3.1

Research on electrical topology generation based on multi-objective merging algorithm

In this method, in order to minimize the line length and the number of intersections of the graph, the above two main objectives are expressed in the form of a W polynomial mathematical model. If necessary, other optimization parameters are added, such as the uniform distribution of primitives. The W weighting method of multiple objective polynomials is integrated, and then centralized optimization is performed. The two main sub objective functions of this method are shown below.

Single line diagrams usually use the mutual distance between nodes to describe the length of the line between nodes. Manhattan distance or Euclid distance are often used. Both distances have been used. W Manhattan distance is an example. The Manhattan distance of the total line length is:

As is shown in formula 1, where parameter i and j are subscripts of two components to be drawn, (x_i, x_j), (y_i, y_j) are the coordinates of two components respectively. When there is a direct topological connection between i and j, the result is 0.

To ensure the minimum line crossing, its mathematical model is shown in Formula 2.

First, after rapid screening, whether the two lines intersect is judged by whether the rectangle determined by the diagonal of the two W line segments intersects. This method can only quickly filter some segments that do not meet the requirements.

Secondly, the remaining unfiltered line segments are judged by the method of judging the span of line segments. If two line segments intersect, then two line segments must span each other. If segment 1 intersects segment 2, then 1 and 2 are distributed on both sides.

However, this method also has obvious shortcomings. For example, it is not easy to select the weights of each target, which requires the definition of the topology of the graph and long-term experience to achieve target selection. In addition, the generation of electrical topology takes a long time. If there are few nodes in the graph, the time required for iterative calculation of mapping is not much. However, for the grid graph with a lot of electrical topology nodes and complex topology connection, because of the need to repeatedly calculate the position relationship between each electrical topology node, each heuristic algorithm iteration takes too much time, sometimes even close to and exceeds the time for manual topology drawing. Nowadays, most of the automatic heuristic methods are prone to be trapped in local optimization, and the mapping effect is not good, which is easy to cause too much cross or too dense connections.

3.2

Research on automatic generation algorithm of electrical topology based on intelligent matching

Electric power single line diagram is to draw the form of points and lines of each entity component in practice. Such a diagram has a special subject of graph theory in mathematics for research, and the set of points and lines is also called graph in graph theory. Both the drawing of single line graphs and the study of graphs in graph theory are based on the relationship between points and lines. The graph to be studied jointly by the two is summarized below. That is, the graph G (V, E) to be studied includes two device sets: vertex set and edge set. V represents vertex set, that is, vertex set with numbers. E is an edge set, which is a subset of set V and refers to the connecting line between two vertices with connection, That is, (x, y). E represents that there is a line between vertices X and Y.

As is shown in figure 1, the multi-objective decomposition method consists of six steps of planarization and compaction, which respectively resolve path retrieval problems in multiple graph theories. In the first step, planarization is used to deal with the target with the least cross connection. The second step is automatically, which deals with the graphic topology with minimal crossing. On the one hand, it achieves uniform distribution of graphic elements, on the other hand, it generates horizontal and vertical orthogonal graphs and minimizes the bending of the connecting lines. The third step is to compress and finally determine the coordinates and connecting positions of each node. The multiple steps are closely linked and correspond to the six objectives of automatic single line drawing, which stand opposite to the multi-objective combination optimization method. This method decomposes and then processes each objective separately, so that the solution has control over all objectives.

Figure 1.

Automatic generation algorithm of electrical topology based on intelligent matching

First, from the first step of planarization research, the existing methods in this part of the research mostly adopt the method of fixing the vertex first and then connecting lines. When there is a crossing between lines, avoid the crossing by detouring. If there is an unavoidable situation, first arrange the position of the line, and then connect the crossed lines after all non-crossed lines are completed. At this time, the crossing points are replaced by virtual nodes, the resulting graph only contains vertices and lines of virtual nodes without crossing.

In the step of planar zing the electrical topology connection, the intersection is reduced by changing the order of edge insertion. Experiments show that changing the order of edge drawing is indeed possible to reduce the intersection. In addition, this paper uses genetic algorithm to find the edge insertion sequence combination that can reduce the intersection. However, due to the addition of heuristic method for optimization, the mapping time increases.

4.1

Data processing for automatic generation of electrical topology

For basic data processing of electrical topology, firstly, the data read in and analyzed by D0M shall be processed by the automatic mapping algorithm, so as to realize automatic drawing of single line diagram. By default, the mapping result is a document file in SVG format. Since SVG files also follow XML syntax, it is allowed to import attached SVG files while importing RDF metafiles.

As is shown in figure 2, in this chapter, the orthogonal automatic mapping program adopts a new mapping method of multi-objective decomposition mode, which is divided into the plane orthogonal compression H steps correspond to the minimum crossing, the average layout and the shortest connecting line H objectives respectively. This method is suitable for the network intensive power grid structure. The overall processing steps of the program are: first, read the basic information such as the points to be drawn and their topological relationships from the EMS or the CIM file imported by the user, then process step by step according to the multi-objective decomposition method, and finally generate the required single line diagram in SVG format.

Figure 2.

Data Processing for Automatic Generation of Electrical Topology

For the first step of planarization, the starting point of the electrical topology node needs to be calculated first. You can select the starting point manually or automatically. There are usually many important nodes in the system, such as the balance node in the transmission ring network and the starting point of the distribution feeder in the distribution network. The important node is preferred as the starting point. If there is no important node, the starting point shall be selected randomly, and the ending point shall be selected according to the farthest point of depth first search. After the selection of W, the starting point shall be connected with a line to start the calculation of topological node relationship. After the calculation is completed, the topological simplification mode shall be reduced according to the node coding order, and the existing nodes shall be matched with the topological tree from bottom to top to finally determine the appropriate point and connection topological relationship to form the largest plane sub graph, Non planar edges are inserted in the way of finding the shortest path in the dual graph to minimize the crossing.

4.2

Realization of automatic generation of electrical topology in distribution network

This chapter establishes a fuzzy comprehensive evaluation model for grid planning of distribution network based on the subjective and objective combination weighting method. On the basis of the grid comprehensive evaluation index system established in Chapter 2, the index system is subject to fuzzy comprehensive evaluation. The fuzzy comprehensive evaluation method is divided into single-layer and multi-layer. Taking the two-layer fuzzy comprehensive evaluation model as an example, its establishment steps are as follows.

First, this chapter introduces the development of the software for automatic generation of electrical topology and the realization process of the function of the generation mode, analyzes the environment and functional structure of the software development, develops the software in accordance with the enterprise standards of the State Grid Corporation of China and other relevant specifications in order to smoothly connect the software to the energy management system, and then introduces the software realization process, analyzes the multi-objective decomposition into the national method process and the hierarchical layout method process respectively, Compare the multi-objective decomposition method to the currently used heuristic multi-objective optimization method, and finally compare the efficiency of electrical equipment topology generation of 1-2000 nodes to process and verify the progressiveness and effectiveness of the multi-objective decomposition method.

As is shown in figure 3, the input is the plane topology diagram formed in the first step. First, determine the relationship between each point and surface. First, determine the point to surface flow path, and then determine the surface to surface flow path. Use the constraints in Chapter H to calculate the initial conditions, substitute the initial conditions into the calculation of the minimum cost flow, and use each path of runoff to determine the relationship between the point and the line.

Figure 3.

Functional process of electrical topology simplification of distribution network

As is shown in figure 4, the sorted input is the orthogonal graph of the second step. First, calculate the vertical and horizontal dual graphs, then search the path from the starting point to the end point, add 1 to each branch flow on the path every time a path is found, finally summarize the data of each path, calculate the flow on each branch, and determine the length of the side passing by according to the size of the flow.

Figure 4.

Functional process of electrical topology simplification of distribution network

This method tests the effectiveness of the orthogonal mapping method by selecting a simple network with nodes that are not easy to cross and manually drawing a more cumbersome electrical topology node network diagram. In order to verify the progressiveness and effectiveness of the multi-objective decomposition method, this paper compares it with the multi-objective optimization method using genetic algorithm. In this paper, the genetic algorithm optimization method is used to select the mathematical model based on the relationship between optimized component coordinates, and the line length is determined according to the distance between components.

As is shown in table 1, for simple electrical topology equipment nodes, genetic algorithm is used to iterate until 2000 times, which takes 511.471s. It can barely draw a suitable understandable single line circle. Through comparison, it is found that the increase of iteration times is gradually reducing the key data such as graph crossing. But it took about 8 minutes to draw manually.

Table 1.

Efficiency analysis of automatic generation of electrical topology

It only takes about 20 minutes for the single line diagram of. For the simple electrical topology node multi-objective decomposition example, it is proved that it is roughly equivalent to the original manual single line drawing, with good mapping effect and no cross, and the density partition is equivalent to the general position of the original manual drawing, the approximate position of each component is close to the original manual drawing position, and the topological relationship is clear. The graph is easy to read, and the time for topology generation is greatly reduced.