SmartUQ and Convergence Science
This study optimizes a diesel engine operating point using an analytical model constructed by sequential design of experiments (DOEs) and then compares the results to a genetic algorithm (GA). Additionally, the analytical model quantifies the uncertainty of the diesel engine operating point.
The DOE method found an optimum with 0.401% improvement in merit and 0.429% improvement in gISFC over the baseline case. The GA method found a slightly better optimum with 0.454% improvement in merit and gISFC, but the DOE method would take a week to run where the GA method would take about two months.
Additionally, the DOE method provided added value in understanding the uncertainty surrounding the optimum point.
Read the full article here.