基于RBF神经网络的沉降槽泥层高度软测量模型研究

Study on Soft-sensing Method of the Height of Mud Interface in Settling Tank on RBF Neural Network

针对赤泥沉降过程中影响泥层高度的各个因素之间存在的复杂未知映射关系,精确数学模型难以建立的问题,提出了基于RBF神经网络的泥层在线预测模型,并和多元线性回归法测量结果进行比较。实验结果表明,基于RBF神经网络的泥层高度软测量模型具有很高的测量精度,可以在实际生产中发挥有效作用。

Since complex mapping relation existed in multi-factors which produce an effect on the height of mud layer,and it is difficult to build accurate mathematics models by traditional methods,a RBF neural network based soft-sensing model for the height of mud interface online detecting is presented in this paper,the measurement result of RBF model and multiple linear regression model is compared.The experimental results show clearly that RBF neural network based soft-sensing model has high measurement precision,and can play efficiency effect in practice producing.