融合语义网的故障检索系统设计与构建

Design and Implementation of a Fault Searching System Combined with Semantic Web

摘要:

针对雷达故障数据和经验性技术支持文档的信息提取,提出了一种融合语义网技术的故障检索系统的设计与实现方法;该系统采用故障树组织结构化的历史故障数据,将故障树转化为二元决策图从而获取顶层故障事件的最小割集,而后通过语义匹配获取精确的故障原因信息;同时,借助开源搜索引擎Lucene检索技术文档库,获得与用户检索条目相关的支持文档;其中,为了提升中文分词和文本匹配的精确性,该系统还利用本体语言构建了专业词汇语义网,考量了词语之间的语义相关度;该系统已应用于某型机载有源相控阵雷达电源单元的故障检索,取得了很好的技术辅助效果。

Abstract:

As for the information retrieval of radar fault data and empirically technical support documents, we propose a design and implementation method of a fault searching system combined with semantic web technology. The system uses fault trees to organize structurized historical fault data, and converts them into binary decision diagrams to obtain minimal cut sets of top fault events, and then acquires accurate fault reasons via semantic matching. Meanwhile, it obtains support documents related to users’ query items by using an open source search engine Lucene to search from technical document database. In the system, it uses ontology language to build a semantic web of terminologies, for improving the accuracy of Chinese words segmentation and text matching, so as to involve semantic correlation between different words. The system has been used in the fault searching of the power unit of an AESA radar and has obtained good effect for technological assistance.