Paper Authors

Abstract

NOTE: The first page of text has been automatically extracted and included below in lieu of an abstract

Session: Instrumentation Design and Applications 2259

Experimental Model-Based Control Design Using Multibody Codes Raffaello D’Andrea Mechanical and Aerospace Engineering 218 Upson Hall, Cornell University Ithaca, NY 14853 rd28@cornell.edu www.mae.cornell.edu/raff

Abstract In this paper we discuss an on-going project at Cornell University aimed at introducing a significant graphical simulation component in the dynamics and control curriculum, and to expose the students to the interplay between simulations and experiments 1 . This is being achieved by incorporating control experiments from Quanser Consulting, MATLAB control software from the Mathworks, and the Working Model 2D and 3D multibody code software from Knowledge Revolution. The benefits of this approach are both economical and pedagogical; only a limited number of control experiments needs to be purchased and maintained, and it exposes to the students to computer simulation and the relationship between simulations and reality.

1 Introduction Due to the great advances in computing power, simulation has become an integral part of most engineering disciplines. It is often easier and cheaper to simulate a physical system than to build a working prototype; even if this is not the case, the incremental costs of simulating an altered version of the system are small compared to the costs of modifying a physical prototype. Since the design process is iterative, it does not take many design cycles before the cost of prototyping overtakes the cost of simulation. There is thus a trend to move as much of the design cycle as possible to a simulated environment. This is certainly true for the design of most high-performance control systems. For many control appli- cations, the design cycle consists of obtaining high fidelity models of the system to be controlled, performing a control design based on reduced order models (for example), verifying the control design on the simula- tion model, altering the design until the required level of performance is achieved, and then testing out the design on the physical system. A typical example of this process is flight control design. In fact, for these applications there are often different levels of simulation employed, since obtaining simulation time on high fidelity simulators is relatively expensive. It should be stressed, however, that simulation is only a tool in the design process, and that extensive experimental verification should always be the end result of the design process; the real world is simply too complex to be fully simulated. Thus, an important goal in engineering education is to expose the students to the obvious advantages of simulation, but at the same time to make strong connections between com- puter simulations and the real devices which are being simulated. The proliferation of powerful computers in education and at home, coupled with the availability of many affordable software packages for control design and dynamic simulation (or multibody simulation), makes it possible to introduce a significant graph- ical simulation component into the dynamics and control education. In particular, including a substantial graphical simulation component results in the following benefits: 1 This project is being supported, in part, by the National Science Foundation, Grant No. DUE-9851406

1

• Citation

• Format
  • APA
  • APA - LaTeX bibitem
  • MLA
  • MLA - LaTeX bibitem
  • Bibtex
  • EndNote - RIS

D'Andrea, R. (1999, June), Experimental Model Based Control Design Using Multibody Codes Paper presented at 1999 Annual Conference, Charlotte, North Carolina. 10.18260/1-2--7662