American Association for Aerosol Research - Abstract Submission

AAAR 37th Annual Conference
October 14 - October 18, 2019
Oregon Convention Center
Portland, Oregon, USA

Abstract View


Predicting the Fuel Consumption and Tailpipe Emissions from Light-Duty Passenger Vehicles using Artificial Neural Networks

Shiva Tarun, Asher Zachary, Johnston Brian, Bradley Thomas, SHANTANU JATHAR, Colorado State University

     Abstract Number: 298
     Working Group: Combustion

Abstract
There is growing evidence that on-road emissions from mobile sources exceed emissions determined during chassis dynamometer tests. In this work, we used a portable emissions monitoring system and artificial neural networks (ANNs) to measure and model on-road fuel consumption and tailpipe emissions from a Tier-2 light-duty gasoline and diesel vehicle. Tests were performed on at least five separate days for each vehicle and each test included a cold start and operation over a hot phase. Fuel consumption and emissions rates were calculated at 1 Hz using information gathered from the vehicle using the onboard diagnostics port and the PEMS measurements. We trained ANN models on part of the data to predict fuel consumption and tailpipe emissions at 1 Hz for both vehicles and evaluated these models against the rest of the data. The ANN models performed best when the training iterations were set to larger than 25 and the number of neurons in the hidden layer was between 7 and 9, although we did not see any specific advantage in increasing the number of hidden layers beyond 1. The trained ANN models predicted the fuel consumption over test routes within 10% and 3% of the measured values for the gasoline and diesel vehicle respectively. The ANN performance varied significantly with pollutant type and we were able to develop satisfactory models only for the diesel vehicle. The ANN models for NOx were able to predict total emissions within 10% of the measured values. The ANN models performed better than multivariable regression models such as those used in mobile source emissions models (e.g., EMFAC). To highlight the value for real world driving, we evaluated the vehicle-specific ANN models to pick routes for lower fuel consumption or tailpipe emissions.