Large-eddy simulation of flow and pollutant dispersion behind a heavy-duty vehicle

Abstract

The deteriorated air quality in urban areas has caused increasing public concern due to its widespread nature, threats to the environment and potential hazard to human health. Traffic is the major source of the urban air pollution and heavy-duty vehicles that makes the greatest contribution compared with other types of vehicles. An efficient method to solve this problem is to monitor the emission level of in-use heavy-duty vehicles. However, the current knowledge on the flow and pollutant dispersion in the near-wake region is still not adequate, which hinders the accurate measurement with on-road vehicle emission sampling techniques, such as remote sensing and plume chasing methods. Large-eddy Simulation (LES) is employed to investigate the near-wake dynamics and plume dispersion behaviours. A practical heavy-duty vehicle model with detailed structures is used rather than an idealised Ahmed body, which is a preferred model in vehicle aerodynamics analysis. The near wake directly behind the vehicle consists of a major recirculation and an upper one. The major recirculation is collectively induced by the side entrainment and underbody jet-like flow with a size comparable to the height of the trunk h, while the upper one is largely driven by the roof-level prevailing flow. The far field is mainly composed of trailing vortices, originated from the near wake. The plume dispersion behaviour is tightly related to the flow characteristics. The plume is carried back to the truck right after emission at first before the recirculation vortices augment the pollutant mixing. Due to the strong mixing, the pollutant distribution tends to be more homogeneous together with a rather high fluctuating concentration (over 20% of the time-averaged component). The plume ascends mildly before being purged out of the major recirculation to the far field by turbulence, resulting in a major reduction (almost an order of magnitude) in pollutant concentration outside the near wake. In the far field, the plume is raised higher than the tailpipe level and the dispersion gradually follows the conventional Gaussian distribution. These findings suggested that remote sensing with a sampling duration longer than h/U∞ would be prone to the underestimation of the tailpipe emission. Sampling within the near wake can provide more accurate measurement results, which is applicable to both remote sensing and plume chasing methods. For plume chasing, sampling duration is one of the important factors to acquire representative time-averaged concentrations, because of the intermittent, rapid pollutant dilution behind the target vehicle. The detailed spatio-temporal dispersion data provided by the LES can help identify the effect of sampling duration on the measurement accuracy. The sample mean concentration gradually converges to the population mean with increasing sampling duration, but this effect is less pronounced in long sampling duration. Sampling in the region x>0.6h could degrade the sampling accuracy to a large extent. The frequency analysis also unveils that sampling at the dominant frequency lowers the uncertainty in sample mean.(479 words)