An investigation into the effect of road gradient and driving style on NOX emissions from a diesel vehicle driven on urban roads

https://doi.org/10.1016/j.trd.2019.05.002Get rights and content

Highlights

  • A diesel vehicle was driven at various levels of aggressiveness on public roads.

  • Instantaneous vehicle energy was determined taking road gradient into account.

  • Aggregate and instantaneous energy was shown to correlate well with NOX emissions.

  • With road gradient, the correlation remained regardless of driver aggressiveness.

  • Potential application in developing better emissions models.

Abstract

This study explores the influence of different driving styles and road gradient profiles on NOX emissions in a diesel passenger vehicle on urban driving. Driving dynamics parameters were correlated with cumulative NOX emissions measured during on-road driving on urban roads. In this work, the vehicle was driven on two different urban routes, one with mostly hilly roads and the other with predominately flat roads to assess the effect of road gradient on NOX emissions. Each route was driven six times, the first drive on each route was driven very timidly, each subsequent drive systematically became more aggressive with the sixth drive being very aggressive. From the vehicle speed and road gradient data, the instantaneous vehicle energy was estimated and correlated against the instantaneous NOX emission. In order to investigate for monotonic relationships, Spearmans rank correlation coefficient was used to investigate potential correlations between NOX emissions and driving parameters. A strong positive correlation was observed between instantaneous NOX emissions and instantaneous vehicle energy irrespective of the driving behaviour. The correlation of driving dynamics parameters with NOX emissions also showed a similar trend indicating that driving aggressiveness and vehicle NOX emissions have a strong relationship. Also there is evidence that the influence of road gradient on NOX emissions decreases with an increase in driving aggressiveness.

Introduction

It is well established that automobiles are significant contributors of gaseous emissions, deteriorating urban air quality (Smit et al., 2017). Diesel vehicles are notable contributors of nitrogen oxide (NOX) (Cox and Blaszczak, 1999), recognised by the US Environmental Protection Agency (EPA) as a severe pollutant affecting the respiratory system (Masum et al., 2013). As a control measure, governments around the world have enforced emissions regulations which require automotive manufacturers selling vehicles in respective regions to ensure that the NOX emissions from their vehicles are well within defined legal limits (Giechaskiel et al., 2014).

There are different NOX formation mechanisms during combustion in a diesel engine. When ionised-nitrogen is present in the fuel, it can oxidise and fuel NOX is produced. Prompt NOX is the result of nitrogen present in the inlet air which reacts with the fuel and oxidises along with it. Thermal NOX is the most common mechanism of NOX production in diesel engines, it is produced when the nitrogen present in the combustion air undergoes oxidation due to high temperatures (from approximately 1300°C) (Cox and Blaszczak, 1999). Turbocharging technology, which enhances the power output and mileage of the diesel engine, causes a significant increase in thermal NOX emissions as the inlet air is heated during the compression process (Carslaw et al., 2011).

Emission regulations vary based on the norms defined by different governments. For EURO 6, the permissible limits of NOX emission are established for vehicles based on their gross vehicle weight(GVW). For a light duty vehicle (GVW less than 3500 kg), the acceptance limit of NOX emissions (EURO 6) is 0.08 g/km (Department of Infrastructure, Regional Development and Cities, 2003, Mock, 2017). In Europe, until recently, the type approval process for vehicles involved only laboratory-based tests to verify its emission characteristics. The tests were conducted using a chassis-dynamometer and the vehicle was driven through a predefined test cycle (new European driving cycle, NEDC). Throughout the cycle, the test vehicle was subjected to different operating conditions and the exhaust flow passed through gas analysers such as VOEMlow systems and sophisticated constant volume sampling systems, such as those used in MIRA and IDIADA, to estimate the quantity of different pollutants present in the exhaust stream (Pelkmans and Debal, 2006).

Laboratory tests are advantageous because of their repeatablility. Drive cycles are mainly categorised into steady-state and transient cycles. Steady-state cycles are capable of evaluating the vehicle behaviour at specific engine running conditions and will have minimal variations in the speed time profile. Whereas, transient cycles are often created directly from captured on-road data. Driving cycles are generally depicted as a speed-time profile along with gear position. The measured emissions are influenced by both internal factors, such as engine operating condition and exhaust after-treatment systems and external factors, such as ambient air temperature (Barlow et al., 2009).

For a driving cycle to effectively assess vehicle behaviour, it needs to be defined in a manner which represents the vehicle’s on-road dynamics. Vehicle driving parameters, such as: acceleration, deceleration, gear utilisation, stopping and cruising manoeuvres influence emissions (André et al., 2006). Many transient cycles, such as the worldwide harmonised light vehicles test cycle (WLTC), are a direct representation of on-road data (Giakoumis and Zachiotis, 2018). However, not all vehicles are the same and one cycle may not be representative of the on-road emissions for any vehicle (Giakoumis and Zachiotis, 2018).

The most common vehicle parameter employed in emission modelling is average speed (Sonntag and Gao, 2009). It has been established that the average emission factor of a single pollutant for a class of vehicles relies heavily on the average speed of the vehicle in a driving trip (Kousoulidou et al., 2013). In order to classify different driving patterns, the logged emission values from a driving cycle are fitted with an average speed function. Emission functions created using average speed models are used to develop prediction models which can estimate the local air quality based on traffic flow in an urban region (Achour et al., 2011). NAEI (National Atmospheric Emissions Inventory, UK) (National Atmospheric Emissions Inventory, UK, 2018) and COPERT are well-known models based on vehicle average speed (Smit and Ntziachristos, 2012).

A major drawback in using average speed models is its normalising effect on instantaneous peak emission conditions during sudden acceleration or deceleration events–which occurs frequently, particularly in urban driving (Smit and Ntziachristos, 2012). One solution to improve driving cycles is to develop models based on traffic and vehicle behaviour from on-road data. Vehicle kinetic and kinematic parameters can be used to develop such cycles, which also establishes the driving dynamics of the vehicle in the identified route. Using these parameters, the aggressiveness and transient behaviour of a particular drive can be evaluated. However, this approach requires a significant amount of on-road driving data in the identified route to develop the emission models. Another parameter which can be used to capture vehicle behaviour is vehicle specific power (VSP). VSP is the instantaneous energy the vehicle possesses at a particular time during the drive and can be computed by estimating the overall vehicle resistance (road and aerodynamic resistance) as well as the acceleration it experiences during the drive (Franco et al., 2013). The vehicle resistances can be estimated by coast down testing and the acceleration can be obtained from the vehicle speed data. VSP has already been used in developing models to predict vehicle emission in simulators such as MOVES (motor vehicle and equipment emission system). MOVES uses a curve fitting method to predict the NOX emissions from available drive data. However, the influence of VSP and road gradient on vehicle NOX emissions in an urban driving scenario with significant differences in road profiles has yet to be studied comprehensively (Nam, 2003).

Since a single vehicle can exhibit different dynamic behaviour owing to traffic and driver behaviour, the effect of these changes in the driving dynamics on the instantaneous and cumulative emission levels can be understood by real-time emission measurements (Gallus et al., 2017, Luján et al., 2018). As a consequence of recent studies which discovered that there is considerable variation between laboratory-based emission behaviour and on-road measurements, an on-road emission verification method was added on to the existing regulations (Mock, 2017). The Joint Research Committee (JRC) published a study which stated that there were significant variations between NOX emission levels obtained from laboratory cycles and that of on-road measurements (Weiss et al., 2011). Acting on such studies, the European Commission revised the EURO 6 norms by adding a clause for on-road emission testing of vehicles, called Real Driving Emissions (RDE) (Mock, 2017). This necessitates a need to understand the influence of different driving styles on emission traits of a vehicle. NOX being a significant air pollutant deserves to be investigated and the behavioural changes of NOX emissions from a diesel vehicle based on changes in driving styles therefore needs to be explored. The driving dynamics of a vehicle can be assessed by evaluating both kinetic and kinematic parameters of the vehicle in motion. These parameters are currently used to review different vehicle driving cycles used world-wide and to develop emission models of an urban region (Barlow et al., 2009). Although some attention has been given to exploring the influence of traffic on on-road emissions, there is currently no comprehensive study investigating the influence of road gradient on urban NOX emissions. Moreover, the combined impact of driver aggressiveness and road gradient during urban driving has also yet to be explored.

Section snippets

PEMS setup

The NOX sensor used in this study is a ceramic exhaust sensor manufactured by ECM (Engine Control and Monitoring). These sensors have a manufacturer specified precision of ±5 ppm (0–200 ppm), ±20 ppm (200–1000 ppm), ±2.0% (>1000 ppm). Prior to the work shown here, a confirmation experiment was performed and found that the sensor was working within the 2.0% manufacturer specified precision. The NOX sensor was used within the ECM miniPEMS system (comprising of: NOX, O2, exhaust pressure,

Coast down test results

From the data collected during the coast down test, the time taken for the vehicle to decelerate from 80 km/h to 10 km/h was extracted in 5 km/h intervals. The coast down force experienced by the vehicle was determined using:Fcoast=W·Δv·β1.8·Δtwhere Fcoast is the coast down force in N, W is the weight of the test vehicle in kg, Δv is the change in velocity (5 km/h) in m/s, β is a correction factor to account for inertia of rotating parts (1.035 for four wheelers) and Δt is the time in seconds.

Conclusion

The NOX emissions from a diesel vehicle during urban driving were studied on a predominately flat route and a hilly route at varying degrees of driver aggressiveness. The on-road test results have shown that there is a strong positive correlation between all the driving dynamics parameters and the cumulative NOX emissions, regardless of whether the route was predominately flat or contained significant and frequent changes in road gradient. The slight decrease in correlation levels in the

Acknowledgements

The authors wish to acknowledge the support of the School of Engineering at Deakin University for supplying the funding to purchase the emission equipment. Special thanks are also given to Alfred Deakin Professor Marcel Klaassen for the loan of his vehicle for this work.

References (28)

Cited by (41)

  • Analysis of the Euro 7 on-board emissions monitoring concept with real-driving data

    2024, Transportation Research Part D: Transport and Environment
  • Predicting Vehicle Behavior Using Multi-task Ensemble Learning

    2023, Expert Systems with Applications
    Citation Excerpt :

    The correlations between driving behaviors and NOx emissions, considering selective catalytic reduction (SCR) and ammonia creation and conversion technology (ACCT) systems, have been studied in Gao et al. (2021). In Prakash and Bodisco (2019), researchers analyzed different levels of aggressive driving and measured NOx emission when the vehicles were driven on a hilly and calm route. Similar considerations were done in Choi and Kim (2017) to expose the effect of aggressive acceleration on fuel consumption.

View all citing articles on Scopus
View full text