Systems Level Trade Studies with CAE Data

This paper describes the method of performing systems level architecture trade studies using data from high fidelity CAE calculations. The method is applied to a SysML model of an IRS Rear Axle driveline with the aim of accurately predicting the lowest possible weight of the half shafts for a driveline configuration while considering complex requirements. First the parametric equations defined in the SysML model that estimate the impact torque are solved. Next a response surface model is created that takes the impact torque as an input and outputs the optimal weight of the half shaft axle. In order to do this, a CAE model of the half shaft axle is developed as an automated workflow. It uses a parametrized CAD model to define the geometry which is then evaluated for maximum stress and deflection with an FEA solver. The performance of all possible axle geometries is sampled using a large DOE and approximated using an RSM. An optimization workflow is then created that uses this RSM to find the optimal geometry for a specified torque load. This workflow is evaluated for a range of torque values and is in turn approximated using an RSM to create a single executable component that provides optimal half shaft axle weight as a function of applied torque. This axle performance component is connected to the SysML model together with a catalog of available CV Joints to enable the study of tradeoffs between weight and cost for the complete half shaft. The SysML model also integrates a requirement that checks if a chosen axle geometry can fit into a housing defined in CAD. Finally, the paper demonstrates optimization of the half shaft under complex constraints using the complete model.