Software for Modeling and Simulating Control Systems


Software for the simulation of continuous-time systems was first standardized in 1967. The standard committee consisted largely of control engineers. Thus, one would expect that today's simulation languages for continuous system simulation should be particularly well suited for modeling and simulating control systems. This article answers the question of whether this expectation holds true or not.

There has always been a strong link between the control and simulation communities. On the one hand, simulation is an extremely important tool for all and every control engineer who is doing practical control system design in industry. For arbitrarily nonlinear plants, there is often no alternative to designing controllers by means of trial and error, using computer simulation. Thus, there is hardly any control engineer who wouldn't be using simulation at least occasionally. On the other hand, although simulation can be (and has been) applied to virtually all fields of science and engineering (and some others as well), control engineers have always been among the most cherished of its customers - after all, they have paid the butter on the bread of many a simulation software designer for years. Moreover, a good number of today's simulation researchers received their graduate education in control engineering.

There exist on the market many highly successful special-purpose simulation software tools, e.g. for the simulation of electronic circuitry, or for the simulation of multibody system dynamics, and there is (or at least used to be) a good reason for that. However, there is no market to speak of for special-purpose control system simulators, in spite of the fact that control is such an important application of simulation. The reason for this seeming discrepancy is that control systems contain not only a controller, but also a plant, which can be basically anything. Thus, a simulation tool that is able to simulate control systems must basically be able to simulate pretty much anything.

Hence a substantial portion of this article shall be devoted to a discussion of general-purpose simulation software. Yet, control systems do call for a number of special features in a simulation tool, and these features shall be pointed out explicitly.

This article is structured in three parts. In a first section, the special demands of control systems to a general-purpose simulation tool are outlined. In a second part, the article classifies the existing modeling and simulation tools and mention a few of them explicitly. The article ends with a critical discussion of some of the shortcomings of the currently available simulation tools for modeling and simulating control systems.

This article is written with several different customer populations in mind. It should be useful reading for the average practical control engineer who needs to decide which simulation tool to acquire and, maybe even more importantly, what questions to ask when talking to a simulation vendor. It should, however, also be useful for simulation software vendors who wish to upgrade their tools to better satisfy the needs of an important subset of their customer base, namely the control engineers. It should finally appeal to the simulation research community by presenting a state-of-the-art picture of what has been accomplished in control system simulation so far, and where some of the still unresolved issues are that might be meaningful to address in the future.

Interested in reading the full paper? (24 pages, 556987 bytes, postscript)