在总结铝电解过程中电流效率与电解工艺参数之间的关系的基础上,用模式识别的主成分分析方法对高维模式空间降维,寻找模式空间中电流效率较高的区域。然后用计算机模拟方法,从低维空间逆映照到高维空间,设计了若干组优化工艺条件。最后用人工神经网络拟合和预测这些设定的工艺条件可能达到的电流效率,从而找到了一组最优工艺条件。 On the basis of summarizing the relationship between current efficiency and electrolytic process parameters in aluminum electrolysis process, the principal component analysis method of pattern recognition is used to reduce the dimension of high-dimensional pattern space and find the area with high current efficiency in pattern space. Then computer simulation method is used to reflect from low-dimensional space to high-dimensional space, and several groups are designed to optimize the process conditions. Finally, artificial neural network is used to fit and predict the current efficiency that can be achieved by these set process conditions, so as to find a set of optimal process conditions.