Recently, the first domestically developed intelligent optimization control system for alumina calcination, independently developed by Shenyang Aluminum and Magnesium Engineering and Research Institute Company Limited (SAMI) of Chinalco International, has been put into operation at Shanxi Xinfa Chemical Co., Ltd. This system fills the technical gap in the field of intelligent control of high-temperature calcination processes in China.

It is reported that the alumina calcination process can be regarded as the "high-temperature black box" of the manufacturing industry. Key parameters such as the residence time of materials in the furnace and the temperature field distribution have long relied on manual experience for adjustment, resulting in large fluctuations in product quality and high energy consumption. This system innovatively introduces non-stationary time series causal tracing analysis technology and develops the first domestic reinforcement learning control model, realizing autonomous adjustment of the calcination production process and solving the industry pain point that traditional control methods struggle to cope with sudden changes in working conditions.

It is understood that during the full-load trial operation, the system withstood the extreme working condition of significant fluctuations in gas pressure, controlling the temperature fluctuations of the calciner within ±2% of the target value, which is ahead of the international level of ±5%. This improves the overall thermal efficiency and reduces the consumption of system energy sources such as gas. Production data shows that after the system was put into use, the thermal consumption per ton of alumina decreased by 3.2%. Calculated based on the enterprise′s annual production capacity of 3 million tons, the annual production cost savings exceed 10 million yuan.