Synthesis and control of generalized dynamically substructured systems
Synthesis and control of generalized dynamically substructured systems
复制标题
DOI：
10.1243/09596518jsce635
复制
发表时间：
2009-05
期刊：
影响因子：
--
通讯作者：
D. Stoten;J. Tu;Guang Li
中科院分区：
文献类型：
--
作者：
D. Stoten;J. Tu;Guang Li
摘要通过动态子结构方法测试工程系统的实验技术正受到全球的广泛关注，例如在大型结构、航空航天和汽车系统测试领域。动态子结构系统(DSS)使完整系统的全尺寸、关键组件能够在实验室环境中实时进行物理测试，而系统的其余部分则通过数字建模。这样做的目的是让物理和数值相结合的DSS表现得就像它是完整的(或仿真的)系统一样。例如，在理想的机械决策支持系统中，需要在数值部件(或子结构)和物理部件(或子结构)之间的界面上实现位移和力的完美同步。因此，成功的DSS系统的一个关键设计特征是控制动作的高保真度。同样，DSS控制器必须能够处理物理子结构动态中的非线性、时变和不确定参数。本文的主要目的是提出一个通用的DSS框架，以及相关的线性和自适应控制策略，这些策略是专门为实现高同步性能而定制的。因此，对这个问题的初步研究，如Stoten和Hyde早先的一篇论文所描述的，通过推广DSS动力学和控制策略来继续，以包括(A)一些新定义的操作模式和(B)多变量动力学。此外，还包括基于机械系统(平面准摩托车试验台)的DSS测试的比较实现和仿真研究，该测试是专门为突出本研究的主要特征而设计的。对比研究表明，尽管物理子结构的动态特性发生了很大的变化，但DSS控制可以获得良好的控制效果，特别是在控制器中加入自适应元件。
Abstract The experimental technique for testing engineering systems via the method of dynamic substructuring is receiving significant global interest, for example in the fields of large-scale structural, aerospace, and automotive system testing. Dynamically substructured systems (DSSs) enable full-size, critical components of a complete system to be physically tested in real-time, within a laboratory environment, while the remainder of the system is modelled numerically. The intention is that the combined physical-numerical DSS behaves as if it were the complete (or emulated) system. In an ideal mechanical DSS, for example, perfect synchronization of displacements and forces at the interfaces between the numerical and physical components (or substructures) is required. Hence, a key design feature of successful DSS systems is the high fidelity of the control action. Equally, a DSS controller must be able to cope with non-linear, time-varying, and uncertain parameters within the physical substructure dynamics. The main purpose of this paper is to present a generalized DSS framework, together with associated linear and adaptive control strategies, that are specifically tailored to achieve high synchronization performance. The initial studies of this problem, as described in an earlier paper by Stoten and Hyde, are therefore continued by generalizing both the DSS dynamics and the control strategies to include (a) a number of newly defined modes of operation and (b) multivariable dynamics. In addition, comparative implementation and simulation studies are included, based upon the DSS testing of a mechanical system (a planar quasi-motorcycle rig), which was specifically designed to highlight the main features of this research. The comparative studies show that excellent DSS control can be achieved, especially with the addition of an adaptive component to the controller, despite significant changes to the physical substructure dynamics.