteristics of the battery system based on an equivalent circuit model. The digital twin module is used to obtain measurement data of sensors of the battery system and visualize the internal state of each battery in the battery system, and the battery status estimation module predicts and evaluates the state of the battery system based on the JADE algorithm.
[0008] Furthermore, the cloud platform includes cloud data and local data, the cloud data includes historical data, machine learning, hybrid intelligence, remaining life and health cloud diagnosis of the battery system; the local data includes battery model, battery cycle and battery status estimation, and the cloud data and the local data also include remote operation and maintenance, energy optimization, smart charging, cloud balancing and SOX estimation.
[0009] The present invention also provides a cloud battery management method, which adopts the above system and includes:
[0010] 1) Based on the equivalent circuit model, a virtual battery model consistent with the characteristics of the battery system is established to construct a digital twin module, and a battery state estimation module is constructed based on the JADE algorithm. The digital twin module and the battery state estimation module form a digital twin model;
[0011] 2) Battery status estimation and availability prediction based on digital twin models;
[0012] 3) Battery status assessment based on digital twin model;
[0013] 4) If the evaluation result meets the requirements, the battery status is output; if the evaluation result does not meet the requirements, return to step 2).
[0014] Furthermore, the JADE algorithm adopts the “DE / current-to-pbest / 1” mutation strategy to optimize the profile, and the individual v after mutation i By individual x i and its associated mutation factor F iJoint generation converges the generated offspring individuals to the global optimal value.
[0015] Furthermore, the JADE algorithm performs the following operations on the generated initial particles:
[0016]
[0017] in, is the mutated particle, x i,g is the particle currently generated, F i is the factor of variation, is the first 100p% of individuals randomly selected from the contemporary individuals, p∈(0,1], g is the current generation of the particle, x r1,g and x r2,g are individuals randomly selected from the current unmutated individuals, r1,r2∈rand(0,1).
[0018] Furthermore, the JADE algorithm uses an external set A to store the parent individuals eliminated by the offspring, and the original number of individuals P remains unchanged. The offspring replaces the parent if it is better than the parent. The population update strategy with the external set A is:
[0019]
[0020] Here, x i,g 、x r1,g and are all individuals selected from the population P in the same way, and It is randomly selected from the current individual set and the eliminated parent set P∪A.
[0021] Furthermore, the operation of the JADE algorithm on set A includes: first setting it to an empty set, then setting a counting flag for set A, and after each evolution of the individual is completed, the eliminated parent individuals are stored in set A, and the counting flag is increased once an individual is stored. When set A exceeds the set threshold, that is, when set A is full of individuals, some particles are randomly deleted from set A to ensure that individuals can continue to be stored and the update is completed.
[0022] Furthermore, the adaptive update of the crossover factor of the JADE algorithm includes:
[0023] In each generation g, each individual x i The crossover factor CR i They are all generated independently, according to the average value μ CR , a normal distribution with a standard deviation of 0.1 is generated:
[0024] CR i=randn i (μ CR ,0.1)
[0025] CR i Restricted to the interval [0, 1], the set of crossover factors that have evolved successfully in the gth generation is S CR , where the average value μ CR It is initially set to 0.5 and updated at the end of each generation according to the following formula:
[0026] μ CR =(1-c)·μ CR +c·mean A (S CR )
[0027] Where c is a constant between 0 and 1, usually set to 0.5, mean A is the arithmetic mean.
[0028] Furthermore, the adaptive update of the mutation factor of the JADE algorithm includes:
[0029] In each generation g, the variation factor F of each individual i It is also generated independently, according to the position parameter μ F and a scale parameter of 0.1 yields:
[0030] F i =randc i (μ F ,0.1)
[0031] If the generated F i ≥1, then F i Set to 1; if F i ≤0, then it is regenerated once, and the set of mutation factors that have evolved successfully in the gth generation is S F , where the location parameter μ of the Cauchy distribution is F It is initially set to 0.5 and updated at the end of each generation according to the following formula:
[0032] μ F =(1-c)·μ F +c·mean L (S F )
[0033] where mean L Update as follows:
[0034]
[0035] Furthermore, the battery state estimation includes battery state of charge SOC estimation and health state SOH estimation, and the implementation strategy of the JADE algorithm includes: after each T cycle, by running the JADE algorithm, updating the battery parameters and health state SOH according to the field data of the previous cycle, and repeating this process during the entire use process.
[0036] Compared with the prior art, the present invention revolutionizes BMS by applying cloud computing and Internet of Things (IoT) technologies, and the computing and data storage capabilities are growing exponentially. All battery-related data can be measured and seamlessly transmitted to the cloud platform, which is used to build a digital twin for the battery system. The digital twin can not only display the measurement data of the sensors in the battery pack, but also visualize the internal state of each battery through advanced diagnostic algorithms. By collecting big data in one stop, the prediction and optimization of the system can be achieved through machine learning algorithms, which will completely change the understanding of battery aging. In addition, by replacing wiring communication with wireless IoT communication, the system reliability can be improved. Compared with the on-board BMS, the digital twin for the battery system has potential advantages in the following four aspects, which together help to reduce the total cost of ownership and improve system performance. In terms of monitoring and diagnosis, high computing power supports advanced diagnostic algorithms to continuously and accurately monitor the battery status; in terms of lifetime prediction, machine learning based on full life cycle operation data can accurately predict the degradation trend; in terms of fault detection and prediction, early detection of system failures at different levels through big data analysis improves the safety and reliability of the system; in terms of evaluation and optimization, the system design and operation strategy are optimized by evaluating big data in battery systems under different operating scenarios.
[0037] The present invention improves computing power and data storage capacity through cloud computing. With the help of the Internet of Things, all battery-related data will be measured and seamlessly transmitted to the cloud, thereby establishing a digital twin system for the battery system. The battery status detection algorithm will estimate the battery state of charge and evaluate the health status at all times in the system. With the development of the state of charge and health status estimation method based on the cloud platform, the application of the battery system based on the equivalent circuit model in the digital twin system can be explored. Moreover, the battery-based digital twin system can also trace the deficiencies in battery materials, structural design, and use through the blockchain, provide a reference for the construction of the material gene map of the battery and the research on the battery control algorithm, realize the closed-loop posture of the data, and help to realize the rapid classification and reorganization of the battery. The present invention can accurately estimate and predict the battery state, improve the computing power and data storage capacity, and can perform aging prediction and strategy optimization, enhancing the safety, reliability and performance of the battery system. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] Figure 1 It is the system principle diagram of the present invention;
[0039] Figure 2 The hardware schematic diagram of the present invention is
[0040] Figure 3 is a flow chart of the method of the present invention;
[0041] Figure 4 The battery equivalent circuit model diagram of the present invention;
[0042] Figure 5 The JADE algorithm battery state estimation flow chart of the present invention;
[0043] Figure 6 It is the strategy diagram of “DE / current-to-pbest / 1” adopted in the JADE algorithm of the present invention;
[0044] Figure 7 This is a diagram of the implementation strategy for SOC estimation and health state SOH estimation of the JADE algorithm of the present invention. DETAILED DESCRIPTION
[0045] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments 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.
[0046] Battery management is crucial to enhancing the safety, reliability and performance of battery systems. Therefore, on the one hand, an embodiment of the present invention provides a cloud battery management system based on digital twin and blockchain technology, see Figure 1 It includes a cloud platform, which is connected to the battery system through the Internet of Things, and the cloud platform traces the relevant information of the battery system through the blockchain. The cloud platform includes a digital twin module and a battery state estimation module. The digital twin module establishes a virtual battery model consistent with the characteristics of the battery system based on an equivalent circuit model. The digital twin module is used to obtain the measurement data of the sensors of the battery system and visualize the internal state of each battery in the battery system, and predicts and evaluates the state of the battery system based on the JADE algorithm through the battery state estimation module.
[0047] The cloud platform includes cloud data and local data. Cloud data includes historical data of the battery system, machine learning, hybrid intelligence, remaining life and health cloud diagnosis, etc.; local data includes battery models, battery cycles and battery status estimation, etc., and cloud data and local data also include remote operation and maintenance, energy optimization, smart charging, cloud balancing and SOX estimation, etc.
[0048] See also Figure 2 , multiple sensors of the battery system are connected to the Slave-BMS module through CAN line communication, and the Slave-BMS module is connected to the Raspberry Pi module through MQTT wireless communication.
[0049] In this embodiment, the battery system is a battery pack of an electric vehicle, but it may also be a battery pack of other electric equipment.
[0050] Another aspect of the present invention provides a cloud battery management method, based on the above system, see Figure 3 ,include:
[0051] 1) Based on the equivalent circuit model, a virtual battery model consistent with the characteristics of the battery system is established to construct a digital twin module, and a battery state estimation module is constructed based on the JADE algorithm. The digital twin module and the battery state estimation module form a digital twin model; the battery equivalent circuit model can be found in Figure 4 ;
[0052] 2) Battery status estimation and availability prediction based on digital twin models;
[0053] 3) Battery status assessment based on digital twin model;
[0054] 4) If the evaluation result meets the requirements, the battery status is output; if the evaluation result does not meet the requirements, return to step 2).
[0055] Considering the function of digital twin circuits in battery systems, battery dynamics modeling is crucial, which is the basis for accurate battery diagnosis. There are three main types of battery cell models: equivalent circuit models (ECM), electrochemical models, and machine learning models. All of these models have highlights and limitations and will play different roles in the digital twin of the battery system. The present invention applies the equivalent circuit model to the digital twin model to accurately monitor the battery's state of charge and health status, etc.
[0056] The JADE (Adaptive Parameter Differential Evolution Algorithm Based on External Archive) algorithm is improved based on the DE (Differential Evolution) algorithm. The calculation process of the original DE algorithm is:
[0057] The initial population is: {xi,0 =(x 1,i,0 ,x 2,i,0 ,...,x D,i,0 )|i=1,2,...,NP};
[0058] The initial group is based on a certain range of Uniform distribution, where D is the dimension of the particles. After initialization, the mutation crossover selection operation begins.
[0059] In the JADE algorithm: Compared with the original DE algorithm, the mutation strategy is improved. An existing mutation update method is However, under this method, the individuals with the optimal solution will also lead to a decrease in the diversity of the population, resulting in premature convergence, and the solution of the problem tending to the local optimum. Therefore, in view of the characteristics of the existing mutation strategy and the problem of fast convergence but unreliability, a new mutation strategy is adopted.
[0060] The JADE algorithm in this invention adopts the "DE / current-to-pbest / 1" mutation strategy to optimize the profile. The individual v i By individual x i and its associated mutation factor F i Co-generation, the generated offspring individuals converge to the global optimal value. Figure 6 , where the dotted line is the outline of the optimization problem, the middle point Global optimum is the global optimal value, v i is the mutated individual, which is composed of individual x i and its associated mutation factor F i Through co-generation, we can see that under the new mutation strategy, the generated offspring individuals will converge to the global optimal value.
[0061] The JADE algorithm performs the following operations on the generated initial particles:
[0062] in, is the mutated particle, x i,g is the particle currently generated, F i is the factor of variation, is the first 100p% of individuals randomly selected from the contemporary individuals, p∈(0,1], g is the current generation of the particle, x r1,g and x r2,g are individuals randomly selected from the current unmutated individuals, r1,r2∈rand(0,1).
[0063] Compared with the current population, the poorer individuals found can also help the iterative evolution of the population, which is conducive to the population moving towards the global optimal direction. Therefore, the JADE algorithm of the present invention uses an external set A to indicate those individuals that have failed to evolve, that is, those parent individuals that have been eliminated by the offspring in the selection mechanism. These individuals are stored in the set A, while the original number of individuals P remains unchanged and is updated according to the normal strategy, that is, if the offspring is better than the parent, the parent is replaced by the offspring. Therefore, the population update strategy with the external set A becomes:
[0064]
[0065] Here, x i,g 、x r1,g , are all individuals selected from the population P in the same way, and It is randomly selected from a new set P∪A (the current individual set and the eliminated parent set).
[0066] The JADE algorithm's operations on set A include: first setting it to an empty set, then setting a counting flag for set A. After each evolution of the individual is completed, the eliminated parent individuals are stored in set A, which means that the parent individuals that are not as good as the offspring are not completely useless. Then, when an individual is stored, the counting flag is increased once. When set A exceeds a certain threshold, that is, when set A is full of individuals, some particles are randomly deleted from set A to ensure that individuals can continue to be stored and the update is completed. The archiving method of the external set A provides information on the direction of individual evolution, and also improves the diversity of individuals in the population from one aspect.
[0067] The parameter adaptation strategy of the JADE algorithm includes the adaptive update of the crossover factor and the adaptive update of the mutation factor.
[0068] The adaptive update of the crossover factor includes:
[0069] In each generation g, each individual x i The crossover factor CR i They are all generated independently, according to the average value μ CR , a normal distribution with a standard deviation of 0.1 is generated:
[0070] CR i =randn i (μ CR ,0.1)
[0071] CR i Restricted to the interval [0, 1], the set of crossover factors that have evolved successfully in the gth generation is SCR , where the average value μ CR It is initially set to 0.5 and updated at the end of each generation according to the following formula:
[0072] μ CR =(1-c)·μ CR +c·mean A (S CR )
[0073] Where c is a constant between 0 and 1, usually set to 0.5, mean A is the arithmetic mean.
[0074] Adaptive updates of mutation factors include:
[0075] In each generation g, the variation factor F of each individual i It is also generated independently, according to the position parameter μ F and a scale parameter of 0.1 yields:
[0076] F i =randc i (μ F ,0.1)
[0077] If the generated F i ≥1, then F i Set to 1; if F i ≤0, then it is regenerated once, and the set of mutation factors that have evolved successfully in the gth generation is S F , where the location parameter μ of the Cauchy distribution is F It is initially set to 0.5 and updated at the end of each generation according to the following formula:
[0078] μ F =(1-c)·μ F +c·mean L (S F )
[0079] where mean L Update as follows:
[0080]
[0081] The setting of the adaptive parameter strategy records the parameters corresponding to the successfully evolved individuals, selects the corresponding means based on these parameters, constructs a probability distribution according to the means and selects the parameters of the next generation of individuals from it. In this way, the iterative parameters will be adjusted adaptively and will not make the parameters too consistent and lose diversity.
[0082] The battery state estimation of the present invention includes battery state of charge SOC estimation and health state SOH estimation, see Figure 7 ,The implementation strategy of the JADE algorithm includes that after each T cycle, the battery parameters and ,state of health SOH are updated according to the field data of the previous cycle ,by running the JADE algorithm, and this process is repeated throughout the ,use process.
[0083] See Figure 5 ,The method flow of JADE algorithm includes:
[0084] 1) Input initial parameters;
[0085] 2) Initialize particles, archive set A, and cross factor set S CR , the set of mutation factors S F , and the optimal number of individuals in each generation Top;
[0086] 3) Calculate the particle fitness value;
[0087] 4) Perform mutation, crossover, and selection operations on particles;
[0088] 5) Determine whether the current solution is better than the previous value. If it is better than the previous value, proceed to step 6); if it is not better than the previous value, return to step 3);
[0089] 6) Store the parameters corresponding to the particles into set A and the set of cross factors into S CR , the set of mutation factors S F In the , put the best particle into Top;
[0090] 7) Calculate the crossover factor CR and mutation factor F for the next operation based on the updated set;
[0091] 8) Determine whether the number of iterations is met. If so, end the operation; if not, return to step 3).
[0092] The present invention uses digital twin technology to achieve cloud management of batteries, improving the accuracy and efficiency of the battery state estimation method. First, a virtual battery model consistent with the characteristics of the physical battery is established, and then it is integrated with the battery model driven by cloud data to improve the accuracy of battery model construction. Moreover, this solution can also realize the mutual comparison and initial state determination of the same type of battery state estimation of different vehicles.
[0093] The present invention proposes estimating the battery state of charge and health state at different time scales under a new architecture, which can effectively avoid excessive computational load.
[0094] The digital twin-based cloud battery management of the present invention can effectively utilize the historical operating data of batteries of the same type to realize cloud correction of the initial parameters of new battery vehicles and directional energy optimization of old electric vehicles.
[0095] The present invention introduces digital twins into the battery management system, and uses the multi-dimensional virtual battery model and simulation data provided by digital twins to enhance the prediction and evaluation of battery charge state and health state. On the one hand, it can effectively and accurately predict the availability of the battery. The generated battery management plan is matched with its actual state; on the other hand, it can carry out multi-dimensional evaluation and prediction of the battery in advance in the cloud battery management system to ensure that the battery system is maintained in a relatively ideal state and extend the battery life.
[0096] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A cloud battery management system based on digital twin and blockchain technology, It is characterized in that The cloud platform includes a cloud platform, the cloud platform is connected to the battery system through the Internet of Things, and the cloud platform traces the relevant information of the battery system through the blockchain, the cloud platform includes a digital twin module and a battery state estimation module, the digital twin module establishes a virtual battery model consistent with the characteristics of the battery system based on an equivalent circuit model, the digital twin module is used to obtain the measurement data of the sensor of the battery system, and visualize the internal state of each battery of the battery system, and predict and evaluate the state of the battery system based on the JADE algorithm through the battery state estimation module; The cloud platform includes cloud data and local data, the cloud data includes historical data, machine learning, hybrid intelligence, remaining life and health cloud diagnosis of the battery system; the local data includes battery model, battery cycle and battery state estimation, and the cloud data and the local data also include remote operation and maintenance, energy optimization, smart charging, cloud balancing and SOX estimation; The cloud platform is used for: 1) Based on the equivalent circuit model, a virtual battery model consistent with the characteristics of the battery system is established to construct a digital twin module, and a battery state estimation module is constructed based on the JADE algorithm. The digital twin module and the battery state estimation module form a digital twin model; 2) Battery status estimation and availability prediction based on digital twin models; 3) Battery status assessment based on digital twin model; 4) If the evaluation result meets the requirements, the battery status is output; if the evaluation result does not meet the requirements, return to step 2); The JADE algorithm adopts the "DE / current-to-pbest / 1" mutation strategy to optimize the profile. i By individual x i and its associated mutation factor F i Joint generation converges the generated offspring individuals to the global optimal value.
2. A cloud battery management method, using the cloud battery management system based on digital twin and blockchain technology as described in claim 1, It is characterized in that include: 1) Based on the equivalent circuit model, a virtual battery model consistent with the characteristics of the battery system is established to construct a digital twin module, and a battery state estimation module is constructed based on the JADE algorithm. The digital twin module and the battery state estimation module form a digital twin model; 2) Battery status estimation and availability prediction based on digital twin models; 3) Battery status assessment based on digital twin model; 4) If the evaluation result meets the requirements, the battery status is output; if the evaluation result does not meet the requirements, return to step 2); The JADE algorithm adopts the "DE / current-to-pbest / 1" mutation strategy to optimize the profile. i By individual x i and its associated mutation factor F i Joint generation converges the generated offspring individuals to the global optimal value.
3. A cloud battery management method according to claim 2, It is characterized in that The JADE algorithm performs the following operations on the generated initial particles: in, is the mutated particle, x i,g is the particle currently generated, F i is the factor of variation, is the first 100p% of individuals randomly selected from the contemporary individuals, p∈(0,1], g is the current generation of the particle, x r1,g and x r2,g is an individual randomly selected from the current unmutated individuals, r 1 ,r 2 ∈rand(0,1).
4. A cloud battery management method according to claim 2, It is characterized in that The JADE algorithm uses an external set A to store the parent individuals eliminated by the offspring. The original number of individuals P remains unchanged. The offspring replaces the parent if it is better than the parent. The population update strategy with the external set A is: Here, x i,g 、x r1,g and are all individuals selected from the population P in the same way, and It is randomly selected from the current individual set and the eliminated parent set P∪A.
5. A cloud battery management method according to claim 4, It is characterized in that The operation of the JADE algorithm on set A includes: first setting it to an empty set, then setting a counting flag for set A, and after each evolution of the individual is completed, the eliminated parent individuals are stored in set A. When an individual is stored, the counting flag is increased once. When set A exceeds the set threshold, that is, when set A is full of individuals, some particles are randomly deleted from set A to ensure that individuals can continue to be stored and the update is completed.
6. A cloud battery management method according to claim 2, It is characterized in that The adaptive update of the crossover factor of the JADE algorithm includes: In each generation g, each individual x i The crossover factor CR i They are all generated independently, according to the average value μ CR , a normal distribution with a standard deviation of 0.1 is generated: CR i =randn i (μ CR ,0.1) CR i Restricted to the interval [0, 1], the set of crossover factors that have evolved successfully in the gth generation is S CR , where the average value μ CR It is initially set to 0.5 and updated at the end of each generation according to the following formula: m CR =(1-c)·μ CR +c·mean A (S CR ) Where c is a constant between 0 and 1, usually set to 0.5, mean A is the arithmetic mean.
7. A cloud battery management method according to claim 2, It is characterized in that The adaptive update of the mutation factor of the JADE algorithm includes: In each generation g, the variation factor F of each individual i It is also generated independently, according to the position parameter μ F and a scale parameter of 0.1 yields: F i =randc i (μ F ,0.1) If the generated F i ≥1, then F i Set to 1; if F i ≤0, then it is regenerated once, and the set of mutation factors that have evolved successfully in the gth generation is S F , where the location parameter μ of the Cauchy distribution is F It is initially set to 0.5 and updated at the end of each generation according to the following formula: m F =(1-c)·μ F +c·mean L (S F ) where mean L Update as follows:
8. A cloud battery management method according to claim 2, It is characterized in that The battery state estimation includes battery state of charge SOC estimation and health state SOH estimation. The implementation strategy of the JADE algorithm includes: after each T cycle, by running the JADE algorithm, updating the battery parameters and health state SOH according to the field data of the previous cycle, and repeating this process during the entire use process.
