Are there any estimates available on how much fuel you can expect to lose from a fuel tank that is leaking vapours? Leaking vapours would be caused by cracked/missing evaporative emissions hoses, a leaky/missing fuel cap, et cetera.

I realise that the exact amount would depend on a number of variables, like temperature and the amount of sloshing in the tank. That's okay, I'm after an 'order of magnitude' estimate or an upper limit.

Any numbers would be quite helpful, whether they be from reliable documentation, anecdotes, forum posts etc.

The reason for the question is that I have been investigating some very poor mileage on a car with a 60% increase over specification on the highway, but very good power otherwise. I'm waiting for a new EVAP check valve, as the old one was very leaky. I'm wondering how much of an increase in consumption this could have made in a worst case scenario.

  • I didn't exactly measure it, but I remember a few years ago a buddy of mine got horrible mileage and it turned out his fuel cap was stolen. I think he saw a 30% reduction in mileage. I have no stats though, but it was a great difference. Nov 11 '13 at 17:05

There are 4 major types of Evaporative Emissions from the fuel system of a car, these are:

  1. Diurnal: Evaporation caused by the fuel tank being heated by the sun
  2. Running Losses: Heat from the engine and exhaust system causes fuel vaporisation
  3. Heat Soak: Once an engine is stopped heat from the engine and exhaust system causes fuel vaporisation
  4. Refuelling: there are always fuel vapours in the fuel tank, when you refuel these vapours are displaced by liquid fuel (and generally vented into the atmosphere).

The amount of Evaporative Emissions created by a fuel depends on the given fuels Reid Vapour Pressure (RVP). RVP is a measure of a fuels volatility, or put simply, how easily a fuel vaporises. In the summer you want fuel with a lower RVP to avoid losses and more specifically to avoid Vapour Lock, in the winter you want fuel with a higher RVP so that vaporisation occurs more easily when trying to start your vehicle, you want the fuel to vaporise as it’s entering the combustion chamber/cylinder so that it will burn. Interestingly the EPA in the USA regulates the RVP of fuels sold depending on the time of year. This also occurs in other countries but I haven’t found any other references to link to.

You are asking about “leaking vapours” which means the first 3 forms of Evaporative Emissions are relevant, but we can probably discount the 4th as, aside from some methods of minimising refuelling vapour emission, these vapours are still deliberately vented to atmosphere in most/all cars so it can’t really be defined as “leaking”.
If you are interested in Refuelling Emissions Kristy Welsh did some calculations in 2008 to estimate how much Refuelling Loss would occur over a year for an “average American”, her initial answer is “about an 1/8th of a gallon”, she then goes on to multiply by a factor of 10 (just in case you think it’s way too low) and the answer is a little over a gallon or about 3.8 litres per year.

None of this answers the question of an estimate for loss from a fuel system leaking vapours and I’m afraid I can’t find anything that does answer the question. There is an EPA document here which gives a lot of the information required to answer some parts of the question, but aside from being a very a large document, it only deals with stationary Liquid Storage Tanks so does not take into account extra evaporation caused by slosh, nor does it deal with Running Losses or Heat Soak.

I don’t think Diurnal Evaporation could cause anywhere near the same losses as Refuelling. These both operate on a similar principal, the sun must heat and expand the vapour enough for it to escape from the tank through emissions control hoses. If you assume Kristy Welsh is correct with her refuel losses estimate then losses from Diurnal Evaporation should be small enough to discount, particularly when talking about measuring fuel economy.

The other two forms of loss, Running Losses and Heat Soak are much more difficult to estimate. They usually occur in what is a closed system, unless there is a leak. I’m afraid I can provide no evidence but experience that these don’t cause enough loss to be measurable when dealing with fuel economy.

If a damaged/malfunctioning EVAP valve is causing a measurable difference in fuel economy there are 2 more likely reasons for this than Evaporative Emissions

  1. (the less likely option) The vehicles ECU has detected a fault and has put itself into a safer mode that over-fuels the car somewhat.
  2. (the more likely option) The leaky valve is causing a leak in your air intake and your engine is sucking in air that has not been measured by the ECU (MAP/MAF/some other air flow measurer). The ECU puts in enough fuel for the amount of air it thinks is in the system, this isn’t enough fuel and it makes the engine run lean. Running lean makes the engine detonate (also called knock/ping/pink) This triggers a knock sensor which in turn puts the ECU into an mode where it over-fuels the engine deliberately in order to stop the knocking.

The outcome is the same, the ECU is deliberately over-fuelling the engine so as to avoid any damage being done.
Of course, this is just my opinion that’s based on a couple of facts supplied. When your mechanic replaces the EVAP valve I’m sure he (or she) will run a full diagnostics test on your vehicle and a simple fault clear will probably sort the problem out.

(I'm not an expert on any of this so please feel free to edit this answer and improve it)

  • Thanks for the detailed and well researched response. I was trying to steer the question away from helping me trouble shoot my fuel specific problem. This is why I left out the specifics of the vehicle. I just wanted to try and see if a number could be found for worst case evaporative loss on any vehicle. I probably should have left out the last paragraph of the question...
    – kahbou
    Mar 10 '12 at 8:35
  • Just as an aside, MAP based fuel control is largely immune to vacuum leaks -- the ECU can't tell if the air has come in through the throttle or through a gaping hole in the manifold.
    – kahbou
    Mar 10 '12 at 8:43
  • @kahbou No worries. The StackExchange engine excels with these sorts of slightly generic questions when people are willing to put a bit of time into both the question and the answer. And you are 100% correct about the MAP sensor I should not have mentioned it. I also left out option 3 which is the forced induction option where air is leaking in the other direction confusing the ECU in a slightly different way but producing the same outcome.
    – Scott
    Mar 15 '12 at 2:47

If you have an air leak on the EVAP system then the MAP sensor will not be able to determine the air density, and the MAF sensor will not be able to determine the volume of air, entering the engine correctly. This will lead to an over fuelling situation AND an under fuelling situation depending on the size of the air leak and engine operating speeds and load. The end result will be destruction of the catalytic convertor. This is not too mention any fire hazard that may present itself.

  • The conclusion about the catalytic converter is complete bunk. You're completely ignoring the fact that the engine will run in closed loop almost all of the time, and when it's not running closed loop it's running intentionally rich. Any vacuum leak on a MAF based system will result in the engine running leaner than expected in open-loop. Whether it's enough to damage the engine depends on the size of the leak, but it's unlikely to be large at WOT--the air would much rather come in through the gaping throttle valve.
    – kahbou
    Nov 24 '13 at 15:47
  • Also: I assume that when you said '... will not be able to determine air density...' you meant that MAP based systems won't be able to account for the added fuel content of the charge. Old MAP based cars had the EVAP system ported into a just-off-idle spot in the throttle body, so you sucked unmetered fuel vapours whenever you'd tip-in from idle. The enrichment wasn't a problem under the emissions regulations of the time. Newer systems added solenoids and ways of modeling the enrichment.
    – kahbou
    Nov 24 '13 at 15:51
  • If you have a weak running condition exhaust gases will be up on temperature. If you have a rich running mixture you will have fuel burning in the cat. Both conditions will eventually destroy the cat. Any air/vacuum leak past the MAF will cause the ECU to enrichen the fuel. A MAP sensor fault will not report air density correctly, ie oxygen content correctly, and will affect the fuel mixture depending how it has failed. Vehicles enter closed loopmas soon as the oxygen sensor is heated and stays in closed loop uptil around 3500 RPM. This means most of the time. Nov 25 '13 at 9:45

In this particular case replacing the EVAP check-valve didn't make a measurable difference to fuel economy. It seems that the amount of fuel vapour involved is very small and only really of interest from an emissions point of view. Unfortunately, I never to got the crude estimate that I was originally after.

The real reason for the horrible fuel consumption was the engine getting 'stuck' in open-loop mode. The ECU would enter open-loop mode (going rich to protect the engine) at a fairly high throttle setting but would not exit open loop mode until it reached a much lower throttle setting. It turned out that most of my highway driving would occur latched in open-loop mode between these two points. I would end up in open-loop mode when climbing a hill, or passing, and then stay there until I had to slow down. The vehicle had larger off-road tyres which increased the gear ratio somewhat and also increased drag, as well as a rather small engine.

After installing a rich/lean indicator for the O2 sensor, I was able to achieve the advertised fuel consumption despite having the off-road tyres, just by making staying out of open loop mode.

The vehicle was an Australian Daihatsu Feroza. Those came with a MAP-based EFI system.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.