基于数据流检测技术的电气设备状态监测仿真

Simulation of Electrical Equipment Condition Monitoring Based on Data Flow Detection Technology

为了及时评估电气设备的优劣程度或老化程度,在设备发生故障前做出预警,提出基于数据流检测技术设计的电气设备状态监测方法。解析监测任务,架构任务数据ID集合,设定任务解析参数,通过数据抵达频率拟合数据发送时刻与接收时刻的差值频率,构建时延概率密度函数,依据窗口缓冲数据数量分配滑动窗口,实现状态数据的分类、丢弃或上传;基于时延概率密度函数建立滑窗规格约束式,判定当前数据流状态监测情况,经过拟合两个不同中心的核函数,估计数据流密度,采用拟合误差值完成设备状态监测、预警。仿真案例分析得出,所提方法在不断的频率波动变化中,能够准确检测出异常波动,且相同数量数据流下,所提方法具有更高的监测准确率。

In order to evaluate the quality or aging degree of electrical equipment in time,and make early warning before equipment failure,this article put forward a method of monitoring electrical equipment conditions based on data flow detection technology.Firstly,we analyzed the monitoring task and constructed the collection of task data ID.Secondly,we set the task analysis parameters and fitted the difference frequency between the sending time and the receiving time by the data arrival frequency.Thirdly,we constructed the delay probability density function.According to the number of window buffer data,we allocated the sliding window to realize the classification,discarding or uploading of state data.Moreover,we established the sliding window constraints based on the delay probability density function to determine the data flow condition monitoring status.After fitting two kernel functions with different centers,we estimated the data stream density.Finally,we used the fitting error value to complete the equipment condition monitoring and early warning.According to simulation and case analysis,we can see that the proposed method can accurately detect abnormal fluctuation in constant frequency fluctuations.In addition,the proposed method has higher monitoring accuracy under the same number of data streams.