Abstract
SUMMARY: Dynamically substructured system (DSS) techniques separate critical components of a complete structural system to be physically tested in full size; the remaining linear subsystems are tested numerically. Successful and robust DSS tests rely on a high-quality controller to cope with undesired disturbances surrounding the real-time environment and thus ensure synchronised responses of the numerical and physical substructure outputs. Three DSS control systems are compared in this paper, which use dynamics-based ordinary differential equations or geometry-based delay differential equations to model the systems. Even though the control designs are not new, a series of new experimental and analytical results capture the essence of DSS control problems in a simple way, showing that (i) reliable DSS tests depend on well-defined dynamics and numerical computational accuracy in the control design, (ii) dynamics-based methods lay a relatively transparent and systematic foundation for deeper investigation into robustness issues and (iii) an understanding of potential and fundamental real-time difficulties is important in order to give hints for accurate modelling, control redesign and quality improvement. © 2014 John Wiley & Sons, Ltd.