Thus far, plenty of ink has been spilled regarding the Tesla Model S and the fallout from the New York Times article, and it even showed up in our latest podcast. I feel like my perspective on this topic, as someone who has worked in the EV space as an engineer for the last four years, as well as driving a family owned Nissan Leaf for the last year and a half, is a bit different from our editors who discussed it, as well as most of the others who have written about it.

I’m not particularly interested in getting involved with the back and forth or analyzing what either party said - I feel like the Times reporter didn’t really understand how EVs work, nor how to drive them, and I don’t really agree with Elon Musk’s Matlab data infused response (it felt too vitriolic to get across any point other than Tesla being angry about the article). And just to go back to the point about not understanding how EVs should be driven, you have to tailor your driving style to suit the powertrain in order to get the maximum out of the EV - if you don’t want to, you’re going to be disappointed. Consider it like needing to switch keyboard shortcuts when you move from Windows to OS X - it’s a slight mental recalibration that has to happen for you to use the platform to its fullest. But that’s another story for another time. 

What I feel like is getting lost here is actual EV performance in cold weather, or hot weather, or really anything in the way of hard numbers. We all know that battery performance is reduced in more extreme environmental conditions, and that all cars, regardless of powertrain type, consume more energy (fuel or battery) in those extreme climates. Unfortunately, quantifying these general ideas is a bit more difficult. That’s where I come in. 

Technology Background

First, a bit of background. I spent the last few months working in Argonne National Laboratory’s Advanced Powertrain Research Facility (APRF). They’re an extremely knowledgeable group of engineers and scientists whose job it is to test advanced technology vehicles in their temperature controlled dynamometer facility, which can sustain temperatures anywhere between -5 F to 100 F. Most of this research is done as a part of the Department of Energy’s Vehicle Technology Program. 

Everything from pure electrics to plug-in hybrids, normal hybrids, clean diesels, direct engine turbos, and advanced transmissions (dual-clutch, CVT, etc.) go through the labs with an extended set of instrumentation to collect various speed and consumption metrics. The resulting data sets are truly comprehensive (2 million data cells for 20 minutes on the dynamometer) and allow the researchers to draw conclusions about the various powertrains and their behavior. It’s a pretty impressive setup. 

The APRF has released some of the data to the public in something called the Downloadable Dynamometer Database. The D3 pages for the Nissan Leaf includes data from the three different temperatures used in EPA’s 5-Cycle fuel economy test: 20F (with a heater load), 72F (climate control off), and 95F + sun lamps (with an air conditioning load). In all cases involving heating or air conditioning, the climate control system is set to 72F. Using this data, we can take a more in depth look at what the actual penalties of environment are on efficiency and performance, and really see the impact of temperature on energy consumption and range. 

Before we start, I should probably explain the terms used in the graphs that follow. UDDS is an urban driving cycle (it stands for Urban Dynamometer Driving Schedule), CS stands for cold start (the first start after the vehicle has been sitting for a number of hours), HS is hot start (any subsequent start after the vehicle has warmed up), HWY is short for HWFET or Highway Fuel Economy Test (also seen as the Highway Fuel Economy Driving Schedule or the EPA’s highway fuel economy cycle), and US06 is a supplemental federal test procedure (SFTP) to provide a more aggressive driving schedule than the relatively tame HWFET cycle. A comparison of the three driving cycles (speed versus time) is shown above.

Thermal Effects on Energy Consumption and Range

Looking at energy consumption, right off the bat, we can see that cold weather has a significant impact on EC - nearly double in urban driving, 40% more in the highway cycle, and 25% on the US06. It’s worth noting here: the Leaf has an 80 kW electric motor, a 24 kWh battery, and a 5 kW resistive heating element.

When broken down further, into a base load case (driving cycle EC at 72F), the amount of extra energy needed to drive the cycle at 20F (listed as EC cold), and then the heater load required to maintain a cabin temperature of 72F, we can see that a significant portion of the additional energy consumption is down to the heater. As with most gasoline cars, driving in sub-freezing temperatures only makes the car about 10% less efficient before heater loads are considered. The mechanically-driven heating elements in conventional vehicles add another 5% or so in terms of mechanical efficiency loss, while obviously the electric heater in the EV is far more costly from an energy consumption standpoint even though it doesn’t increase mechanical losses at all.

(Note: graphs were created by ANL using the raw 10Hz data instead of the filtered 1Hz data available on D3. Available as part of APRF's AVTA Nissan Leaf testing analysis and summary, as presented by Dr. Henning Lohse-Busch. Full presentation available here.)

And while energy consumption is a big deal, range is the be-all, end-all concern for most consumers. With the heater on, that’s a decrease in range by anywhere from 20-50% depending on your drive cycle. That’s a lot. And you can just look at the APRF’s full charge tests at 20F and 72F to see what I mean: 

Think about that - this is a car that, at launch, Nissan claimed had a 100 mile range. My own real-world driving suggests 80-85 a decent estimate. EPA says that number is closer to 73. The APRF’s instrumented testing backs that up (74.1 mile range at 72F), but the same instrumented testing, in 20F weather? 46 miles. That’s it. 

The Impact on the Model S

So, how does this all relate back to the Tesla Model S? The basic technologies involved in the Leaf and the Model S are the same - you have an electric motor, a lithium-ion battery, and an electric heater. The Model S is a bigger car with a more powerful electric motor (310 kW), a larger battery (85 kWh), and a more robust electric heater, but they’re all the same elements. So our findings with the Leaf, particularly in the HWFET side of things, are definitely applicable here (which isn’t to say that the numbers will match exactly, for a number of reasons.) The 85 kWh edition of the Model S is claimed to have a 300 mile range by Tesla and is rated at 265 miles by the EPA based on their 5-cycle fuel economy testing. In cold or near-freezing weather, with the heater running, I would not be that surprised to see range fall to something in the 180 mile region. 

The world of electric vehicles is still very new to the automotive industry, the tech industry, and the mainstream consumer. This breed of cars is completely different than ones that have come before, and there’s a lot that people are learning and still need to learn about EV technology. Incidents like NYT vs. Tesla (and the previous Top Gear vs. Tesla, which was an outright sham) are just steps along that path. I feel like I am in agreement with Anand here in that the way cars are tested, as we move into the EV age, needs to change radically. 

Thanks to Argonne National Laboratory, the Advanced Powertrain Research Facility, and Dr. Henning Lohse-Busch for testing and analysis of the AVTA Nissan Leaf, as well as Kevin Stutenberg for maintaining the Downloadable Dynamometer Database. All graphs and data used in this post are available publically and are courtesy of Argonne National Laboratories.