基于信息融合的球磨机料位分级与检测研究

Study on the grading theory and soft-sensing of mill fill level based on information fusion

火力发电厂锅炉制粉系统的主要设备之一——球磨机的运行状况优劣在一定程度上决定了电厂的经济运行状况好坏。分析了球磨机磨筒内料位的各种影响因素及影响特性,提出了料位分级理论,将料位分为非经济区、优化区与危险区三个区域,解决了数据融合测料位时无法获得大量料位样本的问题。在此基础上,提出了基于BP人工神经网络测量球磨机料位的软测量方法,对球磨机现场数据分析进一步验证了所提方法的有效性。该方法确定了球磨机最佳运行工况的范围,为实现球磨机系统的优化运行和自动控制奠定了基础。

Factors effecting the mill fill level and the relationships among themselves were analyzed.Mill fill level grading theory was offered which classifies the coal amount into three areas: uneconomical area,optimal area and dangerous area.The difficulty to obtain plenty of fill level samples was solved so that the data fusion method can be properly used to measure the fill level.A soft-sensing method of mill fill level based on BP neural network was proposed and its validity was proved by analyzing the data from the operation field.The method gives the optimal operation range for ball mill and forms the basis for the optimum operation and automatic control of ball mill system.