Ore grinding granularity online prediction system and method

Abstract

The invention provides an ore grinding granularity online prediction system and method, and relates to the technical field of ore grinding. The ore grinding granularity real-time online prediction system receives ore feeding amount, grinder concentration, classifier overflow concentration, pump pool concentration and swirler ore feeding pressure which are output by an ore grinding loop control device; ore grinding granularity is online predicted; an ore grinding granularity optimal setting control device receives an ore grinding granularity prediction result and sends control signals of the new ore feeding amount, the new grinder concentration, the new classifier overflow concentration, the new pump pool concentration and the new swirler ore feeding pressure; the ore grinding loop control device conducts closed-loop control. Through the solution of an online ore grinding granularity prediction mechanism model, response speed of ore grinding granularity prediction is improved on the premise that calculation accuracy of the model is not reduced, and the real-time performance of prediction is improved. The ore grinding granularity online prediction system and method are used for online estimating ore grinding granularity indexes, and closed-loop control of the ore grinding granularity is achieved. The change of the ore grinding granularity in a period in future can be predicted according to fluctuation change of other production factors, and therefore real-time optimization of the ore grinding granularity is achieved.

Description

technical field [0001] The invention relates to the technical field of ore grinding, in particular to an online prediction system and method for ore grinding particle size. Background technique [0002] The ore grinding process is a crucial link in the beneficiation production process. Its main task is to fully grind and classify the ore so that its particle size meets the relevant process indicators, which is beneficial to the subsequent separation processes such as magnetic separation and flotation. . The quality of grinding products is mainly measured by the distribution of particle size. Grinding particle size is the most important quality index in the grinding process. It has an important impact on the next separation process (such as magnetic separation, flotation, etc.) and the quality of the entire mineral processing product. It needs to be strictly monitored and controlled. It determines the grade of the final product and metal recovery. [0003] The usual detect...

AI Technical Summary

Problems solved by technology

The traditional forecasting model is mainly obtained by the method of process identification. Such a model is easy to operate, but the performance of the model is low.

Mainly manifested in 1) the prediction time is very short: only one forecast period after the current period can be predicted (that is, "single-step forecast"); while the granularity change process is slow and the lag is large, so the single-step forecast cannot be used for closed-loop control of granularity

2) The generalization ability of the model is poor: it can only be applied to predict the vicinity of the steady-state operating point. When the working conditions and boundary conditions fluctuate greatly, the prediction accuracy is not ideal

However, it also has disadvantages, especially the high-precision grinding particle size prediction model adopts a complex partial differential-integral equation system, and its solution process consumes a lot of computing resources and takes a long time, resulting in a lag in prediction results, which is difficult to meet the real-time requirements of online applications. sexual needs

Images