Glass furnace temperature intelligent prediction control method based on attention mechanism and auto-encoder

Abstract

The invention provides a glass furnace temperature intelligent prediction control method based on an attention mechanism and an auto-encoder. The glass furnace temperature intelligent prediction control method comprises the following steps: collecting production historical data related to preprocessing glass furnace temperature prediction control; according to an input variable obtained through preprocessing, using the attention mechanism to obtain the input variable with an attention weight of each smelting furnace temperature at each moment; obtaining two representation vectors including relative Euclidean distance and cosine similarity according to the reconstruction error of the depth auto-encoder, and forming a final low-dimensional representation in combination with potential representations generated by an encoder in the depth auto-encoder; according to the low-dimensional representation, using an LSTM prediction model to obtain predicted values of the temperature of the smelting furnace in a plurality of subsequent time steps; according to a control mode of combining the LSTM prediction model and a statistical strategy, intelligently adjusting the natural gas flow and the oxygen flow of the glass furnace on line, so that the temperature fluctuation of the glass furnace is intelligently controlled, the glass product quality is improved, and the energy consumption is reduced.

Description

technical field [0001] The invention relates to the technical field of automatic temperature control of a glass furnace melting pool, in particular to an intelligent prediction and control method for glass furnace temperature based on an attention mechanism and an autoencoder. Background technique [0002] The temperature control effect of the glass furnace melting pool is directly related to the quality of glass products, which in turn affects the yield of glass products. Therefore, the stability of glass furnace temperature control is very important. At present, in glass production, the furnace temperature control has been generally controlled by computer, but the most commonly used control method is still ordinary PID control (including single loop, cascade loop and split range control, etc., all of which use PID as the basic control algorithm. ), some improved methods include Smith predictive compensation plus PID control and fuzzy control. [0003] The analysis of the...

AI Technical Summary

Problems solved by technology

Moreover, during the operation of the kiln, it may also be affected by various disturbance factors such as fluctuations in gas pressure, fluctuations in feed quality, changes in the insulation performance of the kiln, and changes in the temperature of the working environment.

In view of the above-mentioned characteristics of the glass furnace, the current control method is difficult to meet the high performance requirements for the temperature control of the furnace

For example, simple PID control is not effective in controlling kilns with large lag, it is difficult to adapt to changes in object parameters, and there is a contradiction between rapidity and overshoot

The design of the Smith control method relies too much on the precise model of the controlled object, and the control performance deteriorates significantly when the parameters change or the model error is large.

Images