基于PCA的多神经网络软测量模型及其在工业中的应用

Multiple Neural Networks Soft Sensing Model Based on PCA and it's Application in Industry

在复杂工业生产中 ,影响生产的因素非常多 ,使得用于软测量的神经网络模型极其复杂 .针对这个问题 ,利用主元分析法 (PCA)将影响因素重组 ,在此基础上 ,提出了一种多神经网络 (PCA- MNN)模型 .介绍了 PCA- MNN的结构及学习算法 ,并将其应用于氧化铝高压溶出过程中苛性比值及溶出率的软测量 ,利用现场实际运行数据进行仿真 ,结果表明 PCA-

In complex industry process, the neural network model for soft sensing is very complex because that there are a large number of factors will influence the industry process. To this question, principal component analysis (PCA) method was used to reorganize the factors. On this basis, a multiple neural network (PCA MNN) model is proposed. The structure and algorithm of PCA MNN is introduced. The PCA MNN model was applied in the process of high pressure digestion of alumina, simulation result show that PCA MNN model can sense ratio of soda to aluminate and leaching rate online effectively.