An Improved Orthogonal List Based Storage Technique Applied In Short Circuit Calculation

The calculation of short-circuit current is very important for the safety operation of the power system. The node admittance matrixes and its decomposed matrixes are sparse matrixes. The matrixes contain a large number of zero elements. The short-circuit current calculation needs the storage and the query of admittance matrixes and the decomposed matrix especially the data location query according to its value, in order to avoid a large number of zero element storage and improve the storage efficiency, the cross linked table storage method was used. But the cross linked table as a chain structure has its drawbacks, namely the irregular data storage and query efficiency will be reduced. Due to the admittance matrixes and decomposed matrixes generated by short-circuit current calculation are very large, the storage and query efficiency decline will be more obvious. If data sorting can be achieved, the efficiency will be improved. For each data in the admittance matrix and its storage location (ranks) has its special electrical meaning, the storage location can not be disorganized. In order to keep a fixed position of storing elements and also maintain a rapid value query speed according to the ranks of row and the column of the query element, and also achieve a high query speed of the location of the row and the column of the data according to the value, to facilitate the storage, calling, and subsequent processing of the matrixes, A improved cross chain table method combined with highly balanced binary tree is proposed in this paper. Specific work is as follows:1. Analysis of the advantages and disadvantages of sparse matrix storage method in the calculation of current short circuit current. The advantages and defects of the two dimensional linked list storage method, the cross linked list storage method and the cross linked list storage area distribution method are analyzed. It provides a reference for innovation.2. On the basis of cross linked list storage, the storage data node pointer field was expanded, each storage node was chained on a height balanced binary tree through a pointer, the height of the tree was maintained within (O(log2n)) and the average search length can also be maintained within (O(log2n)).Experiments showed the operating time has been greatly reduced and the efficiency of the value query has been improved by the rank of the row and the column.3. In order to maximize reducing the impact of time consumption of construction the height balanced binary tree on the overall operation efficiency, the characteristics of the computer hardware itself was used by assigning a block of contiguous memory space to store the cross linked list. Those nodes and its neighbor nodes not in cache memory were loaded into the cache memory, in order to achieve directly to the access and storage operation, which greatly enhance the efficiency of cross linked, greatly reduced time consumption on building highly balanced binary tree according to the row and the column query by the value of the data.In addition, in order to ensure the fairness of the test results, to convenient comparision, the time consumption of the construction of a highly balanced binary tree was included in the total time. The results of the experiments proved the high efficiency of the proposed method.