Surrogate-Assisted Compressor Geometry Optimization for Fuel Economy and Emissions Trade-Off in a Turbocharged Diesel Engine

Turbocharger compressor geometry affects the pressure ratio, flow range, and efficiency distribution of the compressor map, but the influence pathway from compressor geometry to engine fuel economy and emissions trade-off has not been sufficiently quantified. This study develops a surrogate-assisted multi-objective optimization framework that couples compressor and engine models. A calibrated one-dimensional compressor model generated a geometry-specific map for each design, which was imported into a validated one-dimensional engine model to calculate brake-specific fuel consumption and brake-specific NOx emissions at four loads. The results for the four loads were aggregated into weighted brake-specific fuel consumption (BSFCw) and weighted brake-specific NOx emissions (BSNOxw). Extreme gradient boosting (XGBoost) surrogate models are trained to predict these weighted responses. The non-dominated sorting genetic algorithm II (NSGA-II) identified Pareto-optimal compressor geometries. Representative solutions were re-evaluated using the original coupled models. The Pareto front reveals a clear trade-off between BSFCw and BSNOxw, with the minimum-BSFCw and minimum-BSNOxw solutions defining the respective lower bounds within the investigated design space. For the selected trade-off solution, BSNOxw was 1.18 g/(kW·h) lower than the baseline value, while BSFCw was 0.82 g/(kW·h) higher. Compressor geometry alters the distribution of engine operating points on the compressor map under different loads, thereby affecting the trade-off between fuel economy and NOx emissions.