Your mileage may vary: This is how the range of electric cars is tested

As a wide range of electric cars were shown off at the Geneva Motor Show this month, it became increasingly obvious how difficult it is to definitively say how far each can travel between trips to the charger.

For example, when reporting on the new Jaguar I-Pace, we explained how the car has a range of 234 miles, 298 miles, or 336 miles depending on how the car's performance is measured. The first figure is recorded by the US Environmental Protection Agency (EPA), the second is the figure given by the Worldwide Harmonised Light Vehicle Test Procedure (WLTP), and the final number is the old New European Driving Cycle (NEDC).

These are not mere rounding errors, or results which smartly take into account different speed limits, road surfaces, weather conditions and driving styles in the US and Europe (if only). There is a difference of 28 percent between the smallest and largest figure. It's a situation not far off your current car sputtering to a halt when the fuel dial claims you still have quarter of a tank left.

It might even be worse than that, because the range of electric cars is affected more by ambient temperature and gradient than their internal combustion relatives, and these figures do not take into account how the battery's health might fade over time, or how EV drivers are extremely unlikely to drive their car from 100 percent to zero, then recharge to full again.

The biggest issue with how the current cycles test electric vehicles is the difference in maximum speed. Where the American EPA cycle only takes cars to a top speed of 60mph, Europe's NEDC and its replacement, the WLTP, take cars up to 75mph and 82mph respectively — speeds at which electric vehicles quickly deplete their battery.

To give an idea of the problem, here are the EPA-rated ranges of some common electric cars:

• BMW i3 - 114 miles
• Chevrolet Bolt - 238 miles
• Nissan Leaf - 107 miles
• Tesla Model S 75D - 259 miles

And here are those same cars, as tested on the European NEDC cycle:

• BMW i3 - 186 miles
• Chevrolet Bolt - 323 miles
• Nissan Leaf - 155 miles
• Tesla Model S 75D - 304 miles

The estimated range for these cars using the new WLTP cycle falls in-between these two sets of data.

Here is an overview of how each test cycle works:

United States EPA's Federal Test Procedure

As with all of three tests highlighted in this article, The Environmental Protection Agency's cycle is conducted in a laboratory and on vehicles driven on a rolling road.

Established in 1978 and last updated in 2009, the main test is 11.04 miles long and is conducted at an average speed of 21.1mph.

The EPA also conducts a series of other tests, including one designed to simulate urban driving with frequent stops over the course of 7.5 miles. This test has a maximum speed of 57mph and an average speed of 19.6mph. A highway driving test averages 48mph over the course of 10 miles and has a maximum speed of 60mph, lower than the highway speed limit of most countries.

The NEDC and WLTP, which have maximum speeds of 75mph and 82mph respectively, produce higher range estimates than the EPS cycle, even though that has a far lower maximum speed.

New European Driving Cycle

Despite its name, the NEDC was designed in the 1980s and has become out of date as driving conditions and vehicles have changed. As such, it will soon be replaced by the WLTP, which is just a year old.

The NEDC includes a single cycle which is performed on the car in a laboratory, with the wheels turning a rollers used to simulate different gradient and driving conditions. The cycle lasts 20 minutes and during that time the car covers 6.8 miles. There are two phases to the cycle, with 66 percent used to measure performance during slow, stop-start urban driving, while the remaining 34 percent is used to test non-urban driving.

The average speed of the cycle is 21mph and the maximum speed is 75mph, just above the UK motorway limit of 70mph, but below the 130km/h (81mph) limit imposed across much of Europe.

What then lets the cycle down is how the car's air conditioning is not used, and nor are any other systems which would have a noticeable affect on the range of the vehicle.

Finally, the test is conducted with an ambient temperature of between 20 and 30 degrees centigrade - the equivalent to a warm summer's day.

Electric cars lose a lot of range when their batteries are cold, and lose more energy while warming them up (unless still plugged in, of course). A cold electric car also does not properly harvest energy through its regenerative braking system. In short, cold weather has a drastic effect on the range of an electric car, yet this is not addressed in the NEDC range estimates provided by manufacturers.

World Harmonized Light Vehicle Test Procedure

The WLTP is designed to test vehicles more broadly and produce range statistics which are closer to what buyers of electric vehicles can reasonably expect to achieve.

Firstly, the WLTP includes a more dynamic cycle designed to be more representative of how people drive in the real world. Despite this, the tests are still carried out in a laboratory so your mileage will — quite literally — vary. Another issue is that the cycle does not take gradient into consideration; hills cause excess battery use on the way up, then help recharge the battery when coasting or braking down the other side.

The new cycle is 30 minutes long, 50 percent longer than the NEDC, and spans a distance of 14.5 miles. There are four phases to the WLTP test cycle, with 52 percent simulating urban driving and 48 percent designed to recreate non-urban driving. The average speed of the test cycle is 29mph and the maximum speed is 82mph. While these are more representative of real-world driving, the cycle's fastest acceleration event sees a car taken from zero to 30mph in a slow 15 seconds.

Although this is difficult to standardize due to different cars offering different features, the WLTP cycle takes into account things like air conditioning, as well as features offered by the manufacturer to help improve a car's range.

The WLTP cycle is performed at 23 degrees centigrade, which may seem like a more realistic temperature for a globally standardised test, but when you remember that Norway has far-and-away the most EVs per capita - and an average summer temperature of 18 degrees centigrade - the problem remains.

It is difficult to see how the situation can be changed any time soon. The chemistry of electric car batteries means they perform poorly in the cold and struggle to hold charge when cruising at high speed; the importance of range for consumers will remain until it can match that of an internal combustion car, or until there are more chargers than gas stations. Both of these situations are likely many years away for now.

The best we can hope for is that manufacturers will adopt the the WLTP cycle globally, then - while the range stats may not be fully accurate - consumers can accurately compare one car with another safe in the knowledge they are using the same metric.