基于混合模型的间歇生产过程运行优化

Nowadays, model-based optimization is the predominant methodology on optimal control of batch processes. However, due to the complexity of the practical industrial processes, the models are often not well-matched and the off-line optimization results are difficult to implement. Based on the hybrid model and real-time optimization technique, this project proposes operational optimization for batch processes. The main research contents include: (1) By studying the effect of outlier on the hybrid model, a parallel hybrid model of batch process is established adopting robust learning algorithm and support vector regression; (2) Aim at the difficulties of applying the hybrid model constructed based on the offline history data to adapt the time variant characteristic of the process and the blind correction of the traditional correction method, an intelligent correction approach of hybrid model based on model performance assessment is proposed; (3) To solve the problem of the mismatch of the necessary conditions of optimality of the plant and model, the gradient information is used to compensate the objective function and constraints of the optimization model. Then simultaneous method and sequential quadratic programming are used to solve the optimization model; (4) The proposed methods are validated and improved on the simulation platform and gradually applied to 1-2 typical batch industrial processes. The project will help to improve the product quality of batch process.

基于过程模型的优化方法是目前间歇过程优化控制研究的主流方法，然而由于实际生产过程的复杂性，往往导致模型失配及优化结果难以实施等问题。本项目基于混合建模思想，利用实时优化技术开展间歇过程运行优化研究。主要研究内容包括：（1）通过研究离群点对混合模型的影响，利用鲁棒学习和支持向量机方法建立间歇过程的并联混合模型；（2）针对基于历史数据离线建立的混合模型难以适应过程变化和目前校正方法存在一定盲目性的问题，提出基于模型性能评价的智能校正方法；（3）利用梯度信息在线补偿优化模型的目标函数和约束条件，解决过程-模型最优性条件不匹配问题，然后通过联立算法和序列二次规划方法求解优化问题；（4）将所提出的相关方法在计算机仿真平台上进行验证与完善，并逐步将其推广到1-2个典型间歇生产过程。通过本课题的研究，对提高间歇过程产品的质量具有指导意义。

针对间歇生产过程特性以及优化运行中需要解决的几个问题，本项目基于混合建模思想，提出了一种实现间歇生产过程运行优化的方案。1.间歇生产过程混合模型建立。基于并行结构，提出间歇生产过程混合建模的框架，其中数据补偿模型利用高斯混合回归模型建立。同时为了降低模型运算复杂度提高预测效率，项目采用灵敏度分析为模型选取合适的辅助变量。2.间歇过程混合模型在线校正。提出一种基于模型性能评价的混合校正策略。利用高斯混合回归模型中条件方差评价混合模型性能，在此基础上，对引起模型性能下降的过程变化特性进行监测，智能选择模型校正方法，避免因盲目校正导致的混合模型性能下降。3.基于混合模型的间歇生产过程运行优化。利用间歇过程批次间重复特性，研究批次间优化控制策略。为了提高优化问题计算，提出了一种改进的混合组搜索智能优化算法。针对过程与模型存在不匹配的问题，提出一种基于数据的生产指标非线性迭代补偿方法。4.设计开发湿法冶金合成过程仿真实验平台。所搭建的湿法冶金合成过程仿真实验平台可以模拟实际生产过程并验证项目所提出的优化控制方法。