Abstract
City dwellers are regularly exposed to nanoparticle (i.e. particles < 300 nm in diameter), emitted by fossil fuelled vehicles, whilst commuting by transport modes such as taxis and buses. Exposure to these nanoparticles can lead to significant adverse effects on human health. This study aims to investigate spatial distribution of particle number concentrations (PNCs) and distributions (PNDs) in and outside a car cabin during driving. Possible influences of particle transformation processes on PNC and PNDs in the car cabin are also investigated. Another objective is to predict the PNCs in the car cabin using those measured outside. Measurements of particles in the 5-560 nm size range were conducted using a fast response differential mobility spectrometer (DMS50) in conjunction with an automated switching system. The DMS50 was used to measure size-resolved sequential distributions at: (i) four seats in the car cabin during about 700 minutes of driving, and (ii) two points at the driver's seat, inside and the front bonnet outside the cabin, during about 500 minutes of driving. The emission penetration and spatial distribution in the car cabin through (i) the ventilation system (Vent), and (ii) door/window sealing (CG) was simulated by means of three-dimensional computational fluid dynamics (CFD) using the Fluent code. Standard k-ε turbulence model was employed to simulate turbulence flow in the cabin. Vent/CG emission ratio was altered for the two different scenarios (0.9/0.1 and 0.7/0.3), indicating; (i) no filter fitted Vent and high vehicle sealing efficiency, (ii) filter fitted Vent and reduced sealing efficiency. Four-point measurements indicated that the average PNCs at the front seats (3.96 and 3.85 × 104 cm-3) were almost identical to those found at the rear seats (3.82 and 4.00 × 104 cm-3). The very small differences (∼0.1%) suggest that the car cabin is very close to a well-mixed microenvironment. Two-point measurements revealed that the ratio of average PNCs in (2.72 ± 1.03 × 104 cm-3) and outside the car cabin (3.75 ± 1.62 × 104 cm-3) was about 0.72. A semi-empirical box mode model was introduced to predict PNCs in the car cabin as a function of those measured outside and cabin air exchange rate. Performance evaluation of the box model against statistical measures was within the recommended guidelines for urban air quality modelling. Overall, PNCs calculated by the model demonstrate a satisfactory correlation with the measured values. CFD simulations indicate that away from the Vent, emission is dispersed almost uniformly in the car cabin. Vent / CG ratios indicated that despite changes of emission filtering into the cabin, the dispersion characteristics remained almost identical at passengers' breathing height (i.e. 1.2 m from the floor).