A Concentrate Grade Prediction Method for Zinc Flotation Process Based on Feedback Compensation Mechanism Optimization

Abstract

The present invention provides a zinc flotation process concentrate grade prediction method optimized based on a feedback compensation mechanism, which includes the following steps: firstly collect the zinc flotation process image feature data and the corresponding concentrate grade data as sample data, and collect The preprocessed data is preprocessed; the preprocessed sample data is divided into five independent sub-sample spaces according to the concentrate grade and time difference is made respectively; KPCA is used to extract features with high contribution rate as key features; LSSVM is trained based on key feature samples , establish the relationship between image features and concentrate grade; use the improved flora algorithm to optimize the two parameters of LSSVM, the penalty factor ξ and the kernel width σ; establish a model error feedback compensation mechanism, and start feedback compensation when the model error is not controlled The mechanism compensates the model; the invention can be directly programmed on the computer, has low cost, high precision and good timeliness, and is of great significance for guiding on-site production.

Description

technical field [0001] The invention belongs to the technical field of froth flotation, and in particular relates to a zinc flotation process concentrate grade prediction method based on model error estimation and feedback compensation mechanism optimization [0002] technical background [0003] Froth flotation is one of the most important beneficiation methods in current lead-zinc smelting. Flotation method is a method of sorting minerals by using the different physical and chemical properties of the surface of mineral particles to lead to different hydrophilicity. practical value. However, because the flotation process is long, the internal mechanism is not clear, there are many influencing factors, many variables are involved and the nonlinearity is serious, many process indicators cannot be detected online. Predicting the obtained concentrate grade and using it to complete the on-site operation is highly subjective and relies too much on experience and knowledge. It is ...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the difficulty of on-line detection of concentrate grade in the flotation process and the deficiency of the prior art in zinc flotation concentrate grade prediction, the present invention utilizes various characteristics related to zinc concentrate grade to propose a method based on an improved flora algorithm Data-driven concentrate grade prediction method for zinc flotation process

Images