Machine Learning-Assisted optimization of acetylene Dual-Fuel combustion using Copper oxide (CuO) Nano-Enhanced waste Plastic Oil in Compression Ignition (CI) Engine

In this study, an experimental and machine learning-assisted investigation was conducted on a modified single-cylinder direct-injection Compression Ignition (CI) engine operating in acetylene dual-fuel mode using copper oxide (CuO) nanoparticle-enhanced waste plastic oil (WPO) blends. The proposed approach integrates waste-to-energy fuel utilization, catalytic nano-additives, high-reactivity gaseous combustion, and data-driven predictive modelling within a unified sustainable combustion framework. Experimental investigations were performed using WPO10 and WPO20 blends containing 50 and 100 ppm CuO nanoparticles under acetylene flow rates of 5 and 10 LPM. Among all tested fuel combinations, A5WPO10Cu100 exhibited the best overall performance, achieving a maximum brake thermal efficiency of 33.12 %, corresponding to an 8.27 % improvement over diesel, together with a 21.03 % reduction in brake specific energy consumption. The optimum fuel also reduced HC, CO, and smoke opacity by 41.46 %, 47.62 %, and 43.58 %, respectively, while A10WPO10Cu100 produced the highest combustion intensity with a peak heat release rate of 68.9 J/°CA, maximum cylinder pressure of 90.5 bar, and minimum ignition delay of 8.75 °CA, indicating enhanced premixed combustion and catalytic oxidation. Although NOx emissions increased under higher acetylene induction due to elevated combustion temperatures, the overall combustion efficiency and emission trade-off were significantly improved. To further establish predictive capability, Linear Regression, Decision Tree, and Random Forest models were developed using the experimental dataset. Among these, the Random Forest model demonstrated the highest prediction accuracy with R 2 values ranging from 0.995 to 0.999 and prediction accuracy exceeding 98 %, confirming its robustness in predicting nonlinear combustion, performance, and emission characteristics. The integrated experimental and machine learning framework demonstrates that CuO nano-enhanced WPO coupled with acetylene induction is a promising strategy for improving CI engine efficiency while simultaneously advancing intelligent combustion optimization for sustainable transportation applications. To further enhance predictive capability and intelligent engine diagnostics, advanced machine learning (ML) algorithms including Linear Regression (LR), Decision Tree (DT), and Random Forest (RF) were developed using experimentally measured combustion, performance, and emission parameters. The Random Forest model demonstrated superior nonlinear prediction capability with prediction accuracy exceeding 98 %, coefficient of determination (R 2 ) approaching 0.995–0.999 for major engine responses.