A flexible control method and system for dynamically reconfiguring a battery module

Abstract

The application provides a flexible control method and system for dynamically reconstructing a battery module, and relates to the technical field of battery module control. The application comprises the following steps: collecting historical state parameters of each single battery in the battery module, and determining the number of neighbors of each single battery by using a game analysis algorithm based on a regionalized neighbor interaction mechanism to obtain a neighbor battery group; evaluating the aging degree of the battery by using a support vector regression method based on the current state parameters of each single battery in the battery module, and calculating the aging collaborative deviation of each single battery by using a spatial proximity principle in combination with the neighbor battery group; and dynamically reconstructing and flexibly controlling the single batteries in the battery module based on the aging collaborative consistency principle according to the aging collaborative deviation of the single batteries. The application can more accurately grasp the state and aging condition of the battery, so as to formulate a more scientific and reasonable control strategy and improve the efficiency of battery management.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of battery module control, in particular, to a flexible control method and system for dynamically reconstructing a battery module. BACKGROUND

[0002] With the continuous development of energy storage systems, various requirements for battery management systems are also increasing, including high performance, high reliability, high cost performance, and intelligent control.

[0003] As an important part of the battery management system, the battery module is usually composed of multiple fixed single batteries. In the existing technical framework, the charge / discharge voltage of the battery module is fixedly set as the sum of several single battery voltages corresponding to the number of battery series, and cannot be flexibly adjusted according to actual needs. During the charge / discharge operation, due to the change range of the single battery voltage, the battery module voltage has to fluctuate in a large range to complete the charge / discharge work.

[0004] The disadvantages of this fixed connection method are obvious. On the one hand, it greatly reduces the energy conversion efficiency of the battery system, and a large amount of energy is wasted during voltage fluctuation, resulting in low operation efficiency. On the other hand, fixed connection leads to a lack of flexible coordination mechanism in the battery module. When some batteries have problems such as aging and uneven performance, the performance of the entire battery module cannot be optimized by adjusting the connection relationship between the batteries. This makes the control process extremely complex and difficult to achieve efficient management.

[0005] In addition, the state of the battery changes constantly during actual use, the aging degree gradually differs, and the voltage fluctuates frequently. However, the existing technology cannot perceive these dynamic changes in real time and adjust the connection method and working state of the battery module accordingly. This makes it difficult for the battery module to maintain the best operating state at all times, and it is difficult to fully exert its performance advantages. Moreover, it may even accelerate the aging of the battery and shorten the service life of the battery module.

[0006] At present, there is no effective solution to the problems in the related art. SUMMARY

[0007] Therefore, the present application provides a flexible control method and system for dynamically reconstructing a battery module to solve the above-mentioned problems.

[0008] In order to solve the above problems, the specific technical solutions adopted by the present application are as follows:

[0009] According to one aspect of the present application, a flexible control method for dynamically reconstructing a battery module is provided, comprising the following steps:

[0010] S1, collect historical state parameters of each single battery in the battery module, and determine the number of neighbors of each single battery based on a regionalized neighbor interaction mechanism and using a game analysis algorithm to obtain a neighbor battery group;

[0011] S2, based on the current state parameters of each single battery in the battery module, evaluate the aging degree of the battery using a support vector regression method, and combine the neighbor battery group to calculate the aging collaborative deviation of each single battery using a spatial proximity principle;

[0012] S3, according to the single battery aging collaborative deviation, based on the aging collaborative consistency principle, dynamically reconstruct and flexibly control the single battery in the battery module.

[0013] Preferably, the collection of historical state parameters of each single battery in the battery module, and based on the regionalized neighbor interaction mechanism, the number of neighbors of each single battery is determined by using the game analysis algorithm to obtain the neighbor battery group includes the following steps:

[0014] S11, based on the historical state parameters of each single battery in the battery module, the multi-dimensional evaluation index of the single battery is analyzed, and the importance of the single battery in the battery module is determined by combining the ecological niche theory;

[0015] S12, based on the importance of the single battery in the battery module, the single battery is analyzed by using an independent cascading game model, and the neighbor relationship strength between each single battery and other single batteries is determined according to the analysis result;

[0016] S13, based on the neighbor relationship strength between each single battery and other single batteries, the number of neighbors of each single battery is determined to form a neighbor battery group.

[0017] Preferably, the importance of the single battery in the battery module is determined by analyzing the multi-dimensional evaluation index of the single battery and combining the ecological niche theory based on the historical state parameters of each single battery in the battery module includes the following steps:

[0018] S111, the historical state parameters of each single battery in the battery module are preprocessed by using range standardization to obtain importance indexes, and the weight values of each importance index are calculated by using an entropy method;

[0019] S112, according to the weight value of each importance index, the current state index and the state evolution index corresponding to each importance index are determined;

[0020] S113, based on the current state index and the state evolution index corresponding to each importance index, the importance of the single battery in the battery module is calculated.

[0021] Preferably, the formula for calculating the importance of the single battery in the battery module based on the current state index of the battery corresponding to each importance index and the state evolution index of the battery is:

[0022] ;

[0023] wherein N i represents the importance of the i-th single battery in the battery module, S i represents the current state index of the i-th battery, P i represents the state evolution index of the i-th single battery, A i and A j both represent dimension conversion coefficients, S j represents the current state index of the j-th battery, P j represents the state evolution index of the j-th single battery, and n represents the number of single batteries.

[0024] Preferably, based on the importance of the single battery in the battery module, the independent cascade game model is used to perform propagation game analysis on the single battery, and the neighbor relationship strength between each single battery and other single batteries is determined according to the analysis result, including the following steps:

[0025] S121, based on the importance of the single battery in the battery module, the single batteries in the battery module are divided into excited state battery groups and non-excited state battery groups;

[0026] S122, the single batteries in the excited state battery groups are configured as initial source nodes for information propagation, and according to the rules of the independent cascade model, at each time step, the excited state battery propagates information to its non-excited state neighbor batteries;

[0027] S123, based on the excited state of the single battery, the benefit of each single battery in the propagation process is calculated, and the state transition probability of the battery is adjusted according to the benefit matrix to obtain the game analysis result;

[0028] S124, according to the game analysis result, the neighbor relationship strength between each single battery and other single batteries is determined.

[0029] Preferably, based on the excited state of the single battery, the benefit of each single battery in the propagation process is calculated, and the state transition probability of the battery is adjusted according to the benefit matrix to obtain the game analysis result, including the following steps:

[0030] S1231, based on the excited state of the single battery, the power cooperation benefit, voltage stability benefit and fault risk reduction benefit of each single battery in the propagation process are calculated respectively to obtain the benefit calculation result;

[0031] S1232, based on the revenue calculation result, the total revenue of each single battery is calculated by weighted average method, and a revenue matrix is constructed;

[0032] S1233, according to the revenue matrix, the state probability of each single battery from the non-excited state to the excited state is determined, and the final game analysis result is obtained.

[0033] Preferably, based on the current state parameters of each single battery in the battery module, the aging degree of the battery is evaluated by using the stacked neural network, and the aging collaborative deviation of each single battery is calculated by using the spatial proximity principle in combination with the neighbor battery group, including the following steps:

[0034] S21, based on the current state parameters of each single battery in the battery module, the aging evaluation index of the single battery is determined, and each aging evaluation index is standardized;

[0035] S22, an integrated stacked neural network prediction model is constructed, and the aging degree of each single battery is evaluated based on the standardized aging evaluation index;

[0036] S23, based on the neighbor battery group of each single battery, the aging collaborative deviation of each single battery is calculated according to the aging degree evaluation result of the current single battery by using the spatial proximity principle.

[0037] Preferably, the construction of the integrated stacked neural network prediction model and the evaluation of the aging degree of each single battery based on the standardized aging evaluation index includes the following steps:

[0038] S221, neural network structures with different hidden layer structures are constructed, and the neural network structures with different hidden layer structures are spliced and fused to obtain a stacked neural network prediction model;

[0039] S222, based on the standardized aging evaluation index, a data input matrix is constructed;

[0040] S223, the data input matrix is feature extracted and predicted by using the stacked neural network prediction model, and the aging degree prediction value of the battery is output;

[0041] S224, according to the aging prediction value output by the model, the aging grade of the single battery is classified and evaluated, and the aging degree of each single battery is obtained.

[0042] Preferably, the dynamic reconstruction and flexible control of the single battery in the battery module based on the aging collaborative consistency principle according to the single battery aging collaborative deviation includes the following steps:

[0043] S31, compare the monomer battery aging cooperative deviation with the preset threshold value, evaluate the aging cooperation of the monomer battery in the battery module according to the comparison result, and obtain the aging cooperation state;

[0044] S32, based on the aging cooperation state, and combined with the aggregation, separation and alignment mechanism, the expected aging speed of each monomer battery is calculated;

[0045] S33, according to the expected aging speed and the preset control strategy, the working state of the monomer battery is dynamically adjusted, and the charging and discharging scheduling of the battery module is optimized.

[0046] According to another aspect of the present application, a flexible control system for dynamically reconstructing a battery module is provided, comprising:

[0047] A neighbor battery determination module is configured to collect historical state parameters of each monomer battery in the battery module, and determine the number of neighbors of each monomer battery based on a regional neighbor interaction mechanism and using a game analysis algorithm, to obtain a neighbor battery group.

[0048] A cooperative deviation calculation module is configured to evaluate the aging degree of the battery based on the current state parameters of each monomer battery in the battery module using a support vector regression method, and calculate the aging cooperative deviation of each monomer battery using a spatial proximity principle in combination with the neighbor battery group.

[0049] A battery reconstruction control module is configured to dynamically reconstruct and flexibly control the monomer battery in the battery module based on the aging cooperative deviation of the monomer battery and the aging cooperation consistency principle.

[0050] The present application has the following advantages:

[0051] 1. The present application comprehensively considers the historical state parameters, current state parameters of the battery and the information of the neighbor battery group, realizes fine management of the battery module, can more accurately grasp the state and aging of the battery, and thus formulates a more scientific and reasonable control strategy, improving the efficiency of battery management.

[0052] 2. The present application can more accurately grasp the state and mutual relationship of each monomer battery in the battery module through comprehensive and scientific evaluation and dynamic game analysis, and can optimize the connection relationship and working state of the batteries in the battery module by reasonably determining the neighbor battery group, realize the cooperative work between the batteries, improve the capacity, power output and other performance indicators of the battery module, and enhance the stability and reliability of the battery module.

[0053] 3. The present application realizes accurate evaluation of the aging degree of the battery, calculation of the aging cooperative deviation, and dynamic reconstruction and flexible control by using advanced neural network model and scientific calculation method, improves the intelligent level of battery management, and makes battery management more accurate and efficient. BRIEF DESCRIPTION OF DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings described below only show some of the embodiments of the present application, and other drawings can be obtained by those of ordinary skill in the art without any creative effort based on these drawings.

[0055] Figure 1 is a flow chart of a flexible control method of a dynamically reconfigured battery module according to an embodiment of the present application;

[0056] Figure 2 is a principle block diagram of a flexible control system of a dynamically reconfigured battery module according to an embodiment of the present application;

[0057] Figure 3 is a principle schematic diagram of a flexible battery management system in a flexible control method of a dynamically reconfigured battery module according to an embodiment of the present application;

[0058] Figure 4 is a state editing schematic diagram of a battery module in a flexible control method of a dynamically reconfigured battery module according to an embodiment of the present application;

[0059] Figure 5 is a state detection schematic diagram of a battery module in a flexible control method of a dynamically reconfigured battery module according to an embodiment of the present application.

[0060] In the drawings:

[0061] 1, neighbor battery determination module; 2, cooperative deviation amount calculation module; 3, battery reconfiguration control module. DETAILED DESCRIPTION

[0062] In order to enable those skilled in the art to better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be described clearly and completely in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without any creative effort should be within the scope of protection of the present application.

[0063] According to an embodiment of the present application, a flexible control method and system of a dynamically reconfigured battery module are provided.

[0064] The present application will be further described in conjunction with the drawings and specific embodiments, such as Figure 1As shown, according to one embodiment of the present application, a flexible control method for dynamically reconstructing a battery module is provided, comprising the following steps:

[0065] S1, collect the historical state parameters of each single battery in the battery module, and determine the number of neighbors of each single battery based on the regionalized neighbor interaction mechanism and using a game analysis algorithm, to obtain a neighbor battery group;

[0066] It should be noted that for the battery module, such as Figure 3 As shown, the flexible battery management system (FX-BMS) is used for battery management, and the hardware part of the flexible battery management system mainly includes a sampling unit, a BMS master control unit, a communication unit, an isolation DCDC (50V<->400V) bidirectional conversion unit, and a display unit (external) and other parts, an isolation DCDC is internally provided with a control unit (DCC) responsible for its operation control and state management, and is connected with the BMS master control unit through the UART TTL protocol. The sampling unit completes the voltage and current collection of the battery; the BMS master control unit completes system information collection, data processing and information interaction and other functions; the communication unit completes data interaction with other devices, and the interaction mode is CAN bus or RS485, etc., the standard baud rate supported by the CAN bus includes 10 kbps to 1 Mbps, when monitoring the battery state (such as the dynamic balancing of electric vehicles), the refresh rate is ≥10 Hz; the isolation DCDC bidirectional conversion unit completes high-low voltage conversion, collection and protection, etc.; the display unit completes the state indication of the BMS, such as working state, fault state and charging and discharging state (external LED indicator board, which can show the running state through the indicator light), the PCS (energy storage converter) connects the isolation DCDC and the power grid (or load), realizes AC-DC conversion and bidirectional energy flow control, and the cascaded CAN2 can support multi-device networking expansion. The composition modules of the FX-BMS use the SH367309 integrated battery management chip for the sampling unit to complete the voltage and current collection of the single battery; the BMS master control unit uses the STM32G0B1 ARM processor to complete system information collection, data processing and information interaction, etc. The DCC master control unit uses the TMS320F28035PAGT DSP processor to complete high-low voltage conversion, voltage and current collection, etc. The isolation DCDC hardware circuit has a short-circuit protection function. In addition, the regionalized neighbor interaction mechanism is a management strategy for dynamically determining the neighbor relationship of the battery through the game analysis algorithm combined with the physical position and state similarity of the battery and realizing efficient interaction, through regional division and neighbor relationship strength analysis, precise cooperation between the batteries is realized, and the overall performance of the battery module (such as power balance, thermal management, fault isolation) is improved, through the pre-set neighbor relationship strength threshold, only when the neighbor relationship strength between two single batteries is greater than the threshold, they are considered to have a neighbor relationship.

[0067] As a preferred embodiment, the historical state parameters of each single battery in the battery module are collected, and the number of neighbors of each single battery is determined based on the regional neighbor interaction mechanism and using a game analysis algorithm to obtain a neighbor battery group, including the following steps:

[0068] S11, based on the historical state parameters of each single battery in the battery module, the importance of the single battery in the battery module is determined by analyzing the multi-dimensional evaluation index of the single battery and combining the niche theory;

[0069] It should be noted that in the battery module, the historical state parameters of the single battery are various, such as voltage, current, temperature, internal resistance, capacity, etc.

[0070] As a preferred embodiment, the historical state parameters of each single battery in the battery module are collected, and the importance of the single battery in the battery module is determined by analyzing the multi-dimensional evaluation index of the single battery and combining the niche theory, including the following steps:

[0071] S111, the historical state parameters of each single battery in the battery module are preprocessed using range standardization to obtain importance indicators, and the weight values of each importance indicator are calculated using an entropy method;

[0072] It should be noted that the historical state parameters of each single battery in the battery module have different dimensions and value ranges, and direct comparison and analysis will be affected by the dimension. Range standardization can convert these parameters into dimensionless relative values, making different parameters comparable. After standardization, the values of all parameters are within the range [0, 1].

[0073] The entropy method is an objective weighting method that determines the weight according to the variation degree of the index data. Information entropy is an index used to measure the uncertainty of information in information theory. The greater the information entropy, the greater the uncertainty of the data and the less the amount of information contained; on the contrary, the smaller the information entropy, the smaller the uncertainty of the data and the more the amount of information contained. In the entropy method, the information entropy of each index is calculated to obtain the difference coefficient of each index. The greater the difference coefficient, the greater the variation degree of the index, the more the amount of information contained, and the greater the weight.

[0074] S112, according to the weight value of each importance indicator, determine the battery current state index and battery state evolution index corresponding to each importance indicator;

[0075] Specifically, for each single battery, multiply the standardized value of each importance indicator by the corresponding weight, then perform weighted summation to obtain the importance sub-value of the battery based on the current state, which can be used as part of the battery current state index or directly as the battery current state index.

[0076] Secondly, the state evolution index is calculated by analyzing the rate of change of the battery historical state parameters. For example, for a certain importance index, the ratio of the change amount of the parameter value in the adjacent time period to the initial parameter value is calculated to obtain the change rate of the index, and then the weighted sum is obtained by combining the weights to obtain the battery state evolution index.

[0077] S113, based on the battery current state index and the battery state evolution index corresponding to each importance index, the importance of the single battery in the battery module is calculated.

[0078] As a preferred embodiment, the calculation formula for calculating the importance of the single battery in the battery module based on the battery current state index and the battery state evolution index corresponding to each importance index is:

[0079] ;

[0080] In the formula, N i represents the importance of the i-th single battery in the battery module, S i represents the current state index of the i-th battery, reflecting the contribution of the current performance and health state of the battery to the importance, P i represents the state evolution index of the i-th single battery, reflecting the influence of the change trend of the battery state on the importance, A i and A j both represent dimension conversion coefficients, which are used to uniformly process the current state index and the state evolution index of different dimensions to ensure the rationality of the calculation result, S j represents the current state index of the j-th battery, P j represents the state evolution index of the j-th single battery, and n represents the number of single batteries.

[0081] S12, based on the importance of the single battery in the battery module, the independent cascade game model is used to perform propagation game analysis on the single battery, and the neighbor relationship strength between each single battery and other single batteries is determined according to the analysis result;

[0082] As a preferred embodiment, the independent cascade game model is used to perform propagation game analysis on the single battery based on the importance of the single battery in the battery module, and the neighbor relationship strength between each single battery and other single batteries is determined according to the analysis result, which includes the following steps:

[0083] S121, based on the importance of the single battery in the battery module, the single batteries in the battery module are divided into excited state battery groups and non-excited state battery groups;

[0084] It should be noted that the purpose of dividing the excited state battery group and the non-excited state battery group is to simulate the starting condition of the information propagation in the battery module.

[0085] Among them, the excited state battery group is equivalent to the seed node of information propagation, which can actively propagate information to other batteries. The division standard is usually determined according to the importance of the single battery, for example, an importance threshold can be set, and the single battery with an importance higher than the threshold is classified into the excited state battery group, and the single battery with an importance lower than the threshold is classified into the non-excited state battery group.

[0086] S122, configure the single battery in the excited state battery group as the initial source node of information propagation, and according to the independent cascade model rule, at each time step, the excited state battery propagates information to its non-excited state neighbor battery.

[0087] It should be noted that before the information propagation, the single battery in the excited state battery group needs to be configured as the initial source node of information propagation, so that these batteries play a leading role in the information propagation process. It can actively deliver the information (such as battery state, performance data, etc.) carried by itself to the surrounding neighbor battery.

[0088] In addition, the independent cascade model is a commonly used information propagation model, and the information propagation model includes the following rules:

[0089] (1) When each excited state battery propagates information to its non-excited state neighbor battery, there is a propagation probability p, and the propagation probability parameter takes a value in the range of [0, 1]. This probability can be set according to the actual situation, for example, based on the physical connection strength between the batteries, the battery type similarity and other factors. The greater the propagation probability, the easier the information is to propagate between the batteries.

[0090] (2) The propagation process is advanced in time steps. At each time step, the excited state battery will try to propagate information to its non-excited state neighbor battery. If the propagation is successful, the neighbor battery will become excited and can continue to propagate information to its own non-excited state neighbor battery in the next time step.

[0091] (3) The process of each excited state battery propagating information to neighbor batteries is independent of each other, that is, when a battery propagates information to multiple neighbor batteries, whether each neighbor battery is excited is not affected by whether other neighbor batteries are excited.

[0092] Through this rule, the propagation process of information in the battery module can be simulated. For example, in a battery module, the excited state battery D propagates information to its neighbor batteries B and C, and the propagation probabilities are p AB and p ACAt a certain time step, battery D has a certain probability of successfully propagating information to battery B or C, making it excited. Then at the next time step, battery B and C (if excited) will continue to propagate information to their neighbor batteries.

[0093] S123、based on the excited state of the single battery, calculate the benefit of each single battery in the propagation process, and adjust the state transition probability of the battery according to the benefit matrix to obtain the game analysis result;

[0094] As a preferred embodiment, the game analysis result based on the excited state of the single battery, the benefit of each single battery in the propagation process is calculated, and the state transition probability of the battery is adjusted according to the benefit matrix, which includes the following steps:

[0095] S1231, based on the excited state of the single battery, the power cooperation benefit, voltage stability benefit and fault risk reduction benefit of each single battery in the propagation process are calculated respectively to obtain the benefit calculation result;

[0096] Specifically, in the battery module, there is a power cooperation effect between single batteries. When some single batteries are in an excited state and participate in information propagation, they can better coordinate with other batteries for power allocation and cooperation. For example, the excited state battery can more accurately understand the power demand of the entire module, so as to reasonably allocate its own power output and avoid local battery overcharge or overdischarge. The power cooperation benefit is calculated by comparing the change of the power utilization efficiency of the entire battery module before and after the excited state battery participates in the propagation. For example, the power utilization efficiency of the module before the propagation information is 70%, and after the propagation information, it is increased to 75%, and the power cooperation benefit is 5%.

[0097] Wherein, the excited state of the single battery helps to maintain the voltage stability of the battery module. The excited state battery can monitor the voltage of itself and the surrounding battery in real time, and coordinate the charging and discharging process of other batteries through information propagation to reduce voltage fluctuation. Specifically, the voltage fluctuation amplitude can be used as a measure. The standard deviation of the module voltage before and after the excited state battery propagates information is calculated, and the reduction of the standard deviation is the voltage stability benefit. For example, the standard deviation of the module voltage before the propagation information is 0.2V, and after the propagation information, it is reduced to 0.1V, and the voltage stability benefit is 0.1V.

[0098] In addition, the excited state battery can timely spread its own state information, so that other batteries can learn about the potential failure risk in advance and take corresponding preventive measures, thereby reducing the failure risk of the entire battery module. Then, a failure risk assessment model is established according to historical failure data and the current battery state. The probability of failure of the entire module before and after the excited state battery spreads information is calculated. The reduction of the failure probability is the failure risk reduction benefit. For example, the module failure probability is 10% before the information is spread, and is reduced to 5% after the information is spread, and the failure risk reduction benefit is 5%.

[0099] S1232, based on the benefit calculation result, the total benefit of each single battery is calculated by a weighted average method, and a benefit matrix is constructed;

[0100] It should be noted that the power coordination benefit, the voltage stability benefit and the failure risk reduction benefit are different in importance to the battery module, and the weights can be determined according to the actual application scene and requirements. For example, if more attention is paid to the voltage stability of the battery module, the weight of the voltage stability benefit can be appropriately increased. It is assumed that the weight of the power coordination benefit is w1, the weight of the voltage stability benefit is w2, and the weight of the failure risk reduction benefit is w3, and w1+w2+w3=1.

[0101] For each single battery, the power coordination benefit, the voltage stability benefit and the failure risk reduction benefit are multiplied by the corresponding weight, and then the results are added to obtain the total benefit of the single battery. The calculation formula is:

[0102] Total benefit = w1 x power coordination benefit + w2 x voltage stability benefit + w3 x failure risk reduction benefit.

[0103] S1233, according to the benefit matrix, the state probability of each single battery from the non-excited state to the excited state is determined, and the final game analysis result is obtained.

[0104] Specifically, the benefit matrix reflects the benefit relationship between different batteries. The higher the benefit, the greater the positive influence of one battery on another battery, and the greater the possibility of the battery from the non-excited state to the excited state.

[0105] In specific application, the state probability of each single battery from the non-excited state to the excited state is calculated according to the data in the benefit matrix by using the logistic regression method. For example, the elements in the benefit matrix are used as input features by using a logistic regression model to obtain the state probability of each battery. Specifically, for single battery i, the state probability M i is expressed as:

[0106] ;

[0107] In the formula, Mi represents the state probability of the monomer battery i, β0represents the intercept term, β j represents the state probability of the monomer battery i, β0represents the intercept term, β ji represents the corresponding regression coefficient.

[0108] Therefore, the state probability of each monomer battery is recorded to form the final game analysis result. This result intuitively reflects the possibility of each battery changing from a non-excited state to an excited state during information propagation, providing a basis for subsequent analysis of the neighbor relationship strength between batteries.

[0109] S124, according to the game analysis result, determine the neighbor relationship strength between each monomer battery and other monomer batteries.

[0110] It should be noted that the neighbor relationship strength represents the closeness of two monomer batteries in terms of information propagation and state influence. The higher the state probability of a battery pair, the greater the neighbor relationship strength, indicating that the information propagation and cooperation between them are stronger. For each monomer battery, according to the influence degree of other batteries on its state transition (i.e. state probability) in the game analysis result, the neighbor relationship strength between it and other monomer batteries is determined.

[0111] S13, based on the neighbor relationship strength between each monomer battery and other monomer batteries, determine the number of neighbors of each monomer battery, and form a neighbor battery group.

[0112] It should be noted that in order to determine the number of neighbors of each monomer battery, a neighbor relationship strength threshold needs to be set. Only when the neighbor relationship strength between two monomer batteries is greater than the threshold, they are considered to have a neighbor relationship. And the threshold can be determined according to the actual application requirements and the size of the battery module. Then the monomer batteries with neighbor relationship can be aggregated into a neighbor battery group. Neighbor groups can also overlap (e.g. batteries A and B, C are neighbors, but B and C have no direct neighbor relationship).

[0113] S2, based on the current state parameters of each monomer battery in the battery module, evaluate the aging degree of the battery using support vector regression method, and combine the neighbor battery group to calculate the aging collaborative bias of each monomer battery using the spatial proximity principle;

[0114] As a preferred embodiment, the method for evaluating the aging degree of the battery based on the current state parameters of each monomer battery in the battery module using a stacked neural network, and combining the neighbor battery group to calculate the aging collaborative bias of each monomer battery using the spatial proximity principle comprises the following steps:

[0115] S21, based on the current state parameters of each monomer battery in the battery module, determine the aging evaluation index of the monomer battery, and standardize each aging evaluation index;

[0116] Specifically, the current state parameters of the monomer battery can reflect its aging degree, and different state parameters reflect the performance changes of the battery from different angles. For example, the capacity decay of the battery is one of the important indicators to measure the aging of the battery, and its capacity will gradually decrease as the use time of the battery increases; the increase of internal resistance is also a common performance of battery aging, and the increase of internal resistance will cause the increase of energy loss of the battery in the process of charging and discharging.

[0117] For example, the monomer battery aging evaluation indicators include: capacity decay rate (%), internal resistance growth rate (%), charging and discharging efficiency (%), self-discharge rate (% / month), cycle number, etc. A 200Ah lithium iron phosphate (LiFePO4) module containing 16 monomer batteries (4x4 array) battery module can be used, and then simulated experiments are carried out through high temperature (45°C) and fast charging and discharging (1C) to accelerate aging to obtain indicator data.

[0118] Because the dimensions and numerical ranges of different evaluation indicators are quite different (such as the capacity decay rate is a percentage, and the internal resistance is milliohms), standardization processing is needed to eliminate the influence of dimensions.

[0119] S22, constructing an integrated stacked neural network prediction model, and evaluating the aging degree of each monomer battery based on the aging evaluation indicators after standardization processing;

[0120] As a preferred embodiment, the step of constructing an integrated stacked neural network prediction model and evaluating the aging degree of each monomer battery based on the aging evaluation indicators after standardization processing includes the following steps:

[0121] S221, constructing neural network structures with different hidden layer structures, and splicing and fusing the neural network structures with different hidden layer structures to obtain a stacked neural network prediction model;

[0122] Specifically, different hidden layer structures have different feature extraction capabilities and learning abilities. For example, shallow neural networks may be more suitable for extracting simple features, while deep neural networks can learn more complex nonlinear relationships. By constructing neural networks with different hidden layer structures, the battery aging evaluation indicators can be extracted from multiple angles.

[0123] In constructing neural network structures with different hidden layer structures, neural networks with different numbers of layers and different numbers of neurons need to be constructed. For example, a neural network with 2 hidden layers, each with 10 and 20 neurons respectively, is constructed; and a neural network with 3 hidden layers, each with 15, 25 and 30 neurons respectively, is constructed. These neural networks can use different types of network structures such as fully connected neural networks, convolutional neural networks (if the state parameters have spatial structure characteristics), etc. Among them, in order to enable the network to learn complex patterns, an activation function introduces a nonlinear function such as ReLU (Rectified Linear Unit) or its variants (such as LeakyReLU).

[0124] Finally, the neural networks with different hidden layer structures are spliced and fused. In splicing and fusion, the output layers of different neural networks are connected to form a new output layer. For example, assuming that there are two neural networks, the output layer of neural network T has m neurons, and the output layer of neural network U has n neurons, the output layer of the spliced and fused stacked neural network will have m+n neurons.

[0125] S222, constructing a data input matrix based on the standardized aging evaluation indicators;

[0126] S223, using a stacked neural network prediction model to perform feature extraction and prediction on the data input matrix, and outputting a battery aging degree prediction value;

[0127] Among them, the stacked neural network prediction model performs feature extraction on the data input matrix through its hidden layers; the neurons of each hidden layer perform nonlinear transformation on the input data, converting the original aging evaluation indicators into higher-level feature representations. After feature extraction by multiple hidden layers, the output layer of the stacked neural network prediction model will predict the aging degree of each single battery based on the extracted features. The number of neurons in the output layer is usually related to the prediction target. If the target is to predict the continuous value of the battery aging degree, the output layer can have one neuron; if the target is to classify the aging level, the number of neurons in the output layer can be equal to the number of classes of the aging level.

[0128] S224, classifying and evaluating the aging level of each single battery according to the aging prediction value output by the model, to obtain the aging degree of each single battery.

[0129] S23, based on the neighbor battery group of each single battery, calculating the aging coordination deviation of each single battery according to the aging degree evaluation result of the current single battery using the spatial proximity principle.

[0130] It should be noted that the spatial proximity principle defines the neighbor relationship based on physical location or spatial distance, and the division of neighbor battery groups directly depends on the layout of battery groups in physical space (such as adjacent battery modules, racks or energy storage units). Battery groups close in physical location are easily affected by similar environmental conditions (such as temperature, vibration), thus showing similar aging characteristics or performance deviations. This spatial correlation provides the basis for "regionalized neighbor interaction".

[0131] The calculation of the aging coordination deviation quantity aims to measure the difference between the aging degree of a single battery and the aging degree of its neighbor battery group, which reflects the spatial distribution of the aging state in the battery module. Specifically, the following steps are included:

[0132] First, the neighbor battery group of the current single battery is obtained, which is determined based on the neighbor relationship strength and other information in the previous step. Then, the average value of the aging degree evaluation results of all batteries in the neighbor battery group is calculated;

[0133] The aging degree evaluation result of the current single battery is compared with the average aging degree of the neighbor battery group, and the difference between them is calculated, which is the aging coordination deviation quantity of the current single battery.

[0134] Specifically, the aging coordination deviation quantity reflects the difference between the aging degree of the current single battery and the aging degree of the surrounding neighbor battery. If the aging coordination deviation quantity is positive, it means that the aging degree of the current single battery is higher than the average aging degree of the neighbor battery group, which may mean that the battery is aging faster in the current local spatial environment or is affected by some special factors; if the aging coordination deviation quantity is negative, it means that the aging degree of the current single battery is lower than the average aging degree of the neighbor battery group.

[0135] S3, according to the aging coordination deviation quantity of the single battery, based on the aging coordination consistency principle, dynamically reconstruct and flexibly control the single battery in the battery module.

[0136] As a preferred embodiment, the dynamically reconstructing and flexibly controlling the single battery in the battery module based on the aging coordination deviation quantity of the single battery and the aging coordination consistency principle includes the following steps:

[0137] S31, comparing the aging coordination deviation quantity of the single battery with a preset threshold, and evaluating the aging coordination of the single battery in the battery module according to the comparison result to obtain the aging coordination state;

[0138] Specifically, according to the comparison result, the aging coordination state of the single battery needs to be divided into two categories: good coordination and abnormal coordination, to form the aging coordination state evaluation result of the battery module.

[0139] S32, based on the aging coordination state, and combined with the aggregation, separation, and alignment mechanisms, calculate the expected aging speed of each single battery;

[0140] It should be noted that the adjustment mechanism of the battery aging speed (aggregation, separation, and alignment) is mainly used in the reversible capacity decay stage. It adjusts the stress level of the battery through charge and discharge scheduling (such as balanced charging and dynamic power distribution), thereby indirectly affecting the aging speed.

[0141] The aggregation mechanism aims to keep the battery group with good aging coordination at a similar aging speed, so as to maintain the consistency of the overall aging state of the battery module. Through the aggregation mechanism, it can avoid the over-aging or over-youngness of some batteries, thereby ensuring the stability of the performance of the battery module.

[0142] The goal of the separation mechanism is to make the abnormal battery different from the surrounding batteries in aging speed, so as to prevent the aging problem of the abnormal battery from spreading to the entire battery module, and at the same time provide conditions for subsequent individual processing of the abnormal battery.

[0143] The alignment mechanism comprehensively considers the overall performance target of the battery module and the actual situation of the single battery, adjusts the aging speed of the abnormal battery to a reasonable range of the aging speed of the well-coordinated battery group, and realizes the overall alignment of the aging state of the battery module.

[0144] In addition, for the single battery i with good aging coordination, its expected aging speed V i,desired is determined according to the average aging speed V avg of the battery module, for example, V i,desired =V avg , and the average aging speed of the battery module is calculated by averaging the aging speeds of all well-coordinated batteries.

[0145] For the single battery j with abnormal aging coordination, it is necessary to first determine the adjustment direction of its aging speed. If the aging is too fast, the expected aging speed V j,desired should be less than the average aging speed of the battery module to slow down its aging process; if the aging is too slow, the expected aging speed V j,desired should be greater than the average aging speed of the battery module to speed up its aging speed and gradually coordinate with the surrounding batteries. Then the expected aging speed is calculated by the following formula:

[0146] V j,desired =V avg +k×D j ;

[0147] wherein k represents an adjustment coefficient, which can be set according to the characteristics and control requirements of the battery module, and is used to control the amplitude of the aging speed adjustment. D jThis indicates the amount of aging synergistic deviation.

[0148] S33. Based on the desired aging rate and preset control strategy, dynamically adjust the working state of individual cells and optimize the charging and discharging scheduling of the battery module.

[0149] It should be noted that the operating state of a single battery cell (such as charging / discharging current and voltage) directly affects its aging rate. By dynamically adjusting the operating state of a single battery cell according to the desired aging rate, the battery aging process can be controlled.

[0150] Specifically, if a slower aging rate is desired, the charging and discharging current can be appropriately reduced to decrease the rate of chemical reactions inside the battery, thereby slowing down the aging process; conversely, if a faster aging rate is desired, the charging and discharging current can be appropriately increased.

[0151] In addition, adjusting the charge and discharge cutoff voltage of a single battery cell can also affect its aging rate. For example, for batteries where a slower aging rate is desired, the charge cutoff voltage can be appropriately increased and the discharge cutoff voltage decreased, allowing the battery to operate at a shallower charge and discharge depth, thereby reducing the number of battery cycles and the degree of aging.

[0152] Secondly, optimizing the charging and discharging schedule of the battery module aims to achieve coordinated control of the aging rate of individual cells while meeting the overall performance requirements of the battery module, thereby improving the service life and reliability of the battery module. Based on the expected aging rate of individual cells, the charging and discharging sequence and time of each cell in the battery module are rationally arranged, allowing cells with faster aging rates more rest time and cells with slower aging rates to have their working time appropriately increased, thus achieving a balance in the overall aging rate of the battery module.

[0153] In addition, such as Figures 4-5 The image shows the detection and editing panel for Lithium Valley batteries. It can detect the SOC of the batteries in the photovoltaic-energy storage integrated unit, as well as the voltage, power, and load current of the mains power. Then, through preset editing templates, it can assign different control modes to each battery according to the status of the battery module, such as shutdown mode, charging mode, follow mode, and discharging mode. It also includes specific start and end times, as well as the upper limit of charging power and the upper limit of discharging power. It also records the current electricity sales price and purchase price of the battery module.

[0154] like Figure 2 As shown, according to another embodiment of the present invention, a flexible control system for dynamically reconfiguring a battery module is provided, comprising:

[0155] A neighbor battery determination module 1 is configured to collect historical state parameters of each single battery in a battery module, and determine the number of neighbors of each single battery based on a regional neighbor interaction mechanism and a game analysis algorithm, to obtain a neighbor battery group;

[0156] A cooperative deviation amount calculation module 2 is configured to evaluate the aging degree of the battery based on the current state parameters of each single battery in the battery module, and calculate the aging cooperative deviation amount of each single battery based on the neighbor battery group and a spatial proximity principle.

[0157] A battery reconstruction control module 3 is configured to dynamically reconstruct and flexibly control the single batteries in the battery module based on the aging cooperative deviation amount and an aging cooperative consistency principle.

[0158] In summary, by means of the above technical solutions of the present application, the historical state parameters, the current state parameters and the information of the neighbor battery group are comprehensively considered, the fine management of the battery module is realized, the state and the aging condition of the battery can be more accurately mastered, and therefore a more scientific and reasonable control strategy can be formulated, and the efficiency of the battery management is improved. By means of comprehensive and scientific evaluation and dynamic game analysis, the state and the mutual relationship of each single battery in the battery module can be more accurately mastered, the connection relationship and the working state between the batteries in the battery module can be optimized by reasonably determining the neighbor battery group, the cooperative work between the batteries can be realized, the performance indexes such as the capacity and the power output of the battery module are improved, and the stability and the reliability of the battery module are enhanced. By means of the advanced neural network model and the scientific calculation method, the accurate evaluation of the aging degree of the battery, the calculation of the aging cooperative deviation amount and the dynamic reconstruction and flexible control are realized, the intelligent level of the battery management is improved, and the battery management is more accurate and efficient.

[0159] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage, optical storage, etc.) containing computer usable program code.

[0160] The above specific embodiments further illustrate the purpose, technical solutions and beneficial effects of the present application, and it should be understood that the above description is only for specific embodiments of the present application and is not used to limit the protection scope of the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A flexible control method of dynamically reconfiguring a battery module, characterized by, The method comprises the following steps: S1, collecting historical state parameters of each single battery in the battery module, and determining the number of neighbors of each single battery by using a game analysis algorithm based on a regionalized neighbor interaction mechanism, to obtain a neighbor battery group; S2, evaluating the aging degree of the battery by using a support vector regression method based on the current state parameters of each single battery in the battery module, and calculating the aging collaborative deviation of each single battery by using a spatial proximity principle in combination with the neighbor battery group; S3, dynamically reconstructing and flexibly controlling the single batteries in the battery module based on the aging collaborative consistency principle according to the aging collaborative deviation of the single batteries; The collecting of the historical state parameters of each single battery in the battery module and the determination of the number of neighbors of each single battery by using a game analysis algorithm based on a regionalized neighbor interaction mechanism to obtain a neighbor battery group comprises the following steps: S11, determining the importance of the single battery in the battery module by analyzing the multi-dimensional evaluation indexes of the single battery and combining the niche theory based on the historical state parameters of each single battery in the battery module; S12, determining the neighbor relationship strength between each single battery and other single batteries by using an independent cascade game model to perform propagation game analysis on the single battery based on the importance of the single battery in the battery module, and according to the analysis result; S13, determining the number of neighbors of each single battery based on the neighbor relationship strength between each single battery and other single batteries, to form a neighbor battery group.

2. The flexible control method of dynamically reconfiguring battery modules according to claim 1, wherein, The determination of the importance of the single battery in the battery module by analyzing the multi-dimensional evaluation indexes of the single battery and combining the niche theory based on the historical state parameters of each single battery in the battery module comprises the following steps: S111, preprocessing the historical state parameters of each single battery in the battery module by using range standardization to obtain an importance index, and calculating the weight value of each importance index by using an entropy method; S112, determining the battery current state index and the battery state evolution index corresponding to each importance index according to the weight value of each importance index; S113, calculating the importance of the single battery in the battery module based on the battery current state index and the battery state evolution index corresponding to each importance index.

3. The flexible control method of dynamically reconfiguring battery modules according to claim 2, wherein, The calculation formula of the importance of the single battery in the battery module based on the battery current state index and the battery state evolution index corresponding to each importance index is: ； where N i represents the importance of the i-th single battery in the battery module, S i represents the current state index of the i-th battery, P i represents the state evolution index of the i-th single battery, A i and A j both represent dimension conversion coefficients, S j represents the current state index of the j-th battery, P j represents the state evolution index of the j-th single battery, and n represents the number of single batteries.

4. The flexible control method of dynamically reconfiguring battery modules according to claim 2, wherein, The determination of the neighbor relationship strength between each single battery and other single batteries by using an independent cascade game model to perform propagation game analysis on the single battery based on the importance of the single battery in the battery module, and according to the analysis result comprises the following steps: S121, dividing the single batteries in the battery module into an excited state battery group and a non-excited state battery group based on the importance of the single battery in the battery module; S122, configuring the single batteries in the excited state battery group as initial source nodes of information propagation, and propagating information from the excited state battery to its neighbor battery in the non-excited state at each time step according to the rules of the independent cascade model; S123, based on the excitation state of the single battery, calculate the benefit of each single battery in the propagation process, and adjust the state transition probability of the battery according to the benefit matrix to obtain the game analysis result; S124, according to the game analysis result, determine the neighbor relationship strength between each single battery and other single batteries.

5. The flexible control method of a dynamically reconfigurable battery module according to claim 4, wherein, The step of calculating the benefit of each single battery in the propagation process based on the excitation state of the single battery, and adjusting the state transition probability of the battery according to the benefit matrix to obtain the game analysis result includes the following steps: S1231, based on the excitation state of the single battery, calculate the power cooperation benefit, voltage stability benefit and fault risk reduction benefit of each single battery in the propagation process respectively to obtain the benefit calculation result; S1232, based on the benefit calculation result, calculate the total benefit of each single battery by weighted average method, and construct a benefit matrix; S1233, according to the benefit matrix, determine the state probability of each single battery from non-excited state to excited state, and obtain the final game analysis result.

6. The flexible control method of dynamically reconfiguring battery modules according to claim 1, wherein, The step of evaluating the aging degree of each single battery by using stacked neural network based on the current state parameters of each single battery in the battery module, and calculating the aging collaborative deviation of each single battery by using the spatial proximity principle combined with the neighbor battery group includes the following steps: S21, based on the current state parameters of each single battery in the battery module, determine the aging evaluation index of the single battery, and standardize each aging evaluation index; S22, construct an integrated stacked neural network prediction model, and evaluate the aging degree of each single battery based on the standardized aging evaluation index; S23, based on the neighbor battery group of each single battery, calculate the aging collaborative deviation of each single battery according to the aging degree evaluation result of the current single battery by using the spatial proximity principle.

7. The flexible control method of a dynamically reconfigurable battery module according to claim 6, wherein, The step of constructing an integrated stacked neural network prediction model and evaluating the aging degree of each single battery based on the standardized aging evaluation index includes the following steps: S221, construct neural network structures with different hidden layer structures, and splice and fuse the neural network structures with different hidden layer structures to obtain a stacked neural network prediction model; S222, based on the standardized aging evaluation index, construct a data input matrix; S223, use the stacked neural network prediction model to extract features and predict the data input matrix, and output the aging degree prediction value of the battery; S224, according to the aging prediction value output by the model, classify and evaluate the aging grade of the single battery to obtain the aging degree of each single battery.

8. The flexible control method of dynamically reconfiguring battery modules according to claim 1, wherein, The step of dynamically reconstructing and flexibly controlling the single battery in the battery module based on the aging collaborative consistency principle according to the single battery aging collaborative deviation includes the following steps: S31, compare the single battery aging collaborative deviation with the preset threshold value, evaluate the aging collaboration of the single battery in the battery module according to the comparison result, and obtain the aging collaborative state; S32, based on the aging collaborative state, and combined with the aggregation, separation and alignment mechanism, calculate the expected aging speed of each single battery; S33, dynamically adjust the working state of the single battery according to the expected aging speed and the preset control strategy, and optimize the charge and discharge scheduling of the battery module.

9. A flexible control system for dynamically reconfiguring a battery module, for implementing the method of any one of claims 1-8, wherein, Comprise: A neighbor battery determination module is used to collect the historical state parameters of each single battery in the battery module, and determine the number of neighbors of each single battery based on the regional neighbor interaction mechanism and the game analysis algorithm, to obtain a neighbor battery group; A collaborative deviation calculation module is used to evaluate the aging degree of the battery based on the current state parameters of each single battery in the battery module, and calculate the aging collaborative deviation of each single battery based on the space proximity principle combined with the neighbor battery group by using the support vector regression method; A battery reconstruction control module is used to dynamically reconstruct and flexibly control the single battery in the battery module based on the aging collaborative consistency principle according to the single battery aging collaborative deviation.

Images