Co-simulation as effective method for flexible structure vibration control design validation and optimization

In control problems of complex dynamical systems the controller design is usually carried out on order-reduced and otherwise simplified models. This is done to keep the controller complexity manageably low or to ease the application of computationally expensive design methods. If the model reduction was carried out carefully, the resulting controllers will perform equally well on the original, complex system, but this has to be verified by suitable tests and simulations. In this paper, the use of co-simulation for effective and efficient verification and validation of the control design and its performance is proposed. The term co-simulation refers to the parallel simulation of complementary computation tasks, which run synchronized and exchange process data. Here we consider the simulation-based verification and optimization of a vibration control system for metro rail vehicle car bodies, so the two computation tasks are the control algorithm and a multi-body dynamics simulation. The control system uses structure strain sensors and piezoelectric actuators to introduce force / moment pairs into the car body structure. The goal is to minimize the vertical accelerations at performance measurement positions, thereby maximizing passenger ride comfort. Using co-simulation as a system validation tool, the control design methodology including actuator placement and model reduction could be validated and the actuator positioning could be optimized. This yielded outstanding improvements in passenger ride comfort for the considered class of vehicles with a state-of-the-art control approach.