Time-weighted MR for switched LPV systems via balanced realisation
Time-weighted MR for switched LPV systems via balanced realisation
复制标题
DOI：
10.1049/iet-cta.2017.0481
复制
发表时间：
2017-08
影响因子：
2.6
通讯作者：
Huiyan Zhang;C. Lim;Ligang Wu;P. Shi
中科院分区：
文献类型：
--
作者：
Huiyan Zhang;C. Lim;Ligang Wu;P. Shi
针对连续时间切换线性变参系统，提出了一种基于时间加权谱的模型约简方法。利用时间加权的能量函数和参数无关的多重Lyapunov函数，得到了时间加权的可控性和可观测性格律。通过最优导出的平衡变换矩阵，作者证明了原始切换LPV系统可以实现平衡。通过状态截断或奇异摄动技术得到了一系列不同权重因子的低阶模型。降阶模型能保持系统的稳定性。最后以一个三弹簧-质量系统为例说明了该方法的可行性和有效性。
A new model reduction (MR) method is developed for continuous-time switched linear parameter-varying (LPV) systems based on time-weighted gramians. Using time-weighted energy functionals and parameter-independent multiple Lyapunov functions, the time-weighted controllability and observability gramians are obtained. With a balanced transformation matrix optimally derived, the authors show that a balanced realisation can be achieved for the original switched LPV systems. A series of low-order models for different weighting factors are yielded by state truncation or singular perturbation techniques. The reduced-order models can preserve the stability. The feasibility and effectiveness of the proposed MR method are illustrated by two examples including a three spring-mass system.