一种解决蕴含不确定性信息的氧化铝配料问题的智能优化方法

针对原料质量不稳定和成分检测大滞后带来的信息不确定性,提出了一种两级智能优化方法实现氧化铝配料过程中生料浆质量的优化控制.该方法通过引入中间优化同标,将优化问题分解为原料配比优化和料浆调配优化,逐步弱化不确定信息对生料浆质量的影响.配比优化基于入槽生料浆质量预测模型,设计了专家分级推理机制,实现多质量指标约束条件下的配比优化设定:调配优化将不确定的生料浆质量信息引入调配优化模型约束中,采用改进遗传算法求解最优渊配方案,配制高质量的生料浆.将提出的方法应用于国内某厂氧化铝配料过程,实现了生料浆质量指标的优化控制,简化了工艺流程,为存在信息不确定的长流程工业过程的优化控制提供了范例.

Intelligent optimization of raw material blending for alumina production with information uncertainty

Considering the uncertainty of raw material quality and time-lagging in composition measurement, a twostage intelligent optimization method is proposed to realize the optimal control of slurry quality for the raw material blending process in alumina production. By introducing an intermediate optimization objective, the blending optimization problem is decomposed into two stages, i.e. the optimization of the mixture ratio and the optimization of slurry combination, to reduce the effect of uncertainty step by step. In the mixture-ratio optimization, an expert hierarchical reasoning strategy based on the quality prediction model is proposed to optimize the mixture ratio with multi-index constraints. Then, an optimal combination model with uncertainty is built by incorporating the uncertainty of raw slurry quality into constraints, and an improved genetic algorithm is used to solve it. The proposed approach has been applied to the blending process of an alumina factory of China, and the optimal control of slurry quality is realized and the blending process is simplified.