Prediction of emissions and performance of a diesel engine fueled with n-octanol/diesel blends using response surface methodology

n-Octanol (C8H17OH) is an advanced biofuel derived from ligno-cellulosic biomass that is suitable for compression ignition technology with several properties closer to fossil diesel. This study analyses the performance and emissions of a direct-injection (DI) diesel engine fueled with n-octanol/diesel blends containing 10% (OCT10), 20% (OCT20) and 30%(OCT30) by volume of n-octanol using a 3 x 3 full-factorial experimental design matrix that considers blend composition of n-octanol in diesel, exhaust gas recirculation (EGR) rates of 10%, 15% and 20% and injection timings of 19, 21 and 23 crank angle (CA) before top dead centre (bTDC) as factors. Models for oxides of nitrogen (NOx), smoke, brake thermal efficiency (BTE) and brake specific fuel consumption (BSFC) were developed using response surface methodology (RSM) and were found to be significant statistically. The variation of EGR had a considerable effect on both BTE and BSFC of the engine followed by blend composition and injection timing. Best performance (BTE = 37.06%, BSFC = 0.23kg/kWh) was delivered by OCT10 at 10% EGR and 23 CA while the lowest performance (BTE = 30.95%, BSFC = 0.28kg/kWh) was by OCT30 at 20% EGR and 19 CA. Injection timing was found to have the highest effect on NOx emissions while EGR affected smoke opacity to the maximum. NOx was found to decrease from 1790 ppm (for OCT10 at 10% EGR and 23 CA) to as low as 410 ppm (for OCT30 at 20% EGR and 19 CA). Smoke opacity was found to decrease from 94.2% (for OCT10 at 20% EGR and 19 CA) to as low as 43% (for OCT30 at 10% EGR and 23 CA). Desirability approach was used to determine the best combination of blend composition of n-octanol, EGR and injection timing for minimising smoke, NOx and BSFC simultaneously. 17% by volume of n-octanol/diesel blend injected at 20 CA bTDC and 10% EGR was predicted to be optimum which delivered a simultaneous reduction of NOx (-47.4%), smoke (-21.08%) and BSFC (-8%) during confirmatory tests with a reasonable accuracy of within 4%. This method is robust and could be employed to other small engines for developing models that can predict engine characteristics with reasonable accuracy. (C) 2018 Elsevier Ltd. All rights reserved.