基于两种探尺数据融合的高炉料位检测方法

Stock Line Detection of Blast Furnace based on Data Fusion of Mechanical Gauge Rod and Radar Gauge Rod

通过分析机械探尺和雷达探尺在高炉料位检测上的优缺点,采用模糊GK聚类算法,对雷达探尺测量数据进行聚类处理,并把聚类获得的参数用于构建一个RBF网络。利用机械探尺数据训练已建立的RBF网络,建立了基于机械探尺数据的修正模型,通过修正模型对雷达探尺测量数据的逐一修正,实现雷达探尺和机械探尺测量数据的有机融合。仿真结果和工业数据证明,基于机械探尺数据建立的修正模型具有较高的精度和较好的实用价值。

By analyzing the Characteristics of the mechanical gauge rod and the radar gauge rod on the stock line measurement of the blast furnace, the Fuzzy GK Clustering Method is firstly adopted to realize the measurement data clustering of the radar gauge rod. Then, the parameters obtained by the Fuzzy GK Clustering Method is used to establish a RBF neural network （RBFNN）, and the RBFNN is trained by the measurement data of the mechanical gauge rod. Finally, a correction model based on mechanical gauge rod data is built to correct the radar gauge rod data, which can realize the effectively fusion of the mechanical gauge rod data and radar gauge rod data. The proposed method overcomes the discontinuous measurement of the stock line of the mechanical gauge rod, and weak anti-disturbance ability, high accuracy fluctuations and poor stability of the measurement of radar gauge rod. Both the simulation results and industrial validation show that the proposed correction model has high accuracy and good practical value for industrial production.