基于RBF网络的浮选过程技术指标智能预测系统的研究

STUDY ON INTELLIGENT SYSTEM FOR PREDICTION OF TECHNOLOGICAL INDEXES OF FLOTATION PROCESS BASED ON RBF NEURAL NETWORK

利用ZJWNNC标准函数建立了浮选过程技术指标与矿浆性质、药剂用量之间的关系模型 ,并对该模型进行了仿真试验。仿真结果表明 ,模型输出能以较高精度逼近样本数据。基于上述模型开发了西林铅锌矿浮选过程技术指标智能预测系统 ,并进行了工业试验。工业试验的结果表明 ,该系统能够以较高的精度预测浮选过程作业回收率和精矿品位。该系统的成功开发为浮选过程的优化控制奠定了基础。

A correlative model in which technological indexes of flotation process depend on the ore character and reagents consumption has been proposed using the ZJWNNC function and the simulation using this model has been conducted in the laboratory.Simulation results show that the output of the model matches the sample data very well.On the basis of above model,an intelligent system suitable to prediction of technological indexes of flotation process in Xilin Lead & Zinc Mine has been developed and used in commercial test.The results indicate that the intelligent system can be well used for prediction of the recovery and concentrate grade in flotation process.It has laid the foundation for optimal control of flotation process.