Modeling And Correction Of Soft Sensor For Hydrometallurgy Leaching Process

In view of our country’s energy shortage,especially the high grade ore is decreasing continually.To make full use of mineral resources is particularly important.Hydrometallurgical can be more efficient in the extraction of gold from low grade nonferrous metal mineral resources,and the environmental pollution is small.Leaching process is an important working procedure in the process of Hydrometallurgical.The rate of leaching is directly related to the economic benefit of the production process,however,the leaching rate is an important index to measure the leaching process,but it cannot be measured online.In this thesis,based on the problem that the leaching rate cannot be measured online,the soft sensor modeling technology is used to realize the online prediction of leaching rate.1.Based on the deep analysis of the leaching process of the Hydrometallurgical process,the steady-state mechanism model of the leaching process which is based on the material balance and the reaction kinetics equation is mastered.In the modeling process of some simplification and approximation of the mechanism model of the leaching process cannot achieve the desired accuracy,this thesis proposes a soft sensor modeling method that the leaching mechanism combined with the data predict the leaching rate.2.During in the process of Hydrometallurgical,there are many characteristics of multi working conditions.This thesis discusses the method of division of the "hard" and "soft" and the method of recognition.In this thesis,the basis of the multi condition integrated modeling method based on the "hard clustering division" and the "soft clustering classification" are carried out with experiment and simulation.The simulation results show that the precision of the soft sensing model is improved by the multi model integration technology based on FCM soft clustering,and the result is more satisfactory.3.Model calibration is an effective method to solve the problem of soft sensor model mismatch.The correction and updating of the soft sensing model is realized by the combination of short term and long term correction.In this thesis,we adopt the linear dependence of the newly collected samples and historical data to determine the short-term or long-term correction of the model.When the judgment result is linear dependence of the new sample,the short-term correction of the soft sensing model is short.If the judgment result is nonlinear,the long-term correction of the soft measurement model is used.4.In this thesis,a simulation test and analysis of the method presented in this thesis is carried out by using the data from the Hydrometallurgical leaching process.The simulation results show the effectiveness of the proposed method.