Iterative learning from suppressing vibrations in construction machinery using magnetorheological dampers

Abstract The aim of the paper is to propose algorithms for suppressing vibrations of massive components of machinery, e.g., operators' cabins in bucket wheel excavators (BWEs). Repetitive perturbations provide the opportunity to learn a control system to suppress or reduce their influence by applying a repetitive control scheme. A new version of the iterative learning control approach, called iterative learning of optimal control (ILOC), is derived. This approach is able to learn a control signal that approximately minimizes the squared acceleration of a system, e.g., the cabin. The resulting algorithms can be used alone or in parallel with a hybrid version of the classic proportional-derivative (PD) controller. The simulations indicate that the ILOC algorithm provides a large improvement in suppressing quasi-periodic disturbances in comparison to the cases of no control and PD control, while the hybrid version provides even slightly better results. The proposed algorithms are intended to be used with magneto-rheological (MR) dampers as semi-active control devices. Subtle properties of MR dampers, such as their hysteresis, are omitted here.