Combustion and emission performance of biodiesel–ethanol renewable fuels: Experimental study and machine learning prediction

Considering the complementary fuel properties between biodiesel and ethanol, biodiesel–ethanol renewable fuels with different blending ratios were formulated. Under various operating conditions, the effects of different blending ratios of renewable fuels on combustion and emission performances were experimentally investigated. Results showed that equivalent specific fuel consumption (ESFC) was highly sensitive to operating conditions: fuel economy was inferior to diesel under low and medium loads but superior under high loads. Under 50% load, ethanol's cooling effect reduced in-cylinder temperature and delayed combustion, lowering indicated mean effective pressure (IMEP); under 100% load, ethanol's higher volatility improved fuel-air mixing, accelerating combustion and increasing IMEP. Emissions were governed by load-dependent thermal conditions and fuels' oxygenated nature: low and medium loads saw ethanol's high latent heat create a low-temperature environment, suppressing nitrogen oxides (NOx) but increasing hydrocarbon (HC) and carbon monoxide (CO); high loads enhanced combustion via oxygen, slightly raising NOx but significantly reducing HC and CO. Machine learning evaluation indicated that the artificial neural network (ANN) and support vector machine (SVM) performed best for ESFC prediction; SVM and multiple linear regression were most effective for IMEP; and ANN yielded the best results for NOx prediction. This study provides a theoretical basis and practical support for the intelligent prediction and optimization of renewable fuel combustion performance.