Leaching rate prediction in cyanide leaching process based on multi-unit average trajectory method

作者：Liu Yan^*; Chang Yuqing; Wang Fuli; Wang Shu; Tan Shuai

来源：Chinese Journal of Scientific Instrument, 2012, 33(10): 2220-2227.

• Continuous process
• Copper ore
• Cyanide leaching
• Generalization ability
• Input datas
• Integrated effects
• Interpretability
• Leaching process
• Leaching rates
• Modeling data
• Multi-unit
• Partial least square (PLS)
• Physical characteristics
• Prediction model
• Predictive values
• Process data
• Process information
• Regression model
• Trajectory matrix
• Trajectory method
• Weighted Sum

Because the average trajectory matrices of modeling data are undefined and the integrated effect of multi-unit production to leaching rate is not considered in traditional partial least squares (PLS) method, so it is not possible to accurately establish the correlation between the process information and the quality variable. According to the multi-unit and time-consuming characteristics of the rich copper ore cyanide leaching process, we propose a leaching rate prediction method based on multi-unit average trajectory for continuous process. The average trajectory matrices for the process data are acquired; on this basis the regression model of leaching rate for each unit is established. The similarity between input data and modeling data for each unit is defined; the weight for each prediction model is defined as well. The weighted sum of the unit prediction results is formulated as the final predictive value. The method was used to predict the leaching rate in high grade cyanide leaching process. Simulation results verify that the leaching rate prediction model based on multi-unit average trajectory method reflects the actual physical characteristics of leaching process; the method improves the interpretability of the prediction model and enhances its generalization ability.

• 出版日期2012
• 单位东北大学