Numerical Linearisation Method to Efficiently Optimise the Oscillation Damping of an Interdisciplinary System Model

Abstract

This paper presents a method to numerically linearise a complex mechanicalsystem. The method is applied to the parameter optimisation of a controlled elastic hydro-mechanical system. The virtual spring damper control concept is used to damp the oscillating system. The system dynamics of the full system are modelled as an interdisciplinary model. From this a mechanical substitution model is derived and linearised at a given operating point. The system matrix of the linearised system is used to efficiently calculate an optimisation criterion that leads to optimal control parameters for rapid oscillation damping. To verify the results a classical, but very time consuming, optimisation is performed. The quality of the optimal parameter sets found is evaluated by comparing simulation results of the complex interdisciplinary model. The system response is almost the same in both cases. The parameters found using the linear model are only slightly different to the parameters found using the non-linear and interdisciplinary model. The performance of the linear model based optimisation is good – it reduces the computational cost by a factor of more than one hundred.