基于主元相似度的间歇过程操作曲线递推优化

发布时间：2018-02-26 01:14

本文关键词： 间歇过程操作轨迹优化主元分析余弦相似度递推算法　出处：《化工学报》2017年07期 　论文类型：期刊论文

【摘要】：针对一类较长周期的间歇过程操作优化问题,提出了一种基于正常运行批次的数据驱动型操作曲线递推优化方法。首先采用分段离散化方法将原非线性优化问题转化为线性优化问题,再利用主元分析对离散化后的高维时段变量进行降维处理,然后在降维后的主元平面中,基于时段变量与最终产品指标间的余弦相似度,实现对原操作曲线的摄动优化。考虑到时段变量方差和相似度随批次会发生变化,建立了递推算法以实现操作曲线的递推更新。最后将该方法应用于某化工产品的间歇结晶过程中,结果验证了所提方法的有效性。
[Abstract]:For a class of long-period batch process operation optimization problems, A recursive optimization method of data-driven operation curve based on normal operation batch is proposed. Firstly, the original nonlinear optimization problem is transformed into a linear optimization problem by using piecewise discretization method. Then using principal component analysis to reduce the dimension of the discretized high-dimensional time interval variables, then in the reduced principal component plane, based on the cosine similarity between the time period variables and the final product indicators, The perturbation optimization of the original operation curve is realized. Considering that the variance and similarity of the period variables vary with the batch, A recursive algorithm is established to update the operation curve recursively. Finally, the method is applied to the batch crystallization process of a chemical product, and the effectiveness of the proposed method is verified.
【作者单位】：江南大学轻工过程先进控制教育部重点实验室;
【基金】：国家自然科学基金项目(61473137)~~
【分类号】：TQ02

【相似文献】