Research On Information Retrieval Method Based On XML Document Structural Semantics And Its Application

With the development of electric power information management, many enterprises have accumulated large amounts of data, the research of query data from large amounts of electric power information fast and flexibly is beneficial for making full advantage of information resources and for manager's decision. This paper pays attention to the deep research on XML index structure, XML retrieval method based on structural semantics, the extended CIM model of substation equipment information and its expression by XML, the integration and query of substation equipments'information based on XML, the clustering analysis in condition evolution of transformer in family. The main achievements are as the following:1. A new XML index structure is put forward, which includes inversed element tag index (ETI), inversed element content index (ECI) and node level-path index (NLPI). The index structure considers both content and structure information of XML document; more over, it is fit for the structural semantic retrieval of XML documents.2. The structural semantic concept of XML document is extended, the rule of judging many nodes's semantic correlated relation is put forward and proved, which supplies the theory basis for XML structural semantic retrieval algorithm. A new tag-keyword semantic retrieval algorithm is put forword, which avoids judging many nodes's semantic correlated relation and improves the retrieval speed greatly.3. An extended CIM model of substation equipment information and the translation rule from CIM model to XML document are put forword, the frame of substation equipment information search system based on XML is given and the key technology used by each part of the frame is analyzed. The XML documents of substation equipment information based on CIM standard is compatible with other electric power information model according with CIM standard; with the standard XML document criterion, substation equipment information from different electric power enterprise expressed by XML document may have same semantic, which is beneficial to improve the retrieval efficiency of XML search engine.4 . Using clustering technology to research the condition evolution rule of transformer in family is first put forward, which is used for deciding the influence of family quality on transformer's condition evaluation. An improved agglomerate hierarchical clustering algorithm based on value distance and curve slope distance is put forward to analyze transformer condition evolution rule. The example shows that the algorithm is better than traditional agglomerate hierarchical clustering algorithm. Using clustering result to decide the influence of family quality defect on transformer's condition evaluation is provided. At last, this paper analyzes another transformer's condition in family according to the clustering result and gets accurate result, which shows that research of condition evolution rule of transformer in family is important to integrated condition evaluation and fault forecast.