摘要
提出一种基于操作轨迹LPV模型的非线性辨识方法:根据调度变量的操作轨迹,选取若干个典型工作点;在各个典型工作点,进行测试与辨识相应的线性模型;然后,根据工作点测试数据以及各工作点间的过渡数据,辨识出全局插值LPV模型.它降低了测试辨识成本,且算法简单可靠,仿真结果表明了该算法的有效性.
Nonlinear process identification based on operating-trajectory LPV model is proposed. Firstly, typical working-points are selected based on operating-trajectory of scheduling variable. Secondly, linear models are identified using data sets at various working-points exclusive transition data; then the LPV model is identified by interpolating the linear models using total data. The method uses low cost tests and is numerically simple and reliable. Simulation study will be used to verify the effectiveness of the method.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
2008年第7期155-159,共5页
Systems Engineering-Theory & Practice
基金
国家自然科学基金(60503065)
国家科技支撑计划项目(2007BAF22B05)
关键词
非线性辨识
线性时变参数模型
调度变量
nonlinear process identification
linear parameter varying model（LPV）
scheduling variable
