摘要
随着传感器技术、通信技术的高速发展,实时监测工业生产中加工状况成为可能。为了实时监测生产车间中的异常,根据生产现场感知到情境信息的特点,提出异常信息表示方法。分别构建情境模型和事件模型,结合两者的特点,提出Event-Context异常信息表示方法;根据事件的不同类型将异常归结为七种模式,并转换为复杂事件形式,应用复杂事件处理引擎Esper进行异常识别;对事件流进行预处理解决了因同源事件的干扰而未能识别出全部异常和正常识别为异常的问题。实验结果表明,该方法能够更准确识别出生产中异常情况。
With the rapid development of sensor technology and communication technology, it is possible to monitor the processing status in industrial production in real time. In order to monitor the abnormity in the workshop in real time, the method of abnormal information representation is proposed according to the characteristic of the scene information. The context model and event model were respectively constructed, and an Event-Context method described anomaly was proposed according to their characteristics. Depending on the different types of events, the anomaly was classified into seven patterns, and converted to the form of complex events. Anomaly could be recognized by a complex event processing product Esper. Through preprocessing of the event streams, the problems of failing to recognize all anomaly and the normal identified as anomaly because of the interference of homologous events were solved. The experimental results show that the method can accurately identify the abnormal situation in production.
出处
《计算机应用与软件》
2017年第12期88-95,共8页
Computer Applications and Software
基金
国家自然科学基金项目(71271038)
关键词
事件
情境
异常表示
异常监测
事件流预处理
Event
Context
Anomaly representation
Anomaly monitoring
Preprocessing
