The Impact of Temperature on Powertrain Efficiency

"Torquelife" (torquelife) 03/16/2015 at 11:10 • Filed to: None

7

2

!!!error: Indecipherable SUB-paragraph formatting!!! is a new editorial venture focusing on the technical side of the automotive world, written by current and former automotive engineers. Find out more and follow our latest content at !!!error: Indecipherable SUB-paragraph formatting!!! or on Twitter !!!error: Indecipherable SUB-paragraph formatting!!! .

!!! UNKNOWN CONTENT TYPE !!!

!!! UNKNOWN CONTENT TYPE !!!

Let's talk about vehicle powertrain efficiency. Over the past handful of years, as CAFE standards and federal emission regulations have become increasingly ambitious targets, the powertrain design of mainstream vehicles has been defined by a focus on energy efficiency, sometimes even at the cost of driveability.

This philosophy has profoundly impacted the automotive landscape at the macro level, with the shifting of entire lineups to CVTs and downsized, direct-injected turbocharged engines, but has also filtered down to minutia, like automatic shift points, tire selection, and auxiliary component design.

As we move into more complex powertrain architectures, this becomes very important. A few years ago, in a past life, I wrote !!!error: Indecipherable SUB-paragraph formatting!!! examining the effect of temperature on electric vehicle range. It used data sourced from Argonne National Laboratory's Downloadable Dynamometer Database ( !!!error: Indecipherable SUB-paragraph formatting!!! ), and mostly focused on the cold-weather range anxiety questions initially faced by the Tesla Model S.

Using that type of energy consumption data, it's possible to learn a great deal about the way different types of cars operate. The Advanced Powertrain Research Facility group in Argonne has a temperature controlled dynamometer, from which they can collect a truly impressive amount of data about the way a vehicle's powertrain behaves under various scenarios and loading conditions. The D3 pages contain a filtered subset of this data curated for public release, but the sheer magnitude of raw data generated by the APRF is staggering. (Full disclosure: I did a portion of my graduate research at Argonne in 2012.)

The drive cycle testing is based on the EPA 5-cycle fuel economy tests, with three ambient temperature modes: 20F, 72F, and 95F. The 20F and 95F testing also involve additional auxiliary loads, with the heater being used at 20F and air-conditioning plus sun lamps for the 95F testing.

Driving cycles (speed versus time). Image: AnandTech/Vivek Gowri ( !!!error: Indecipherable SUB-paragraph formatting!!! ).

!!! UNKNOWN CONTENT TYPE !!!

There are three different drive cycles used in formulating the fuel economy numbers: UDDS, HWFET, and US06. UDDS is the common urban driving cycle - it stands for Urban Dynamometer Driving Schedule. HWFET is short for the Highway Fuel Economy Test, but is also seen as the Highway Fuel Economy Driving Schedule, and used to be the standard EPA highway fuel economy cycle. I've listed it on my graphs as HWY. The US06 cycle is a supplemental federal test procedure (SFTP) to provide a more aggressive driving schedule than HWFET - it involves more periods of heavy acceleration and higher overall speeds, topping out at 80 mph instead of 60. A comparison of the three driving cycles (speed versus time) is shown above.

Let's take some cars from the D3 list and run a simple analysis to see if we can gain some insight on how their powertrain architectures impact their behaviour in varying conditions. I picked five vehicles, representing different powertrain types: the Nissan Leaf (BEV), Chevy Volt (PHEV), Ford C-Max Hybrid (HEV), VW Jetta TDI 2.0 (diesel), and the Ford Focus (conventional gasoline). In the table below, you may notice that the Volt has two sets of results. It has two operating modes - charge depleting, in which the powertrain tries to stay in EV-only mode for as long as possible, and charge sustaining, which tries to maintain battery level. Given the difference in powertrain behaviour, it makes sense to separate the results.

There are three distinct energy sources involved in that list: gasoline, diesel, and electricity. So converting them to a common unit of energy is a necessary first step before we can begin comparing the vehicles.

The fuel consumption of conventional vehicles is pretty straightforward - consumption is the inverse of economy, so instead of miles per gallon, we flip it and have gallons per mile. This may seem a little odd, as consumption is a rarely quoted metric in the Americas, but in Europe it's pretty common to see the liters/100km unit. And then you can convert the fuel consumption to energy consumption using the energy content of fuel. This is essentially the same process that the EPA uses to determine MPGe (miles per gallon equivalent) for plug-in vehicles, but in reverse.

For electric vehicles, it's even easier: D3's results table includes net energy consumption in kWh per mile, and even the EPA sticker includes a kWh/100mi rating. But for hybrids, where the combined efforts of the internal combustion engine and the electric machines is reported in fuel consumption, this is a more difficult equation. Thankfully, the APRF testing methodology measures the separate power sources individually, so once the numbers are converted to a common unit, you can just add them to determine total energy consumption.

Generalized formula for calculating MPGe. Image: Wikipedia ( !!!error: Indecipherable SUB-paragraph formatting!!! ).

!!! UNKNOWN CONTENT TYPE !!!

The energy content of fuel is, of course, different for gasoline and diesel, and can also vary depending on the different blends of fuel. For MPGe, the EPA decided on 33.7 kWh per gallon of gasoline. California's Air Resources Board (CARB) set it as 32.6 kWh/gal. According to the GREET (Greenhouse gases, Regulated Emissions, and Energy used in Transportation) model developed by the Department of Energy, the number is 34.02 kWh/gal.

I want to take a quick second to put this in perspective - one single gallon of fuel contains nearly 50% more energy than the entire 24 kWh lithium-ion battery pack in the Nissan Leaf. That's actually a bit staggering to think about. Fossil fuels are incredibly energy dense.

The energy density of fuel is typically measured in BTU (British Thermal Unit, a common Imperial unit of energy) per gallon, with a simple equivalency ratio for the conversion to kWh (3412 BTU/kWh). The BTU/gal figure for gasoline varies, but it's generally accepted to be in the 115,000 range, give or take a percent. Diesel has an energy density of around 130,000 BTU/gal, with biodiesel coming in about 8% lower. For simplicity, in my calculations I will use the values from the GREET model: 34.02 kWh per gallon of gasoline, 37.95 kWh per gallon of diesel.

So through that methodology, we can take our fuel economy table from above and convert it into an energy consumption table.

There's actually a lot of interesting data here. For now I'm just going to look at the UDDS cycle; later on, I will go more in-depth on the other drive cycles and things like the effect of speed on energy consumption for the various powertrain types. But just with this one cycle, there's a number of observations we can make, particularly if we graph the data from the above table.

We can see that in Charge Depleting mode, the Volt behaves very much like the pure electric Leaf, but in Charge Sustaining mode, the powertrain looks much more similar to a standard conventional vehicle. Undoubtedly, the range extender (the fancy name for the 1.4L IC engine) efficiency suffers due to the pair of motor-generators the power is routed through.

We can also see the impacts of auxiliary loads - with the AC on, the vehicles with high voltage electric compressors (Leaf, Volt, C-Max) have less of a penalty than the vehicles with traditional belt driven compressors. But the resistive heating elements used in the electric vehicles has a dramatic impact on the energy consumption (this was the primary thesis of the AnandTech article I linked to earlier). It's much less so for the conventional vehicles, which can use heat generated by the IC engine to heat the cabin.

The Jetta TDI has better fuel economy than the Focus, but has less of an advantage in energy consumption due to the differing fuel energy densities. The energy consumption penalty at 95F for the Jetta TDI is higher than the Focus, but appears similar to the previous generation A5 Jetta with the archaic (and anemic) 1.9L TDI motor. We can explore this more in-depth in another article, but this seems to suggest that either the diesel ICE suffers a bit at higher temperatures or that the AC compressor used in the Jetta is less efficient than the one in the Focus.

The Volt consumes significantly less energy in charge depleting mode. The current (outgoing) generation Volt already has much more electric-only range than most existing PHEVs, and the 2016 model is going even further with 50 miles of EV operation. Given the large efficiency advantage to running in charge depleting mode, it's clear why pushing that number higher is important even when there's a range-extending ICE present.

There's a lot more analysis that can be done with this data, some of which I've hinted at in this article. This includes speed versus consumption, evaluating various types of hybrid architectures, and comparing different types of powertrains as we've done here.

!!! UNKNOWN CONTENT TYPE !!!

This data is from the Downloadable Dynamometer Database ( !!!error: Indecipherable SUB-paragraph formatting!!! ) and was generated at the Advanced Powertrain Research Facility (APRF) at Argonne National Laboratory under the funding and guidance of the U.S. Department of Energy (DOE).

BiTurbo228 - Dr Frankenstein of Spitfires > Torquelife 03/16/2015 at 12:23

0

Very interesting. I wonder how much is to do with the actual temperature affecting the ICE cars. I always thought the lower charge density of hot air was better for economy, although where I got that from I have no idea so I may be way off the mark.

I wonder if the energy saving from a belt-driven compressor to an electric one for AC would be worth changing over on a standard car, or is it only so efficient because the charge for that is already 'there' in electric vehicles...

Torquelife > BiTurbo228 - Dr Frankenstein of Spitfires 03/16/2015 at 12:57

1

So temperature is less the effect we're seeing on the conventional vehicles than the parasitic loads, though this is also impacted by engine temperature (cold start versus hot start behaviour is a whole separate article).

The reason it isn't possible to use an electric air conditioner on a conventional vehicle is because the electric compressors are run at a very high voltage, which is (relatively) easy to integrate with an EV/PHEV/HEV that already has a high voltage, high current circuit built in, but not so easy to do when your main auxiliary circuit runs at 12V (standard for conventional vehicles).