Fuel Economy Testing Part Two
by David Finlay (30 May 2003)
The first part of our investigation into fuel economy testing showed how it is done (measuring exhaust emissions during a specified virtual "journey" on a rolling road in a laboratory) and asked a few key questions arising from that information. In this second part we'll be doing our best to answer those questions.
Q1. How can a rolling road replicate the aerodynamic and friction resistance found in normal motoring?
The simple answer is that you find out what the aerodynamic and friction resistance is and programme it into the rolling road. The resistance is found by taking an example of the car to a large area of flat tarmac on a day when there is no wind.
The car is accelerated up to 130km/h (just over 80mph) and then allowed to coast to a halt. The time it takes to slow down to various target speeds is then measured, the lowest being 20km/h (12.5mph). The combination of aerodynamic and friction resistance is what makes the car slow down, so the various measurements show the total effect they have - not the individual effects, of course, but the total is what matters here.
A "map" of the overall resistance has now been created, and it can be programmed into the rolling road, which therefore provides the required resistance at all the speeds the car will reach during the test. It might seem that the system is open to abuse, but the coasting test is officially monitored, as is the first economy/emission test, as are subsequent tests taken at random during the production run of the car. But there are still ways round this, as we'll see shortly. In the meantime . . .
Q2. Does a car's fuel economy change as it gets older?
Yes, it does. A brand new mass-production engine does not generally perform as well as it will later in life. One of the main reasons for this is that the surfaces inside the engine are at first slightly rough and therefore generate significant amounts of friction. A few thousand miles' worth of wear makes the surfaces much smoother and brings performance, economy and emissions to acceptable levels.
The official testing takes this into account. The car used in the coasting test, and the one used in the first, monitored economy test, must both have covered at least 3000km, by which time their engines (and indeed transmissions, which are just as crucial in this respect) should be running properly. It's out of the question that production cars could be run for 3000km before being tested and then sold as new, which is where the Green Factor comes in.
The Green Factor assumes that an engine will be running 8% better at 3000km than it will when new, so the test results for new cars are improved by 8%. The improved figures are the important ones; the average of all the cars tested (about 1% of production) must be the same as the manufacturer's claim.
A very intriguing fact about the claim is that it is determined at an early stage of the car's design. Let's say that Manufacturer A builds a car which gives 40mpg on the combined cycle. If Manufacturer B then wants to build a similar car, it would be a major disadvantage if it only gave 32mpg. The engineers are therefore told to build a car which will return 40mpg. If they can't do this, they are in big trouble. That leads us nicely into . . .
Q3. Why do some cars seem to achieve their combined economy figures with ease while others get nowhere near them?
Well, now. The first thing to bear in mind here is that, as described in Part One, the virtual route designed for the European economy test does not bear much resemblance to normal motoring. A huge number of external factors including weather, traffic, tyre pressures, the amount of weight being carried in the car and so on can all affect fuel economy to an enormous extent.
It's also worth bearing in mind that even a small change in the state of the car can also make a big difference. A car that passes the economy test will fail it immediately afterwards if it goes through the same test with the sidelights switched on. And sidelights don't use a lot of power. If you're driving through the rain on a cold night with the headlights, windscreen wipers and air-conditioning all working, while also listening to the radio and using the satellite navigation system, the car's fuel consumption will soar.
Still, we've had several e-mails from owners who say that their cars cannot be persuaded to meet the combined economy figures in any driving conditions. Could it be that some manufacturers fiddle the figures?
We couldn't possibly comment. But it can be done. Think of the coasting test mentioned above. The car presented for that test may not be the same as the car you buy in the showroom. It may have a "blueprinted" engine whose components have all been carefully chosen and/or machined so that it works beautifully and produces less resistance. The same may apply to the transmission. The car may be very light - not because standard equipment items have been left out, but because a particular specification has been chosen, for example one for a country where the range of standard equipment is smaller than it is elsewhere. The smallest and least resistant tyres available for the vehicle may be fitted. The alternator may have no internals. A very thin engine and/or transmission oil may have been used.
If any or all of the above are the case, the car will take a lot longer to coast to a halt because there is so little resistance. That reduced resistance will be programmed into the rolling road, so production cars have less to fight against when they are tested and will use less fuel on the virtual route.
The Green Factor may also be brought into play. You'll remember that it's set at 8%, but a manufacturer might put in a claim that the figure should be higher, because its cars improve by more than that in the first 3000km. Such a claim has to be backed up by evidence, and one way of doing this is to build an engine that is completely mismatched ("reverse blueprinted" so to speak) and then run in very carefully so that it works properly at 3000km.
If it's tested at the beginning and end of this process, the car may well be producing figures that are considerably more than 8% better during the second test than during the first. Let's say it's 12% better - its Green Factor is therefore raised to that amount. Figures achieved by tested production cars are improved by 12%, which may bring them to a level that the car could not possibly achieve in the real world.
Other tactics are more blatant still. Imagine a front-wheel drive car on a rolling road. The front wheels are on the rollers, the rear wheels are sitting there doing nothing. The tester applies the handbrake to keep the car steady . . . and the handbrake lever pushes a switch which tells the engine's ECU to use a completely new ignition/injection map designed specifically for economy tests!
That wouldn't work if you had a rear-wheel drive car. But you're allowed to open the bonnet during the test to help keep the engine cool (remember that the car isn't actually going anywhere, so there is no high-speed flow of air). Some cars have engine compartment lights which come on when you open the bonnet. As with a fridge, there's a switch which makes this happen, and once again the same switch could conceivably tell the ECU to read a special ignition/injection map.
To conclude: economy testing is a very complex process, and in some ways a very flawed one, which is nevertheless probably about as good as it can be. The figures it produces can't be taken entirely seriously for individual cars, but they do give a moderately helpful guide (no better or worse than 0-62mph acceleration times, for example) as to how different cars of the same type compare against each other. For potential buyers they provide reasonably useful information, though approaching them with a healthy dose of cynicism is perhaps advisable.
