Advance Data Analytics Methodologies to Solve Diesel Engine Exhaust Aftertreatment System Challenges 2019-01-5035

Recent developments are making powertrain systems more complex day by day. Understanding such system complexities and addressing their specific issues requires advance methodologies. This paper will discuss the approach and implementation of such advance data analytics methodologies using the field of artificial intelligence. The application of artificial intelligence is widely accepted in solving intricate issues. This paper points out methodologies using machine learning and neural network techniques to solve such intricate challenges in diesel engine exhaust aftertreatment system (ATS). Both the supervised and unsupervised learning methods have been used to solve specific powertrain challenges by taking cases from both “classification” and “regression” learnings. This paper discusses the step-by-step approach (from descriptive to predictive to prescriptive analysis) including feature extraction process to deal with such challenges using these advance data analysis methodologies. The following use cases will be discussed in this paper: (a) Self-learning methodology for analysis of powertrain parameters. (b) Prediction of a failure (fault/error code) in the system. (c) Prediction of exhaust temperature shoot-up events during active Diesel Particulate Filter (DPF) regeneration. (d) Diesel Oxidation Catalyst (DOC) outlet temperature sensor modeling using real-world vehicle data. (e) Dynamic calibration methodology for controlling hydrocarbon (HC) dosing during active DPF regeneration event. The main objective of this paper is to provide a direction for solving typical cases in powertrain systems and not to comment or evaluate any existing machine learning methodology.