Operation and maintenance method and apparatus for severe overload and three-phase imbalance of distribution transformer, and medium

Abstract

Disclosed in the present invention are an operation and maintenance method and apparatus for severe overload and three-phase imbalance of a distribution transformer, and a medium. The method comprises: by quantifying the degree of difference between devices as the distance between two points, performing clustering analysis on an operation and maintenance state of a distribution transformer, so as to obtain a classification result; and on the basis of a short-term single-phase severe overload category and a short-term three-phase severe overload category in the classification result, establishing a multi-objective optimization model and solving same, so as to obtain a multi-objective optimization result, and on this basis, formulating a day-ahead source-load distribution transformer operation and maintenance plan and performing operation and maintenance control on a distribution network. The present invention provides an operation and maintenance method and apparatus for severe overload and three-phase imbalance of a distribution transformer, and a medium. Solving a multi-objective optimization model to obtain a multi-objective optimization result can ensure that operational risks caused by three-phase imbalance are reduced while balancing three-phase loads, such that differentiated operation and maintenance can be achieved by means of a day-ahead source-load distribution transformer operation and maintenance plan, thereby solving the problem of difficulty in managing severe overload and three-phase imbalance of a distribution transformer to formulate a targeted operation and maintenance plan.

Description

An operation and maintenance method, device and medium for heavy overload and three-phase imbalance of distribution transformers TECHNICAL FIELD

[0001] The present application relates to the technical field of distribution maintenance, in particular to an operation and maintenance method, device and medium for heavy overload and three-phase imbalance of distribution transformers. BACKGROUND

[0002] The distribution network undertakes the important task of connecting the power grid and directly supplying power to users, and the continuous good operation of the distribution transformer is related to the continuous and reliable power supply of the majority of power users and the safe and reliable operation of the power grid. With the large-scale access of distributed power sources and the substantial growth of seasonal loads such as important holidays, agriculture and tourism, the problems of heavy overload, seasonal heavy overload and three-phase imbalance of distribution transformers are extremely prominent. Exploiting the potential of source and load adjustment and control in the transformer area, and arranging targeted maintenance and differentiated operation and maintenance plans based on heavy overload and three-phase imbalance risk assessment, is conducive to improving operation and maintenance efficiency and reasonably allocating operation and maintenance resources. Traditional operation and maintenance methods are to control the operation and maintenance of the distribution network by strengthening monitoring and maintenance, using technical means to improve three-phase imbalance, and introducing advanced technical means, which can improve the stability and reliability of the distribution network to a certain extent.

[0003] However, under the condition of single-phase access of distributed photovoltaic and charging piles, the problems of heavy overload and three-phase imbalance of distribution transformers are becoming increasingly serious, and even reverse heavy overload phenomenon occurs in transformer areas with high photovoltaic penetration rate. The traditional operation and maintenance method cannot adapt to complex situations, resulting in prominent problems of heavy overload and three-phase imbalance of distribution transformers and low operation and maintenance efficiency. SUMMARY

[0004] The present application provides an operation and maintenance method, device and medium for heavy overload and three-phase imbalance of distribution transformers to solve the problem of being difficult to overcome the heavy overload and three-phase imbalance of distribution transformers and develop targeted operation and maintenance plans.

[0005] To solve the above problems, the present application provides an operation and maintenance method for heavy overload and three-phase imbalance of distribution transformers, comprising:

[0006] obtaining differentiated characteristic data related to the distribution transformer from the distribution network;

[0007] According to the differentiated characteristic data, the operation and maintenance state of the distribution transformer is clustered and analyzed by quantifying the equipment difference as the straight-line distance between two points to obtain a classification result;

[0008] According to the short-term single-phase heavy overload category and the short-term three-phase heavy overload category with source and load control potential in the classification result, a multi-objective optimization model including network loss and imbalance is established;

[0009] Solve the multi-objective optimization model to obtain a multi-objective optimization result; formulate a day-ahead source-load-distribution-transformer operation and maintenance plan according to the multi-objective optimization result, and perform operation and maintenance control on the distribution network according to the day-ahead source-load-distribution-transformer operation and maintenance plan.

[0010] The application can quantitatively evaluate the similarity or difference between different devices by quantifying the device difference as the straight-line distance between two points, and can classify devices with similar characteristics into one category, so as to more reasonably allocate and schedule operation and maintenance resources according to the classification result. Moreover, the multi-objective optimization result obtained by solving the multi-objective optimization model can consider multiple optimization objectives, such as network loss and three-phase load imbalance, which helps to realize comprehensive optimization of the distribution network, reduce network loss, improve energy utilization efficiency, balance three-phase load, and reduce device loss and distribution network operation risk caused by three-phase imbalance. Therefore, constructing a day-ahead source-load-distribution-transformer operation and maintenance plan based on the multi-objective optimization result can reduce the probability of occurrence of distribution network faults and hidden dangers by optimizing source-load-distribution-transformer measures, reduce power grid operation risk, and ensure stable operation of the distribution network.

[0011] Compared with the prior art, the application establishes a multi-objective optimization model including network loss and imbalance according to short-term single-phase overload categories and short-term three-phase overload categories with source-load regulation potential, and then obtains a multi-objective optimization result, which can exert the source-load regulation capacity of data, ensure balancing of three-phase load while reducing device loss and distribution network operation risk caused by three-phase imbalance, and make the finally obtained day-ahead source-load-distribution-transformer operation and maintenance plan have strong pertinence and achieve differentiated operation, so as to solve the problem of difficult-to-overcome distribution transformer overload and three-phase imbalance governance and formulate a targeted operation and maintenance plan.

[0012] As a preferred scheme, according to the differentiated characteristic data, the device difference is quantified as the straight-line distance between two points, and the distribution transformer operation and maintenance state is analyzed by clustering to obtain a classification result, specifically:

[0013] A hierarchical analysis model including a plurality of levels is established according to the differentiated characteristic data; wherein the plurality of levels include a target layer, a criterion layer and a scheme layer;

[0014] A distribution transformer hierarchical clustering model is established according to the evaluation result matrix of the plurality of levels;

[0015] According to the distribution transformer hierarchical clustering model, the similarity of devices is represented by the difference in distance, and the Euclidean distance is selected as the similarity measurement method to solve the classification result.

[0016] The preferred scheme can decompose the complex power distribution network problem into several relatively simple levels by establishing a hierarchical analysis model comprising several levels, so as to systematically understand and analyze the operation state of the power distribution network. By establishing a power distribution transformer hierarchical clustering model, it is beneficial to classify the operation state data of the transformer, classify the transformers with similar operation states into a class, and further obtain the classification result.

[0017] As a preferred scheme, a hierarchical analysis model comprising different levels is established according to the differentiated characteristic data, specifically:

[0018] The elements related to decision in the differentiated characteristic data are decomposed into a target layer, a criterion layer and a scheme layer; wherein the target layer represents the equipment characteristics, the criterion layer represents the characteristic factors of the equipment, and the scheme layer represents the related equipment to be analyzed;

[0019] The hierarchical analysis model is established by combining the heavy overload data and the three-phase imbalance data in the differentiated characteristic data, and the target layer, the criterion layer and the scheme layer.

[0020] In the preferred scheme, the target layer provides a clear focus as the purpose of decision, and can clearly analyze the target and direction; the criterion layer represents the characteristic factors of the equipment, which are the key points and decision criteria to be considered to achieve the target, and by decomposing and refining the characteristic factors of the equipment, the influence of different factors on the equipment operation state can be more accurately evaluated; the scheme layer represents the related equipment to be analyzed, and provides specific decision objects, so that these data can be compared and selected in actual situations.

[0021] As a preferred scheme, a power distribution transformer hierarchical clustering model is established according to the evaluation result matrix of the several levels, specifically:

[0022] The evaluation result matrix is obtained by determining the evaluation result of the criterion layer to the target layer and the evaluation result of the scheme layer to the criterion layer;

[0023] The equipment characteristic weight vector set is calculated according to the evaluation result matrix;

[0024] The power distribution transformer hierarchical clustering model is established according to the calculation result of the transformer area characteristic value and the corresponding weight vector in the equipment characteristic weight vector set.

[0025] In the preferred scheme, the evaluation result matrix provides basic data for subsequent clustering analysis, which is used to calculate the similarity between equipment. Through clustering analysis based on these data, the operation state and classification of the equipment can be further identified, which provides strong data support for operation and maintenance management.

[0026] As a preferred solution, according to the power distribution transformer hierarchical clustering model, the similarity of devices is represented by the degree of difference in distance, and the Euclidean distance is selected as the similarity measurement method to obtain the classification result, specifically:

[0027] According to the power distribution transformer hierarchical clustering model, the Euclidean distance is used to measure the closeness of the characteristic vectors of two devices to obtain distance data;

[0028] The two types of parameters with the highest closeness in the distance data are combined into one type to obtain a hierarchical clustering tree diagram;

[0029] According to the average distance between different categories in the hierarchical clustering tree diagram, the hierarchical clustering tree diagram is segmented to obtain the classification result.

[0030] This preferred solution can quantitatively evaluate the similarity or difference between different devices by calculating the Euclidean distance between device characteristic vectors, and can more reasonably allocate and dispatch operation and maintenance resources by grouping devices with similar characteristics into one category through a hierarchical clustering tree diagram. The average distance is used as the basis for power distribution transformer characteristic clustering to obtain the classification result, which can avoid the phenomenon of excessive compression or expansion between classes, as well as non-monotonicity hindrance, while ensuring the flexibility of the clustering model, reducing result deviation, and ensuring the effectiveness and accuracy of the classification result.

[0031] As a preferred solution, according to the short-term single-phase heavy overload category and short-term three-phase heavy overload category with source-load regulation potential in the classification result, a multi-objective optimization model including network loss and imbalance is established, specifically:

[0032] According to the short-term single-phase heavy overload category and short-term three-phase heavy overload category with source-load regulation potential in the classification result, a day-ahead control layer objective function is established, which includes load rate balance degree, network loss and three-phase imbalance degree;

[0033] According to the upper and lower limits of the charging pile rated power and the output power of the distributed power supply, the charging power feasible region and the upper and lower limit constraints of photovoltaic output are established respectively;

[0034] The multi-objective optimization model is composed of the day-ahead control layer objective function, the charging power feasible region and the upper and lower limit constraints of photovoltaic output.

[0035] This preferred solution ensures that the working state of the charging pile is within the allowed power range by establishing a charging power feasible region, which not only avoids device damage caused by overload, but also ensures the stability and efficiency of the charging process. At the same time, considering the upper and lower limits of the output power of the distributed power supply to establish the upper and lower limit constraints of photovoltaic output can maximize the use of renewable energy, reduce unnecessary energy waste, and improve the overall operation efficiency of the system.

[0036] As a preferred solution, according to the short-term single-phase heavy overload category and the short-term three-phase heavy overload category with source-load regulation potential in the classification result, a day-ahead control layer objective function is established, which includes load rate balance, network loss and three-phase imbalance, and specifically is:

[0037] According to the maximum and minimum values of the load rate of the distribution transformer, a first objective function is established;

[0038] According to the total loss of the transformer and the network loss, a second objective function is established;

[0039] According to the three-phase imbalance of the node, a third objective function is established;

[0040] The first objective function, the second objective function and the third objective function constitute the day-ahead control layer objective function.

[0041] The first objective function of the preferred solution is established according to the maximum and minimum values of the load rate of the distribution transformer, which can ensure that the load of the transformer operates within a reasonable range, and improve the operation efficiency and reliability of the distribution transformer. The second objective function considers the total loss of the transformer and the network loss, which can reduce the power loss by optimizing the operation strategy, not only helps to save power resources, but also reduces environmental pollution caused by power loss; the third objective function is established according to the three-phase imbalance of the node, which can improve the power quality and stability of the system, and ensure that the equipment operates in the best state. By comprehensively considering multiple factors such as load rate balance, loss and three-phase imbalance, a day-ahead control layer objective function is developed, which helps to develop a comprehensive day-ahead source-load distribution transformer operation and maintenance plan.

[0042] As a preferred solution, the multi-objective optimization model is solved to obtain a multi-objective optimization result, and specifically:

[0043] According to the multi-objective optimization model, the optimization target of the particle is set, so that the particle performs optimization according to the preset speed within the executable space range, and the multi-objective optimization result is obtained;

[0044] The optimization target is established according to a logical function.

[0045] In the particle swarm algorithm of the preferred solution, each particle represents a possible solution, and by setting the position and speed of the particle, the algorithm can control the search direction and step length, thereby improving the search efficiency; at the same time, the introduction of the preset scale helps to avoid excessive randomness of the particle in the search process, improves the search accuracy and stability, and greatly shortens the time to obtain the multi-objective optimization result.

[0046] As a preferred solution, the preset speed is specifically:

[0047] The position of the particle in each dimension is set as a preset scale, and the particle parameters are calculated according to the maximum and minimum values of the factors learned in the dimension space according to the binary particle swarm optimization algorithm;

[0048] The preset speed is calculated by combining the particle parameters with the historical optimal position and the current position of the particle individual.

[0049] The application also provides an operation and maintenance device for distribution transformer overload and three-phase imbalance, comprising a data module, a classification module, a model module and an operation and maintenance module.

[0050] The data module is configured to obtain differentiated characteristic data related to the distribution transformer from the distribution network.

[0051] The classification module is configured to cluster analyze the distribution transformer operation and maintenance state by quantifying the equipment differentiation as the straight-line distance between two points according to the differentiated characteristic data, and obtain a classification result.

[0052] The model module is configured to establish a multi-objective optimization model including network loss and imbalance according to the short-term single-phase overload category and the short-term three-phase overload category with source-load regulation potential in the classification result.

[0053] The operation and maintenance module is configured to solve the multi-objective optimization model to obtain a multi-objective optimization result, formulate a day-ahead source-load distribution transformer operation and maintenance plan for overload and three-phase imbalance according to the multi-objective optimization result, and perform operation and maintenance control on the distribution network according to the day-ahead source-load distribution transformer operation and maintenance plan.

[0054] As a preferred scheme, the classification module comprises a hierarchical unit, a matrix unit and a similarity unit.

[0055] The hierarchical unit is configured to establish a hierarchical analysis model comprising several levels according to the differentiated characteristic data, wherein the several levels comprise a target layer, a criterion layer and a scheme layer.

[0056] The matrix unit is configured to establish a distribution transformer hierarchical clustering model according to the evaluation result matrix of the several levels.

[0057] The similarity unit is configured to obtain the classification result by representing the similarity of the equipment by the difference of the distance according to the distribution transformer hierarchical clustering model, and selecting the Euclidean distance as the similarity measurement method.

[0058] As a preferred scheme, the hierarchical unit comprises a hierarchical subunit and a model subunit.

[0059] The hierarchical sub-unit is configured to decompose elements related to decision-making in the differentiated characteristic data into a target layer, a criterion layer and a scheme layer; the target layer represents equipment characteristics, the criterion layer represents characteristic factors of the equipment, and the scheme layer represents related equipment that needs to be analyzed.

[0060] The model sub-unit is configured to establish the hierarchical analysis model by combining the heavy overload data and the three-phase imbalance data in the differentiated characteristic data, and the target layer, the criterion layer and the scheme layer.

[0061] As a preferred solution, the matrix unit includes an evaluation sub-unit, a weight sub-unit and a hierarchical sub-unit.

[0062] The evaluation sub-unit is configured to obtain the evaluation result matrix by determining evaluation results of the criterion layer on the target layer and evaluation results of the scheme layer on the criterion layer.

[0063] The weight sub-unit is configured to calculate a set of equipment characteristic weight vectors according to the evaluation result matrix.

[0064] The hierarchical sub-unit is configured to establish the hierarchical clustering model of the distribution transformer according to a calculation result of a variable substation characteristic value and a corresponding weight vector in the set of equipment characteristic weight vectors.

[0065] As a preferred solution, the similarity unit includes a distance sub-unit, a synthesis sub-unit and a segmentation sub-unit.

[0066] The distance sub-unit is configured to use the Euclidean distance to measure the closeness of characteristic vectors in two equipment according to the hierarchical clustering model of the distribution transformer, to obtain distance data.

[0067] The synthesis sub-unit is configured to synthesize two types of parameters with the highest closeness in the distance data into one type, to obtain a hierarchical clustering tree diagram.

[0068] The segmentation sub-unit is configured to segment the hierarchical clustering tree diagram according to average distances between different categories in the hierarchical clustering tree diagram, to obtain the classification result.

[0069] As a preferred solution, the model module includes a short-term unit, a constraint unit and a composition unit.

[0070] The short-term unit is configured to establish a day-ahead control layer objective function including load rate balance, network loss and three-phase imbalance degree according to short-term single-phase heavy overload categories and short-term three-phase heavy overload categories with source-load regulation potential in the classification result.

[0071] The constraint unit is configured to establish a charging power feasible region and photovoltaic output upper and lower limit constraint according to the charging pile rated power and the upper and lower limits of the output power of the distributed power supply.

[0072] The constituting unit is configured to constitute the multi-objective optimization model from the day-ahead control layer target function, the charging power feasible region and the photovoltaic output upper and lower limit constraint.

[0073] As a preferred scheme, the short-term unit comprises a first subunit, a second subunit, a third subunit and a fourth subunit.

[0074] The first subunit is configured to establish a first target function according to the extreme value of the distribution transformer load rate.

[0075] The second subunit is configured to establish a second target function according to the total transformer loss and the network loss.

[0076] The third subunit is configured to establish a third target function according to the three-phase imbalance degree of the node.

[0077] The fourth subunit is configured to constitute the day-ahead control layer target function from the first target function, the second target function and the third target function.

[0078] As a preferred scheme, the operation and maintenance module comprises an optimization unit.

[0079] The optimization unit is configured to obtain the multi-objective optimization result by setting a particle optimization target, so that the particle performs optimization according to a preset speed within an executable space range, according to the multi-objective optimization model.

[0080] The optimization target is established according to a logical function.

[0081] As a preferred scheme, the preset speed is specifically:

[0082] The position of the particle in each dimension is set as a preset scale, and the particle parameter is calculated according to the maximum value and the minimum value of the learning factor in the dimension space according to the binary particle swarm optimization algorithm.

[0083] The preset speed is calculated by combining the particle parameter, the historical optimal position and the current position of the particle individual.

[0084] The application further provides a storage medium, wherein the storage medium stores a computer program, the computer program is called and executed by a computer, and a method for operation and maintenance of distribution transformer overload and three-phase imbalance is realized. BRIEF DESCRIPTION OF DRAWINGS

[0085] Fig. 1 is a flowchart of a method for operation and maintenance of a distribution transformer under heavy overload and three-phase imbalance according to an embodiment of the present application;

[0086] Fig. 2 is a structural diagram of an operation and maintenance device for a distribution transformer under heavy overload and three-phase imbalance according to an embodiment of the present application. DETAILED DESCRIPTION

[0087] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of the present application.

[0088] In the description of the present application, it should be understood that the terms "first", "second", "third" and "fourth" are used only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second", "third" and "fourth" can be explicitly or implicitly included one or more of the features. In the description of the present application, unless otherwise specified, the meaning of "several" is two or more.

[0089] It should be noted that in the embodiments of the present application, DG is the abbreviation of Distributed Generation, which refers to small-scale power generation devices distributed near the load and connected to the distribution network.

[0090] The operation and maintenance method for a distribution transformer under heavy overload and three-phase imbalance provided by the embodiments of the present application is mainly applied to overcome the limitation that the traditional operation and maintenance method cannot develop differentiated operation and maintenance plans for distribution transformers under different conditions. By governing the heavy overload and three-phase imbalance of the distribution transformer, an effective day-ahead source-load-distribution transformer operation and maintenance plan is developed to better match the actual equipment working under different working conditions and improve the operation and maintenance efficiency of the distribution transformer.

[0091] Embodiment one:

[0092] Referring to Fig. 1, the embodiments of the present application provide an operation and maintenance method for a distribution transformer under heavy overload and three-phase imbalance, which includes S1-S5, and the specific implementation steps are as follows:

[0093] S1, obtaining differentiated characteristic data related to the distribution transformer from the distribution network.

[0094] The step S1 of the embodiments of the present application is specifically:

[0095] The differentiated characteristic data such as original data, operation data and maintenance data related to the distribution transformer are obtained from the distribution network.

[0096] The differentiated characteristic data specifically includes distributed photovoltaic in the transformer area, phase sequence and capacity of the charging pile access, photovoltaic charging pile penetration rate in the transformer area, and three-phase distribution imbalance rate, node current overload times, overload time, and maximum overload rate in the transformer area, three-phase imbalance times, three-phase imbalance time, and maximum imbalance rate in the transformer area.

[0097] S2, according to the differentiated characteristic data, the operation and maintenance state of the distribution transformer is analyzed by quantifying the equipment differentiation as a straight line distance between two points, and a classification result is obtained.

[0098] In step S2 of the embodiment of the present application, S2 includes S2.1-S2.3, wherein S2.1 is a process of constructing an analytic hierarchy model, S2.2 is a process of constructing a hierarchical clustering model of the distribution transformer, and S2.3 is a process of obtaining a classification result, specifically:

[0099] S2.1, the elements related to decision in the differentiated characteristic data are decomposed into target layer, criterion layer and scheme layer using analytic hierarchy, and qualitative and quantitative analysis is performed on the data, and the weights of each layer are determined by comparing the related problems in the system two by two; wherein the target layer represents the equipment characteristics, the criterion layer represents the characteristic factors of the equipment, and the scheme layer represents the related equipment to be analyzed;

[0100] The priority of the data is analyzed to obtain a priority analysis result; specifically, when the load rate is high, three-phase imbalance may cause the hot spot temperature to exceed the limit, so the overload data has higher priority than the three-phase imbalance, and the time has higher priority than the number of times;

[0101] According to the priority analysis result, the operation and maintenance information feedback of the distribution transformer, and the actual situation of the equipment, the overload data (number of times, overload time, and maximum overload rate), three-phase imbalance data (three-phase imbalance times, three-phase imbalance time, and maximum imbalance rate), and eight kinds of equipment characteristics of distributed photovoltaic and charging pile access (photovoltaic charging pile penetration rate in the transformer area and three-phase distribution imbalance rate) are obtained.

[0102] According to the target layer, the criterion layer, the scheme layer and the weights thereof, and the eight kinds of equipment characteristics, an analytic hierarchy model is constructed; and the constructed analytic hierarchy model is a three-layer single-target model.

[0103] In the embodiment S2.1, the target layer serves as the purpose of decision, provides a clear focus, and can clearly analyze the target and direction; the criterion layer represents the characteristic factors of the equipment, and the factors are the key points and decision criteria to be considered to achieve the target; by decomposing and refining the characteristic factors of the equipment, the influence of different factors on the running state of the equipment can be more accurately evaluated; the scheme layer represents the related equipment to be analyzed, and provides the specific decision object, so that the data can be compared and selected in the actual situation.

[0104] S2.2, the eight equipment characteristics of the equipment are compared with each other according to the scale table, and the evaluation results of the criterion layer to the target layer and the evaluation results of the scheme layer to the criterion layer are determined in sequence to obtain an evaluation result matrix A; wherein A=(a ij )(i,j=1,2,3,4,5);

[0105] The geometric mean method is used to calculate the equipment characteristic weight vector set w i according to the evaluation result matrix A.

[0106] The characteristic value of each characteristic of the distribution transformer area is multiplied by the weight vector corresponding to each characteristic in the equipment characteristic weight vector set w i to obtain an equipment characteristic vector.

[0107] A hierarchical clustering model of the distribution transformer is established according to the equipment characteristic vector.

[0108] The equipment characteristic weight vector set is as follows:

[0109] Wherein a ij is a parameter of the evaluation result matrix.

[0110] For application of the embodiment, please refer to Table 1, which is a scale table provided by the embodiment of the application, and represents the importance of different influence factors in the equipment characteristics corresponding to the scale.

[0111] Table 1 Scale table

[0112] In the embodiment S2.2, the evaluation result matrix provides basic data for subsequent clustering analysis, which is used to calculate the similarity between the equipment, and through the clustering analysis based on the data, the running state and classification of the equipment can be further identified, thereby providing strong data support for operation and maintenance management.

[0113] S2.3, according to the hierarchical clustering model of the distribution transformer, the Euclidean distance is used to measure the closeness of the characteristic vectors in two equipment by using the distance to represent the difference between objects to indirectly reflect the similarity between sample objects, and distance data is obtained.

[0114] The two classes with the highest closeness in the distance data are combined into one class to obtain a hierarchical clustering tree diagram;

[0115] The hierarchical clustering tree diagram is segmented according to the average distance s between different classes in the hierarchical clustering tree diagram to obtain a classification result;

[0116] wherein, the characteristic vectors of two devices are x i ={x i1 ,x i2 ,...,x i5}, x j ={x j1 ,x j2 ,...,x j5}, and the Euclidean distance between the two is:

[0117] The segmentation process is specifically as follows:

[0118] First, the average distance s is used as the basis for clustering the characteristics of distribution transformers; wherein, the average distance s is expressed as: let class D r be merged from classes D p and D q , then the distance between class D r and another class D s can be represented by the average of the distance from D p to D s and the distance from D q to D s , that is, the average distance s can be obtained;

[0119] Then, the obtained hierarchical clustering tree diagram is observed, and a suitable average distance s is selected to segment to obtain the final class;

[0120] wherein, the average distance is:

[0121] wherein, D ps is the distance from D p to D s , and D qs is the distance from D qs to D s .

[0122] The embodiment S2.3 can quantitatively evaluate the similarity or difference between different devices by calculating the Euclidean distance between the device characteristic vectors, and can more reasonably allocate and schedule operation and maintenance resources by classifying devices with similar characteristics into a class through a hierarchical clustering tree. The average distance is used as the basis for clustering the characteristics of distribution transformers to obtain the classification result, which can avoid the phenomenon of excessive compression or expansion and non-monotonicity hindrance between classes, while ensuring the flexibility of the clustering model, reducing the deviation of the result, and ensuring the effectiveness and accuracy of the classification result.

[0123] Overall, the embodiment S2 can decompose the complex distribution network problem into several relatively simple levels by establishing a hierarchical analysis model containing several levels, so as to systematically understand and analyze the operation state of the distribution network. By establishing a distribution transformer hierarchical clustering model, it is beneficial to classify the operation state data of the transformer, classify the transformers with similar operation states into a class, and then obtain the classification result.

[0124] S3, according to the short-term single-phase heavy overload category and the short-term three-phase heavy overload category with source and load regulation potential in the classification result, a multi-objective optimization model including network loss and imbalance is established.

[0125] In the embodiment S3 of the present application, S3 includes S3.1-S3.3, wherein S3.1 is a process of constructing a day-ahead control layer objective function, S3.2 is a process of constructing a charging power feasible region and photovoltaic output upper and lower limit constraint, and S3.3 is a process of constructing a multi-objective optimization model, specifically:

[0126] S3.1, according to the short-term single-phase heavy overload category and the short-term three-phase heavy overload category with source and load regulation potential in the classification result, a day-ahead control layer objective function including load rate balance degree, network loss and three-phase imbalance degree is established; specifically:

[0127] According to the maximum and minimum values of the load rate of the distribution transformer, a first objective function F1 is established;

[0128] According to the total loss of the transformer and the network loss, a second objective function F2 is established;

[0129] According to the three-phase imbalance degree of the node, a third objective function F3 is established;

[0130] The first objective function F1, the second objective function F2 and the third objective function F3 constitute the day-ahead control layer objective function;

[0131] The first objective function is:

[0132] The second objective function is: F2=P loss.tot +P loss.net

[0133] The third objective function is:

[0134] The day-ahead control layer objective function is:

[0135] For the parameter VUF i There are:

[0136] Wherein, F is the objective function value; F1', F2' and F3' are the normalized values of objective functions F2, F2 and F2, which eliminate the influence of different dimensions on the optimization result; s2 and s3 are scale factors, p max is the maximum value of the distribution transformer load rate, p min is the minimum value of the distribution transformer load rate, P loss.tot is the total transformer loss, P loss.net is the network loss, VUF i is the three-phase imbalance degree of node i, is the set of nodes in the distribution network, Φ is the three-phase set, U neg,i and U pos,i are the negative sequence and positive sequence voltages of node i, U a,i , U b,i and U c,i are the a, b and c phase voltages of node i,

[0137] The first objective function of the present embodiment S3.1 is established according to the maximum and minimum values of the distribution transformer load rate, which can ensure that the load of the transformer runs within a reasonable range, improves the operation efficiency and reliability of the distribution transformer. The second objective function considers the total transformer loss and network loss, which can reduce the power loss through optimization of the operation strategy, not only helps to save power resources, but also reduces environmental pollution caused by power loss; the third objective function is established according to the three-phase imbalance degree of the node, which can improve the power quality and stability of the system, and ensure that the equipment runs in the best state. By comprehensively considering multiple factors such as load rate balance degree, loss and three-phase imbalance degree, the day-ahead control layer objective function is formulated, which is helpful to formulate a comprehensive day-ahead source and load distribution transformer operation and maintenance plan.

[0138] S3.2, without considering the satisfaction loss, according to the rated power of the charging pile, the charging power feasible region is established;

[0139] According to the upper and lower limits of the output power of the distributed power supply, the upper and lower limits of the photovoltaic output are established;

[0140] The charging power feasible region is:

[0141] The upper and lower limits of the photovoltaic output are:

[0142] wherein, is the rated power of the charging pile; t i,a is the start charging time; t i,d is the departure time; E i,targ is the target electric quantity; and are the upper and lower limits of the output power of the i-th DG, respectively, is the output power of the i-th DG at time t, P i,t is the power of the charging pile at time t, U i,t is the operating state of the i-th DG at time t.

[0143] The embodiment S3.2 ensures that the working state of the charging pile is within the allowable power range by establishing a charging power feasible region, thereby avoiding damage to equipment caused by overload and ensuring the stability and efficiency of the charging process; at the same time, the upper and lower limits of photovoltaic output are established by considering the upper and lower limits of the output power of the distributed power source, so that renewable energy can be maximized, unnecessary energy waste can be reduced, and the overall operation efficiency of the system can be improved.

[0144] S3.3, a multi-objective optimization model is formed by the day-ahead control layer target function, the charging power feasible region and the upper and lower limits of photovoltaic output.

[0145] S4, a multi-objective optimization model is solved to obtain a multi-objective optimization result; a day-ahead source load distribution transformer operation and maintenance plan is formulated according to the multi-objective optimization result, and the power distribution network is operated and maintained according to the day-ahead source load distribution transformer operation and maintenance plan.

[0146] In the embodiment step S4 of the present application, S4 includes S4.1-S4.4, wherein S4.1 is a process of calculating a preset speed, S4.2 is a process of obtaining a multi-objective optimization result, and S4.3 is a process of formulating a day-ahead source load distribution transformer operation and maintenance plan and performing operation and maintenance control, specifically:

[0147] S4.1, respectively, set the position of the particle in each dimension to 0 and 1 two scales, and calculate the particle parameter ω according to the maximum and minimum values of the learning factor in the dimension space according to the binary particle swarm optimization algorithm; wherein the particle parameter ω is obtained by setting through a dynamic adjustment strategy;

[0148] The preset speed v is calculated by combining the particle parameter ω with the historical optimal position and the current position of the particle individual;

[0149] wherein the particle parameter is:

[0150] The preset speed is: v=ωv i+c i k i (x-x i )

[0151] wherein c imax and c imin are the maximum and minimum values of the learning factor of the binary particle swarm optimization algorithm in the dimension i space; t max is the maximum time consumption of the executable space range search at the v speed;

[0152] v i is the current speed of the particle i, c i is the learning factor, k i is a random number between 0 and 1, x is the individual historical optimal position of the particle, and x i is the current position of the particle i.

[0153] S4.2, according to the multi-objective optimization model, by setting the optimization target of the particle, the particle is optimized in the executable space range according to the preset speed v to obtain a multi-objective optimization result;

[0154] wherein the optimization target is established according to a sigmoid function;

[0155] wherein the formula of the sigmoid function is:

[0156] wherein x is an input value.

[0157] In the embodiment S4.2, the sigmoid function can map any real number value to between 0 and 1, and this characteristic enables the sigmoid function to limit the range of the output value, so that the optimization target of the particle is set by using the sigmoid function, which can avoid the situation of infinite growth or excessive saturation of the value.

[0158] S4.3, based on the next day photovoltaic output in the multi-objective optimization result, the power of the charging station is set to form a day-ahead source-load regulation strategy of this category, realize the multi-objective optimization setting of the distribution transformer operation mode, thereby formulating a day-ahead source-load distribution transformer operation and maintenance plan for heavy overload and three-phase imbalance, and performing operation and maintenance control on the distribution network according to the day-ahead source-load distribution transformer operation and maintenance plan;

[0159] wherein the category refers to the "short-term single-phase heavy overload category and short-term three-phase heavy overload category" with source-load regulation potential.

[0160] In addition, for the long-term single-phase heavy overload category and the long-term three-phase heavy overload category, a maintenance plan is formulated according to the maintenance specification and the distribution transformer operation regulation, which is as follows:

[0161] 1) Long-term single-phase heavy overload: Add power electronic devices such as capacitors to compensate the circuit, improve the power factor, and reduce overload situations;

[0162] 2) Long-term three-phase overload: The power distribution capacity can be increased according to economic conditions to avoid overload. At the same time, a transformer substation switching system can be installed according to the power supply conditions of the surrounding transformer substations.

[0163] Overall, in the particle swarm optimization algorithm of this embodiment S4, each particle represents a possible solution. By setting the position and velocity of the particles, the algorithm can control the direction and step size of the search, thereby improving the search efficiency. At the same time, the introduction of a preset scale helps to avoid excessive randomness of particles in the search process, improves the accuracy and stability of the search, and greatly shortens the time to obtain multi-objective optimization results.

[0164] Overall, this embodiment has the following beneficial effects:

[0165] This invention quantifies equipment differences as the straight-line distance between two points and performs cluster analysis on the operation and maintenance status of distribution transformers. This allows for a quantitative assessment of the similarities or differences between different devices, grouping devices with similar characteristics into one category. This enables more rational allocation and scheduling of operation and maintenance resources based on the classification results. Furthermore, the multi-objective optimization results obtained by solving the multi-objective optimization model can simultaneously consider multiple optimization objectives, such as network losses and three-phase load imbalance. This facilitates comprehensive optimization of the distribution network, reducing network losses, improving energy efficiency, and balancing three-phase loads to reduce equipment losses and operational risks caused by three-phase imbalance. Therefore, based on the multi-objective optimization results, a day-ahead source-load distribution transformer operation and maintenance plan can be constructed for distribution network operation and maintenance control. By optimizing source-load distribution transformer measures, the probability of distribution network faults and hidden dangers can be reduced, operational risks can be mitigated, and the stable operation of the distribution network can be ensured.

[0166] Furthermore, by utilizing the differentiated characteristic data such as original data, operation data, and maintenance data of distribution transformers, the study focused on mining the characteristics related to heavy overload and three-phase imbalance. The hierarchical analysis clustering method was used to extract distribution transformer categories with source-load regulation potential, which can accurately achieve differentiated operation and maintenance and better match the actual equipment working under different conditions. In addition, with load rate balance, network loss, and three-phase imbalance as multiple objectives, the binary particle swarm optimization algorithm was used to optimize the objective function, emphasizing the accuracy of the regulation scheme.

[0167] Example 2:

[0168] Please refer to Figure 2. An embodiment of the present invention provides an operation and maintenance device for distribution transformer overload and three-phase imbalance, including a data module 10, a classification module 20, a model module 30 and an operation and maintenance module 40.

[0169] The data module 10 is configured to obtain operation data related to the distribution transformer from the power distribution network.

[0170] The classification module 20 is configured to perform cluster analysis on the operation and maintenance state of the distribution transformer by quantifying the equipment difference as a straight-line distance between two points according to the operation data, and obtain a classification result.

[0171] The model module 30 is configured to establish a multi-objective optimization model including network loss and imbalance according to the short-term single-phase heavy overload category and the short-term three-phase heavy overload category with source-load regulation potential in the classification result.

[0172] The operation and maintenance module 40 is configured to solve the multi-objective optimization model to obtain a multi-objective optimization result, formulate a day-ahead source-load distribution transformer operation and maintenance plan for heavy overload and three-phase imbalance according to the multi-objective optimization result, and perform operation and maintenance control on the power distribution network according to the day-ahead source-load distribution transformer operation and maintenance plan.

[0173] In one embodiment, the data module 10 is specifically configured as follows:

[0174] The data module 10 is configured to obtain differentiated characteristic data such as raw data, operation data, and maintenance data related to the distribution transformer from the power distribution network.

[0175] The differentiated characteristic data specifically includes distributed photovoltaic in the transformer area, charging pile access phase sequence and capacity, transformer area photovoltaic charging pile penetration rate and three-phase distribution imbalance rate, transformer area node current heavy overload times, overload time, maximum overload rate, transformer area node current three-phase imbalance times, three-phase imbalance time, and maximum imbalance rate.

[0176] In one embodiment, the classification module 20 includes a hierarchy subunit, a model subunit, an evaluation subunit, a weight subunit, a hierarchy subunit, a distance subunit, a synthesis subunit, and a segmentation subunit. The hierarchy subunit and the model subunit are processes for constructing an analytic hierarchy model, the evaluation subunit, the weight subunit, and the hierarchy subunit are processes for constructing a distribution transformer hierarchical clustering model, and the distance subunit, the synthesis subunit, and the segmentation subunit are processes for obtaining a classification result, which are specifically configured as follows:

[0177] The hierarchy subunit is configured to use analytic hierarchy to decompose elements related to decision-making in the differentiated characteristic data into target layer, criterion layer, and scheme layer, and perform qualitative and quantitative analysis on the data, and determine the weight of each layer by comparing the relevant problems in the system two by two. The target layer represents equipment characteristics, the criterion layer represents equipment characteristic factors, and the scheme layer represents relevant equipment that needs to be analyzed.

[0178] The model subunit is configured to perform priority analysis on the data to obtain a priority analysis result. Specifically, when the load rate is high, three-phase imbalance can cause hotspot temperature to exceed the limit, and therefore, the priority of overload data is higher than that of three-phase imbalance, and the priority of time is higher than that of frequency.

[0179] The model subunit is further configured to obtain eight device characteristics, including overload data (frequency, overload time, and maximum overload rate), three-phase imbalance data (three-phase imbalance frequency, three-phase imbalance time, and maximum imbalance rate), and distributed photovoltaic and charging pile access (photovoltaic charging pile penetration rate and three-phase distribution imbalance rate), according to the priority analysis result, distribution transformer operation and maintenance information feedback, and actual situation of the device.

[0180] The model subunit is further configured to construct an analytic hierarchy process model according to the target layer, the criterion layer, the scheme layer, and the weights thereof, and the eight device characteristics. The constructed analytic hierarchy process model is a three-layer single-target model.

[0181] In the hierarchy subunit and the model subunit of the embodiment, the target layer serves as the purpose of decision-making, provides a clear focus, and can clearly analyze the target and direction. The criterion layer represents the characteristic factors of the device, and these factors are the key points and decision criteria to be considered to achieve the target. By decomposing and refining the characteristic factors of the device, the influence of different factors on the running state of the device can be more accurately evaluated. The scheme layer represents the related devices that need to be analyzed, and provides specific decision objects, so that these data can be compared and selected in actual situations.

[0182] The evaluation subunit is configured to compare the eight device characteristics of the device with each other according to a scale table, sequentially determine the evaluation results of the criterion layer on the target layer and the evaluation results of the scheme layer on the criterion layer, and obtain an evaluation result matrix A. A=(a ij )(i,j=1,2,3,4,5).

[0183] The weight subunit is configured to calculate a device characteristic weight vector set w i according to the evaluation result matrix A by using a geometric mean method.

[0184] The hierarchy subunit is configured to multiply each characteristic value of the distribution transformer area with the weight vector corresponding to each characteristic in the device characteristic weight vector set w i to obtain a device characteristic vector.

[0185] The hierarchy subunit is further configured to establish a distribution transformer hierarchy clustering model according to the device characteristic vector.

[0186] The device characteristic weight vector set is as follows:

[0187] The device characteristic weight vector set is as follows: ijTo evaluate the parameter of the result matrix.

[0188] For the application of the embodiment of the present application, please refer to Table 1, which is a scale table provided by the embodiment of the present application, indicating the importance of different influencing factors in the device characteristics corresponding to the scale.

[0189] Table 1 Scale table

[0190] In the evaluation subunit, weight subunit and hierarchy subunit of the embodiment, the evaluation result matrix provides basic data for subsequent cluster analysis, which is used to calculate the similarity between devices. Through cluster analysis based on these data, the running state and classification of the devices can be further identified, which provides strong data support for operation and maintenance management.

[0191] The distance subunit is used to measure the closeness of the characteristic vectors of two devices by using Euclidean distance according to the power transformer hierarchical clustering model, so as to obtain distance data;

[0192] The synthesis subunit is used to synthesize the two types of parameters with the highest closeness in the distance data into one type, so as to obtain a hierarchical clustering tree diagram;

[0193] The segmentation subunit is used to segment the hierarchical clustering tree diagram according to the average distance s between different categories in the hierarchical clustering tree diagram, so as to obtain a classification result;

[0194] Wherein, the characteristic vectors of two devices are x i ={x i1 ,x i2 ,...,x i5}, x j ={x j1 ,x j2 ,...,x j5}, and the Euclidean distance between them is:

[0195] The segmentation process is as follows:

[0196] First, the average distance s is used as the basis for power transformer characteristic clustering; wherein, the average distance s is represented as: let D r be merged from D p and D q , then the distance between D r and another class D s can be represented as the distance from D p to D s and the distance from D q to D sThe average distance s is obtained by taking the average of the distances of the two points.

[0197] Then, the obtained hierarchical clustering tree diagram is observed, and a suitable average distance s is selected to obtain the final category.

[0198] The average distance is:

[0199] D ps is the distance between D p and D s , and D qs is the distance between D qs and D s .

[0200] The distance subunit, the synthesis subunit and the segmentation subunit of the embodiment can quantitatively evaluate the similarity or difference between different devices by calculating the Euclidean distance between the device characteristic vectors, and can classify the devices with similar characteristics into one category through the hierarchical clustering tree diagram, so that the operation and maintenance resources can be more reasonably allocated and dispatched; the average distance is used as the basis for clustering the characteristics of distribution transformers to obtain the classification result, which can avoid the phenomenon of excessive compression or expansion and non-monotonicity hindrance between categories, while ensuring the flexibility of the clustering model, reducing the deviation of the result, and ensuring the effectiveness and accuracy of the classification result.

[0201] Overall, the classification module 20 of the embodiment can decompose the complex power distribution network problem into several relatively simple levels by establishing a hierarchical analysis model containing several levels, so as to systematically understand and analyze the operation state of the power distribution network. By establishing the hierarchical clustering model of the distribution transformer, it is beneficial to classify the operation state data of the transformer, classify the transformers with similar operation states into one category, and then obtain the classification result.

[0202] In one embodiment, the model module 30 includes a short-term unit, a constraint unit and a composition unit, wherein the short-term unit is a process of constructing a day-ahead control layer objective function, the constraint unit is a process of constructing a charging power feasible region and upper and lower limit constraints of photovoltaic output, and the composition unit is a process of constructing a multi-objective optimization model, specifically:

[0203] The short-term unit is configured to establish a day-ahead control layer objective function including load rate balance degree, network loss and three-phase imbalance degree according to the short-term single-phase heavy overload category and the short-term three-phase heavy overload category with source-load regulation potential in the classification result; specifically:

[0204] According to the maximum and minimum values of the load rate of the distribution transformer, a first objective function F1 is established;

[0205] According to the total loss of the transformer and the network loss, a second objective function F2 is established;

[0206] According to the three-phase imbalance degree of the node, a third objective function F3 is established;

[0207] The day-ahead control layer objective function is composed of the first objective function F1, the second objective function F2 and the third objective function F3;

[0208] The first objective function is:

[0209] The second objective function is: loss.tot +P loss.net

[0210] The third objective function is:

[0211] The day-ahead control layer objective function is:

[0212] For the parameter VUF i , there is:

[0213] Wherein, F is the objective function value; F1', F2' and F3' are the normalized values of the objective functions F2, F2 and F2, which eliminate the influence of different dimensions on the optimization result; s2 and s3 are scale factors, p max is the maximum value of the distribution transformer load rate, p min is the minimum value of the distribution transformer load rate, P loss.tot is the total transformer loss, P loss.net is the network loss, VUF i is the three-phase imbalance degree of the node i, is the set of nodes in the distribution network, Φ is the three-phase set, U neg,i and U pos,i are the negative sequence and positive sequence voltages of the node i, U a,i , U b,i and U c,i are the a, b and c phase voltages of the node i,

[0214] The short-term unit of the embodiment establishes a first target function according to the maximum and minimum values of the load rate of the distribution transformer, can ensure that the load of the transformer runs in a reasonable range, and improves the operation efficiency and reliability of the distribution transformer. The second target function considers the total loss of the transformer and the network loss, which can reduce the power loss through the optimization of the operation strategy, not only helps to save the power resources, but also reduces the environmental pollution caused by the power loss; the third target function is established according to the three-phase unbalance degree of the node, which can improve the power quality and stability of the system and ensure that the equipment runs in the best state. By comprehensively considering multiple factors such as load rate balance degree, loss and three-phase unbalance degree, the day-ahead control layer target function is formulated, which is helpful for formulating a comprehensive day-ahead source and load distribution transformer operation and maintenance plan.

[0215] The constraint unit is configured to establish a charging power feasible region according to a rated power of the charging pile without considering satisfaction loss;

[0216] The constraint unit is further configured to establish photovoltaic output upper and lower limit constraints according to upper and lower limits of output power of the distributed power source.

[0217] The charging power feasible region is:

[0218] The photovoltaic output upper and lower limit constraints are:

[0219] wherein, is the rated power of the charging pile; t i,a is a start charging time; t i,d is a driving-off time; E i,targ is a target electric quantity; and are an upper limit and a lower limit of output power of the i th DG, respectively, is output power of the i th DG at time t, P i,t is power of the charging pile at time t, U i,t is an operating state of the i th DG at time t.

[0220] The constraint unit of the embodiment establishes the charging power feasible region to ensure that the operating state of the charging pile is within the allowed power range, which not only avoids damage to the equipment caused by overload, but also ensures the stability and efficiency of the charging process; at the same time, the photovoltaic output upper and lower limit constraints are established by considering the upper and lower limits of the output power of the distributed power source, which can maximize the use of renewable energy, reduce unnecessary energy waste, and improve the overall operation efficiency of the system.

[0221] The constituent unit is configured to constitute a multi-objective optimization model from the day-ahead control layer target function, the charging power feasible region and the photovoltaic output upper and lower limit constraints.

[0222] In one embodiment, the operation and maintenance module 40 includes a speed unit, an optimization unit and an operation and maintenance unit, wherein the speed unit is a process of calculating a preset speed, the optimization unit is a process of obtaining a multi-objective optimization result, and the operation and maintenance unit is a process of formulating a day-ahead source-load-distribution-transformation operation and maintenance plan and performing operation and maintenance control, specifically:

[0223] The speed unit is configured to set the position of a particle in each dimension to two scales of 0 and 1, respectively, to learn the maximum value and the minimum value of a factor in a dimension space according to a binary particle swarm optimization algorithm, and to calculate a particle parameter ω; wherein the particle parameter ω is obtained by setting through a dynamic adjustment strategy.

[0224] The preset speed v is calculated by combining the particle parameter ω with the historical optimal position and the current position of the particle individual.

[0225] The particle parameter is:

[0226] The preset speed is: v = ωv i +c i k i (x-x i )

[0227] wherein c imax and c imin are the maximum value and the minimum value of the factor in the dimension i space learned by the binary particle swarm optimization algorithm; t max is the maximum time consumption of the space range search executable at the speed v;

[0228] v i is the current speed of the particle i, c i is the learning factor, k i is a random number between 0 and 1, x is the historical optimal position of the particle individual, and x i is the current position of the particle i.

[0229] The optimization unit is configured to set an optimization target of a particle according to a multi-objective optimization model, to make the particle perform optimization in an executable space range according to the preset speed v, and to obtain a multi-objective optimization result.

[0230] The optimization target is established according to a sigmoid function.

[0231] The formula of the sigmoid function is:

[0232] wherein x is an input value.

[0233] In the optimization unit of the embodiment, the sigmoid function can map any real number value to between 0 and 1, and this feature enables the sigmoid function to limit the range of output values, so that setting the optimization target of the particle by using the sigmoid function can avoid the situation of infinite growth or excessive saturation of the value.

[0234] The operation and maintenance unit is configured to form a day-ahead source-load regulation strategy of this category based on the next-day photovoltaic output and the power setting of the charging station, to realize multi-objective optimization setting of the distribution transformer operation mode, to formulate a day-ahead source-load distribution transformer operation and maintenance plan for heavy overload and three-phase imbalance, and to perform operation and maintenance control on the distribution network according to the day-ahead source-load distribution transformer operation and maintenance plan.

[0235] The category refers to the "short-term single-phase heavy overload category and short-term three-phase heavy overload category" with source-load regulation potential.

[0236] In addition, for the long-term single-phase heavy overload category and the long-term three-phase heavy overload category, a maintenance plan is formulated according to the maintenance specification and the distribution transformer operation regulation, and the specific formulation is as follows:

[0237] 1) Long-term single-phase heavy overload: increase the power electronic devices such as capacitors to compensate the circuit, improve the power factor, and reduce the overload situation;

[0238] 2) Long-term three-phase heavy overload: the distribution capacity can be improved according to the economic situation to avoid the overload phenomenon, and the distribution area exchange system can be installed according to the surrounding distribution area power supply situation.

[0239] Overall, in the particle swarm algorithm of the operation and maintenance module 40 of the embodiment, each particle represents a possible solution, and by setting the position and speed of the particle, the algorithm can control the search direction and step length, thereby improving the search efficiency; at the same time, the introduction of the preset scale helps to avoid excessive randomness of the particle in the search process, improves the search accuracy and stability, and greatly shortens the time for obtaining the multi-objective optimization result.

[0240] Overall, the embodiment has the following beneficial effects:

[0241] The application can quantitatively evaluate the similarity or difference between different devices by quantifying the device difference as the straight-line distance between two points, and clustering the distribution network operation and maintenance state, and classifying the devices with similar characteristics into a class, so as to more reasonably allocate and schedule operation and maintenance resources according to the classification results. Moreover, the multi-objective optimization results obtained by solving the multi-objective optimization model can simultaneously consider multiple optimization objectives, such as network loss and three-phase load imbalance, which helps to realize comprehensive optimization of the distribution network, reduce network loss, improve energy utilization efficiency, balance three-phase load, and reduce device loss and distribution network operation risk caused by three-phase imbalance. Therefore, the construction of the day-ahead source-load-distribution transformer operation and maintenance plan based on the multi-objective optimization results can control the operation and maintenance of the distribution network, which can reduce the probability of occurrence of distribution network faults and hidden dangers, reduce the operation risk of the power grid, and ensure the stable operation of the distribution network by optimizing the source-load-distribution transformer measures.

[0242] Moreover, by using the differentiated characteristic data such as the original data, operation data and maintenance data of the distribution transformer, the related characteristics of heavy overload and three-phase imbalance are focused on, the analytic hierarchy process clustering method is used to extract the distribution transformer categories with source-load regulation potential, and the differentiated operation and maintenance can be accurately realized, and the actual devices working in different working conditions can be better matched. In addition, the load rate balance degree, network loss and three-phase imbalance degree are taken as multiple objectives, the binary particle swarm optimization algorithm is used to realize the optimization of the objective function, and the accuracy of the regulation scheme is emphasized.

[0243] Embodiment three

[0244] The embodiment of the application provides a computer readable storage medium, the computer readable storage medium comprises a stored computer program, wherein when the computer program runs, the device where the computer readable storage medium is located is controlled to execute the operation and maintenance method for heavy overload and three-phase imbalance of distribution transformer.

[0245] The operation and maintenance method for the variable transformer overload and three-phase imbalance can be stored in a computer readable storage medium if it is implemented in the form of a software function unit and used as an independent product. Based on this understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware. The computer program can be stored in a computer readable storage medium. When the computer program is executed by a processor, the steps of the above-mentioned various method embodiments can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0246] The above is the preferred embodiment of the present application. It should be pointed out that for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, which are also considered within the scope of protection of the present application.

Claims

1. An operation and maintenance method for variable heavy overload and three-phase imbalance, characterized in that, The application comprises the following steps: obtaining differentiated characteristic data related to distribution transformers from a power distribution network; performing cluster analysis on the operation and maintenance state of the distribution transformer by quantifying the equipment difference as the straight-line distance between two points according to the differentiated characteristic data, and obtaining a classification result; establishing a multi-objective optimization model including network loss and imbalance according to the short-term single-phase heavy overload category and the short-term three-phase heavy overload category with source-load regulation potential in the classification result; solving the multi-objective optimization model to obtain a multi-objective optimization result; formulating a day-ahead source-load distribution transformer operation and maintenance plan for heavy overload and three-phase imbalance according to the multi-objective optimization result, and performing operation and maintenance control on the power distribution network according to the day-ahead source-load distribution transformer operation and maintenance plan.

2. The operation and maintenance method according to claim 1, wherein, According to the differentiated characteristic data, the operation and maintenance state of the distribution transformer is quantified as the straight-line distance between two points, and cluster analysis is performed to obtain a classification result, which is specifically as follows: a hierarchical analysis model including several levels is established according to the differentiated characteristic data; wherein the several levels include a target layer, a criterion layer and a scheme layer; a distribution transformer hierarchical clustering model is established according to the evaluation result matrix of the several levels; the classification result is obtained by using the difference degree of distance to represent the similarity of equipment and selecting Euclidean distance as the similarity measurement method according to the distribution transformer hierarchical clustering model.

3. The operation and maintenance method according to claim 2, wherein, The hierarchical analysis model including different levels is established according to the differentiated characteristic data, which is specifically as follows: elements related to decision in the differentiated characteristic data are decomposed into a target layer, a criterion layer and a scheme layer; wherein the target layer represents equipment characteristics, the criterion layer represents equipment characteristic factors, and the scheme layer represents related equipment that needs to be analyzed; the hierarchical analysis model is established by combining heavy overload data and three-phase imbalance data in the differentiated characteristic data with the target layer, the criterion layer and the scheme layer.

4. The operation and maintenance method for power transformation according to claim 2, wherein, A distribution transformer hierarchical clustering model is established according to the evaluation result matrix of the several levels, which is specifically as follows: the evaluation result matrix is obtained by determining the evaluation result of the criterion layer to the target layer and the evaluation result of the scheme layer to the criterion layer; a set of equipment characteristic weight vectors is calculated according to the evaluation result matrix; the distribution transformer hierarchical clustering model is established according to the calculation result of the corresponding weight vector in the set of equipment characteristic weight vectors and the characteristic value of the transformer substation area.

5. The operation and maintenance method for power transformation according to claim 2, wherein, The classification result is obtained by using the difference degree of distance to represent the similarity of equipment and selecting Euclidean distance as the similarity measurement method according to the distribution transformer hierarchical clustering model, which is specifically as follows: Euclidean distance is used to measure the closeness of the characteristic vectors of two equipment according to the distribution transformer hierarchical clustering model, and distance data is obtained; the two types of parameters with the highest closeness in the distance data are combined into one type to obtain a hierarchical clustering tree diagram; the hierarchical clustering tree diagram is segmented according to the average distance between different categories in the hierarchical clustering tree diagram to obtain the classification result.

6. The operation and maintenance method for power transformation according to claim 1, wherein, According to the short-term single-phase heavy overload category and the short-term three-phase heavy overload category with source-load regulation potential in the classification result, a multi-objective optimization model including network loss and imbalance is established, specifically: According to the short-term single-phase heavy overload category and the short-term three-phase heavy overload category with source-load regulation potential in the classification result, a day-ahead control layer objective function including load rate balance degree, network loss and three-phase imbalance degree is established, specifically: According to the upper and lower limits of the charging pile rated power and the distributed power output power, a charging power feasible region and photovoltaic output upper and lower limit constraint are respectively established; The multi-objective optimization model is composed of the day-ahead control layer objective function, the charging power feasible region and the photovoltaic output upper and lower limit constraint.

7. The operation and maintenance method according to claim 6, wherein, According to the short-term single-phase heavy overload category and the short-term three-phase heavy overload category with source-load regulation potential in the classification result, a day-ahead control layer objective function including load rate balance degree, network loss and three-phase imbalance degree is established, specifically: According to the maximum and minimum values of the distribution transformer load rate, a first objective function is established; According to the transformer total loss and network loss, a second objective function is established; According to the three-phase imbalance degree of the node, a third objective function is established; The day-ahead control layer objective function is composed of the first objective function, the second objective function and the third objective function.

8. The operation and maintenance method for power transformation according to claim 1, wherein, Solving the multi-objective optimization model obtains a multi-objective optimization result, specifically: According to the multi-objective optimization model, by setting the optimization target of the particle, the particle is optimized in the executable space range according to the preset speed to obtain the multi-objective optimization result; The optimization target is established according to a logical function.

9. The operation and maintenance method according to claim 8, wherein, The preset speed is specifically: The position of the particle in each dimension is set as a preset scale, and the particle parameters are calculated according to the maximum value and the minimum value of the learning factor in the dimension space according to the binary particle swarm optimization algorithm; The preset speed is calculated by combining the particle parameters, the historical optimal position and the current position of the particle individual.

10. An operation and maintenance device for variable heavy overload and three-phase imbalance, characterized in that, It comprises a data module, a classification module, a model module and an operation and maintenance module; The data module is configured to obtain differentiated characteristic data related to distribution transformers from a distribution network. The classification module is configured to perform cluster analysis on the operation and maintenance state of the distribution transformers by quantifying the equipment difference as the straight-line distance between two points according to the differentiated characteristic data to obtain a classification result. The model module is configured to establish a multi-objective optimization model including network loss and imbalance according to the short-term single-phase heavy overload category and the short-term three-phase heavy overload category with source-load regulation potential in the classification result. The operation and maintenance module is configured to solve the multi-objective optimization model to obtain a multi-objective optimization result, formulate a day-ahead source-load distribution transformer operation and maintenance plan according to the multi-objective optimization result, and perform operation and maintenance control on the distribution network according to the day-ahead source-load distribution transformer operation and maintenance plan.

11. The operation and maintenance device for power transformation and heavy overload and three-phase imbalance according to claim 10, characterized in that, The classification module comprises a hierarchical unit, a matrix unit and a similarity unit. The hierarchical unit is configured to establish a hierarchical analysis model including a plurality of levels according to the differentiated characteristic data; the plurality of levels include a target layer, a criterion layer and a scheme layer. The matrix unit is configured to establish a power distribution transformer hierarchical clustering model according to the evaluation result matrix of the several levels. The similarity unit is configured to represent the similarity of devices by using the difference degree of distance according to the power distribution transformer hierarchical clustering model, and solve by selecting the Euclidean distance as the similarity measurement method, to obtain the classification result.

12. The operation and maintenance device for power transformation and heavy overload and three-phase imbalance according to claim 11, characterized in that, The hierarchy unit includes a hierarchy subunit and a model subunit. The hierarchy subunit is configured to decompose elements related to decision in the differentiated characteristic data into a target layer, a criterion layer, and a scheme layer, wherein the target layer represents device characteristics, the criterion layer represents characteristic factors of the device, and the scheme layer represents related devices that need to be analyzed. The model subunit is configured to combine the overload data and the three-phase imbalance data in the differentiated characteristic data, and the target layer, the criterion layer, and the scheme layer, to establish the analytic hierarchy model.

13. The operation and maintenance device for power transformation overload and three-phase imbalance according to claim 11, characterized in that, The matrix unit includes an evaluation subunit, a weight subunit, and a hierarchy subunit. The evaluation subunit is configured to obtain the evaluation result matrix by determining the evaluation result of the criterion layer on the target layer and the evaluation result of the scheme layer on the criterion layer. The weight subunit is configured to calculate a device characteristic weight vector set according to the evaluation result matrix. The hierarchy subunit is configured to establish the power distribution transformer hierarchical clustering model according to the calculation result of the transformer substation characteristic value and the corresponding weight vector in the device characteristic weight vector set.

14. The operation and maintenance device for power transformation overload and three-phase imbalance according to claim 11, characterized in that, The similarity unit includes a distance subunit, a synthesis subunit, and a segmentation subunit. The distance subunit is configured to use the Euclidean distance to measure the closeness of the characteristic vectors in two devices according to the power distribution transformer hierarchical clustering model, to obtain distance data. The synthesis subunit is configured to synthesize the two types of parameters with the highest closeness in the distance data into one type, to obtain a hierarchical clustering tree diagram. The segmentation subunit is configured to segment the hierarchical clustering tree diagram according to the average distance between different categories in the hierarchical clustering tree diagram, to obtain the classification result.

15. The operation and maintenance device for power transformation and heavy overload and three-phase imbalance according to claim 10, characterized in that, The model module includes a short-term unit, a constraint unit, and a composition unit. The short-term unit is configured to establish a day-ahead control layer objective function including load rate balance degree, network loss, and three-phase imbalance degree according to short-term single-phase overload categories and short-term three-phase overload categories with source-load regulation potential in the classification result. The constraint unit is configured to establish a charging power feasible region and photovoltaic output upper and lower limit constraint according to the upper and lower limits of the output power of the charging pile rated power and the distributed power supply. The composition unit is configured to constitute the multi-objective optimization model from the day-ahead control layer objective function, the charging power feasible region, and the photovoltaic output upper and lower limit constraint.

16. The operation and maintenance device for power transformation and heavy overload and three-phase imbalance according to claim 15, characterized in that, The short-term unit includes a first subunit, a second subunit, a third subunit, and a fourth subunit. The first subunit is configured to establish a first objective function according to the extreme value of the power distribution transformer load rate. The second subunit is configured to establish a second objective function according to the transformer total loss and network loss. The third subunit is configured to establish a third objective function according to the three-phase imbalance degree of the transformer. The fourth subunit is configured to establish a fourth objective function according to the load rate balance degree of the transformer. The third subunit is configured to establish a third objective function according to the three-phase unbalance degree of the node; The fourth subunit is configured to constitute the day-ahead control layer objective function from the first objective function, the second objective function and the third objective function.

17. The operation and maintenance device for power transformation overload and three-phase imbalance according to claim 10, characterized in that, The operation and maintenance module comprises an optimization unit. The optimization unit is configured to set a particle optimization target according to the multi-objective optimization model, so that the particle performs optimization according to a preset speed within an executable space range, and obtain the multi-objective optimization result. The optimization target is established according to a logic function.

18. The operation and maintenance device for power transformation and heavy overload and three-phase imbalance according to claim 17, characterized in that, The preset speed is specifically: The position of the particle in each dimension is set as a preset scale, and the particle parameters are calculated according to the maximum and minimum values of the learning factor in the dimension space according to the binary particle swarm optimization algorithm; The preset speed is calculated by combining the particle parameters, the historical optimal position and the current position of the particle individual.

19. A storage medium, characterized by The storage medium stores a computer program, the computer program is called and executed by the computer, and the operation and maintenance method for the distribution transformer overload and three-phase imbalance according to any one of claims 1 to 9 is realized.

Images