A method for formulating an operation and maintenance plan for an energy storage battery insertion box

Abstract

The present invention discloses a method for formulating an operation and maintenance plan for a storage battery cassette, belonging to the field of formulating an operation and maintenance plan for a storage battery cassette; specifically, based on various characteristic data of the storage battery, scatter plots are generated between two data categories, and each generated scatter plot is sampled with replacement. Each scatter plot can be sampled to obtain multiple sub-scatter plots, and an abnormal state vector of the storage battery cassette obtained by sampling is obtained according to the minimum spanning tree generated from the sub-scatter plots. The abnormal state vectors of the storage battery cassette generated from each sub-scatter plot are statistically analyzed, and finally, the abnormal state of each storage battery cassette is judged based on statistical thinking, so as to formulate an operation and maintenance plan according to the grading of the abnormal state of the storage battery cassette; the present invention combines the principle of the minimum spanning tree to achieve a comprehensive and accurate sorting of the overhaul priorities.

Description

Technical Field

[0001] The present invention belongs to the field of formulating operation and maintenance plans for energy storage battery cassettes, and specifically relates to a method for formulating an operation and maintenance plan for an energy storage battery cassette. Background Art

[0002] In recent years, the proportion of new energy power generation in the power grid has gradually increased. Due to its high efficiency, low cost, and environmental friendliness, it has received a large amount of investment. However, new energy also has significant drawbacks: the power generation is volatile. There are many factors that determine the amount of new energy power generation, such as sunny or cloudy weather, temperature, and meteorology, etc. Its randomness and volatility are relatively strong, resulting in large fluctuations in a high-penetration power grid, and it is necessary to suppress such fluctuations to maintain the stable operation of the power grid. Therefore, the proportion of energy storage devices in the power grid has also been continuously increasing.

[0003] Energy storage is a new field. With the increasing proportion of new energy power generation forms, energy storage has developed rapidly in many fields such as transportation and household use, and energy storage technologies have become increasingly mature. The commissioning of more and more energy storage power stations has also helped the power grid obtain better stability and higher economic benefits. Energy storage power stations have gradually become an important device for transferring electrical energy on the time scale that is indispensable in the power grid. Subsequently, the operation and maintenance and overhaul of energy storage power stations have also become gradually important.

[0004] Good operation and maintenance can maintain the operating state of the energy storage battery cassette, so as to achieve the purpose of extending the life of the energy storage battery cassette, and can also effectively reduce the failure rate of the energy storage battery cassette; while poor operation and maintenance effects of the energy storage battery cassette may lead to very serious energy storage power station accidents.

[0005] For the operation and maintenance of energy storage power stations, the formulation of its operation and maintenance plan fundamentally affects the operation and maintenance effect. Since the operating states of energy storage battery cassettes are different, when formulating a maintenance plan at the same time, there is a sequence in terms of importance. However, this sequence cannot be observed intuitively, and it can only be calculated through various characteristic indicators of the energy storage battery cassette to obtain the abnormal state information of the energy storage battery cassette. Summary of the Invention

[0006] Aiming at the lack of a better scientific formulation method for the current operation and maintenance plan of energy storage power stations, the present invention discloses a method for formulating an operation and maintenance plan for an energy storage battery cassette, aiming to use the idea of statistical minimum spanning tree to study the accurate judgment of the abnormal state of the energy storage battery cassette, and provide a reference for the problem of formulating the operation and maintenance plan of the energy storage battery cassette during the maintenance of the energy storage battery cassette.

[0007] The method for formulating the operation and maintenance plan of the energy storage battery cassette is specifically as follows:

[0008] Step 1: For n energy storage battery cassettes, collect the time period T NAmong them, there are m kinds of characteristic data for each energy storage battery chassis;

[0009] The data characteristics include voltage, current, temperature, etc.;

[0010] Step 2: For each kind of characteristic, generate a characteristic data matrix for this kind of characteristic of n energy storage battery chassis, and calculate the eigenvector of the characteristic data matrix. Combine the m kinds of eigenvectors to form an eigenvalue table λ.

[0011] The q-th characteristic data matrix is represented by λ q It means;

[0012] The change rate of each characteristic of each energy storage battery chassis within the time period T N forms the eigenvector corresponding to each characteristic.

[0013] Step 3: In the eigenvalue table λ, traverse and take out two columns of eigenvectors to form a scatter plot, and obtain a total of different scatter plots;

[0014] Step 4: Conduct multiple sampling with replacement for each scatter plot. According to the sub-scatter plots obtained by sampling with replacement, generate the minimum spanning tree of each sub-scatter plot.

[0015] Specifically: For the p-th scatter plot H p , Conduct k times of sampling with replacement for all scatter points. Each time, draw n - 2 scatter points to form a sub-scatter plot. The sub-scatter plot obtained by the x-th sampling is represented by H p,x It means;

[0016] For the sub-scatter plot H p,x , generate the minimum spanning tree T p,x of all scatter points in the graph.

[0017] The minimum spanning tree T p,x adopts a binary tree to connect all scatter points in the graph, while ensuring that the connected path is the shortest and the number of paths is the smallest.

[0018] Repeat to obtain the corresponding to all minimum spanning trees of scatter plots.

[0019] Step 5: For each minimum spanning tree, calculate the average distance of all paths in the minimum spanning tree, and discriminate each path in the minimum spanning tree to obtain the abnormal energy storage battery chassis marked in all minimum spanning trees

[0020] Specifically:

[0021] First, for the minimum spanning tree Tp,x , the tree corresponds to n - 2 nodes, generates n - 3 paths, numbers each path, and represents them with the set L = {L p,x,1 ,..., L p,x,n-3}; and calculates the average distance L p,x,ave of all paths.

[0022] Then, compare the length of each path with the average distance L p,x,ave respectively, and three situations occur:

[0023] Situation 1: The length of the y-th path L p,x,y is less than the average distance. At this time, the two nodes at both ends of the path L p,x,y are regarded as normal nodes, that is, the corresponding two energy storage battery cassettes are considered normal energy storage battery cassettes.

[0024] Situation 2: The length of the y-th path L p,x,y is greater than the average distance. At this time, the two nodes at both ends of the path L p,x,y are regarded as isolated nodes, unless these two nodes are connected to other paths with a length less than the average distance, otherwise the corresponding energy storage battery cassettes will be marked as abnormal energy storage battery cassettes.

[0025] Situation 3: The length of the y-th path L p,x,y is equal to the average distance. At this time, the two nodes at both ends of the path L p,x,y are regarded as normal nodes, that is, the corresponding two energy storage battery cassettes are considered normal energy storage battery cassettes.

[0026] Finally, repeat the above steps for all minimum spanning trees to obtain the abnormal energy storage battery cassettes marked in all minimum spanning trees.

[0027] Step Six: Statistically analyze the abnormal energy storage battery cassette vectors marked by each minimum spanning tree to form the final abnormal judgment result of the energy storage battery cassettes.

[0028] Specifically: For the scatter plot H p , the abnormal energy storage battery cassette vector marked by the minimum spanning tree T p,x of its sub - scatter plot H p,x is J p,x ;

[0029] Arrange all the abnormal energy storage battery cassette vectors J p of this scatter plot H p,1 , J p,2 ,..., J p,x ,..., J p,k in rows to form the judgment matrix J p, where the element 0 in the matrix represents normal and 1 represents abnormal.

[0030] Similarly, for all scatter plots, repeat the above steps respectively, and sum the corresponding elements of the judgment matrices corresponding to all scatter plots to obtain the total judgment matrix J.

[0031] Finally, judge all the elements in the total judgment matrix J:

[0032] 1) Find the maximum value element J Max,1 in the total judgment matrix J. The corresponding energy storage battery cassette is the most abnormal energy storage battery cassette B Max,1 , and add it to the energy storage battery cassette maintenance sequence, that is, R = {B Max,1};

[0033] 2) If the number of elements in the sequence R is less than the preset number G of energy storage battery cassettes to be repaired in this operation and maintenance plan, then remove the maximum value element J Max,1 from the total judgment matrix J, and continue to select the next maximum value element.

[0034] 3) If the number of elements in the sequence R is equal to the preset number G, at this time the energy storage battery cassette maintenance sequence is full, and the maintenance is carried out according to the time sequence of the energy storage battery cassette maintenance sequence, that is, the maintenance plan for the energy storage battery cassette is realized.

[0035] The advantages of the present invention are as follows:

[0036] 1), A method for formulating an operation and maintenance plan for an energy storage battery cassette, aiming at the problem that it is difficult to select the maintenance priority due to a large number of battery cassettes, combines the minimum spanning tree principle, and realizes a comprehensive and accurate maintenance priority ranking.

[0037] 2), A method for formulating an operation and maintenance plan for an energy storage battery cassette, improves the accuracy of the minimum spanning tree result and reduces its contingency by combining statistical thinking, so the maintenance result is more reasonable.

[0038] 3), A method for formulating an operation and maintenance plan for an energy storage battery cassette, comprehensively considers different characteristics of the battery cassette, can detect and discover potential safety hazards of the battery cassette in all directions, and the maintenance sorting result of the cassette is more reliable. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a flowchart of a method for formulating an operation and maintenance plan for an energy storage battery cassette of the present invention;

[0040] Figure 2 is a schematic diagram of the p-th scatter plot H p of the present invention;

[0041] Figure 3 Sub - scatter plot H in the p - th scatter plot of the present invention p,x Schematic diagram;

[0042] Figure 4 For the corresponding sub - scatter plot H of the present invention p,x Minimum spanning tree T p,x Schematic diagram. Detailed implementation mode

[0043] For the convenience of those of ordinary skill in the art to understand and implement the present invention, the following further describes the present invention in detail and in - depth with reference to the accompanying drawings.

[0044] A method for formulating an operation and maintenance plan for an energy - storage battery insertion box of the present invention uses a minimum spanning tree to find abnormal information of the energy - storage battery insertion box, and uses statistical principles to calculate the abnormal state of the energy - storage battery insertion box, thereby providing a judgment basis for formulating an operation and maintenance plan for the energy - storage battery insertion box. The method includes the following steps: S1. Based on the characteristic values calculated from the voltages, currents, temperatures, etc. of multiple energy - storage battery insertion boxes within a certain time period, a characteristic value table of all battery insertion boxes is established. For different types of characteristic values in the characteristic value table, a scatter plot is formed by taking every two non - repeating characteristic values; S2. For each scatter plot, multiple sampling with replacement is performed. Each sampling with replacement can obtain a sub - scatter plot composed of some scatter points in the scatter plot. A minimum spanning tree is generated for each sub - scatter plot, and the average distance of the minimum spanning tree is calculated; S3. Each path in the minimum spanning tree is discriminated based on the average distance of the minimum spanning tree. For a path with a distance less than the average distance, the battery insertion boxes represented by the nodes at both ends are regarded as normal battery insertion boxes. The remaining nodes not connected to the qualified paths are considered as abnormal nodes with a relatively large distance; S4. For each scatter plot, the abnormal battery insertion boxes isolated by all the minimum spanning trees generated by all the samplings with replacement are counted. Then, the abnormal judgment vectors of each battery insertion box under multiple characteristics are obtained by adding the abnormal judgment vectors counted for each scatter plot, and operation and maintenance and early - warning strategies are specified according to different abnormal levels.

[0045] As Figure 1 shown, based on various characteristic data of energy - storage batteries, scatter plots are generated between every two data categories. Multiple samplings with replacement are performed on each generated scatter plot. Each scatter plot can sample multiple sub - scatter plots. The abnormal state vectors of the energy - storage battery insertion boxes obtained by sampling are obtained according to the minimum spanning trees generated from the sub - scatter plots. The abnormal state vectors of the energy - storage battery insertion boxes generated from each sub - scatter plot are statistically analyzed. Finally, the abnormal states of each energy - storage battery insertion box are judged according to statistical thinking, and an operation and maintenance plan is formulated according to the classification of the abnormal states of the energy - storage battery insertion boxes; the specific steps are as follows:

[0046] Step 1. For n energy - storage battery insertion boxes, the time period T is respectively collectedN Among them, there are m kinds of characteristic data for each energy storage battery chassis;

[0047] The data characteristics include voltage, current, temperature, etc.;

[0048] Step 2: For each type of characteristic, generate a characteristic data matrix for this type of characteristic of n energy storage battery chassis, and calculate the eigenvector of this characteristic data matrix. Compose the eigenvectors of m kinds of characteristics into an eigenvalue table λ.

[0049] The q-th characteristic data matrix is represented by λ q ;

[0050] For each characteristic, calculate the change rate of this characteristic of each energy storage battery chassis within the time period T N to form the eigenvector corresponding to this characteristic.

[0051] Step 3: In the eigenvalue table λ, traverse and take out two columns of eigenvectors λ q,ave and λ q+1,ave to form a scatter plot, obtaining a total of different scatter plots;

[0052] Step 4: Conduct multiple sampling with replacement for each scatter plot. According to the sub-scatter plots obtained from sampling with replacement, generate the minimum spanning tree of each sub-scatter plot.

[0053] Each sampling with replacement can obtain a sub-scatter plot composed of some scatter points in the scatter plot;

[0054] Specifically: For the p-th scatter plot H p , using the eigenvector λ q,ave as the horizontal axis and the eigenvector λ q+1,ave as the vertical axis, this scatter plot H p illustrates the subsequent minimum spanning tree operation.

[0055] Conduct k times of sampling with replacement for all scatter points. Each time, n - 2 scatter points are drawn to form a sub-scatter plot. The sub-scatter plot obtained from the x-th sampling is represented by H p,x ; 0 < x ≤ k.

[0056] For the sub-scatter plot H p,x , generate the minimum spanning tree T p,x of all scatter points in the graph.

[0057] The minimum spanning tree T p,x adopts a binary tree to connect all scatter points in the graph, while ensuring that the connected path is the shortest and the number of paths is the smallest.

[0058] Repeat to obtain the corresponding all scatter plots minimum spanning trees

[0059] Step 5: For each minimum spanning tree, calculate the average distance of all paths in the minimum spanning tree, and discriminate each path in the minimum spanning tree to obtain the abnormal energy storage battery insertion boxes marked in all minimum spanning trees

[0060] Specifically:

[0061] First, for the minimum spanning tree T p,x , which corresponds to n - 2 nodes, generate n - 3 paths, number each path, and use the set L = {L p,x,1 ,..., L p,x,n-3} to represent; and calculate the average distance L p,x,ave .

[0062] Then, compare the length of each path with the average distance L p,x,ave respectively, and three situations occur:

[0063] Situation 1: The length of the y-th path L p,x,y is less than the average distance. At this time, the two nodes at both ends of the path L p,x,y are regarded as normal nodes, that is, the corresponding two energy storage battery insertion boxes are considered normal energy storage battery insertion boxes.

[0064] Situation 2: The length of the y-th path L p,x,y is greater than the average distance. At this time, the two nodes at both ends of the path L p,x,y are regarded as isolated nodes, unless these two nodes are connected to other paths with a length less than the average distance, otherwise the corresponding energy storage battery insertion boxes will be marked as abnormal energy storage battery insertion boxes.

[0065] Situation 3: The length of the y-th path L p,x,y is equal to the average distance. At this time, the two nodes at both ends of the path L p,x,y are regarded as normal nodes, that is, the corresponding two energy storage battery insertion boxes are considered normal energy storage battery insertion boxes.

[0066] Finally, repeat the above steps for all minimum spanning trees to obtain the abnormal energy storage battery insertion boxes marked in all minimum spanning trees.

[0067] Step 6: Statistically analyze the vectors of abnormal energy storage battery insertion boxes marked in each minimum spanning tree to form the final abnormal judgment result of the energy storage battery insertion boxes.

[0068] Specifically: For the scatter plot H p , its sub-scatter plot H p,xMinimum spanning tree T p,x The marked abnormal energy storage battery cassette vector is J p,x ; In k sub-scatter plots, each sub-scatter plot has obtained a judgment vector on whether the energy storage battery cassette is abnormal.

[0069] For the scatter plot H p All the abnormal energy storage battery cassette vectors J p,1 , J p,2 ,..., J p,x ,..., J p,k , arranged by rows, to form a judgment matrix J p , where the element 0 in the matrix indicates normal and 1 indicates abnormal.

[0070] Similarly, for all scatter plots, repeat the above steps respectively, and sum the corresponding elements of all the judgment matrices corresponding to the scatter plots to obtain the total judgment matrix J.

[0071] Then, for all scatter plots, repeat the above steps respectively, sum the corresponding elements of all the judgment matrices, and obtain the total judgment matrix J of all the energy storage battery cassettes.

[0072] Finally, judge all the elements in the total judgment matrix J to obtain the final judgment result of the abnormal energy storage battery cassette:

[0073] 1) Find the maximum value element J Max,1 in the total judgment matrix J, and the corresponding energy storage battery cassette is the most abnormal energy storage battery cassette B Max,1 , add it to the energy storage battery cassette maintenance sequence, that is, R = {B Max,1};

[0074] 2) If the number of elements in the sequence R is less than the preset number G of energy storage battery cassettes to be repaired in this operation and maintenance plan, then remove the selected energy storage battery cassettes from the total judgment matrix J and continue to select the next maximum value element.

[0075] 3) If the number of elements in the sequence R is equal to the preset number G, at this time the energy storage battery cassette maintenance sequence is full, and perform maintenance according to the time sequence of the energy storage battery cassette maintenance sequence, that is, the maintenance plan for the energy storage battery cassette is realized.

[0076] After obtaining the total judgment matrix J after summation, according to the given judgment logic and the number of battery cassettes to be repaired in this plan, select the full number of battery cassettes with the largest summation value. The larger the value in the total judgment matrix, the more abnormal the corresponding battery cassette is, and the higher the priority of arranging its maintenance.

[0077] Example:

[0078] Step 1, T N During the time period, for m types of characteristic data such as voltage, current, and temperature of n energy storage battery cassettes, a characteristic data matrix is generated for each type of characteristic of all energy storage battery cassettes;

[0079] The qth characteristic data matrix is denoted by λ q as shown in the formula:

[0080] In the formula, λ q,1,1 represents the first type of characteristic data of the first energy storage battery cassette in the qth characteristic data matrix; λ q,1,TN represents the T N th type of characteristic data of the first energy storage battery cassette in the qth characteristic data matrix; λ q,n,1 represents the first type of characteristic data of the nth energy storage battery cassette in the qth characteristic data matrix; λ q,n,TN represents the T N th type of characteristic data of the nth energy storage battery cassette in the qth characteristic data matrix; T N represents the length of the data acquisition time period of the energy storage battery cassette.

[0081] Step 2, for all m generated characteristic data matrices, calculate the change rate of each characteristic data matrix corresponding to each energy storage battery cassette during its T N time period and form an eigenvalue vector, thereby forming an energy storage battery cassette eigenvalue matrix λ;

[0082] Using the eigenvalue vector λ q of the qth type of characteristic of all energy storage battery cassettes, i.e., the matrix λ q,ave is illustrated by calculation, and its calculation formula is as shown in the formula:

[0083] In the formula, λ q,ave，1 represents the change rate of the qth type of characteristic data of the first energy storage battery cassette; λ q,ave,n represents the change rate of the qth type of characteristic data of the nth energy storage battery cassette; λ q,ave represents λ q matrix eigenvalue vector.

[0084] Use m eigenvalue vectors to form an energy storage battery cassette eigenvalue matrix λ, as shown in the formula:

[0085]

[0086] In the formula, λ 1,ave represents the eigenvalue vector of the λ1 matrix; λ m,ave represents the eigenvalue vector of λ m matrix;

[0087] Step 3: In the eigenvalue matrix λ, take out two columns of eigenvalue vectors λ q,ave and λ q+1,ave to form a scatter plot; the two columns of eigenvalue vectors each time cannot be repeated, so a total of different scatter plots are obtained.

[0088] Step 4: Conduct multiple sampling with replacement for each scatter plot. Each time of sampling with replacement can obtain a sub-scatter plot composed of some scatter points in the scatter plot. Generate the minimum spanning tree for each sub-scatter plot and calculate the average distance of the minimum spanning tree.

[0089] Specifically: For the p-th scatter plot H p , take the eigenvalue vector λ q,ave as the horizontal axis and the eigenvalue vector λ q+1,ave as the vertical axis to illustrate the subsequent minimum spanning tree operation with this scatter plot H p . This scatter plot is as shown in Figure 2 .

[0090] Conduct k times of sampling with replacement for all the scatter points in the scatter plot H p . Each time, draw n - 2 scatter points to form a sub-scatter plot. The sub-scatter plot obtained from the x-th sampling is denoted by H p,x , as shown in Figure 3 .

[0091] For the sub-scatter plot H p,x , generate the minimum spanning tree T p,x of all the scatter points in the graph, connect all the scatter points in the graph, and at the same time ensure that the connected path is the shortest and the number of paths is the smallest, as shown in Figure 4 .

[0092] Repeat the above steps to obtain the minimum spanning trees corresponding to all scatter plots.

[0093] Step 5: Discriminate each path in the minimum spanning tree according to the average distance of the minimum spanning tree. For the paths with a distance less than the average distance, the battery insertion boxes represented by the nodes at both ends are regarded as normal battery insertion boxes. For the remaining nodes not connected to the qualified paths, these nodes are considered as abnormal nodes with a relatively large distance.

[0094] Specifically: Use the minimum spanning tree T p,x in the p-th scatter plot to illustrate this step.

[0095] Number each path in the minimum spanning tree T p,x , and represent it with the set L = {L p,x,1 , …, L p,x,n-3}; calculate Tp,x Average distance of all paths in

[0096] In the formula, i represents the variable inside the summation symbol; L p,x,i represents the i-th path in T p,x ; n represents the total number of scatter points in the p-th scatter plot; L p,x,ave represents the average path distance in T p,x .

[0097] Then, compare the lengths of each path in the minimum spanning tree T p,x with the average distance L of this minimum spanning tree p,x,ave . At this time, three situations will occur.

[0098] Situation 1: The length of the y-th path L p,x,y is less than the average distance. At this time, the two nodes at both ends of L p,x,y are regarded as normal nodes, that is, the corresponding two energy storage battery cassettes are considered normal energy storage battery cassettes.

[0099] Situation 2: The length of the y-th path L p,x,y is greater than the average distance. At this time, the two nodes at both ends of L p,x,y are regarded as isolated nodes. Unless these two nodes are connected to other paths with lengths less than the average distance, otherwise their corresponding energy storage battery cassettes will be marked as abnormal energy storage battery cassettes.

[0100] Situation 3: The length of the y-th path L p,x,y is equal to the average distance. At this time, the two nodes at both ends of L p,x,y are regarded as normal nodes, that is, the corresponding two energy storage battery cassettes are considered normal energy storage battery cassettes.

[0101] Step 3.3 performs Steps 3.1 to 3.3 on all minimum spanning trees of the scatter plots. Therefore, the abnormal energy storage battery cassettes marked in all minimum spanning trees of the scatter plots can be obtained.

[0102] Step 6, Determine whether each energy storage battery cassette is abnormal based on the minimum spanning trees.

[0103] For the p-th scatter plot H p , among the k sub-scatter plots generated by it, each sub-scatter plot has obtained a judgment vector on whether the energy storage battery cassette is abnormal. For all scatter plots of the Statistical analysis of the individual scatter plot judgment vectors can obtain the final abnormal judgment result of the energy storage battery cassette.

[0104] Specifically, for the p-th scatter plot H p , among the k sub-scatter plots obtained by sampling it with replacement, the sub-scatter plot H p,x generated by the x-th sampling p,x is used to judge the abnormal state of the energy storage battery cassette, and a judgment vector J p,x can be obtained:

[0105] In the formula, J p,x,1 represents a 0-1 variable. If the judgment result of T p,x for the first energy storage battery cassette is normal, then J p,x,1 is 0. If the judgment result of T p,x for the first energy storage battery cassette is abnormal, then J p,x,1 is 1; J p,x,n represents a 0-1 variable. If the judgment result of T p,x for the n-th energy storage battery cassette is normal, then J p,x,n is 0. If the judgment result of T p,x for the n-th energy storage battery cassette is abnormal, then J p,x,n is 1; J p,x represents the judgment vector obtained by judging the abnormal state of the energy storage battery cassette by the minimum spanning tree T p,x generated by the sub-scatter plot H p,x obtained by the x-th sampling.

[0106] Then, repeat the above steps for the p-th scatter plot H p to obtain the judgment matrix corresponding to all the sub-scatter plots generated by sampling H p with replacement. Add the corresponding rows of these judgment matrices to obtain the scatter plot judgment matrix J p of the p-th scatter plot H p :

[0107] In the formula, J p,1,1 represents the abnormal judgment made by the first sub-scatter plot in the p-th scatter plot for the first energy storage battery cassette, which is 1 if abnormal and 0 if normal; J p,k,1 represents the abnormal judgment made by the k-th sub-scatter plot in the p-th scatter plot for the first energy storage battery cassette, which is 1 if abnormal and 0 if normal; J p,1,n represents the abnormal judgment made by the first sub-scatter plot in the p-th scatter plot for the n-th energy storage battery cassette, which is 1 if abnormal and 0 if normal; J p,k,nIndicate the abnormal judgment made by the k-th sub-scatter plot in the p-th scatter plot for the n-th energy storage battery chassis. If it is abnormal, it is 1; if it is normal, it is 0;

[0108] J p,1 Indicate the sub-scatter plot H obtained from the first sampling p,1 The minimum spanning tree T generated p,1 The judgment vector obtained by judging the abnormal state of the energy storage battery chassis; J p,k Indicate the sub-scatter plot H obtained from the k-th sampling p,k The minimum spanning tree T generated p,k The judgment vector obtained by judging the abnormal state of the energy storage battery chassis; J p Indicate the scatter plot H of the p-th scatter plot p The scatter plot judgment matrix of

[0109] Finally, for all Repeat the above steps for the scatter plots, and sum the corresponding elements of all scatter plot judgment matrices to obtain the total judgment matrix J of all energy storage battery chassis:

[0110] In the formula, J1 indicates the scatter plot judgment matrix of the first scatter plot H1; Indicate the scatter plot of the scatter plot judgment matrix; Indicate the abnormal judgment made by the first scatter plot H1 for the first energy storage battery chassis; Indicate the scatter plot of the abnormal judgment made for the first energy storage battery chassis; Indicate the abnormal judgment made by the first scatter plot H1 for the n-th energy storage battery chassis; Indicate the scatter plot of the abnormal judgment made for the n-th energy storage battery chassis; J indicates the total judgment matrix of all energy storage battery chassis; J 1 Indicate the total judgment value of all characteristic data for the first energy storage battery chassis; J n Indicate the total judgment value of all characteristic data for the n-th energy storage battery chassis;

[0111] Step 7. Assume that G energy storage battery chassis are to be repaired in this operation and maintenance plan, then judge all elements in J. The judgment follows the following logic:

[0112] 1) Find the maximum value element in J, and use J Max,1 to indicate this element. J Max,1 The corresponding energy storage battery chassis is the most abnormal energy storage battery chassis B Max,1, add it to the maintenance sequence of the energy storage battery cassette, that is, R = {B Max,1}.

[0113] 2) If the number of elements in R is less than G, then remove the selected energy storage battery cassettes from J and repeat the actions in 1).

[0114] 3) If the number of elements in R is equal to G, at this time the maintenance sequence of the energy storage battery cassettes is full, and perform maintenance according to the time sequence of the maintenance sequence of the energy storage battery cassettes, that is, the formulation of the maintenance plan for the energy storage battery cassettes is realized.

Claims

1. A method for formulating an operation and maintenance plan for a storage battery insertion box, characterized in that The specific steps are as follows: First, for n energy storage battery cassettes, collect m types of characteristic data of each energy storage battery cassette respectively within the time period T N ; for each type of characteristic, generate a characteristic data matrix of this type of characteristic for the n energy storage battery cassettes, and calculate the eigenvalue vector of the characteristic data matrix. Compose the m eigenvalue vectors into an eigenvalue table λ; Then, in the eigenvalue table λ, traverse and extract two columns of eigenvalue vectors to form a scatter plot, obtaining a total of different scatter plots; perform multiple sampling with replacement on each scatter plot, and generate the minimum spanning tree of each sub-scatter plot based on the sub-scatter plots obtained from sampling with replacement. Next, for each minimum spanning tree, calculate the average distance of all paths in the minimum spanning tree, and discriminate each path in the minimum spanning tree. For a path with a distance less than the average distance, the battery insertion boxes represented by the nodes at both ends are regarded as normal battery insertion boxes; otherwise, they are regarded as abnormal battery insertion boxes; Specifically: For the minimum spanning tree T p,x For the n - 2 nodes, generate n - 3 paths, number each path, and use the set L = {L p,x,1 ,..., L p,x,n-3} to represent; and calculate the average distance L p,x,ave ; Then, compare the lengths of each path with the average distance L p,x,ave There are three situations: Situation 1: The y-th path L p,x,y has a length less than the average distance. At this time, the two nodes at both ends of the path L p,x,y are regarded as normal nodes, that is, the corresponding two energy storage battery cassettes are considered normal energy storage battery cassettes; Condition 2: The y-th path L p,x,y has a length greater than the average distance. At this time, for the path L p,x,y the two nodes at both ends are regarded as isolated nodes. Unless these two nodes are connected to other paths with a length less than the average distance, the corresponding energy storage battery cassette will be marked as an abnormal energy storage battery cassette; Situation 3: the y-th path L p,x,y has a length equal to the average distance. At this time, for the path L p,x,y the two nodes at both ends are regarded as normal nodes, that is, the corresponding two energy storage battery cassettes are considered to be normal energy storage battery cassettes; Finally, for each scatter plot, count the abnormal battery insertion boxes isolated by all the minimum spanning trees generated by sampling with replacement. Add up the abnormal judgment vectors of the battery insertion boxes under multiple features, so as to obtain the abnormal judgment vectors of each battery insertion box under different features, and specify the operation and maintenance and early warning strategies according to different abnormal levels; Specifically: for the scatter plot H p , its sub-scatter plot H p,x 's minimum spanning tree T p,x marks the abnormal energy storage battery cassette vector as J p,x ; The scatter plot H p of all abnormal energy storage battery cabinet vectors J p,1 , J p,2 ,..., J p,x ,..., J p,k , arranged by rows to form a judgment matrix J p , where the element 0 in the matrix represents normal and 1 represents abnormal; Similarly, repeat the above steps for all scatter plots, add the corresponding elements of the judgment matrices corresponding to all scatter plots to obtain the total judgment matrix J; Judge all elements in the total judgment matrix J: 1) Find the maximum element $J$ in the total judgment matrix $J$ Max,1 , and the corresponding energy storage battery cassette is the most abnormal energy storage battery cassette $B$ Max,1 , and add it to the energy storage battery cassette maintenance sequence, that is, $R = \{B$ Max,1 $\}$; 2) If the number of elements in sequence R is less than the preset number G of energy storage battery cassettes to be repaired in this operation and maintenance plan, then remove the maximum value element J Max,1 from the total judgment matrix J, and continue to select the next maximum value element; 3) If the number of elements in the sequence R is equal to the preset quantity G, the maintenance sequence of the energy storage battery insertion box is full at this time, and maintenance is carried out according to the time sequence of the maintenance sequence of the energy storage battery insertion box, that is, the formulation of the maintenance plan for the energy storage battery insertion box is realized.

2. The method for formulating an operation and maintenance plan for an energy storage battery insertion box according to claim 1, wherein The data features include voltage, current and temperature.

3. The method for formulating an operation and maintenance plan for an energy storage battery insertion box according to claim 1, characterized in that The eigenvalue vector refers to: for each feature, the change rate of this feature within the time period T for each energy storage battery chassis N is used to form the eigenvalue vector corresponding to this feature.

4. A method for formulating an operation and maintenance plan for an energy storage battery plug-in box, as described in claim 1, characterized in that The minimum spanning tree for generating each sub-scatter plot is specifically as follows: for the p-th scatter plot H p , perform k times of sampling with replacement on all scatter points, and each time draw n - 2 scatter points to form a sub-scatter plot, where the sub-scatter plot obtained from the x-th sampling is denoted by H p,x . 0 < x ≤ k; For the sub-scatter plot H p,x , generate the minimum spanning tree T of all the scatter points in the figure p,x ; Minimum spanning tree T p,x Use a binary tree to connect all the scattered points in the connected graph, while ensuring that the connected path is the shortest and the number of paths is the smallest; Repeatedly obtain the corresponding all scatter plots minimum spanning trees

Images