Task-aware multi-stream convolutional network based estimation of cylinder pressure and heat release rate in a directinjected corn oil methyl ester biodiesel engine

Le résumé fourni par la source

With the increasing global focus on renewable energy sources, biodiesel demonstrates a valuable and feasible alternative to traditional fossil fuels, offering significant environmental benefits. This study introduces a novel predictive framework based on Task-Aware Multi-Stream Convolutional Network (TAMSCN) to estimate net heat release and pressure measurements during engine operation across various scenarios with a one-cylinder configuration and a four-stroke petroleum engine running on biodiesel mixtures with direct injection of corn oil methyl ester biodiesel engine. The research includes a comprehensive comparative analysis of pure ester compounds (ethyl and methyl), ester-based fuel blends, and ultra-low sulfur diesel across different engine loads and speeds at a fixed 1500 rpm. Ethyl ester fuels exhibited the highest ignition timing advance, while pure methyl esters and regular diesel showed the most significant delay in combustion onset. Regarding thermal efficiency, ester blends generally experienced a maximum decrease of 9%, with ethyl ester blends showing a much smaller reduction of around 1%. At low engine loads, ethyl ester fuels resulted in decreased fuel consumption. Notably, blends containing 20% diesel in ethyl esters achieved the second-highest efficiency, with about a 7% reduction in fuel consumption, with diesel alone consistently showing the lowest values during testing. Keywords: convolutional network, prediction, diesel, ester, cylinder pressure, methyl ester

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.