论文部分内容阅读
提出了一种新的回转窑烧成带温度的软测量方法。首先利用大津方法从烧结图像中分割出物料区和火焰区,并提取相应的特征值,然后采用最速下降法选择支持向量回归(SVR)模型的参数,基于这些参数建立了烧成带温度的软测量模型,最后利用前向滑动平均进行滤波提高拟合优度。将该方法应用于中国铝业公司山西铝厂的3#回转窑的氧化铝烧结过程,通过大量实验,验证了烧成带温度的软测量结果与实测值的拟合优度达到0.927,表明了该方法具有很强的实用性和准确性。
A new soft measurement method of sintering zone temperature for rotary kiln was put forward. Firstly, the material area and the flame zone were separated from the sintered images by the Otsu method, and the corresponding eigenvalues were extracted. Then the parameters of support vector regression (SVR) model were selected by steepest descent method. Based on these parameters, Measurement model, and finally use the forward sliding average filtering to improve the goodness of fit. The method was applied to the alumina sintering process of 3 # rotary kiln in Shanxi Aluminum Plant of China Aluminum Corporation. Through a large number of experiments, the goodness-of-fit of the soft measurement result and the measured value of the sintering zone temperature reached 0.927, The method has strong practicability and accuracy.