A Mixed Model Based Concentration Prediction Method for Thickener Underflow

Abstract

The invention provides a method for predicting the underflow concentration of a thickener based on a mixed model. Aiming at the problem that the underflow concentration is difficult to measure online in the thickening washing process of hydrometallurgy, on the basis of in-depth analysis of the characteristics of the thickening washing process, the mechanism modeling and overall based The hybrid modeling method combined with the three-layer ELM error compensation model improved by the distribution optimization algorithm realizes the accurate measurement of the underflow concentration in the dense washing process.

Description

Background technique: [0001] With the large-scale, centralized and continuous production of hydrometallurgical industry, there is an urgent need for efficient and stable automatic production lines. However, the overall level of automation in my country's hydrometallurgical production process is low, and its automation technology greatly restricts the development of my country's hydrometallurgical industry. At present, it is difficult to detect the concentration of the underflow of a concentrate thickener, and the operators rely on production experience to draw ore, resulting in sharp fluctuations in the production of the downstream filter press process, and it is difficult to meet the moisture content of the filter cake product. The tailings thickener is controlled by the experience of the operator, which is highly random. If the optimal control can be done well, the pressure of the tailings pond will be reduced and the production efficiency will be improved. [0002] Due to ...

AI Technical Summary

Problems solved by technology

At present, most of the following problems exist in the application process of thickeners in my country's mineral processing plants: many key variables in the production process have not yet been detected online; the production process of thickeners is still in a state of manual operation, and most production personnel rely on their own experience and feelings to judge Production conditions, so as to operate; the work load of the thickener, the concentration of the underflow water, the turbidity of the overflow water and other key links cannot be effectively controlled, resulting in relatively large fluctuations in its concentration and flow rate, which will affect the production indicators of the subsequent beneficiation process At the same time, it is likely to increase the dosage of chemicals in the subsequent flotation process, increase the cost of ore dressing, and seriously restrict the improvement of the production efficiency of the thickener

[0005] The advantage of mechanism modeling is that it reflects the law of the process, high reliability, good extrapolation, and interpretability. The disadvantage is that the modeling process is cumbersome and depends on prior knowledge. For some complex processes, reasonable assumptions are required. A simplified mechanism model of the controlled process is obtained, however, the accuracy of the simplified mechanism model cannot be guaranteed

The advantage of data modeling is that the process model can be directly established according to the input and output data of the process, without prior knowledge of the process object, and the analysis of complex mechanisms is avoided. The disadvantage is that the model has poor generalization performance, is not interpretable, and is prone to overfitting. fit phenomenon, and may even fit noise into it, resulting in instability of the model

Images