Prediction Method Of Gold Grade In Whole Process Of Hydrometallurgy

With the rapid development of China's economy and society,the requirement of automation technology for industry is getting higher and higher.Especially in the field of hydrometallurgy,high-grade mineral resources are scarce.How to obtain gold efficiently from low-grade mineral resources,how to obtain gold under the premise of ensuring green environment protection,how to use mineral resources efficiently to obtain the greatest economic benefits are the main issues that metallurgical enterprises pay close attention to.Among them,the gold mud grade in the whole hydrometallurgical process is the key quality index to determine the whole production process.However,due to technical reasons,online monitoring cannot be realized.Therefore,it is of great practical significance to realize the modeling and prediction of gold grade.The actual complex industrial process has a bad operating environment,high detection cost,and simple basic automation facilities,resulting in the coexistence of quantitative and qualitative information in the process.In the hydrometallurgical process,the process is complex,the production conditions are harsh,and there are numerous qualitative and quantitative variables.Traditional model prediction based on single data is difficult to achieve ideal prediction results.In this paper,the gold mud grade prediction method is studied for the characteristics of coexistence of hydrometallurgical qualitative information and quantitative information.While discussing the structure of complex industrial key parameter prediction model driven by qualitative information and quantitative information,several hydrometallurgical gold mud grade prediction models driven by qualitative information and quantitative information are proposed.Aiming at the problem that the traditional modeling method based on quantitative data can not make full use of qualitative information in hydrometallurgy,a qualitative information quantification method based on D-S evidence theory and cloud model is proposed.Based on D-S theory,multi-expert information is fused to reduce the uncertainty of qualitative information and avoid the influence of expert misinformation on prediction model.For quantitative information-based modeling methods that cannot handle qualitative information,a cloud model is proposed to quantify qualitative information,which can maximize the randomness and fuzziness of uncertain information while achieving quantitative transformation.Aiming at the complex industrial process without multi-process intermediate quality index,a predictive structure and method driven by qualitative information and quantitative information based on single model are proposed.Using the qualitative information quantification method based on D-S evidence theory and cloud model proposed in this paper to quantify qualitative auxiliary variables,combined with the remaining quantitative auxiliary variables as input of the data prediction model to establish the prediction model of key parameters.Realizing the prediction of key parameters when qualitative and quantitative information coexist.In view of the complex industrial process with intermediate quality index in multi-process.A key parameter prediction model driven by qualitative information and quantitative information based on multi-model is proposed.According to the characteristics of sub-process variable information,the combination of knowledge modeling and data modeling is used to realize key parameter prediction.Finally,this paper uses the data generated by the hydrometallurgical process simulation platform,combined with the proposed prediction model to predict the grade of gold mud,and then carries out simulation test and analysis on the proposed method,and verifies the effectiveness of the proposed method.