钴湿法冶炼萃取过程中的组分含量软测量

Soft sensing for component content in cobalt hydrometallurgy extraction process

提出了一种钴湿法冶炼萃取组分含量混合建模方法.该模型由基于物料衡算关系的动态机理模型与基于萃取平衡实验数据的RBF神经网络模型组成.机理模型作为描述过程动态行为的整体框架,RBF神经网络用来辨识机理模型中的未知函数关系.在上述混合模型的基础上,还提出了一种模型校正策略,进一步提高了模型的精确性.将所建立的混合模型应用于实际湿法冶炼生产过程中,结果表明该方法具有良好的估计性能.

A hybrid modeling scheme is presented in this paper for component content in cobalt hydrometallurgy extraction process, which is composed of first principle model based on material balance and RBF neural networks models based on the data of extraction balance experiment. First principle model describes dynamic behavior of process as the rough framework. RBF neural networks are employed to identify the unknown function relationship. Based on the hybrid model above, a model correction strategy is also proposed to make the model more accuracy. Applying the hybrid model to practical industrial process of hydrometallurgy, the results show that the method has satisfactory estimation performance.