IEA raises forecast global oil consumption in 2013 to 90.8 mb/d
Chinese automakers bring hybrids and plug-ins to Detroit, Geneva auto shows

ORNL researchers quantify the effect of increasing highway speed on fuel economy

John F. Thomas, Brian H. West and Shean P. Huff
Fuels, Engines and Emissions Research Center, Oak Ridge National Laboratory

Figure 1. Vehicle installed on the chassis dynamometer in the ORNL vehicle research laboratory. Click to enlarge.

Oak Ridge National Laboratory (ORNL) staff have been performing vehicle research and testing in support of the website. This website, jointly maintained by the US Department of Energy and the US Environmental Protection Agency (EPA), provides information such as official EPA “window label” fuel economy estimates for city, highway, and combined driving for all U.S.-legal light-duty vehicles from 1984 to present. It also offers consumer information and advice pertaining to vehicle fuel economy and energy-related issues such as driving tips. One tip is that drivers should obey the speed limit since the fuel economy of most vehicles decreases above 50 mph [80 km/h].

ORNL staff members John Thomas, Shean Huff and Brian West have worked to quantify this trend through analysis of dynamometer testing results for 74 vehicles at steady-state speeds from 50 to 80 mph [80 to 129 km/h].

Data has been collected for 23 light-duty vehicles at ORNL’s vehicle research laboratory and a valuable data set for 51 vehicles was loaned to ORNL by Chrysler, LLC under a non-disclosure agreement. Vehicles were tested in dynamometer laboratories at steady speeds from 40 to 80 mph [64 to 129 km/h], with the proper road-load applied. Analysis has focused on speeds of 50, 60, 70 and 80 mph.[1] The data resulting from these tests simulates steady highway cruising on flat roads at moderate temperatures (SAE J2263, J2264).

The study includes various sizes of sedans, wagons, and SUVs, as well as pickup trucks, minivans and a few “muscle” and sports cars. Vehicles from model years 2003 to 2012 with a wide variety of powertrains were represented and included two hybrid sedans and a diesel sedan. The combined data from the 74 vehicles gives insight into the effect of cruising speed on fuel economy.

Results are quantified in the summary table, showing the general effect of increased speed on fuel economy. For example, the last column in the table for the row with 70 to 80 mph results reveals that most vehicles will have 12.5-17.5% drop in fuel economy due to traveling 80 mph rather than 70 mph. No obvious pattern for specific vehicle types in terms of fuel economy percent change with speed has yet emerged (for example, results for SUVs were similar to small sedans or pickup trucks in terms of percent change in mpg).

Brief summary of vehicle data
Speed increase Percent mpg decrease for a given 10 mph increase based on 74 vehicles.
Average Data range Std. deviation Middle 2/3s of vehicle data
50 to 60 mph 12.4 6.9-18.3 2.2 10.0-14.3
60 to 70 mph 14.0 8.8-19.5 2.6 11.2-16.1
70 to 80 mph 15.4 10.8-26.0 3.0 12.5-17.5
All three speed increments 13.9 6.9-26.0 2.9 N/A

The results are summarized in histogram form in Figure 2, which shows the distribution of fuel economy penalties for each 10-mph increase in speed from 50 mph to 80 mph. A comparison of the three histograms shows a slight shift toward higher mpg penalties for each 10-mph speed increase. In other words, the mph penalty for increasing your speed from 70 mph to 80 mph is slightly greater than the penalty for increasing from 60 mph to 70 mph, which is slightly greater than the penalty for increasing from 50 mph to 60 mph.

Figure 2. The fuel economy penalty (for a 10-mph increase in speed) becomes more severe at higher speeds.[2] Click to enlarge.

The bottom histogram highlights some further interesting observations that were made from the data. There are explanations for four of the five vehicles represented in the “warm” colored bars showing the high fuel economy drop values (> 21%) for the 70 to 80 mph case.

Three V8 sedans with cylinder deactivation were included in the vehicle set, and were responsible for some of the largest values seen in the table and Figure 2. These vehicles conserve fuel by only powering 4 cylinders at lower speeds, and then switch to 8 cylinders when needed to meet the power demand at higher speeds. The switch from 4 to 8 cylinders was seen to occur between 60 and 70 mph or 70 and 80 mph, and this change causes a large percent change in fuel economy over that speed increment.

An ORNL tested vehicle was observed to transition from stoichiometric operation to protective enrichment between 75 and 80 mph, and this change caused an unusually steep drop in fuel economy. Protective enrichment occurs when a vehicle’s engine is at a high load point such that the exhaust temperatures may become great enough to damage the catalyst or other components: damage is avoided by injecting extra fuel (rich fueling) which produces lower temperature exhaust. The Chrysler data included a vehicle that appears to employ protective enrichment at 80 mph. These two vehicles are responsible for the two highest fuel economy drops in the data set (~25 and 26% when comparing 70 to 80 mph).

Testing was not conducted beyond 80 mph, but it is reasonable to think more vehicles would transition into protective enrichment operation at speeds above 80 mph. This may be a valid consideration for setting maximum speed limits.


  • [1] Road-loads are determined by on-road coastdown testing, and are simulated by the chassis dynamometer according to SAE Standards J2263, J2264.

  • [2] 68 vehicles are represented in the histogram for traveling 80 mph versus 70 mph. The other 2 histograms contain data for 74 vehicles. Six vehicles were tested only to 70 mph.


  • J.F. Thomas, H-L. Hwang, B. West, S. Huff, Predicting Light-Duty Vehicle Fuel Economy as a Function of Highway Speed, SAE technical paper 2013-01-1113, SAE 2013 World Congress, Detroit, MI, April, 2013 (in press)



Fuel consumption should be given for various:

1. Highway speeds, 60, 70, 80 mph etc
2. Highway Terrains, flat, hilly, very hilly etc
3. Individual City driving skills, light foot, medium foot, heavy foot etc.
4. On-board loads..300 lbs. 600 lbs etc

This way, EPA's estimates could be closer to reality and buyers would know what to expect?


With all those promised connected cars, we'll have a good tool to measure precisely what an efficient car is and what is an efficient car driver is.

I don't expect car companies to allow benchmarks as proposed by HarveyD, with regrets though.

Meanwhile reality versus mpg tags will diverge more and more.


You know if enough of us insist driving at 60 mph on the highway than other people would have to drive at the same speed as well.

Nick Lyons

I would be happy if drivers were content to drive only 5 mph over the limit, frankly. 80mph is flow of traffic speed here on many CA freeways. (Except when it's ~5 mph, of course)


5 over our limits would be either 80 or 85.

Anyway, drag goes as v**2 and power required goes v**3


It would be so easy to impose speed limits with a few hundred/thousand automated radar camera stations, at no cost to the States.


This is something every driver should test for themselves.. and many cars will react differently as the transmissions drop in-and-out of locked torque converter status. The point is to keep the engine highly loaded to its optimum level, without damaging anything.


I think the non-linear effects of wind resistance at higher speeds was overlooked (or not explicitly mentioned) by using a dynamometer in a closed environment? This implies even greater fuel consumption in real world driving.


What a waste of time.

1. Who didn't know this already.

2. Drive both ways on an open road on a low wind day and REAL air resistance is included (I assume - I hope- they at least included CALCULATED wind resistance).

The results are going to be generalized anyway.


"Anyway, drag goes as v**2 and power required goes v**3" means higher speed, MUCH worse gas mileage.

TT, by all means deny physics, logic, and generalized reality in EVERY comment anyway.

Nick Lyons

As for road loads, please note the first footnote above:

[1] Road-loads are determined by on-road coastdown testing, and are simulated by the chassis dynamometer according to SAE Standards J2263, J2264.



Of course these dynamometers are made to account for the laws of physics.


Oh, and that comment goes for TT as well. Dynamometers are much more sophisticated than you seem to realise.


1. I have no complaint with dynamometer sophistication.
But I do wonder what all this reveals that is not already well known.
Are they going to provide detailed mpg for endless combinations of speed and distance for each individual vehicles?

2. Although drag does go up as V^2 - and this does mean the irreversible losses (and gas mileage) are MUCH worse at higher speeds - WASTED POWER only goes up the same as drag; with v^2 not V^3.
This is because while the higher mph increases gallons/hour it also provides more miles/hour, so power required is cubed (and limits max speed), gallons per mile is only squared.

Hybrids obviously provide remarkably improved mpg in the city because of the momentum recovery.

Hybrids can provide improved mpg at low highway speeds IF they cycle the ICE at high manifold pressure, (which would emulate an ICE auto with the ability to run efficiently at very low rpm - a very high overdrive and high manifold pressure).

But I believe mainstream hybrids and EREVs are not designed with enough excess HP to cycle the ICE at HIGH highway speeds and consequently get increasingly worse mileage (than ICE autos) on long trips as cruising speed increases – the ICE is more or less flat out, and the batteries and electric motors are just costly excess weight and volume.

As battery technology improves, this disadvantage should vanish.


Do any modern, naturally aspirated engines actually transition from stoichiometric operation to enriched just to protect the engine or converter?

Max torque/power is achieved by enrichment, and this makes some sense (compared to a larger engine) as long as it enriches only moderately and occurs rarely.



My car does regularly drop out of closed loop to add extra enrichment specifically for catalytic converter protection. MY 2004 so modern enough I suppose?


Yes a 2004 should be modern enough but I question that it is specifically for catalytic converter protection.

Since the 1940s until recently, an enrichment valve or metering rod was employed to provide more fuel for WOT operation because a richer mixture will produce more torque/power.
But this was WAY BEFORE catalytic converters.

Why would they use excess fuel to prevent knock and cool valves, engines and cat converters in a naturally aspirated (low specific output) engine. More torque, sure. To cool the CAT, why? Sounds like poor CAT design.

The comments to this entry are closed.