数据融合在球磨机料位检测中的应用

Application of data fusion in ball mill material level detection

针对火电厂球磨机料位用常规方法难以准确检测的难题,将多传感器数据融合技术与BP神经网络相结合,采用多传感器数据融合技术对球磨机的磨音信号、出入口差压和出入口温差三个参数采样,结合BP神经网络法对采样数据进行融合,实现对球磨机料位的测量。结果表明,采用数据融合技术获得的融合结果比单独采用磨音检测料位获得结果要好得多,若能获得足够多的训练数据,则融合结果必定能够非常接近实际值,同时也能消除干扰,为球磨机优化控制奠定基础。

Aiming at the problem that the detection of the ball mill material level using conventional methods is ont accurate in power plants,this paper combines the multi-sensor data fusion technology with the BP neural network,samples three parameters of the ball mill which include grinding sound signal,pressure difference between export and import and temperature difference between export and import with multi-sensor data fusion technology,inosculates the sampling data through the BP neural network method,achieves the detection of the ball mill material level.The results show that the fusion results using data fusion technology are much better than the results using grinding sound signal,if enough training data can be obtained,then the fusion result must be very close to the actual value,at the same time the method can eliminate the interference,establish the foundation for optimizing the control of ball mill.