Why your car does not achieve the fuel economy it is supposed to

I have posted several times about energy ratings for electrical appliances, computers and so on and I have complained that the energy ratings for appliances do not relate directly to how much energy they use. Sadly, car fuel consumption specifications are also misleading and I have been reading a report [1] published earlier this year which investigates the issue. They compared the official fuel consumption figures - from tests in a laboratory - with on-the-road consumption statistics from a variety of sources from 2001 to 2011. They found not only do the official figures underestimate real fuel consumption but the difference is getting bigger. In 2001 the difference was about 10% but the most recent data finds 15-35% difference. Some of the sources they used for on the road fuel consumption are available on the internet so you can look up your own car and see what other drivers are getting.

Do you look at fuel consumption figures when you are choosing which car to buy? Maybe you have already lost all faith in them. There is a lot of pressure on manufacturers to improve fuel consumption, at least in terms of the New European Driving Cycle (NEDC) test, the only measure in current use.

Customers are interested in car running costs. As well as fuel costs, the car's fuel use generates carbon emissions which govern the price of road tax. For a Prius T3 with emissions 89 gCO2/km, tax costs nothing. A Ford Focus can have emissions up to 149 gCO2/km, depending on model: that would cost you £140/year [2].

As for regulators, the EU has set a target for manufacturers such that fleet average emissions must be 89 gCO2/km or less by 2020. The manufacturers need to sell lots of cars with low test emissions to meet this requirement, but that doesn't necessarily mean the cars will have quite such low on-the-road emissions.

Of course you can never expect your fuel economy to match the real standard because it depends on many factors such as the speed at which you drive, how often you go up and down hills, how often you get stuck in traffic, how often you use the air conditioner, how much stuff you keep in the boot and your driving style. Maybe it is even too much to expect that the test would be accurate for an average driver in average conditions. You might expect that the consumption figures could be compared reliably between different cars - so a car with 10% higher test results would use 10% more fuel - but even that is not true. Here are some of the reasons for the discrepancies.

Fundamental aspects of the test

Firstly, the test is the same for all vehicles, which means it involves relatively low accelerations that all vehicles can achieve. In practice if you have a powerful car you are likely to accelerate and drive faster - so the test is biased to give low emissions for high performance cars.

Also, in order to get consistent results, the tests are done under repeatable but unlikely conditions. For example, the test starts with the battery fully charged. This means it isn't using any power to charge the battery - in extreme conditions this can make 30% difference though typically it would be much less. Also the test starts with a cold engine but in an ambient temperature of 20-30C : a lot warmer than your average UK winter morning.

The tests are conducted with optional features such as air conditioning turned off. AC takes a lot of power and more and more cars now have it. Presumably people use it. The same goes for seat heaters. This helps to explain why the discrepancy between test and real world results is getting bigger.

Designing the car for the test

The urban test has a 25% idle time and a good number of stops and starts. This means that cars that switch the engine off when idle (start-stop) or that have advanced automatic transmission do very well on the test.

Optimising the car for the road load test

The test is conducted in a lab with the car on rollers - it doesn't actually move. Therefore there has to be adjustment for the 'road load' which is missing in the lab - this is a combination of rolling resistance of the tyres on a real road
and air resistance which depends on the shape of the car. The road load is measured in another test which involves getting the car running on a straight road and then letting it coast to a stop. The road load is calculated from the rate at which it decelerates at different speeds. However, this test is also highly dependent on a number of factors which are likely to vary in the real world.

Weight is obviously important - and this depends not just on the passengers and baggage but also on extra features. The road load test is done on one car per model - so it will be done on a car without any non-standard features which add weight.

The rolling friction also depends on the type of tires and the tire pressure - friction is lower if the tires are pumped up more than usual. Also careful adjustment of the wheel alignment will reduce friction. It is even possible that the car will have its brakes pushed back to avoid 'parasitic drag', caused by friction from the brakes when they are not engaged, sometimes due to brake pads swelling when hot.

The researchers performed tests to determine the road load for typical cars and found discrepancies between their results and the official values of up to 30% at high speeds and 70% at low speeds.

Why the difference is increasing

We have already mentioned some reasons why the difference between test and real world consumption is increasing. Partly this is due to more use of air conditioning and other power hungry extra features. Another issue is that where adding extra weight, for example, adds to the fuel consumption, this is an increasing proportion when the test fuel consumption is low. If the test result is 150 gCOe/kg adding another 9 for extra weight is 6%.  Starting from 90 gCO2/kg, adding the same amount is a 10% increase.

However, it is also likely that the manufacturers are becoming more canny.

A replacement for the New European Driving Cycle test called the worldwide harmonized light vehicles test procedure (WLTP) is under development and should be defined in 2014. It will address some of these issues, in particular it will tighten the rules on the road load assessment and have a more representative drive cycle.

Real fuel economy data sources

The real fuel economy data that the report used came from several sources. From the UK the sources were WhatCar tests in 2011/2012, based on their own tests with professional test drivers and the Honestjohn database, which contains data entered voluntarily by mainly private car owners. Currently this contains 43000 entries for a wide range of makes and models, but this is still small compared to some of the other databases. If you are a driver you could add yours and help make the data more accurate.

It is interesting to note that the database recording the highest extra fuel consumption was from a cohort of company cars in Germany (Leaseplan). These drivers had bigger cars than average (lots of BMWs and Volkswagen) and they were not paying for their own fuel - it was paid for by their employer. The Leaseplan car emissions were on average 33% higher than the official figures in 2011, up from 22% in 2008/2009. The lowest fuel consumption data was also from Germany: the Spritmonitor.de database, which holds data collected like Honestjohn from private car owners volunteering their information. This was only 23% over the official figures.

[1] From laboratory to Road (May 2013) International Council on Clean Transportation
[2] Vehicle tax rate tables www.gov.uk