Machine learning based prediction of the performance and emission characteristics of CRDI diesel engine using diethyl ether and carbon nanotube additives with Spirulina platensis as a third-generation biofuel

Alternative fuels are required to provide the world's energy demands due to excessive fossil fuel use, harmful petrol emissions, environmental pollution, growing demand, rising costs, and fossil fuel degradation. The additives are utilized in biodiesel-diesel blends because they partially meet the physicochemical and thermal properties, resulting in improved combustion, performance, and reduced emissions. This research aims to enhance performance metrics and minimize emissions by incorporating diethyl ether (DEE) and single walled carbon nanotube (CNT) nanoparticles (NPs) into a Spirulina platensis (SP) microalgae-based biodiesel-diesel blend at various injection pressures (IPs). The prepared blends are homogeneous and consistent, as analyzed by characterization using FTIR, XRD, FESEM, EDX, and Raman spectrum analysis. The prepared samples are tested on a two-cylinder, four-stroke, Common Rail Direct Injection (CRDI) diesel engine. The findings revealed that the Brake Thermal Efficiency (BTE) is improved by 15% while the Brake Specific Fuel Consumption (BSFC), Hydrocarbon (HC), Carbon Monoxide (CO), Nitrogen Oxide (NOx), and smoke opacity are reduced by 13.5, 20.7, 39.5, 20.6, and 9.7%, respectively, for the TF + CNT50 sample at higher IP and Brake Mean Effective Pressure (BMEP). The fuel sample TF + CNT50 was shown to be more sustainable and suitable for use in multi-cylinder, CRDI diesel engines without engine modifications. Furthermore, Machine Learning (ML) methods such as Support Vector Regression (SVR), Random Forest (RF), and Decision Tree (DT) are used for accurately predicting engine performance and emission characteristics by analyzing the correlations between input and output data. Simulating these interactions improves engine design and reduces experimentation costs.