Methods for reconstruction of transient emissions from heavy -duty vehicles

Abstract

Methods for Reconstruction of Transient Emissions from Heavy-Duty Vehicles Madhava R. Madireddy Emissions measurement analyzers give out a response that may not reflect the true instantaneous engine-out emissions. Currently, the heavy-duty diesel engines are being certified for emissions measured in a thirty second time window with certain specification requirements for the analyzers. Since these measured emissions values may not be the same as the true instantaneous emissions, integrated values for the thirty second windows may be affected by analyzer response. This document presents and examines reconstruction techniques to estimate instantaneous heavy-duty engine-out emissions. These techniques will take as the input, the continuous set of emissions data and approximate dispersion characteristics of the analyzer employed in measuring the continuous data. For this purpose, this research dealt with understanding and modeling the transient dynamics (dispersion function) of the analyzers and the sampling system to establish a relationship between the measured and instantaneous heavy-duty emissions. Four methods of reconstruction were presented in this study: Sequential Inversion Technique (SIT), Differential Coefficients Method (DCM), Inverse Fast Fourier Transform (IFFT) and Modified Deconvolution Technique (MDT). The application of each method in reconstructing real-time emissions data was presented. While SIT failed in practical applications, each of the other three methods was shown to offer advantage in the post-processing of the measured emissions data. DCM accounted for the small errors

Investigation of Emissions and Fuel Economy for the Integrated Bus Information System

“…For each test run, the time shifts were determined using the highest cross-correlation coefficients between emissions data and the target speed-time trace [98]. In addition, dispersion could affect continuous data [92]. Then, the diesel equivalent fuel consumption rates were computed based on carbon balance using fuel properties and measured carbon emissions.…”

“…When continuous emissions data were prepared, dispersion effect was not considered while timealignment was performed. The dispersion of the signal could affect emissions rates [92].…”

Effects of Boundary Parameters on the Work-Based Window (WBW) and Not-to-Exceed (NTE) In-Use Emissions Quantification Methods

“…The minimum duration for the event to be valid in the NTE zone is another limitation for the engine operation. The time constraint for the region has been developed to account for signal dispersion [66]. If the driver frequently changes gears or continuously varies the throttle position, then the data collected during that period of operation will not be accountable for evaluation due to the time limitation for the validation of the event [36].…”

Development of an Artificial Neural Network to Predict In-Use Engine Emissions

“…Equation 6.9.2 depicts the confidence interval. For a 95 percent confidence level the value for Z a/2 was 1.96 [56]. The confidence interval implied that 95% of the data will occur in the calculated range, or in other words there is 95% confidence that the percent difference between the emissions predicted by the ANN and the measured emissions will be in the range presented.…”

“…The values of these constants were then obtained by calculating the least-squared error at each point, and minimizing this error for the best fit. Once the constants were determined, the input to the analyzer for any output can be determined [56]. After determining the C i values, they are normalized, and a curve is generated.…”