Effects of Vehicle Speed and Engine Load
on Emissions from In-Use Light-Duty Vehicles
Principal Investigator:
Robert Harley
Civil & Environmental Engineering
University of California, Berkeley
510-643-9168
harley@ce.berkeley.edu
The goals of this study are to determine how emissions from light-duty passenger vehicles depend on vehicle speed and engine load in an on-road setting, and to assess trends in vehicle emission rates over time.
While much research has been conducted to characterize vehicle emissions over standardized city and highway driving cycles, it is not clear how well these data represent what takes place on-road. The small sample of vehicles used in laboratory testing may not represent accurately the contribution from malfunctioning and high-emitting vehicles that are responsible for the majority of exhaust emissions from the vehicle fleet. A key gap in our understanding is the effect of changes in vehicle speed and engine load on average emission rates for the on-road vehicle fleet.
Vehicle fuel consumption and emissions are sensitive to a wide variety of factors including vehicle speed, acceleration, roadway grade, wind resistance, tire-roadway friction, use of accessories such as air conditioning, etc. Transportation engineers seeking to reduce travel delay may also influence air pollutant emissions through interventions such as traffic signal timing, on-ramp meters, posting and enforcement of speed limits, and high-occupancy vehicle lanes. Traffic congestion and aggressive driver behavior are also likely to influence vehicle emissions. Unfortunately, large gaps remain in the quantitative description of vehicle speed and engine load effects on vehicle emissions.
Progress has been made in developing modal emissions models (Barth et al., Transportation Research Record, 1520: 81-88, 1996; Washington et al., International Journal of Vehicle Design, 20: 351-359, 1998; Fomunung et al., Transportation Research Part D, 4: 333-352, 1999). These studies mostly rely on results of laboratory dynamometer testing of individual vehicles. However, it has proved costly and difficult to acquire a large enough random sample of vehicles to ensure the overwhelming contribution to total fleet emissions from malfunctioning/high-emitting vehicles is represented accurately. Measurements of vehicle emissions in roadway tunnels are a useful and complementary method for addressing this question (Pierson et al., Atmos. Environ. 30: 2233-2256, 1996; Kirchstetter et al., Environ. Sci. Technol., 33: 318-328, 1999). In this research, we use on-road measurements of vehicle emissions in a California highway tunnel to understand changes in emission factors for oxides of nitrogen (NOx), non-methane organic compounds (NMOC), and carbon monoxide (CO) as a function of changing vehicle speed and engine load.
Many changes to vehicle engine and emission control technology and fuel properties have occurred in the last decade. A second objective of this research is to continue an assessment of long-term trends in emissions from light-duty motor vehicles.
The sampling site for this investigation is a highway tunnel in northern California. The Caldecott Tunnel is located on Highway 24 in the San Francisco Bay area and connects the inland communities of Contra Costa county with Oakland, Berkeley, and San Francisco. The tunnel is 0.97 km (0.60 miles) long on a 4.1% grade. The tunnel has three bores, each with two lanes of traffic. Heavy-duty vehicles are restricted from the center bore of the tunnel, allowing measurements of light-duty vehicle emissions in this bore. The traffic direction through the center bore is changed to accommodate the dominant flow direction, so emissions due to both low-load conditions (downhill and westbound) and higher-load conditions (uphill and eastbound) can be measured at one site. The volume of traffic through the tunnel also changes over the course of the day, which in turn affects the speed and engine load of vehicles in the tunnel, particularly during the afternoon.
We measure concentrations of CO, NOx, NMOC, and carbon dioxide (CO2) at the tunnel entrance and exit. The tunnel’s mechanical ventilation system is not used during our measurements, so air flow is entirely longitudinal and is driven by the flow of vehicles through the tunnel and prevailing winds. We subtract tunnel entrance concentrations from those measured at the tunnel exit, and then divide the measured concentration difference for each pollutant by the increase in total carbon concentrations (dominated by CO2) inside the tunnel. Since the source of carbon measured in the tunnel is the carbon present in gasoline, by carbon balance we compute emission factors in units of mass of pollutant emitted per unit of fuel burned.
Field measurements were completed in July and August of 2001. Downhill (low-load) driving was measured over 5 weekday morning sampling periods from 5 to 11 AM. Emissions from vehicles traveling uphill in the afternoon and evening were measured over 9 sampling days. Fridays were not included in the uphill driving analysis because afternoon traffic volumes and driving conditions on this day are different from the rest of the work week. On most days, afternoon sampling took place from 2 to 8 PM, though sampling was extended as late as 10 PM on some days to quantify emissions occurring at night under higher-speed driving conditions.
Observed driving conditions included a range of highway speeds for both downhill and uphill traffic. For downhill driving in the morning, speeds were typically 80-100 km h-1, with a minimum occurring during rush-hour when the traffic volume was highest. For uphill driving in the afternoon, speeds were typically 60-90 km h-1. The minimum vehicle speeds again occurred when traffic volume was highest (4-6 PM).
Emissions of CO were typically 16-34 g L-1 during downhill driving and ranged from 27 to 75 g L-1 during uphill driving. The CO emission factor for downhill driving is not significantly different from the uphill measurement for 4-6 PM, despite the fact that these driving conditions involve different engine loads. Effectively, there is an engine-load threshold for CO emissions, below which the CO emission factor was relatively constant.
Average NOx emission factors for downhill driving in the morning range from 1.1 to 3.3 g L-1. For uphill driving, NOx emission factors typically range between 3.8 and 5.3 g L‑1. In contrast to what was observed for CO, downhill driving has a lower NOx emission factor compared to all uphill driving conditions. Similar to results for CO during the afternoon, NOx emission factors are high before rush-hour, reach a minimum during the 4-6 PM period, and are at a maximum at the end of the day. While both CO and NOx emission factors follow this pattern, the CO emission factors vary more dramatically than NOx. Fuel-based emission factors for NOx showed a larger change between downhill and uphill driving, whereas fuel-based CO emission factors showed a larger change between lower-speed and higher-speed uphill driving.
The hydrocarbon (technically, the nonmethane organic compound or NMOC) emission factor for downhill driving was 5.0 ± 0.4 g L-1, compared to 1.3 ± 0.1 g L-1 for uphill driving. Therefore, the hydrocarbon emission factor for low-load downhill driving is 4 times greater than the corresponding emission factor for uphill driving. Note these emission rates are expressed per unit of fuel burned, with a higher rate of fuel consumption measured for uphill driving. When expressed per unit distance traveled, hydrocarbon emission rates were similar for uphill and downhill driving, whereas CO and NOx emission factors were both higher for uphill driving. The reason for this behavior is not clear.
Emissions of NOx, NMOC, and CO have been measured at this site in summers 1994-97, 1999, and 2001. From 1994 to 2001, emission factors decreased by 49 ± 4% for NOx,
62 ± 5% for CO, and 67 ± 7% for NMOC. These reductions are due to factors including replacement of older vehicles with less-polluting new vehicles, and reformulation of gasoline that occurred mainly between 1995 and 1996. Over the entire period from 1994 to 2001, the effect of continuing turnover in the vehicle fleet appears to dominate the emissions trends reported here.
Understanding how emissions depend on vehicle speed and engine load is necessary for assessing impacts of transportation projects that alter the flow of traffic. All three pollutants studied here were sensitive to driving mode.
Light-duty vehicle emission factors are decreasing more rapidly than statewide use of gasoline is increasing. Therefore, the local and regional-scale air pollution burden imposed by passenger vehicle travel has decreased despite increased numbers of vehicles on the road and increased vehicle use.
Vehicle operating conditions not represented in the tunnel measurements include low-speed/idle, high-speed (>100 km h-1) and high-load (>20 W kg-1) conditions, and cold starts. Additional on-road measurement sites and sampling methodologies are needed to complete the assessment of the effects of changes in driving conditions on emissions rates. The high hydrocarbon emission rates observed during downhill driving in this study are surprising and deserve further attention.
Conference Presentations
Light-Duty Vehicle Emissions and Their Relation to Engine Load at Highway Speeds.
Coordinating Research Council (CRC) Workshop on On-Road Vehicle Emissions, San Diego, CA, Apr. 2002. Presented by Andrew Kean (Ph.D. candidate at UC Berkeley).
Trends in Exhaust Emissions from In-Use California Light-Duty Vehicles, 1994-2001. SAE Spring Fuels and Lubrications Meeting. Reno, NV, May 2002. Presented by Andrew Kean.