Prediction Of Hydrometallurgical Gold Grade Based On Interval Neuron Network

With the development of science and technology,industry is quietly experiencing the transition from industrial automation to information and knowledge automation,and people,machines and data will be perfectly integrated.At the same time,the current situation of mineral resources such as the shortage of mineral resources,the decline of the number of high-quality minerals and the scarcity of high grade ore seriously restricts the development of our national economy.How to predict the key quality parameters that can not be measured online during the process of industrial production has become a hot topic in the industrial and control field.In this thesis,the prediction of the gold mud grade in the hydrometallurgical process is carried out to realize the on-line prediction of the gold mud grade.According to the multi-modal characteristics of hydrometallurgy production process and some process data with the characteristics of interval expression,this thesis use a prediction model of gold mud grade based on multi-interval neural network method to achieve its online forecast.According to the characteristics of hydrometallurgical process,this thesis deeply analyzes and discusses the multi-modal characteristics,the characteristics of the process and the key factors affecting the grade of gold mud in the hydrometallurgical process.Based on the analysis of the multi-modal characteristics of the hydrometallurgical process and the interval expression of working condition data,the thesis analyzes and calculates the weight of the influence of the process data on the modal partition by using AHP and the interval de-clustering algorithm is used to realize the modal division of off-line data and the modal identification of online data.In a certain modality,according to the interval characteristic of process data,a prediction modeling of gold mud grade is proposed based on interval RBF network method,the use of interval RBF network is fitted the nonlinear relationship between the process data and the gold mud grade.Finally,the method proposed is validated and analyzed in the hydrometallurgical simulation data.Simulation results verify the effectiveness of the method proposed.