Modeling NO Emissions of an Off-road Diesel Engine Based on Emission Tests

Abstract

Emissions values determined by the ISO 8178 emission certification tests do not necessarily represent emissions of a tractor in operation (Hansson et al., 2001). Rather than using ISO 8178 tests solely for certification, data collected during the tests may be suitable for predicting nitrogen oxide (NO_x) emissions of an engine operating at constant loads and speeds. Linear multiple regression (LMR) and nonlinear polynomial network (NPN) models were developed with data collected from ISO 8178-4 (1996) test cycle B-type tests (ISO) and an expanded set of tests (EXP) to predict NO_x emissions from a diesel engine. LMR using the ISO training data (R² = 0.94) resulted in over-training of the model, as applied to the evaluation data (R² = 0.51). LMR based on the expanded data (R² = 0.60) was a better LMR model, when applied to the evaluation data (R² = 0.73). NPN using the ISO training data (R² = 0.99) resulted in a considerable improvement over the LMR models for the evaluation data (R² = 0.81). NPN using the EXP training data (R² = 0.96) resulted in the best model when applied to the evaluation data (R² = 0.95). When applied to the evaluation data, the mean absolute error of the NPN EXP based model was significantly less than from the NPN ISO based model. The NPN model based on EXP data is recommended for predicting NO_x. Data could be collected during ISO 8178-4 emission tests that included additional test modes and modeled with NPN to predict NO_x emissions for a diesel engine operating at various constant speeds and loads.