摘要： 摘 要： 针对工业过程中由于时延问题造成软测量模型预测精度不高的问题，给出了一种通过差分进化算法估计时延的动态软测量方法。通过偏最小二乘法建立合适的适应度函数，将软测量系统的时延参数估计问题转化为一个多维非线性优化问题，然后利用差分进化算法的全局搜索能力求解该优化问题。针对PTA精制过程中的PTA平均粒径大小建模研究，结果表明，时延参数估计的引入大大提高了软测量模型的预测精度，证实了所提方法的有效性和可行性。

中图分类号： TP274
文献标识码： A
DOI： 10.19358/j.issn.2096-5133.2022.02.008
引用格式： 郭储磊. 基于差分进化算法的PTA过程平均粒径的动态软测量[J].信息技术与网络安全，2022，41(2)：46-52.

Dynamic soft sensor of average particle size in PTA process based on differential evolution algorithm

Abstract： Aiming at the problem that the prediction accuracy of soft sensor model is not high due to the delay problem in industrial process, a dynamic soft sensor method using differential evolution algorithm to estimate the delay is proposed. The problem of time delay estimation of soft measurement system is transformed into a multi-dimensional nonlinear optimization problem by establishing a suitable fitness function by partial least square method, and then the global search ability of differential evolution algorithm is used to solve the optimization problem. Aiming at the modeling of PTA mean particle size in PTA refining process, the results show that the introduction of delay parameter estimation greatly improves the prediction accuracy of soft sensor model, and the validity and feasibility of the proposed method are verified.

Key words : soft sensor; differential evolution; time delay estimation; partial least squares