中文摘要：

分解率梯度是铝酸钠溶液碳酸化分解过程中的重要工艺指标．对产品质量和产量有直接的影响，必须对操作参数进行优化设定，保证分解率梯度达到要求。在深入分析碳分过程机理的基础上，基于现场检测数据，综合考虑影响分解率的因素，应用主成分分析法（PCA）对影响因素进行降维处理，采用神经Y网络建立了分解率梯度模型，提高了模型的精度和泛化能力。通过调整影响分解率的操作参数．使中间槽分解率满足约束范围，末槽分解率达到最大值。采用遗传算法获取操作参数的优化设定值。

英文摘要：

The decomposition ratio gradient in the process of continual carbonation precipitation of sodium aluminate solutions determines the outputs and their quality, which is the important process index. It is necessary to optimize the operating parameters to make decomposition ratio are met. On the basis of analyzing the decomposition mechanism of the ACCDP and the measuring data from industrial field, the model of decomposition ratio is build using neural network. The variables are reduced by principal component （PC1A） which makes the model more accurate and reasonable. By adjusting the operating parameters which influence the decomposition ratio to make sure the ratio of intermediate tank in within the target range, and the ratio of the last tank is maximum. The Genetic Algorithm（GA） is used to get the optimization settings.