Temperature Mode Recognition of Metallurgical Slag Based on KPCA and NN

As fusion and crystallization temperature is an important physical and chemical characteristic of metallurgical slag, an efficient way of soft computation is presented in the paper to recognize temperature mode in order to replace complicated hardware measurement device. First, in terms of sample data of image information in nonlinear correlation, kernel principal component analysis is adopted to get characteristics that adversely rule out nonlinear correlation information. Then, neural network is adopted to recognize fusion and crystallization temperature of metallurgical slag with help of unrelated sample data. Finally, the method is proved efficient after it is inspected in practical application of metallurgical slag’s temperature recognition.