Car fuel economy claims are miles from the truth

“EU tests are performed on engines at full operating temperature.” Not so – only the extra-urban cycle, that is continued after the urban, uses an initially hot engine. However, tests are done in high ambients. Dispel the myth that the EU tests represent real life- they don’t and are not claimed to. Just an attempt at comparing vehicles in a relative way. Use other data, as above, to get an idea of real life mpg (l/km).
“Urban cycle
The urban test cycle is carried out in a laboratory at an ambient temperature of 20°C to 30°C on a rolling road from a cold start where the engine has not run for several hours. The cycle consists of a series of accelerations, steady speeds, decelerations and idling. The maximum speed is 31 mph (50 km/h). The average speed 12 mph (19 km/h) and the distance covered is 2.5 miles (4 km).”

Thanks Malcolm. I stand corrected. I see the EU as part of the problem here.

Many of us are aware that the EU fuel economy figures are not representative of real life. That does not stop car dealers using these figures to sell cars. As I reported earlier, an experienced sales rep tried that one on me and he received a lecture about dishonest selling.

I agree, Flocker. Even if the simulation is not perfect we could still have figures that are attainable.

One point I disagree about. Even when fuel is not being injected, the engine acts as a brake and thus affects fuel consumption. The way that I can sometimes achieve fuel economy figures similar to the published values is to allow the car to run down long gentle slopes out of gear so that there is no braking effect of the engine. The car remains in control and I don’t inconvenience other motorists.

The NEDC (New European Driving cycle) was devised when cars were less powerful and so acceleration was gentler than now. So the cycle needs updating. There is a report produced by the Joint Research Centre (JRC) that supports EU policy making that you may find interesting:
http://publications.jrc.ec.europa.eu/repository/bitstream/111111111/22474/1/co2_report_jrc_format_final2.pdf
The current cycle appears, on average, to underestimate fuel consumption by around 10-15% for petrol, and 12-20% for diesel compared with “in use” driving. For drag (including air resistance) it seems the dynamometer settings that simulate this are based in ideal data that underestimates its effect. The report says:
“5.6 Type-Approval optimization
Although this is not directly an outcome of the study, this is an important conclusion from relevant work that should be re-iterated. Type-approval tests of fuel consumption are conducted on chassis dynamometer using resistance settings provided by the manufacturer. These settings are derived from coast-down vehicle tests. It appears that resistance of actual vehicles measured by independent test centres are higher than the ones submitted by the manufacturers for the type-approval tests. There are several reasons why this can be happening, i.e. manufacturers test vehicles in ideal conditions (tarmac condition, weather, vehicle run-in, configuration such as
tyre dimensions, trained drivers to perform the test, etc.). Unfortunately, type-approval resistance settings are confidential. Using of real vehicle resistances instead of type-approval resistances has been shown to lead to fuel consumption increases of up to 17%. This is even beyond the in-use over type-approval fuel consumption ratio developed in this report. As a minimum impact this means that maybe the NEDC is not a bad (underpowered) cycle to report fuel
consumption but that maybe the actual test is an idealistic one. It can be recommended that vehicle resistance settings become public together with the type-approval fuel consumption value, so that independent authorities can check both whether these represent reality and whether the type-approval test has been conducted as required.”
I would suggest that it is the test rules and conditions that need strengthening, both to reflect the performance of modern vehicles and to avoid the loopholes that can, and will, be exploited by all manufacturers.

I didn’t buy my Mitsubishi Outlander PHEV on the basis of the manufacturer’s claim – only a fool would do that, surely. I looked at the claims, read a bit about the type of driving I would do, investigated further and made my decision.. I am currently getting 140mpg (excluding electricity costs, of course).

Andrew, as far as I am aware the EU figures are not portrayed as mpg (l/km) “claims”. They are results of tests under very specific and consistent test conditions for a particular purpose, which is not to tell you what you will get in practice.

The DfT state “The fuel consumption figures quoted in this guide are obtained under specific test conditions, and therefore may not necessarily be achieved under ‘real life’ driving conditions. A range of factors may influence actual fuel consumption – for example, driving style and behaviour, as well as the environment and conditions under which the vehicle is operated. Furthermore, since several different specifications (variants or versions) of a given model may be grouped together in the list, the figures used in this guide should be treated as indicative only.

A definitive figure for a given specification of vehicle will be available at the point of sale.

The most helpful results for the motorist would be to produce a spread of values between which most drivers would expect to find their own consumption. I don’t see multiplying an unrepresentative EU figure by a single factor makes them any more representative. Some do already get mpg (l/km) near the published values – I do on my Espace – but others don’t. Real life testing has rather different objectives from the EU test cycle. Another way is to test a car over a day or a weekend before you buy to see if the it lives up to expectations.

The fact that the present EU test is unrepresentative of real life is the problem. Having now seen the magazine, Which? refers to it as an obsolete test. Until we have official tests that are more representative of what motorists are likely to achieve, scaling the EU figures as I have suggested would give more realistic figures. Like the ‘up to’ figures for broadband speeds, we need more realistic figures and I am grateful that Which? publishes these.

What I have never understood is the effect of temperature on fuel economy, which I have seen with both petrol and diesel cars. I can understand why a car performs less efficiently until the engine has reached operating temperature, but after that there is still a marked difference. Yesterday I made a journey of 50 miles, mainly on A-roads, and at the end of the journey the meter showed 73 mpg. For the same journey in January, the figure was about 55 mpg.

First, the EU test conditions lag behind vehicle development – for example acceleration capabilities are much higher than when the test cycle was devised. So the test certainly needs updating; a problem with all standards is the time they take to agree paticularly in an industry that can changes fairly quickly.

That said, the mistake is to portray the EU consumption test as something they do not set out to be – suggesting that is simply to mislead. They have a particular purpose. Use published “real life” figures to get nearer what you will achieve – but don’t expect those to be accurate for you either.

According to the US Dept of Energy:
Cold weather affects your vehicle in more ways than you might expect:
•Engine and transmission friction increases in cold temperatures due to cold engine oil and other drive-line fluids.
•It takes longer for your engine to reach its most fuel-efficient temperature. This affects shorter trips more, since your car spends more of your trip at less-than-optimal temperatures.
•Heated seats, window defrosters, and heater fans use additional power.
•Warming up your vehicle before you start your trip lowers your fuel economy—idling gets 0 miles per gallon.
•Colder air is denser, increasing aerodynamic drag on your vehicle, especially at highway speeds.
•Tire pressure decreases in colder temperatures, increasing rolling resistance.
•Winter grades of gasoline can have slightly less energy per gallon than summer blends.
•Battery performance decreases in cold weather, making it harder for your alternator to keep your battery charged. This also affects the performance of the regenerative braking system on hybrids.

Hard to see that increasing your fuel consumption by 25% – perhaps you had a heavier foot!.

It is a good illustration of the difficulty in giving “real life” fuel consumption, isn’t it 😀

The trouble is that the EU test figures are used in the real world to mislead us.

When I last bought a car I told the dealer that I was most interested in two models (and that I would not consider anything without a spare wheel). Within five minutes I was being told what mpg I could expect. He received a pre-prepared lecture about his dishonesty and I referred him to the sales brochure which briefly explained that the published figures were not representative of what a driver could expect to achieve.

The fact that the sales brochure explained that the EU tests were not “real world” is honest. Hard to legislate for an individual being either ignorant or dishonest, and not grounds to condemn a test regime designed with a particular purpose.

Malcolm – Ask motorists if they are happy with the official fuel economy figures and most people will say that they are not. We need something more realistic, fit for the 21st century.

The mpg figures I quoted were just for a recent comparison of fuel economy. The effect of ambient temperature is not usually so great but I do wonder why the mpg improves so much in warm conditions. I had assumed that engine oil temperature might be significantly different between summer and winter motoring, but not according to the gauge in my present car. I check my tyre pressures weekly but temperature could still affect rolling resistance.

“Malcolm – Ask motorists if they are happy with the official fuel economy figures and most people will say that they are not. We need something more realistic, fit for the 21st century.”

Then don’t perpetuate the myth, by articles like this, that the EU tests are somehow dishonest. They give results from standardised tests that manufacturers must publish. They have a particular purpose, including emission data, but were not designed to generate real life claimed performance. New tests are being devised to be more representative of current car technology and driving situations but will no doubt still give comparative data by virtue of their need for standardisation.

It is a pity Which? resorts to being “economical with the facts” to apparently gain publicity. I wish it operated in a more objective way otherwise it is in danger of damaging its credibility as an independent consumer watchdog.

It would be better if Which? gave an objective explanation of the facts so people can know to look at either their results, Honest john’s, or other motoring press sites for so-called “real life” consumptions. But will you achieve these? Of course not – your driving style, journey type, topgraphy, tyre inflation, etc will all influence significantly the consumption you get. The range will be quite large. No such thing as a real life consumption – it will be a real life range within which you have a probability of falling.

We will have to agree to differ Malcolm. I support Which? in alerting us to the large difference between published fuel economy figures and what we are likely to achieve. It’s clear to me that the EU is listening to the manufacturers rather than consumers, otherwise we might have standard tests that better reflect average use.

DfT state the following, literature states it, so “alerting us” as if it was a calculated deception is a distortion.

“How Representative of Real Life Driving are the Standard Tests?
Because of the need to maintain strict comparability of the results achieved by the standard tests, they cannot be fully representative of real-life driving conditions. Firstly, it is not practicable, nor is it viable to test each individual new car. Only one production car is tested as being representative of the model and this may produce a slightly better or worse result than another similar vehicle. Secondly, there are infinite variations in driving styles, as well as road, car and weather conditions, all of which can have a bearing on the results achieved. For these reasons the fuel consumption achieved on the road is unlikely to be the same as the official test results. The purpose of the official fuel consumption test is to provide data that will permit a comparison of the fuel consumption of different cars, rather than to provide an estimate of average, on-the-road, fuel economy.

It is recognised that, for a variety of reasons, the fuel consumption achieved by the majority of motorists is poorer than that suggested by the standard tests, and work is going on with the intention of introducing a new test cycle which will better represent the way in which most people actually use their cars.”

Of course there are numerous reasons why there will be substantial variation in the fuel economy achieved by drivers. The problem is that the way that the official tests are carried out mean that they generally produce figures that are substantially above what most motorists would achieve in practice.

Why not design tests so that they correspond to the average or the median of what motorists achieve? If they are carried out under standard conditions then the comparison between models should remain valid.

In defending the motor industry, can you really justify manufacturers over-inflating tyres to help improve the fuel economy figures for their cars?

It would make us all feel a bit better if the tests were designed as you suggest, or even to a slightly lower benchmark, so that we could derive some satisfaction from out-performing the standardised performance test results. So long as the tests were truly standardised and regulated to prevent distortion the actual values would become less important [while at the same time becoming more meaningful] but the relative positions of different cars would become more credible and reliable thus assisting purchasers to make their choice.

Apart from the point that no two vehicles off the same production line will perform identically, no car will perform the same in two or three years’ time as it did when new and might well have diverged markedly from any proximity to the declared fuel consumption figures for that model. Since sales of used cars far outstrip sales of new cars there could be a case for checking whether there is a ‘standard’ deviation over time for any particular model from its norm in comparison with the deviations of models that were orginally similar. Or it might just all be too random and abstruse and not worth bothering about, letting anecdote and hearsay prevail; those who keep fuel and mileage records at least have something to report when the question is asked.

John wrote: “It would make us all feel a bit better if the tests were designed as you suggest, or even to a slightly lower benchmark, so that we could derive some satisfaction from out-performing the standardised performance test results.”

I agree. It is always a pleasant surprise when something proves to be better than expected, for example a really good meal, a tradesman’s bill that is less than the estimate, an unexpected act of generosity. In many respects, cars can exceed our expectations but fiddling economy tests is as bad as advertising one price and charging more.

10th/11th line should read “than the claimed less economical car”.

AQ – I have no problem with a manufacturer that designs a car to perform well in fuel economy tests. If it is more streamlined than other cars I can check if that affects space inside the car or the boot. If they decide to leave out a spare wheel, that’s fine as long as I am made aware, so I can avoid buying the car.

What I object to is carrying out tests on any car that has been modified in any way.

As you say, motoring journal tests may be focusing on performance rather than reflecting normal driving. The ‘real life’ tests may be of highly variable standards. Some drivers will take great care to base their measurements on careful long-term measurements whereas others will report what is shown on their car computer from a single fill-up. That’s why we need proper testing of unmodified vehicles under controlled conditions. As has been said several times there are many factors relating to how the car is used that will determine what figures you and I can achieve.

As I see it, preparing a car by over-inflating the tyres or taping-up the doors is equivalent to changing the price label in a shop, so that the sales assistant charges less than the intended price. One of the reasons that we have bar codes and tills that show a description of the goods is to tackle the fact that we have some dishonest customers. Most customers are honest (I hope) and would not resort to this or other form of dishonesty. I hope we have some honest car manufacturers that test cars that have not been prepared in any way. The problem is that we hear about the ones that cheat.

Oops – that should read: “The ‘real life’ tests may be of highly variable MEASUREMENTS.”

I accept that emissions present a problem and that pragmatic solutions are needed to make progress. If there is a problem with implementing new emission standards then their introduction could be delayed, as we have seen with the planned phase out of halogen bulbs.

What I will not accept is manufacturers taking cars that you or I could buy and preparing them in a way that will return artificially high fuel economy figures.

I agree that our policy makers deserve criticism, but that is not a reason for modifying cars before testing. I would like to know if there are any manufacturers that do not modify their tests. That would gain a lot of my respect.