Efficient Modular Modeling of Interconnected Dynamic Systems



Modeling and numerical simulation of dynamic systems represented as ordinary differential equations is easily done through in-house developed and commercially available integration routines (e.g. MATLAB). Therefore, for most common industrial systems, modeling is not a technical challenge any longer. However, modeling does consume resources. In today's cost conscious industry, the challenge is how to efficiently build models. This typically results in building libraries of "re-usable", "modular" components.

One of the issues related to building modular components is that the resultant dynamic equations of inter-connected systems are dependent on the architecture of the connections between the modular components. This implies that component models need to be able to comprehend the different types of interactions that they can have when integrated into a system.

We will discuss this issue for a class of nonlinear systems affine in the input, and illustrate the proposed methodology on a practical problem of interest in the auto industry.