Enhancing CI engine performance and emission control using a hybrid RSM–Rao algorithm for ZnO-doped castor–neem biodiesel blends

Traditional optimization methods, like Response Surface Methodology (RSM) alone, have failed to capture the complexities of engine operation, leading to the adoption of a more efficient algorithmic approach that requires no algorithm-specific parameters. Additionally, biodiesel produced from a single oil has not only failed to enhance performance effectively but has not maintained a benign environment, resulting in high emission tail, most importantly NOx. For the first time, the study adopts RSM and the Rao-3 Algorithm, a population-based metaheuristic, to optimize and model the performance and emissions attributes of a CI engine running on a ZnO-doped composite biodiesel (CB) blend, adopting neem-castor biodiesel as CB. The former algorithm possesses the aforementioned features. Rao's algorithm does not mimic the behaviour of swarms, animals, birds, or any physical or chemical phenomenon, but it provides effective solutions to complicated problems. For the first time, the study adopts the hybridization of the RSM and Rao 3 algorithm to model and predict the IC engine operated on composite biodiesel/castor-neem seed oil methyl ester (CNSOME)/diesel/ZnO blends. This involves the impact of input engine variable ranges such as fuel blends (10–40 v/v%), ZnO dosage (20–100 ppm). engine speed (800–2000 rpm), and engine load (10–40 %) on performance indices (PI) such as brake power (BP), and brake thermal efficiency (BTE) and emission attributes (EAs) namely carbon monoxide (CO), unburnt hydrocarbon (UHC), oxide of nitrogen (NOx) as responses. RSM was used to plan the design of the experiment and detect the optimal engine conditions for enhancing PI and reducing EAs while 3 algorithms were used to solve a multi-objective function. The results show that the RSM model determined optimal condition resulted experimentally the BP of 7.3 kW, and BTE of 25.96 % and CO of 64.5 ppm, UHC of 2.64 ppm, and NOx of 705.4 ppm were attained at a fuel blend of 21.34 % vol., ZnO additive of 61.35 ppm, engine load of 21.5 %, and engine speed of 1380 rpm correspond to case 5 determined highest desirability function value of 0.9553. Rao-3 algorithm provides the optimal parametric condition (21.34 vol% of blend fuel, 63.43 ppm of nanoparticle, 1435.18 rpm of engine speed, and 24.35 % for engine load) with a desirability function approach (DFA) value equal to 0.9556, and were subjected to conduct practical experiments. Experiments corresponding to case 5 (21.34 vol% of blend fuel, 64 ppm of nanoparticle, 1435 rpm of engine speed, and 24 % for engine load) resulted in BP, BTE, CO, UHC, and NOx equal to 8.2 kW, 29.4 %, 66.2 ppm, 2.82 ppm, 701.4 ppm, respectively. The performance of the Rao algorithm seems to be superior to the RSM because the deviation in Rao algorithm predictions with experimental values resulted in an absolute percent error equal to 4.63 % for BP, 3.98 % for BTE, 2.46 % for CO, 5.32 % for UHC, and 0.85 % for NOx emissions, respectively. The hybrid models using RSM-Rao-3 algorithms can be a robust model to enhance performance and reduce the emission profile of an IC engine powered by a novel green diesel derived from a myriad of generations without being trapped at local minima solutions. It obtains the best solution through iteration. This study presents an innovative framework for biodiesel engine optimization, contributing to the advancement of sustainable and efficient green diesel technologies.