Design And Application Of Distributed Full-text Retrieval System Based On SolrCloud For The Electric Power Knowledge Base

With the development of the economy and the construction of the power grid,the quality of power supply is improving day by day,but due to bad weather,aging equipment and other reasons,power failures still occur from time to time.Reaching the fault site accurately and quickly and repairing the fault are not only the service goal of State Grid,but also the ardent hope of customers.Therefore,the customer’s repair address is particularly important in power emergency repair work,but at present,there is no platform that can accurately provide customer’s repair addresses,and the traditional address information retrieval capabilities are also very poor.Therefore,this dissertation studies the method of constructing the knowledge graph of the repair address based on the ontology,and the full-text retrieval model based on the preference of the repair address,improved the retrieval method for confirming the customer’s repair report address,and finally designs and develops a distributed full-text retrieval system based on Solr Cloud for the electric power knowledge base to help the customer service staff guide the electric repair personnel to the accurate address for repairs and restore power to customers in time.The specific research contents of this dissertation are as follows:Construction of knowledge graph of repairing address.This dissertation first analyzes the knowledge characteristics of the repair address;then according to the characteristics of the repair address,it improves on seven-step methodology and skeletal methodology,and proposes the iterative method of constructing the repair address ontology,and completes the construction of the repair address ontology.Finally,based on the constructed address ontology,the construction of the knowledge graph of the repair address is completed.Combining the preference characteristics,construction of the retrieval model of the entire repair address is realized.This dissertation first studies the construction of a repair address preference model based on the probability of a power outage,and then analyzes an algorithm for full-text retrieval.Combining the above two points,a full-text retrieval framework based on the repair address preference model is proposed,which enables the system to be in multiple names.The result of retrieving the repair address with high probability of power failure in the address greatly reduces the interference of the duplicate address and improves the accuracy of retrieval of the repair address.Construction of a distributed full-text retrieval system for the power knowledge base.Based on the above research results,this dissertation designs and implements a Solr Cloudbased distributed full-text retrieval system for electric power knowledge base.The system is divided into two parts.One is the repair address knowledge platform,where the administrator an manage the ontology knowledge,add concepts,attributes and rules,and maintain the repair address entity;the other is the full-text search system,where the administrator can create and delete the index,and the customer service staff can retrieve the repair address.The realization of the electric power knowledge base distributed full-text retrieval system.Firstly,it manages ontology knowledge through the Jena framework.Secondly it maintains the entity information of the repair address in the system.Then it elaborates on the construction of the repair address full-text retrieval system.Finally,the system operating results are demonstrated and analyzed.