基于变压器运行声音识别的智能运维辅助决策技术  

Intelligent Operation and Maintenance assistant Decision Technology based on Sound Recognition of Transformer Operation

作　　者：王剑强 彭涛 Wang Jianqiang;Peng Tao(State Grid Jinhua Power Supply Company,Jinhua Zhejiang,321000)

机构地区：[1]国网浙江省电力有限公司金华供电公司,浙江金华321000

出　　处：《电子测试》2021年第20期33-36,共4页Electronic Test

摘　　要：根据现场采集到的主变运行时的声音、负荷潮流、中性点电流、环境及冷却系统运行方式等数据形成主变运行声音数据库,对数据进行初步探索后,利用BP神经网络算法对变压器运行声音的特征量进行预测分类,结合路况实时信息、人员配置、气候条件等数据,利用if-then规则快速生成运维策略,将预测结果及运维策略按权限等级推送至相关人员。According to the collected data of main transformer operation sound,load flow,neutral point current,environment and cooling system operation mode,the main transformer operation sound database is formed.After preliminary exploration of the data,the BP neural network algorithm is used to predict and classify the characteristic quantity of transformer operation sound,combined with the real-time information of road condition,personnel configuration,power supply and so on Climate conditions and other data,use if then rules to quickly generate operation and maintenance strategy,and push the prediction results and operation and maintenance strategy to relevant personnel according to the authority level.

关 键 词：声音数据库 BP神经网络算法 IF-THEN规则 运维策略 

分 类 号：TN912.34[电子电信—通信与信息系统] TM406[电子电信—信息与通信工程]