Supplemental Information Emissions, Performance and Design of UK Passenger Vehicles

ABSTRACT Consumer, legal, and technological factors influence the design, performance, and emissions of light-duty vehicles (LDVs). This work examines how design choices made by manufacturers for the UK market result in emissions and performance of vehicles throughout the past decade (2001–2011). LDV fuel consumption, CO2 emissions, and performance are compared across different combinations of air and fuel delivery system using vehicle performance metrics of power density and time to accelerate from rest to 100 km/h (62 mph, tz-62). Increased adoption of direct injection and turbocharging technologies helped reduce spark ignition (SI, gasoline vehicles) and compression ignition (CI, diesel vehicles) fuel consumption by 22% and 19%, respectively, over the decade. These improvements were largely achieved by increasing compression ratios in SI vehicles (3.6%), turbocharging CI vehicles, and engine downsizing by 5.7–6.5% across all technologies. Simultaneously, vehicle performance improved, through increased engine power density resulting in greater acceleration. Across the decade, tz-62 fell 9.4% and engine power density increased 17% for SI vehicles. For CI vehicles, tz-62 fell 18% while engine power density rose 28%. Greater fuel consumption reductions could have been achieved if vehicle acceleration was maintained at 2001 levels, applying drive train improvements to improved fuel economy and reduced CO2 emissions. Fuel consumption and CO2 emissions declined at faster rates once the European emissions standards were introduced with SI CO2 emissions improving by 3.4 g/km/year for 2001–2007 to 7.8 g/km/year thereafter. Similarly, CI LDVs declined by 2.0 g/km/year for 2001–2007 and 6.1 g/km/year after.

SI Figure 1: Relative availability of LDVs by air-delivery systems, fuel type and year. CI-Naturally Aspirated (NatA) vehicles are represented in the "Other" category. T = Turbocharged.
SI Figure 2: Relative availability of LDVs by fuel-delivery system, fuel type and year. SI -NotAv refers to SI vehicles whose fuel delivery system was not specified.
The mean bore-to-stroke ratio are shown by air and fuel delivery system for 2001 to 2011 in SI Figure 3. SI Figure 3: Annual mean bore-stroke ratios for SI and CI LDVs by air (dashed line) and fuel (solid line) systems. Erratic regressions for SI -Supercharged, CI -NatA, CI -II and SI -DI attributed to sample bias.

Mean Piston Speed
Mean piston speed (S p ̅ ) of the engine is influenced by inertial loads, resistance to airflow and engine friction. Equation 3 (Chon & Heywood, 2000), where denotes stroke length and denotes engine speed. Heywood & Welling (2009) note that the speed at which SI maximum power occurs ( p ̅,max ) is influenced by the choking limit during the intake stroke. Brake power is restricted to a maximum value at p ̅,max , which has led to additional valves, variable valve timing or lift to increase this limit. Additionally, frictional losses increase with engine speed, which is represented with the combination of Equations 2 and 3 within Equation 4.
Increasing the number of valves and introducing variable valve timing improve the breathing capabilities of non-boosted engines (Heywood, 1988). Improved breathing reduces the work required to pump air into the engine and increase the maximum operating speed of the engine. The maximum speed that the piston can travel is limited by the rate at which air can be inducted. Vehicle manufacturers modified SI-NatA engines in this manner to increase maximum mean piston speed, S p ̅ (MMPS) as shown in SI Table 2. Naturally aspirated vehicles saw an increased MMPS of 8.6% over the decade. Conversely boosted engines (CI-T and SI-T) slightly decreased over the decade as airflow is not limited within such systems (Heywood & Welling, 2009). The MMPS is lower in CI LDVs, since SI engines require higher engine speeds to achieve equivalent levels of power and torque. The increased engine MMPS of naturally aspirated SI engines allows for higher power output, thus enabling engine downsizing or better vehicle performance.  Figure 4 shows a correlation between vehicle mass and fuel consumption. Here, each 100 kg reduction in vehicle mass led to reduced fuel consumption by an mean 0.62 L/100 km in CI LDVs (orange diamonds) and 0.42 L/100 km in SI LDVs (green circles).

SI
SI Figure 4: Scatter of vehicle mass against fuel consumption for SI (green circles) and CI (orange diamonds) vehicles. The plot is based on the representation of all available vehicles, including high performance LDVs, where mass may influence the correlations. The exclusion of vehicles over 1,800 kg in mass reduces the coefficient of determination (R 2 ) to 0.6 for both SI and CI.

Comparison of Performance using Normalised Relationships
Normalised engine design metrics are discussed to compare technological LDV advancements, independent of engine size. This includes an analysis of BMEP and specific power by AFD system from 2001 to 2011.
Maximum BMEP and mean piston speed are almost fixed for particular AFD system. Assuming equal bore and stroke lengths for engines of similar geometry ( = , with c = Number of cylinders in Equation 5

(SI 4)
Vehicle performance trends were reviewed for the most popular configurations of AFD system across model year. The assumptions of linear trends were assumed valid as R 2 estimates exceeded 0.96 for all AFD configurations. SI Table 3 shows that relationships for NatA LDVs have the highest concurrence whilst the relationships for turbocharged LDVs have lowest. This reflects engine designers' capacity to utilise boosted systems to simultaneously increase performance and reduce engine displacements.
Engine displacement was plotted against maximum power (SI Figure 5-A) and torque (SI Figure 5-B) to represent power density and BMEP at maximum torque. The gradients indicate that power density is highest for SI-T DI vehicles at 71 kW/l and validates the use of scaling relationships when quantifying LDV performance. For example, it is reasonable to expect engine designers to use SI-T DI engines when designing high-performance LDVs (Heywood & Welling, 2009), and for these vehicles to have highest estimates of power. Moreover, the gradients (see SI Table 3) in SI Figure 5-C show mean BMEP at maximum torque is highest for the CI over the comparable SI, DI over the comparable II and boosted over the comparable un-boosted engine. Differences between BMEP at maximum torque and power are attributed to frictional losses. Larger deviations occurred in CI vehicles, where friction is higher because larger engine sizes are required to accommodate higher compression ratios (Heywood, 1988, Gupta, 2006. Two further correlations are plotted in SI Figure 6 based on SI Equations 3 and 4. The first relates engine power, size and cylinder numbers to the inverse of BMEP which showed the largest variability of all the scaling relationships presented. Variability is lowest for the relationship between acceleration time and mass-to-power ratios at R 2 ≥ 0.98 for all systems. Here, deviations between the slopes (which represent the inverse of mean velocity) are attributed to differences between mean AFD-system velocities and the higher mass-to-power ratios of CI over SI vehicles.  Figure 5: Scaling relationships for (A) maximum power versus engine displacement, (B) maximum torque versus engine displacement and (C) brake power and mean piston speed versus total piston area.