Search

Gaseous-liquid fuel combinations of biogas as primary fuel and diesel as pilot fuel were employed to power a compression ignition (CI) engine in this study. The effects of variable input factors such as engine load, compression ratio, and pilot fuel injection advance on the dual-fuel engine's combustion, thermal, and emission performance were investigated. The data acquired during this lab-based testing was used to develop a prognostic model using a Bayesian optimization strategy for hyperparameter optimization and a contemporary ensemble Gaussian process regression (GPR) technique. During model testing, the error analysis revealed that the developed model could predict the experimental data rather correctly, with R2 values ranging from 0.9995 to 0.9999 and MSE values ranging from 0.00018 to 0.1334. The mean absolute error was discovered to be between 0.0091 and 0.3065. Consequently, the developed GPR model can effectively simulate the on-board performance, emission, and combustion characteristics of a diesel-biogas dual-fuel setting. As a consequence, the current work established the GPR model in the existing CI engine meta-modeling framework as a consistent, trustworthy, and robust system analytical approach for diesel-biogas dual fuel mode of operation. • Waste-derived biogas was used to power the diesel engine in dual-fuel mode. • Gaussian process regression was employed to develop a prognostic model. • Bayesian approach was used for hyper parameter optimization for best results. • Used model could predict emission and performance of the engine with 99.9% accuracy. • A precise metamodel framework to simulate the input–output paradigm was developed. [ABSTRACT FROM AUTHOR]