Research on Combustion, Emissions, and Fault Diagnosis of Ternary Mixed Fuel Marine Diesel Engine

This study aimed to investigate the effects of diesel/ethanol/n-butanol mixed fuel on the marine diesel engine combustion and emissions at different ethanol blending ratios, different single injection times, and pre-injection times. In addition, this study takes the injector fault phenomenon as an example, simulates the three fault phenomena of the injector, and uses a variety of algorithms to optimize the probabilistic neural network model to achieve the fault state identification and diagnosis of the injector. The results of research showed that, with the increase in the ethanol blending ratio, the peak cylinder pressure shows a decreasing trend. The ignition delay period is extended, and the peak instantaneous heat release rate increases. Compared with D100, the nitrogen oxide (NOx) emissions of D50E40B10 mixed fuel are reduced by 12.3%, soot emissions are reduced by 29.18%, and carbon monoxide (CO) emissions are increased by 5.7 times. With the injection time advances, the peak values of cylinder pressure and heat release rate show an increasing trend, soot emissions gradually decrease, and NOx and CO emissions gradually increase. The peaks of the cylinder pressure and heat release rate in the pilot injection stage gradually decrease as the pilot injection time advances, while the peak heat release rate in the main injection stage increases. In terms of emissions, NOx emissions first decrease and then increase as the pilot injection time advances, while soot emissions gradually increase. The average accuracy of the PSO-PNN neural network model reaches 90%, and the average accuracy of the WOA-PNN neural network model reaches 95%. Therefore, the WOA-PNN neural network model is determined to be the optimal injector fault diagnosis model, which can be applied to the identification and diagnosis of injector fault states of diesel engines.