It seems axiomatic that improved aerodynamics will result in better fuel economy - and that smoothing the underside of a vehicle is one way of gaining that increased slipperiness. However, we found this is simply not always the case...
The car in question was a Honda Insight.
Already boasting excellent aerodynamics (a drag coefficient of just 0.25 and a small frontal area), the hybrid Honda has been successfully modified to improve its already world-beating fuel economy.
We’ve altered the amount of part-throttle Exhaust Gas Recirculation (EGR) that occurs (this reduces pumping losses - see Tweaking the EGR, Part 1 and Tweaking the EGR, Part 1); altered the way in which the ECU monitors throttle position (so allowing the engine management to stay longer in its ultra-lean ‘lean cruise’ mode – see Giving the Insight a Good Driver); and changed the intake to give a slight positive pressure through part of the intake system (see We Have a Record!). In addition we’ve fitted dashboard LEDs that show when the air/fuel ratio is ultra lean or very rich, and another LED that indicates the action of the EGR valve (see Monitoring Factory Oxygen Sensors, Part 2). These LEDs allow the driver to slightly alter driving behaviour, so improving fuel economy.
In the past we’ve also tried aerodynamic vortex generators across the trailing edge of the hatchback, something that measurably worsened fuel economy (see Blowing the Vortex, Part 4).
So what about trialling some more aero changes?
Inspection under the car shows that while the Honda runs more underbody panelling than many cars, the rear half of the underside still looks rather poor. The torsion beam axle is fully exposed, and there are panel gaps between the fuel tank, spare wheel well and axle location.
On the other hand, the front underbody panels are smoother and more continuous – although gaps still exist.
Trial Rear Undertray
Using ‘Corflute’ lightweight sign material, held in place with cable ties and heavy duty duct tape, a rear undertray was formed that totally enclosed the rear axle, leaving exposed only the springs and lower parts of the dampers.
For safety reasons, the undertray did not enclose the exhaust pipe or rear muffler.
In addition, two ramp-shaped fairings were made...
...and then positioned in front of the rear wheels.
The rear underside of the car then looked like this. Installing the fairings and undertray took about three hours of work with scissors, tape and Corflute signs.
The easiest way of assessing whether or not a real world reduction in drag has occurred is to measure open-road fuel economy. If the drag has been lessened by an amount that results in a measurable improvement in fuel economy, clearly the modification has been successful.
Testing was undertaken on a freeway. A 64-kilometre loop was driven, half heading in one direction and the other half in the other direction. The car was driven at 105 km/h for about half the route and at 100 km/h for the other half. The same driving style was used for each test and the traffic was such that very similar runs could be made.
So, how to assess the fuel economy? The Insight’s trip fuel economy readout was used; importantly, this reads in litres/100km to only one decimal place. This means that to try to gain a feel for very small changes, some estimates needed to be made of the missing second decimal place.
The first test was conducted with the rear undertray and fairings in place. This resulted in a displayed fuel economy of 2.9 litres/100km, but importantly, as the car exited the freeway and slowed to a stop, this dropped to 2.8 litres/100km. Therefore, the fuel consumption with the undertray and fairings in place was very close to the 2.8/2.9 changeover point – say around 2.84 litres/100km.
The next test was made with the undertray in place but the fairings removed. This resulted in a measured 2.9 litres/100km but this time the figure stayed unchanging as the car came to a halt. Clearly then, the fuel consumption in this test was a little higher. Driving off in urban conditions (ie in conditions that give poorer fuel economy) and watching how long it took for the display to change to 3.0 litres/100km gave an indication of the second decimal place (ie if the consumption had actually been 2.94, then it would very quickly ratchet up to a displayed 3.0 once the car moved off again.) But the figure was slow to increase – I gave it a guesstimate of about 2.90 litres/100km.
The final test was with the underside of the car standard – both the fairings and the undertray were removed. Again after the 64 kilometres, the display showed 2.9 litres/100km; this time it was even slower to increase once thirstier driving was undertaken. Guesstimate? Say, 2.88 litres/100km.
So, going on both the displayed data (ie to one decimal place) and the data guessed on the basis of how long it took the display to increase once thirstier driving was undertaken, the results look like this:
So over the three tests we’re talking an estimated variation of 0.06 litres/100km, or 2 per cent. Even driving carefully over the same route, I’d expect at least this variation in fuel economy - just by driving differences. Therefore, I think that the impact of the aero mods was negligible.
I also could not pick any differences in noise or stability.
A couple of important points.
Obviously, it would be nice to have fuel consumption actually displayed to two decimal places, rather than having to estimate the final decimal place. However, after watching the display for many thousands of kilometres, I know it’s possible to get a good feel as to whether a displayed 2.9 litres/100km is near to changing to 3.0 litres/100km – or to 2.8. In the testing described here, the terrain and driving conditions also meant the fuel consumption improved as each drive progressed – so another way of getting a feel for that second decimal place was to observe whether the number had only just changed to 2.9 litres/100, or had done so much earlier in the drive. I am confident that the second decimal place guesstimates are in the right ballpark.
To be worthy of pursuing further, I think the aero mods should have delivered a clear improvement, one that would cause say a 5 per cent change in fuel economy. A change of this magnitude would have been directly observable on the fuel consumption display.
Finally, testing conducted over thousands of kilometres may show clearer results. However, if an improvement isn’t going to be shown in the sort of testing that was actually undertaken, I don’t think it’s worthwhile doing much more extensive testing to try to find gains that probably aren’t there!