Abstract
The present study investigates the possibility of increasing efficiency and lowering pollutant emissions from a syngas-powered engine by modulating operational parameters such as engine load and syngas composition. Although artificial intelligence-based technologies are becoming more common for modeling single-fuel mode engines, they are rarely employed to simulate a dual-fuel syngas/diesel engine. In the first of its kind Endeavor in the area of syngas fueled engines, a hybrid adaptive neuro-fuzzy inference system (ANFIS)-response surface methodology (RSM) is investigated. The performance of syngas (H2+CO), a novel synthetic gaseous fuel, was studied in four different combinations. ANFIS and RSM-based prediction models were developed using engine performance and emissions data collected over the whole load range. While ANFIS surpassed RSM in model prediction, RSM was useful in establishing mathematical links between engine input and output. The ANFIS model developed had a good correlation between R (0.996–0.9998) and R2 (0.992–0.9972), as well as low model errors as determined by the root means square error range (0.0086–5.936) and mean absolute percentage error range (0.0028–0.0194). Theil's U2 was used to calculate the model's uncertainty, which was estimated to be 0.0065–0.0439. The superior forecasting abilities of ANFIS models were proved by low errors and uncertainty. The performance of the syngas-powered engine was optimized using the desirability approach to achieve optimum efficiency while emitting the least amount of pollution. The ideal engine load and syngas compositions for maximum production were 67.99% engine load and 72.4:27.6 as H2:CO syngas mix.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call