High-Dimensional Sensitivity Analysis Applied at Vehicle Component and System Level in the Context of CO ₂ Exhaust Emissions

This paper demonstrates how the validation and verification phase of prototype development can be simplified through the application of the Model Development Suite (MoDS) software by integrating advanced statistical and numerical techniques. The authors have developed and present new numerical and software integration methods to support a) automated model parameter estimation (model calibration) with respect to experimental data and, b) automated global sensitivity analysis through using a High Dimensional Model Representation (HDMR).

These methods are demonstrated at 1) a component level by performing systematic parameter estimation of various friction models for heavy-duty IC engine applications, 2) at a sub-component level by performing a parameter estimation for an engine performance model, and 3) at a system level for evaluating fuel efficiency losses (and CO₂ sources) in a vehicle model over 160 ‘real-world’ and legislated drive cycles.