A battery equalization charge and discharge control device and method based on fuzzy clustering

Abstract

This invention discloses a battery equalization charge / discharge control device and method based on fuzzy clustering, comprising: S1. Inputting all individual cells; S2. Acquiring charge / discharge data sets of all individual cells in the series-connected battery pack; S3. In response to at least one individual cell having a capacity greater than / less than a corresponding set capacity threshold, initiating the current round of charge / discharge equalization; S4. Statistically calculating the similarity between individual cells based on the charge / discharge data sets; S5. Determining cluster centers, and sequentially clustering all individual cells according to several similarity thresholds from largest to smallest until all are merged into one class, obtaining several categories C1~Cw; S6. Sequentially cutting out individual cells from non-Cw-1 categories to the first category C1 / second category C2, and repeating steps S1~S6. This scheme uses fuzzy clustering to realize a battery equalization charging device, avoiding energy waste, effectively balancing the capacity differences of individual cells in the battery pack, and improving the energy utilization rate of the battery pack.

Description

Technical Field

[0001] This invention belongs to the field of power electronics technology, and in particular relates to a battery equalization charge and discharge control device and method based on fuzzy clustering. Background Technology

[0002] Battery system balancing management aims to balance the capacity and energy differences of individual cells in the battery pack, thereby improving the energy utilization rate of the battery pack. Balancing circuits are required during the charging and discharging process of the battery pack. Balancing circuits are divided into two main categories: energy dissipation circuits and non-energy dissipation circuits. Energy dissipation circuits dissipate excess energy as heat, resulting in energy waste and increased load on the thermal management system. Non-energy dissipation circuits transfer or convert excess energy to other batteries, resulting in much less energy waste compared to energy dissipation circuits, but their circuit structure is relatively more complex.

[0003] Current battery balancing solutions primarily rely on battery pack voltage to determine battery capacity. Achieving proper battery pack balancing requires high accuracy and precision in voltage detection, and the leakage current in the voltage detection circuit also affects the consistency of the battery pack. Currently, it is difficult to design a simple and efficient voltage detection circuit. Furthermore, voltage is not the only measure of battery capacity; internal resistance and contact resistance in the connection method can also cause variations in battery charge, easily leading to over-balancing and wasted energy. Summary of the Invention

[0004] The purpose of this invention is to provide a battery equalization charging device and method based on fuzzy clustering to address the above-mentioned problems.

[0005] To achieve the above objectives, the present invention adopts the following technical solutions:

[0006] A battery equalization charge-discharge control method based on fuzzy clustering, the method comprising:

[0007] S1. Switch all individual cells to bring the entire series-connected battery pack into a charging / discharging state;

[0008] S2. Obtain the charge and discharge data set of all individual cells in the series battery pack;

[0009] S3. In response to the condition that the charge of at least one individual battery cell is greater than / less than the corresponding set charge threshold, start the current round of charge / discharge balancing;

[0010] S4. Based on the aforementioned charge / discharge data set, statistically analyze the similarity between individual battery cells;

[0011] S5. Using the cell with the least / most power as the cluster center, cluster all cell cells according to several similarity thresholds from large to small until all are merged into one class, resulting in several categories C1~Cw, where w represents the number of categories.

[0012] Category C1 contains the fewest / most individual cells;

[0013] S6. Based on the settings, cut out individual cells from non-Cw-1 type to first type C1 / second type C2 in sequence. This round ends. Repeat steps S1 to S6.

[0014] Each round of charging / discharging corresponds to a set power threshold, and the set power threshold corresponding to the previous round is less than that corresponding to the later round;

[0015] S7. Complete the charging / discharging of the series-connected battery pack.

[0016] In the above-mentioned battery equalization charge and discharge control method based on fuzzy clustering, step S4 specifically includes:

[0017] A data matrix of size m*n is constructed based on the charge and discharge data sets of each individual battery, where m represents the number of data sets in the charge and discharge data sets and n represents the number of individual batteries.

[0018] Each element in the data matrix is ​​denoted as x. ij , i={1,2,…,n},j={1,2,…,m}, x ij This represents the j-th charge / discharge data set of the i-th individual battery cell;

[0019] Perform standard deviation transformation and normalization on each element;

[0020] Establishing a fuzzy similarity matrix based on the correlation coefficient method;

[0021] Step S5 involves clustering based on the similarity matrix.

[0022] In the above-mentioned battery equalization charge and discharge control method based on fuzzy clustering, when the circuit is in the charging state, the single cell with the least charge is the cluster center and is located in all classes;

[0023] When the circuit is in a discharging state, the single cell with the highest charge is the cluster center and is located in all classes.

[0024] In the aforementioned battery equalization charge-discharge control method based on fuzzy clustering, when the circuit is in a charging state, the charge-discharge data set for each individual battery includes the charging voltage v. 充 Charging current i 充 The derivative of the charging voltage with respect to time, dv 充 The derivative of the charging current with respect to time, di 充 ;

[0025] When the circuit is in a discharge state, the charge / discharge data set for each individual battery cell includes the discharge voltage V. 放 Charging current i 放 The derivative of discharge voltage with respect to time, dv 放 The derivative of the discharge current with respect to time, di 放 .

[0026] In the above-mentioned battery equalization charge and discharge control method based on fuzzy clustering, the capacity of a single battery is determined based on the voltage and / or current of the individual battery.

[0027] Each individual battery cell is connected in parallel with a battery switch. By controlling the opening and closing of the corresponding battery switch, the corresponding individual battery cell can be switched in or out.

[0028] In the above-mentioned battery equalization charge and discharge control method based on fuzzy clustering, in step S4, the similarity between the cell with the least / most charge and the remaining cells is calculated at least.

[0029] In step S5, the similarity thresholds are L1, L2...Lw, where w is the number of clusters = the number of categories, L1 > L2 > ... > Lw. L1 is used for the first clustering to obtain C1, L2 is used for the second clustering to obtain C2, ..., and Lw is used for the wth clustering to obtain Cw.

[0030] Step S5 specifically includes:

[0031] Select the cell with the least / most capacity as the first cell;

[0032] Using the first cell as the cluster center and with a similarity threshold of L1, all individual cells are clustered for the first time to obtain the first class.

[0033] Using the first cell as the cluster center and with a similarity threshold of L2, a second clustering is performed on all individual cells to obtain the second class;

[0034] This process continues until all individual cells are merged into one category to obtain the wth category.

[0035] In the above-mentioned battery equalization charge and discharge control method based on fuzzy clustering, in step S6, individual cells from non-Cw-1 to the first type C1 / second type C2 are sequentially cut out according to the power status of each type of individual cell.

[0036] In the above-mentioned battery equalization charge and discharge control method based on fuzzy clustering, step S6 specifically includes:

[0037] In response to meeting the preset power state, the non-Cw-1 type single cell is switched off, and the remaining single cells continue to charge / discharge;

[0038] In response to non-Cw-2 type single cell being charged / discharged to meet set conditions, the corresponding single cell is switched off, and the charging / discharging of the remaining single cells continues;

[0039] When a non-Cw-3 type single cell is charged / discharged to meet the set conditions, the corresponding single cell is disconnected and the charging / discharging of the remaining single cells continues.

[0040] This process continues until all individual cells are cut out or the individual cells of the first type C1 are charged / discharged to meet the set conditions.

[0041] Reconnect all individual cells;

[0042] Repeat steps S2 to S6 until the series battery pack is fully charged / discharged.

[0043] In the above-mentioned battery equalization charge and discharge control method based on fuzzy clustering, the setting condition is that the corresponding single cell is charged / discharged to the state of charge of any single cell that is not of the Cw-1 class.

[0044] The preset power status includes all individual batteries (excluding Cw-1 type) being greater than / less than or equal to the corresponding set power threshold.

[0045] A battery equalization charging and discharging device based on fuzzy clustering achieves equalization of charging / discharging of individual cells in a series battery pack. The device includes a control unit and several battery switches connected in parallel to each individual cell. The control unit controls the closing or opening of each battery switch based on a fuzzy clustering algorithm to cut off or into the charging and discharging circuit of the corresponding individual cell.

[0046] The advantages of this invention are: it uses fuzzy clustering to realize the battery equalization charging device, which solves the problem of over-equalization caused by voltage not being the only measure of battery capacity, avoids wasting energy, effectively balances the capacity and energy differences of individual cells in the battery pack, and improves the energy utilization rate of the battery pack.

[0047] This device is a non-energy dissipation type of equalization charging, which will not cause energy consumption and heat generation problems. It also avoids the problem of high accuracy and precision requirements for voltage detection in existing equipment, and will not cause over-equalization and extreme situations. It can be widely used in the equalization management of power batteries. Attached Figure Description

[0048] Figure 1 The circuit structure diagram of the battery equalization charging device based on fuzzy clustering provided for this solution;

[0049] Figure 2 A flowchart of the battery equalization charge and discharge control method based on fuzzy clustering provided for this solution;

[0050] Figure 3 This is a schematic diagram of charging clustering provided in an embodiment of the solution;

[0051] Figure 4 This is a schematic diagram of the switch control during the charging process provided in this embodiment of the solution. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of this solution clearer, the technical solution of this solution will be clearly and completely described below in conjunction with specific embodiments and corresponding accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this solution, and not all of them. Based on the embodiments of this solution, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this solution.

[0053] like Figure 1 As shown, the battery equalization charge / discharge control device based on fuzzy clustering provided in this solution includes a control unit, individual battery cells, and several battery switches connected in parallel to each individual battery cell, preferably solid-state switches. The solid-state switches control whether an individual battery cell is connected to the charge / discharge circuit. The control unit controls the closing or opening of each battery switch based on a fuzzy clustering algorithm to disconnect or connect the corresponding individual battery cell to the charge / discharge circuit. The algorithm processes the open-circuit voltage, discharge voltage, charging voltage, and charging current of the individual battery cell to be charged, as well as the corresponding trend data of these variables, and makes solid-state switch control decisions through data standardization, establishing a fuzzy similarity matrix, and clustering steps. Figure 2 As shown, the specific implementation method is as follows:

[0054] S1. Switch all individual cells to bring the entire series-connected battery pack into a charging / discharging state;

[0055] S2. Obtain the charge and discharge data set of all individual cells in the series battery pack;

[0056] S3. In response to the condition that the charge of at least one individual battery cell is greater than / less than the corresponding set charge threshold, start the current round of charge / discharge balancing;

[0057] For charging, the current round of charging equalization begins when at least one individual battery cell has a charge level greater than the corresponding set charge threshold.

[0058] For discharge, the current discharge equalization cycle begins when at least one individual cell has a charge level lower than the corresponding set charge threshold.

[0059] S4. Based on the aforementioned charge / discharge data set, statistically analyze the similarity between individual battery cells;

[0060] S5. Using the cell with the least / most power as the cluster center, cluster all cells according to the similarity level and several similarity thresholds from large to small until all are merged into one class. During the clustering process, categories C1 to Cw are obtained in sequence, where w represents the number of categories. The first category C1 contains the cell with the least / most power. This category will be cut out last in this round of balancing, or it may not be cut out.

[0061] S6. In response to the condition that all individual cells belonging to the Cw class are greater than / less than or equal to the corresponding set power threshold, disconnect the individual cells belonging to the Cw class and continue charging / discharging the remaining individual cells;

[0062] During charging, in response to the condition that all individual cells belonging to the Cw class are greater than or equal to the corresponding set power threshold, the individual cells belonging to the Cw class are switched out.

[0063] During discharge, in response to the condition that all individual cells belonging to the Cw class are less than or equal to the corresponding set capacity threshold, the individual cells belonging to the Cw class are switched off.

[0064] When a single cell belonging to the Cw-1 class is charged / discharged to meet the set conditions, the corresponding single cell is disconnected, and the charging / discharging of the remaining single cells continues.

[0065] When a single cell belonging to the Cw-2 class is charged / discharged to meet the set conditions, the corresponding single cell is disconnected, and the charging / discharging of the remaining single cells continues.

[0066] This process continues until all individual cells are cut out or the first-class individual cells are charged / discharged to meet the set conditions.

[0067] Reconnect all individual cells;

[0068] Repeat steps S2 to S6 until the series battery pack is fully charged / discharged.

[0069] The setting condition is that the corresponding single cell is charged / discharged to the state of charge of any single cell belonging to the Cw class.

[0070] The preferred setting for the final round of equalization is 100% (charging) / 1% (discharging). The number of set equalization thresholds is related to the round and can be preset by those skilled in the art as needed. That is, when the charge of a single cell belonging to the Cw class exceeds the corresponding set equalization threshold, this round of equalization begins, until all single cells are charged / discharged to the charge level of a single cell belonging to the Cw class, then the process waits to enter the next round of equalization or ends the charging / discharging process.

[0071] Here, the charge level of a single battery cell can be obtained based on its voltage and / or current, and the specific value is not limited here. It should be noted that the charge level here is only for clustering purposes. This solution does not directly perform charge / discharge balancing based on single data such as charge level or voltage, but rather balances charge and discharge by classifying charge / discharge data groups. Therefore, it is not necessary to obtain the precise charge level, only the approximate charge level is needed.

[0072] Step S4 specifically includes:

[0073] A data matrix of size m*n is constructed based on the charge and discharge data sets of each individual battery, where m represents the number of data sets in the charge and discharge data sets and n represents the number of individual batteries.

[0074] Each element in the data matrix is ​​denoted as x. ij Let i = {1, 2, ..., n} represent the i-th individual cell, j = {1, 2, ..., m} represent the j-th charge / discharge data set, and x ij This represents the j-th charge / discharge data set of the i-th individual battery cell;

[0075] Perform standard deviation transformation and normalization on each element;

[0076] A fuzzy similarity matrix r is established based on the correlation coefficient method. ab Where a,b={1,2,…,n}, and n represents the number of individual cells.

[0077] The standard deviation transformation is shown in the following formula:

[0078] (1)

[0079] (2)

[0080] After standard deviation transformation, normalization is performed. For a single cell, its standardized data is x. j (j=1,2,…,m).

[0081] x ij Let x represent any element in the data matrix. ij This represents the result after standard transformation; n represents the number of individual cells. This represents the sample mean of the k-th data point.

[0082] Using the correlation coefficient method shown below, the statistic r that measures the similarity between data points of individual cells to be charged is plotted. ab Establish a fuzzy similarity matrix:

[0083] (3)

[0084] (4)

[0085] These represent the k-th data point in the a-th and b-th data groups, respectively.

[0086] This represents the sample variance of the k-th data point;

[0087] This represents the sample mean of the k-th data point.

[0088] Generate fuzzy similarity matrix r ab Then, clustering is performed to obtain a cluster graph.

[0089] When the circuit is in a charging state, the single cell with the least charge is the cluster center and is located in all classes;

[0090] The charge / discharge data set for each individual cell includes the charging voltage (V). 充 Charging current i 充 The derivative of the charging voltage with respect to time, dv 充 The derivative of the charging current with respect to time, di 充 ;

[0091] When the circuit is in a discharge state, the individual cell with the highest charge is the cluster center, located in all classes. The charge / discharge data set for each individual cell includes the discharge voltage V. 放 Charging current i 放 The derivative of discharge voltage with respect to time, dv 放 The derivative of the discharge current with respect to time, di 放 .

[0092] Furthermore, in step S4, the similarity between the cell with the least / most charge and the remaining cells is calculated.

[0093] In step S5, several similarity thresholds are defined as L1, L2…Lw, where w is the number of clustering, and L1 > L2 > … > Lw. Step S5 specifically includes:

[0094] Select the cell with the least / most capacity as the first cell;

[0095] Using the first cell as the cluster center and with a similarity threshold of L1, all individual cells are clustered for the first time to obtain the first class.

[0096] Using the first cell as the cluster center and with a similarity threshold of L2, a second clustering is performed on all individual cells to obtain the second class;

[0097] This process continues until all individual cells are merged into one category to obtain category w. At this point, we have category w, and each category has an inclusion relationship where the next category contains the previous category.

[0098] The following explanation uses charging as an example:

[0099] A battery equalization charging device based on fuzzy clustering is used to charge the battery pack, which contains 12 individual cells BT1~BT12, corresponding to switch numbers S1 to S12. The switches are controlled based on the fuzzy clustering results. After charging begins, all 12 switches are disconnected, and all individual cells are connected in series to enter the charging state. Three capacity thresholds are set: 30%, 70%, and 100%. The battery current / voltage is continuously monitored to obtain the battery capacity. When an individual cell reaches 30% capacity, the first round of equalization charging begins. The fuzzy clustering algorithm uses the individual cell BT1 with the lowest capacity as the cluster center and performs multiple clustering operations until all cells are merged into one class. Ultimately, six clustering calculations are performed to obtain six categories C1~C6, at which point N=6. Figure 3 As shown, C1 class includes BT1, BT2 and BT11; C2 class includes BT7, BT12 and C1 class; C3 class includes BT4 and C2 class; C4 class includes BT5, BT9, BT10 and C3 class; C5 class includes BT3 and C4 class; and C6 class includes all individual cells.

[0100] When all non-C5 class (i.e., only C6 class) individual batteries have a charge level greater than 30%, switches S6 and S8 are closed, disconnecting all non-C5 class batteries from the charging circuit, and the remaining individual batteries continue to be charged in series. Next, when all non-C4 class individual batteries have a charge level greater than 30%, switch S3 is closed again, disconnecting all non-C4 class batteries from the charging circuit, and the remaining individual batteries continue to be charged in series. When all non-C3 class individual batteries have a charge level greater than 30%, switches S5, S9, and S10 are closed again, disconnecting all non-C3 class batteries from the charging circuit. When all non-C2 class individual batteries have a charge level greater than 30%, switch S4 is closed again, disconnecting all non-C2 class batteries from the charging circuit. When all non-C1 class individual batteries have a charge level greater than 30%, switches S7 and S12 are closed again, disconnecting all non-C1 class batteries from the charging circuit. Finally, when all C1 class individual batteries have a charge level greater than 30%, this round of charging equalization is complete. The system reconnects to all individual battery cells to obtain their charge levels. When any cell reaches 70%, a second round of charging balancing begins. The second and third rounds are similar to the first and will not be described in detail here. After the third round, charging is complete.

[0101] The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to substitute them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

[0102] Although this paper uses terms such as single cell, charge / discharge data set, set capacity threshold, charge / discharge balancing, similarity, cluster center, and similarity threshold frequently, the possibility of using other terms is not excluded. These terms are used merely for the convenience of describing and explaining the essence of this invention; interpreting them as any additional limitation would contradict the spirit of this invention.

Claims

1. A battery equalization charge-discharge control method based on fuzzy clustering, characterized in that, The method includes: S1. Switch all individual cells to bring the entire series-connected battery pack into a charging / discharging state; S2. Obtain the charge and discharge data set of all individual cells in the series battery pack; S3. In response to the condition that the charge of at least one individual battery cell is greater than / less than the corresponding set charge threshold, start the current round of charge / discharge balancing; S4. Based on the aforementioned charge / discharge data set, statistically analyze the similarity between individual battery cells; S5. Using the cell with the least / most power as the cluster center, cluster all cell cells according to several similarity thresholds from large to small until all are merged into one class, resulting in several categories C1~Cw, where w represents the number of categories. Category C1 contains the fewest / most individual cells; S6. In response to the condition that all non-Cw-1 class individual cells are greater than / less than or equal to the corresponding set power threshold, cut off the non-Cw-1 class individual cells and continue charging / discharging the remaining individual cells; In response to non-Cw-2 type single cell being charged / discharged to meet set conditions, the corresponding single cell is switched off, and the charging / discharging of the remaining single cells continues; When a non-Cw-3 type single cell is charged / discharged to meet the set conditions, the corresponding single cell is switched off, and the charging / discharging of the remaining single cells continues. This process continues until all individual cells are cut out or the individual cells of the first type C1 are charged / discharged to meet the set conditions. Reconnect all individual cells; Repeat steps S2 to S6 until the series battery pack charging / discharging is completed; The setting condition is that the corresponding single cell is charged / discharged to the state of charge of any single cell other than Cw-1 type. S7. Complete the charging / discharging of the series-connected battery pack.

2. The battery equalization charge and discharge control method based on fuzzy clustering according to claim 1, characterized in that, Step S4 specifically includes: A data matrix of size m*n is constructed based on the charge and discharge data sets of each individual battery, where m represents the number of data sets in the charge and discharge data sets and n represents the number of individual batteries. Each element in the data matrix is ​​denoted as x. ij , i={1,2,…,n},j={1,2,…,m}, x ij This represents the j-th charge / discharge data set of the i-th individual battery cell; Perform standard deviation transformation and normalization on each element; Establishing a fuzzy similarity matrix based on the correlation coefficient method; Step S5 involves clustering based on the similarity matrix.

3. The battery equalization charge and discharge control method based on fuzzy clustering according to claim 1, characterized in that, When the circuit is in a charging state, the single cell with the least charge is the cluster center and is located in all classes; When the circuit is in a discharging state, the single cell with the highest charge is the cluster center and is located in all classes.

4. The battery equalization charge and discharge control method based on fuzzy clustering according to claim 1, characterized in that, When the circuit is in a charging state, the charge / discharge data set for each individual battery includes the charging voltage V. 充 Charging current i 充 The derivative of the charging voltage with respect to time, dv 充 The derivative of the charging current with respect to time, di 充 ; When the circuit is in a discharge state, the charge / discharge data set for each individual battery cell includes the discharge voltage V. 放 Charging current i 放 The derivative of discharge voltage with respect to time, dv 放 The derivative of the discharge current with respect to time, di 放 .

5. The battery equalization charge and discharge control method based on fuzzy clustering according to claim 4, characterized in that, The capacity of a single cell is determined based on its voltage and / or current. Each individual battery cell is connected in parallel with a battery switch. By controlling the opening and closing of the corresponding battery switch, the corresponding individual battery cell can be switched in or out.

6. The battery equalization charge and discharge control method based on fuzzy clustering according to claim 4, characterized in that, In step S4, at least the similarity between the cell with the least / most charge and the remaining cells is calculated. In step S5, the aforementioned similarity thresholds are L1, L2…Lw, where w is the number of clustering, L1 > L2 > … > Lw, and step S5 specifically includes: Select the cell with the least / most capacity as the first cell; Using the first cell as the cluster center and with a similarity threshold of L1, all individual cells are clustered for the first time to obtain the first class. Using the first cell as the cluster center and with a similarity threshold of L2, a second clustering is performed on all individual cells to obtain the second class; This process continues until all individual cells are merged into one category to obtain the wth category.

7. A battery equalization charging and discharging device based on fuzzy clustering, characterized in that, The method described in any one of claims 1-6 achieves charge / discharge balancing of individual cells in a series battery pack, including a control unit and a plurality of battery switches connected in parallel to each individual cell. The control unit controls the closing or opening of each battery switch based on a fuzzy clustering algorithm to cut the corresponding individual cell out of or into the charge / discharge circuit.

Images