Aluminum oxide comprehensive production index decision-making method based on multi-scale deep convolutional network

Abstract

The invention provides an aluminum oxide comprehensive production index decision-making method based on a multi-scale deep convolutional network, and relates to the technical field of aluminum oxide comprehensive production decision-making. The method is mainly composed of several sub-models: a multi-scale deep splicing convolutional network forecasting sub-model for reflecting the influence of abottom production process index on an alumina comprehensive production index; a full-connection neural network forecasting sub-model which reflects the influence of the upper-layer aluminum oxide scheduling index on the comprehensive aluminum oxide production index;a full-connection neural network forecasting sub-model which reflects the influence of the aluminum oxide comprehensive production index on the current aluminum oxide comprehensive production index at the past moment; and a multi-scale information neural network integration model used for cooperatively optimizing sub-model parameters. Through the integrated forecasting model structure, the memory ability of a shallow network and the feature extraction ability of a deep network are utilized at the same time, and accurate decisionmaking of aluminum oxide production indexes is achieved.

Description

technical field [0001] The invention relates to the technical field of alumina comprehensive production decision-making, in particular to a method for decision-making of alumina comprehensive production indicators based on a multi-scale deep convolutional network. Background technique [0002] Aluminum and its alloys have many excellent properties. At the same time, aluminum resources are very rich, so the aluminum industry has developed very rapidly since its inception. Large-scale equipment is conducive to the automatic detection and control of the process, and the production control and management system based on microcomputers and computers provides a huge potential for alumina plants to improve labor productivity, reduce raw material consumption and save energy. [0003] Although relevant enterprises have carried out project improvement and upgrading in terms of alumina smelting technology, there are still problems of poor raw material quality, high project energy consu...

AI Technical Summary

Problems solved by technology

[0003] Although relevant enterprises have carried out project improvement and upgrading in terms of alumina smelting technology, there are still problems of poor raw material quality, high project energy consumption and insufficient product quality. Most of the products are intermediate state alumina, which will affect the overall technical application. Management influences and restricts product structure

[0004] In the traditional alumina production process, many control indicators mainly rely on managers, dispatchers, engineers and other knowledge workers to manually set them based on experience, and the production system cannot operate under optimal conditions.

[0005] At the same time, for the large amount of data generated by alumina production, it is difficult to effectively mine the information in the large amount of data using traditional machine learning algorithms

Images