无模型自适应控制参数输入方法改进及在气体分馏装置操作调整中的应用

摘要

无模型自适应控制（MFAC）不依赖受控系统的机理模型，只利用其输入输出（I/O）数据实现控制，近年来得到快速发展。本文用拟牛顿算法（BFGS）改进了MIMO系统MFAC算法的参数输入方法，以减少达到稳定操作的计算次数和调整时间，某30×104 t·a−1气体分馏装置的应用表明，面对操作波动，改进后的MFAC能较快地实现操作稳定和质量达标；且由于不超调，塔底再沸和塔顶冷却负荷小，因而能耗低。%In recent years, model-free adaptive control (MFAC) harvests a rapid development because it does not require the mathematical model and only uses the input/output data of the controlled system. In this paper the Quasi-Newton Methods (BFGS) was used to revamp the parameter input of MFAC of MIMO systems, resulting in less iterative calculating number of times and adjusting time. The application in a gas fractionation unit with an annual processing capacity of 0.3 Mt showed that MFAC can reach propylene quality target and steady operation rapidly with less energy consumption due to the minimum heating and cooling duties when facing operation fluctuation.