论文部分内容阅读
本文针对一个大型炼油蒸馏塔的控制,设计了一种适用于具有多个确定性扰动对象的多输入预报自校正控制器,其特点是综合了在线辨识、多输入前馈、最小方差和PI规律各自的优点,并采用变遗忘因子的最小二乘估计算法在线修正参数,使控制器的收敛性、稳定性与自适应能力较一般自校正控制器大为提高.实践表明,该控制器很好地克服了生产中的干扰和波动,平稳了操作,成功地实现了对蒸馏塔的产品质量控制.
Aiming at the control of a large oil refining distillation column, a multi-input forecasting self-tuning controller with multiple perturbed objects is designed, which is characterized by online identification, multi-input feedforward, minimum variance and PI rule Their own advantages, and the use of forgetting factor of the least squares estimation algorithm online correction parameters, the controller’s convergence, stability and self-adaptive ability than the general self-tuning controller greatly improved. Practice shows that the controller is very good Overcome the production of the interference and fluctuations, smooth operation, the successful realization of the distillation column product quality control.