Development of a Modelling Tool For Robustness Analysis of Embedded Powertrain Control System

Abstract: Robustness analysis of an Embedded control system is a crucial test stage in quantifying the quality of the system and proposing further improvements on it. In this thesis work a modelling tool for robustness analysis of embedded powertrain control system of truck and bus is developed on top of an in house SIL simulation environment. The aim of the tool is to induce certain degradations modelled as different distributions in the allowed degradation range of sensor signals in the control system. Another goal of the tool is to induce faults into the control system and trigger certain fault handling states. In both the cases, some output signals from the system is recorded and their behavior is analyzed. A custom drive cycle called mini-WHTC was developed and was executed on both cold and warm system conditions. The faults were modelled as structured distribution and random distributions like normal distribution, uniform distribution and custom distributions. The random distributions were sampled from allowed degradation range of sensors using Latin hypercube sampling methodology. On obtaining the results, the following conclusions could be drawn. The control system exhibited higher sensitivity to input signal deviations during the cold cycles compared to the warm cycles. That is the system showed higher degree of robustness in the warm cycle. It was also observed that on running simulations for longer drive cycles, the powertrain control system showed higher degree of adaptations and stabilizations. Finally, it was observed that inducing faults into sensor signals one at a time does not degrade the functionality of the system as a whole.