模型不确定条件下预测控制经济性能评估的研究

摘要

为了解决在模型不确定条件下的预测控制系统经济性能评估分析的问题,本文通过基于二次锥规划的鲁棒线性规划的方法来描述模型不确定性对控制系统经济性能评估造成的影响,并采用约束调整与方差调整的策略来改善控制系统的经济性能.Shell公司提供的重油分馏塔典型案例实验证明该方法的有效性.%In order to solve the problem in evaluating economic performance for predictive control systems in the case of model uncertainty appropriately a robust linear programme method is proposed based on a second-order cone programme to describe the affect on economic performance assessment due to the model uncertainty.Then some constraint tuning and variaility tuning are done to improve the economic performance.Simulation results of a typical control problem of a heavy oil fractionator proposed by Royal Dutch/Shell Group showed the effectivity of the proposed algorithm.