The Research On Station Alarm Correlation Analysis And Health Evaluation Method

With the rapid development of network technology, people are increasingly great for network quality of service of mobile operators. In the entire mobile communication system, base station system is mainly to provide users with access to services,occupies a very important position. Timely analyzing alarms and association with the mining laws, understanding the overall operation of the system and fault locating exclusion, ensuring stable continuous operation of base station system, is the major issue of concern to most mobile operators.The purpose of this paper is to design a set of base stations and base station alarm correlation running state assessment system so that maintenance personnel can locate faults timely, maintain and analyze alarms,to understand the overall operation of the base station system. It mainly includes alarm correlation analysis and evaluation of the health of base station system. Base station alarm correlation analysis:using data mining association rule method to analyze base station law correlation and features on the base station alarm summary, alarm filtering mechanism design (low,redundant, irrelevant field excluded),using sliding window method to transfer relational data into alarm transaction sequence; using APRIORI,CARMA and GRI to mine association rule of power failure and using Sequence to mine important sequence rules of important alarms, using data mining tools to mine rules, getting the original rules;alarm rules post-treatment, pruning the original rules,and rule storage preservation. Health evaluation of base station system:using fuzzy evaluation method to build the health modeling of base station. Establishment of the indicator system,including static and dynamic indicators,dynamic indicators are calculated in alarm correlation rules; the indicator weight is acquired by the similar rights method, order relations act and AHP method, and giving system health evaluation results (alarm statistics of the health of the scores, alarm correlation, the health of the curve, etc.).Through this study, it gives a set of base alarm correlation analysis and health evaluation method program. The system is running tests in a city mobile company and running stably3months, the system can dig out rules of the power failure of base station related and the sequence law of important warning alarm such as stop, channel, etc. This system can calculate daily, monthly statistics of the base station system in various historical and real-time health of curve, and has some theoretical research value and practical significance.