Prediction of emissions and performance from transient driving cycles using stationary conditions: Study of advanced biofuels under the ETC test

Abstract

This paper applies and improves a methodology for estimating engine responses from transient cycles using steady conditions according to a Design of Experiments (DoE). The fuels tested are diesel-farnesane, biodiesel from sugarcane, and diesel fuel S50. A common-rail engine and the European Transient Cycle (ETC) are considered. Two DoEs of 13 runs each were analysed: the 13 modes of the European Stationary Cycle (ESCDoE) and a 5-level Fractional Factorial Design (FFDoE). The mathematical transformation of the engine working region and the experimental data approximation were improved using chord length parameterization and tensor product surfaces, respectively. Both DoEs provide an instantaneous approximation of engine performance responses of high accuracy. However, in general, better results are obtained using the FFDoE (R2 > 0.92, but R2 > 0.84 for exergy rate) compared to the ESCDoE (R2 > 0.87, and R2 > 0.53 for exergy rate). The FFDoE is the most appropriate design for the instantaneous prediction of THC and NOx regulated emissions (R2 > 0.91) as well as for its specific (accumulated) emissions (relative error e < 11%). However, worse results are obtained for CO emissions prediction (R2 > 0.7 and e < 19.1%).