High Dynamic Torque Control for Industrial Engine Test Beds

In this paper, an approach for high dynamic torque control of an industrial engine test bed is proposed. The main goal is to develop a control concept that is capable of periodic reference tracking at high control bandwidth while at the same time sufficiently rejecting disturbances coming from external excitations. For that reason, model predictive control (MPC) is utilized which offers two major benefits as compared to classical controllers. On the one hand, its capability to explicitly consider constraints can be applied to limitations of the actuating variable on the test bed. On the other hand, it can also be used to increase control bandwidth by anticipating future reference trajectories. For MPC, an accurate test bed model is of vital importance. Therefore, a system model is developed and validated using measurements obtained from the engine test bed. Moreover, a robustifying approach is proposed based on an extended system model to overcome hardly observable, weakly damped engine block suspension dynamics which are difficult to model. The proposed control concept is discussed and validated through simulations as well as through experimental results.