An online prediction system and method for blast furnace molten iron quality based on multivariate online sequential extreme learning machine

Abstract

An online forecasting system and forecasting method for blast furnace hot metal quality based on a multivariate online sequential extreme learning machine. The forecasting system consists of a conventional measurement system, a data collector, M‑OS‑ELM online forecasting software and a computer system running the software; The forecasting method includes the following steps: (1) auxiliary variable selection and model input variable determination; (2) training and use of the M‑SVR soft sensor model. This invention uses the online process data provided by the conventional detection system and is based on M-OS-ELM intelligent modeling technology to establish a multi-element molten iron quality prediction model with output self-feedback and taking into account the input and output timing and time lag relationships, while achieving Si content The multivariate online dynamic measurement of the four major molten iron quality indicators, P content, S content and molten iron temperature, has the characteristics of good practicability, more accurate measurement results and stronger generalization ability.

Description

technical field [0001] The present invention relates to an online prediction method of multivariate molten iron quality parameters in blast furnace ironmaking process, in particular to an online prediction method of multivariate molten iron quality parameters in blast furnace ironmaking process based on multivariate online sequential extreme learning machine (M-OS-ELM), belonging to Blast furnace smelting automation control field. Background technique [0002] A blast furnace is a large convective reactor and heat exchanger in the ironmaking process. Blast furnace ironmaking reduces iron from iron ore and other iron-containing compounds through complex gas-solid, solid-solid, and solid-liquid reactions in the furnace, and smelts qualified molten iron. As the most important production index in the blast furnace ironmaking process, the molten iron quality index directly determines the quality of subsequent steel products and the energy consumption state of the blast furnace s...

AI Technical Summary

Problems solved by technology

[0006] The method reported in the above-mentioned patents and other similar methods and technologies related to many other literatures are only for the prediction or soft measurement of a single molten iron quality element (such as molten iron temperature, Si content, S content, etc.), and fail to characterize the main parameters of blast furnace molten iron quality. That is, [Si], [P], [S] and molten iron temperature are multivariate online forecasts at the same time, so it cannot fully reflect the overall level of molten iron quality, and its practicability is poor

At the same time, because these methods do not consider the input and output timing and the time-delay relationship of the process, the established static model cannot well reflect the inherent dynamic characteristics of the blast furnace smelting process

When the smelting conditions change significantly, these methods cannot adapt to the changes in the working conditions and measure the quality parameters of molten iron more accurately

To sum up, at present, there is no multivariate dynamic prediction method for the quality parameters ([Si], [P], [S] and molten iron temperature) of molten iron in the blast furnace smelting process at home and abroad.

Images