An Explainable DL-Based Condition Monitoring Framework for Water-Emulsified Diesel CR Systems

Despite global patronage, diesel engines still contribute significantly to urban air pollution, and with the ongoing campaign for green automobiles, there is an increasing demand controlling/monitoring pollution severity of especially in heavy-duty industries. Emulsified fuels provide a readily available solution engine pollution; however, inherent reduction power, component corrosion, and/or damage poses major concern adoption. Notwithstanding, on-going investigations suggest need reliable condition monitoring frameworks accurately monitor/control water-diesel emulsion compositions inevitable cases. This study proposes use common rail (CR) pressure differentials deep one-dimensional convolutional neural network (1D-CNN) local interpretable model-agnostic explanations (LIME) empirical diagnostic evaluations (and validations) using KIA Sorento 2004 four-cylinder line as case study. CR signals were digitally extracted at various water-in-diesel RPMs, pre-processed, used necessary transient spectral analysis, validations. Results reveal high model trustworthiness average validation accuracy 95.9%.