Synthetic electro gasoline fuel design for ultra lean burn engine: an AI-powered approach for fuel reformulation

Electrification is considered a long-term solution for decarbonizing transportation; however, hybrid electric vehicles (HEVs) are essential for a smoother transition. Ultra lean burn (ULB) engines are a promising HEV technology, offering high efficiency and low emissions. These engines typically operate at an excess air ratio (λ) > 1.8. However, ULB combustion faces challenges, such as instability and increased ignition energy requirements. Fuel design is critical to overcoming these challenges. The molecular structure of fuels impacts their combustion properties, influencing efficiency and emissions. This study examined the influence of various fuel properties on the performance of the ULB engines. A systematic investigation was conducted using a Hyundai ULB engine and 29 distinct fuels to identify the key properties influencing engine operation. This study’s primary contribution is the creation of a novel, artificial intelligence (AI) driven fuel design framework built upon these experimental results. By integrating the quantified effects of boiling point, laminar flame speed, adiabatic flame temperature, carbon-to-oxygen ratio, heat of vaporization, and antiknock properties into a multiobjective merit function, we developed a tool for prescriptive fuel formulation. This methodology was used to conduct a high-throughput screening of approximately 379,500 compounds, identifying short-chain alcohols and their derivatives as ideal additives to enhance ULB engine efficiency, performance, and emissions. Ultimately, this work demonstrates a complete design-to-formulation workflow, culminating in the formulation of five new fuels with optimized properties using an AI-based approach. Validation testing in a ULB engine confirmed the success of this approach; all new fuels achieved significant reductions in NOx emissions (3–35%) compared to those of a reference gasoline. Notably, one of the designed fuels, X4, was identified as the most suitable for aggressive lean-burn applications, demonstrating superior combustion stability and the lowest fuel consumption under highly lean conditions (λ > 1.8), where the reference fuel’s stability deteriorated.