Why does cold weather sap mpg?

[lakedude]

how much lower is the psi from 30f to 10f?

If my math is correct, about 4%

 

[Topweasel]

True, but hitting the gas doesn't necessarily burn more fuel. For instance, driving up a mountain, the air gets thinner. You press the gas farther as you ascend to burn the same amount of fuel because the air is getting thinner, and you need a fixed amount of air for cumbustion. Gas mileage would improve as you go up (assuming you don't have to downshift) because you'd have less and less parasitic vacuum loss as that throttle plate opens more.

EDIT: Am I going beyond the scope of the discussion?

I thought it worked differently you aren't burning less fuel as you get higher, but actually getting worse combustion and therefore need more fuel with a higher throttle to keep up. Lets say at sea level you were getting X amount of energy potential per gallon of gas at 1 Mile above sea level you would get X-(Some crazy calculation) of potential power per gallon.

 

[Yuriman]

It does sound intuitive, but it isn't so. Most modern engines, once fully warmed up and with proper air temperature get something like 98%+ combustion efficiency. It's all of the parasitic losses after that which cause less of that heat energy released to make it into forward kinetic energy. Diesels have three big efficiency advantages over gasoline engines: higher expansion ratios (more effectively extract energy from thermal expansion), more BTU per gallon (by about 15%, a large part of why diesel is more expensive), and no throttle plate (and thus no vacuum generating parasitic loss at part throttle). Thinner air = less throttling, which in turn means less parasitic loss and more of the fuel burned makes it to the wheels.

This is also part of why downsizing engines generally works - a smaller engine will typically have the throttle more open, operating closer to the way a diesel does.

And, additionally, BMW and Honda (and perhaps others, but I believe Honda was first) both have technologies which allow their cars to run "throttle-less", by using valve timing to control how much air is let into the combustion chambers.

 

[JEDI]

Yesterday was 15f.
Garage to work is 30 miles.
Carputer said 35mpg.

Today is 30f.
Same garage to work route.
Carputer said 38mpg.

15f difference= +10% mpg

 

[Topweasel]

Yuriman, I read up more on what you were saying throttle and and vacuum don't seem to be an impact or not as much as your post implies.

You are right that Fuel injection and Direct Injection systems remove a lot of the difference and that basically the engines computer chooses the mix. Might require more throttle but the performance should be the same. But you will see more milage savings from reduced air pressure reducing drag.

When I thought of altitude driving was more about the climb, where forced air induction sounds more like what you are talking about. I that sense you would be better off with a smaller engine and a turbo.

 

[Yuriman]

Base specific fuel consumption charts are basically what I base this off of:

https://x-engineer.org/automotive-e...ormance/brake-specific-fuel-consumption-bsfc/

These charts show how efficiently an engine can produce an amount of power at various RPM and throttle. In the above image, the blue isolines represent fixed power output across rev and rpm range. They tilt downward to the right because engines are typically less efficient at higher RPM, and they curve because engines are typically less efficient at lower load.

If it takes ~20HP to cruise down the highway at 65mph (a very reasonable figure for a mid-sized sedan), running the engine at 2000rpm gives a BSFC of around 240g/kwh. At the same rpm at wide open throttle, it's about 200g/kwh, which means throttling is causing a 20% loss in efficiency here, for various reasons beyond the scope of this post. Of course, at WOT you're producing more like 50hp so you can't maintain a fixed speed and stay in this peak efficiency island.

To make matters worse, a car with a traditional automatic or manual mated to a 4 cylinder is probably not going to be geared for 2000rpm @ 65mph; 2500-3000 is more common, because it gives you extra passing power without the need to downshift, and prevents the car from feeling "underpowered". If in top gear the engine spins at 3000rpm, to produce 20hp, the engine in the chart above is at an abysmal 320g/kwh. Or in other words, it's burning around 60% more fuel than if gearing would allow it to run at 1400rpm @ 65mph (producing the same power)

CVTs are nice from an efficiency perspective because they have effectively infinite variability of RPM and load. They're generally programmed (unless you hit paddle shifters) to keep the engine at its optimal load vs RPM to produce the power you're asking for. The red line in the following image shows where a Prius engine is allowed to run by its orbital gear system:

So, throttling losses can be very significant, and picking an oversized engine with limited gearing can hurt fuel economy tremendously.

If you're using the engine in the first image paired with a traditional auto or manual and driving at 10,000ft vs at sea level, air pressure is only about 66%, meaning when cruising you'd be running much closer to that island efficiency; lower air pressure is effectively the same, from an efficiency standpoint, as downsizing the engine.

 

[JCH13]

I don't think that throttling itself is causing the drop in efficiency that you observe in those BSFC plots. That is to say I do not think that there is a mechanistic relationship between the position of the throttle plate position/MAP and the BSFC that explains the differences in efficiency observed. Specifically, it takes a baseline amount of energy to simply turn an engine at a given rpm with no power output (friction losses in the water pump, oil pump, valve train, crank, etc) so once the energy is paid to spin the engine any more energy pulled out as power tends to increase efficiency. In that sense there is a correlation between throttle position and efficiency, but not the dramatic causal relationship you're indicating.

A completely throttless engine could realize about 10% efficiency improvement at best. An experiment published in an SAE article (not publicly available) showed around a 15% efficiency improvement in using titanium valves and valve springs, reducing drag on the camshaft(s).

The Prius example you posted is a bit misleading because the Prius uses VVT timing to throttle the engine and approximate an Atkinson cycle (as opposed to an Otto cycle) and it's fun gearing system lets the engine spin more slowly, leading to lower overhead losses. But I do not think the efficiency improvements are realized entirely because of throttling losses, or lack thereof, as you seem convinced of.

TL;DR - I think you've got the right overall impression, small engines operating near maximum output are more efficient than big engines operating near idle, but I don't think you have the right mechanistic explanation. Throttling related losses are just one small piece of the puzzle. Overhead losses associated with spinning the engine are generally more significant from the research I've done.