Experimental Investigation and ANN‐Based Prediction of CI Engine Performance Fueled With Waste Plastic Oil–Diesel Blends Containing Al 2 O 3 Nanopar
Dudekula Jamal Basha, Upendra Rajak, Abhishek Agarwal, Balram Yelamasetti, Manoj Panchal, Sudheer Kumar, Abhishek DasoreABSTRACT
The increasing accumulation of plastic waste and depletion of fossil fuel reserves have intensified the search for sustainable alternative fuels for compression ignition (CI) engines. Although waste plastic oil (WPO) demonstrates significant potential as an alternative energy source, its direct utilization is limited by inferior combustion characteristics and unstable engine performance. Furthermore, limited studies have examined the combined influence of Al 2 O 3 nanoadditives and artificial neural network (ANN)‐based prediction under varying injection timings. Therefore, the present work experimentally investigates the performance and combustion characteristics of WPO–diesel blends containing Al 2 O 3 nanoparticles in a single‐cylinder CI engine operated at injection timings ranging from 19° to 25° before top dead center. The WP20PD blend containing 1000 ppm Al 2 O 3 nanoparticles exhibited improved brake thermal efficiency (BTE) and reduced brake‐specific fuel consumption (BSFC) compared with neat diesel under advanced injection timing conditions. The observed variations in maximum cylinder pressure (MCP) and ignition delay (ID) were comparatively small but remained consistent with the measured combustion behavior. ANN modeling using the Levenberg–Marquardt training algorithm demonstrated satisfactory predictive capability, with regression coefficients exceeding 0.95 for all investigated response parameters. Additional fivefold cross‐validation indicated response‐dependent predictive performance, with stronger generalization for MCP and ID, moderate performance for exhaust gas temperature, and comparatively larger variability for BSFC and BTE because of the limited experimental data set. The findings demonstrate the potential of nanoparticle‐assisted WPO blends combined with ANN‐based prediction for improving CI engine performance and supporting sustainable waste‐to‐energy utilization.