摘要
传统检测方法只能对单个启动装置进行故障检测,无法实现多类故障智能检测,为了避免传统方法带来的弊端,提出了基于多Agent的多类故障高效智能检测方法。根据Agent功能划分结果,从实体状态信息中提取运行行为信息,定时查询相关变量,获取完整装置运行情况,并将其转化为状态信息。利用已有状态信息建立低维数特征空间,采用等能量分段方法提取相关特征,获取异常状态信息。利用决策Agent模糊理论融合技术对异常状态信息隶属度条件进行设定,获取最可能出现的故障种类,依据高效智能检测流程,实现启动装置多类故障检测。通过实验结果可知,该方法最小检测误差可达到0.005,为电力设备稳定运行奠定基础。
Traditional detection methods can only detect the faults of a single starter,but can not realize the intelligent detection of multiple types of faults.In order to avoid the drawbacks of traditional methods,a multi- agent based intelligent detection method for multiple types of faults is proposed. According to the result of Agent function partition,the operation behavior information is extracted from entity status information,the relevant variables are queried regularly,and the complete operation of the device is obtained,which is transformed into status information.The existing state information is used to build low- dimensional feature space,and the related features are extracted by equal- energy segmentation method to obtain abnormal state information.The membership condition of abnormal state information is set by using decision agent fuzzy theory fusion technology,and the most likely fault types are obtained.According to the efficient intelligent detection process,multi-type fault detection of startup device is realized.The experimental results show that the minimum detection error of this method can reach 0.005,which lays a foundation for the stable operation of power equipment.
作者
夏景
XIA Jing(China Academy of Information and Communications Technology,The Research Institute of Informatization and Industrialization Integration,Beijing 200001,China)
出处
《电子设计工程》
2019年第13期7-10,共4页
Electronic Design Engineering
基金
国家工业和信息化部财政专项(GC-HG4170634)
关键词
多AGENT
启动装置
多类故障
高效智能检测
低维数特征空间
等能量分段
multi-agent
start-up device
multi-type faults
efficient intelligent detection
lowdimensional feature space
equal-energy segmentation
