A decision-tree-based approach to smoke spike detection in a heavy-duty diesel engine

Abstract

Smoke spikes occurring during transient engine operation have detrimental health effects and increase fuel consumption by requiring more frequent regeneration of the diesel particulate filter. This paper proposes a decision tree approach to real-time detection of smoke spikes for control and on-board diagnostics purposes. A contemporary, electronically controlled heavy-duty diesel engine was used to investigate the deficiencies of smoke control based on the fuel-to-oxygenratio limit. With the aid of transient and steady state data analysis and empirical as well as dimensional modeling, it was shown that the fuel-to-oxygen ratio was not estimated correctly during the turbocharger lag period. This inaccuracy was attributed to the large manifold pressure ratios and low exhaust gas recirculation flows recorded during the turbocharger lag period, which meant that engine control module correlations for the exhaust gas recirculation flow and the volumetric efficiency had to be extrapolated. The engine control module correlations were based on steady state data and it was shown that, unless the turbocharger efficiency is artificially reduced, the large manifold pressure ratios observed during the turbocharger lag period cannot be achieved at steady state. Additionally, the cylinder-to-cylinder variation during this period were shown to be sufficiently significant to make the average fuel-to-oxygen ratio a poor predictor of the transient smoke emissions. The steady state data also showed higher smoke emissions with higher exhaust gas recirculation fractions at constant fuel-to-oxygen-ratio levels. This suggests that, even if the fuel-to-oxygen ratios were to be estimated accurately for each cylinder, they would still be ineffective as smoke limiters. A decision tree trained on snap throttle data and pruned with engineering knowledge was able to use the inaccurate engine control module estimates of the fuel-to-oxygen ratio together with information on the engine control module estimate of the exhaust gas recirculation fraction, the engine speed, and the manifold pressure ratio to predict 94% of all spikes occurring over the Federal Test Procedure cycle. The advantages of this non-parametric approach over other commonly used parametric empirical methods such as regression were described. An application of accurate smoke spike detection in which the injection pressure is increased at points with a high opacity to reduce the cumulative particulate matter emissions substantially with a minimum increase in the cumulative nitrogrn oxide emissions was illustrated with dimensional and empirical modeling.

Measurement and Prediction of Discharge Coefficients in Highly Compressible Pulsating Flows to Improve EGR Flow Estimation and Modeling of Engine Flows

“…While trying to determine the root cause of smoke spikes in electronically controlled diesel engines, previous work by the author and co-workers have found that the flow-nozzle based EGR flow estimate as well as the Volumetric Efficiency estimate can be significantly inaccurate during the turbocharger lag period (Brahma, 2011(Brahma, , 2013(Brahma, , 2014. The inaccurate estimation of in-cylinder Oxygen mass was responsible for the smoke/particulate matter spikes.…”

Analysis and prediction of transient opacity spikes using dimensional modeling

“…The nonspike segments were not simulated in GTPower and simply predicted by simple global regression trained on transient snap-throttle data as per recommendations made in previous work. 3 The point-bypoint R 2 agreement was only 0.1347 but such a comparison is not indicative of predictive accuracy in this particular case because the opacity was measured a few meters downstream of the engine and was not instantaneously equal to the opacity at the exhaust manifold during transients. This is not only because of transport delay (the measured signal was aligned to ECM events by shifting the raw signal by 0.7 s, determined statistically, to account for sensor and transport delay) but also due to volume capacitance.…”

“…This new approach, termed the “nonparametric reduced dimensionality” (NPRD) modeling approach, was developed to address the lack of robustness and generality particularly inherent in empirical smoke or particulate matter (PM) emission models commonly used for transient calibration and control modeling. 1–3 The dual motivations for this work are as follows: (1) The ubiquitous presence of smoke from heavy duty engines in everyday life and (2) the lack of generality exhibited by empirical transient opacity or PM models trained on ECM estimated engine flows.…”

“…Some of the recent transient calibration work in literature include neural networks by Atkinson et al, 51 regression and neural networks by Brahma et al, 2,52 and decision trees for classification. 3 None of the engine calibration–related work mentioned above has used dimensional modeling or phenomenological modeling. Although there exist numerous studies that use one-dimensional models for “what-if” scenarios that include emission predictions, for example, the study by Schwoerer et al, 53 dimensional modeling is by-and-large considered separate from calibration modeling.…”

“…Data were collected for the sole purpose of investigating different kinds of calibration models and classifiers. 3 The operating parameters of this engine were engine speed, fuel injected per stroke, main injection timing, EGR fraction, charge flow, injection pressure, pilot quantity and timing, and post quantity and timing. Random space-filling data were acquired by randomly changing the first six of these parameters while maintaining the pilot and post quantities at their default production calibration values.…”