连续碳酸化分解过程是烧结法生产氧化铝的关键工序之一,其分解率梯度与末槽分解率直接影响产品的产量和质量。针对此过程具有强非线性、强耦舍性、大时滞、难以建立精确的数学模型等特点,在分析过程机理特征和总结专家经验的基础上,提出了分解率梯度专家控制和末槽分解率预测前馈补偿相结合的控制策略,以首槽进料阀门和1#~5#槽CO2通气阀门为控制对象,采用产生式规则建立了由首槽进料量专家控制规则和1#-5#槽分解率专家控制规则共同组成的专家知识库,并引入了末槽分解率预测模型修正控制输出,最后开发了连续碳酸化分解过程分解率梯度专家控制系统。实际运行结果表明,各槽分解率的合格率提高了2%以上,实现了分解率梯度的稳定优化控制,提高了产品的质量。
The continuous carbonation decomposition problem in sintering process of alumina production is discussed, To the characteristics of such processes, an expert control strategy for decomposition ratio gradient is proposed, combined with feed-forward compensation of predictive model for the last-tank decomposition ratio, based on the experiential knowledge and the mechanic characteristics. An expert knowledge base is established by using production rule including expert rules for feed rate of the first-tank and decomposition ratios from first-tank to fifth-tank. The expert control system for decomposition ratio gradient is developed. The practical operation results show that the qualification rate of decomposition ratio in each tank is increased by 2 % , and the stable optimization control for decompo- sition ratio gradient is realized.