Every so often one encounters a myth about electric vehicles that's so odd it makes you scratch your head. When it happens just once, it's easy to dismiss, but when it keeps popping up, it becomes time to address it.
In recent weeks, I've seen multiple people assert that, “Sure, you can take EVs on road trips on major highways now, but I like to take the backroads, and there are few to no chargers on backroads. EVs can’t do backroads trips!” Now, this obviously confused me – not simply because of the fact that it's not true that chargers exist only on major highways, but also for the obvious reason that even if you choose not to drive on a highway, surely you're going to be frequently crossing highways – quite often, near a charger. Yet simply pointing these facts out is generally met with heavy skepticism.
In order to further expound on this, I decided to conduct an experiment. I wrote a Python script to create 50 random road trips (images at the bottom of this article) which were then simulated by A Better Route Planner, an excellent tool that takes into account a wide variety of factors to create a simulation of a drive in any of a variety of EVs, as well as to help plan routes and stops. Most critically, ABRP has an “Avoid highways” option, making it a perfect fit for this task.
Some notes about the various test factors follow (+ = beneficial to the EVs, – = adverse to the EVs, * neutral to the EVs):
(+) The vehicle chosen was a Model 3 LR RWD Aero, an excellent road-tripping EV.
(-) Upgrades to 145kW, in progress, were not considered for the (currently 120kW) V2 Superchargers, and the V3 rollout is only scheduled to begin en masse later this summer/fall.
(-) ABRP uses relatively conservative assumptions about energy consumption
(-) ABRP is extremely aggressive about keeping you off of highways when you tell it to avoid them; it'll prefer to let you almost run out of electricity or slow way, way down than spend even a moment driving on a highway.
(*) The routes were picked randomly within a simple bounding box across the continental United States. Some routes had to be discarded due to being impossible (such as being out to sea).
(-) While people frequently visit or pass through remote areas, road trips are most often done in areas where more people live (as can be seen in road traffic figures), which also tend to have a higher density of chargers. The software is however quite likely to pick trips out in the middle of bloody nowhere.
(*) Vehicles were chosen with a variety of ages (degradation modeled on that of the Model S, although it should be if anything lower for the 3), payload weights, road conditions, and winds (both headwinds and tailwinds).
(*) ABRP was instructed to not let the charge level drop below 10%.
(*) Stock ABRP values were used for everything not specified above, including 3 minutes of overhead (similar to gas station overhead) for every stop. This time is subtracted from the length of a stay, since it's not time you actually spend at a given location.
(-) While in the real world, people road trip most during times of year when there's good weather, and often delay trips when there's bad weather, no attempt was made to emulate this; the vehicles are simulated in sometimes extremely hostile conditions.
(-) Tires become more efficient as they wear down. No attempt was made to account for this.
(+) For third-party chargers, the CHAdeMO adapter may not yield as high speeds as native CCS support; a CCS adapter is coming but not yet available (European Model 3s support CCS natively). This may yield higher speeds at US third-party stations than would actually be achieved. Regardless, usage of third-party stations was relatively rare, and omitting or slowing them did not make a significant difference to the results.
(-) Some trips were long enough that they would almost certainly be done with an overnight stay. This was not modeled. An overnight stay with charging equates to the removal of a long charging stop and part of a subsequent stop (as one charges to full overnight, unlike during a midday stop)
A spreadsheet of the data can be found here. Some graphs summing up the results can be seen below. First, the distribution of trip lengths and trip times is as follows:
As can be seen, most are “all-day single-day trips”, along with some partial day and two-day trips.
As for the cost per mile of these trips, not including charging before you leave or after you arrive (a 30mpg gasoline car at $3.00/gal is listed in red):
The “exceptionally expensive” cases, while sometimes in part involving driving in particularly bad conditions (such as nearly 50mph headwinds or far-below-freezing temperatures), are most commonly due to the use of third party charging stations – particularly Volkswagen's “Electrify America” network, which costs much more than Tesla's Supercharger network.
The number of stops on the trip and the average time between stops can be seen below:
Notice that the average time between stops is surprisingly long – similar to, if not longer than, most people generally take between stops during trips in gasoline cars (one can of course take more frequent stops if they choose). Stopping is thus not an imposition, but a requirement of your need to refresh your bladder, your legs, your stomach and your mind. Part of the reason for the long time between stops is that backroads generally have lower average speeds than major highways; this not only means that it takes longer to travel a given distance, but additionally, EVs become more efficient at lower speeds.
2% – 10 miles – 0 mins (126 kW)
5% – 16 miles – 1 min (250 kW)
20% – 62 miles – 4 mins (250 kW)
21% – 65 miles – 4.5 mins (Taper from peak starts – 248 kW)
30% – 92 miles – 6 mins (218 kW)
40% – 123 miles – 8.5 mins (179 kW)
50% – 153 miles – 11 mins (142 kW)
60% – 184 miles – 14.5 mins (108 kW)
70% – 213 miles – 19 mins (87 kW)
80% – 245 miles – 24.5 mins (56 kW)
90% – 275 miles – 34 mins (36 kW)
This would be expected to cut typical charge times in the simulated trips to 50-70% of what’s listed.
Let's examine our results. First, let us look at the two worst cases. Because unfortunately for us, the program decided to do two routes from eastern North Dakota to eastern Montana – which just happens to be the only place in the continental US with a large hole in the Supercharger network. As a result, the software is forced to make awkward detours to the south, increasing trip times. Additionally, ABRP's strict adherence to “no highway” policies becomes far worse here, causing it to require significant slowdowns and/or lower arrival charge levels (red segments). Waypoint hints even had to be given to ABRP to help it find said routes.
These two cases – applicable only to the “North Dakota Hole” – should indeed be considered failures for the task at hand, and if anybody plans to frequent the backroads of North Dakota and Montana in the present day, this becomes a legitimate argument against doing so in a BEV. That said, charging network growth rates are tremendous – the number of Superchargers has roughly doubled in the past two years, while the Electrify America network didn't even exist two years ago – and this hole will likely be plugged soon enough.
There are a couple cases elsewhere where ABRP's strict adherence to a “no highways, even for a moment” policy led to “red line” slowdown situations:
Additionally, there were two cases where ABRP reported a nominal “slowdown”, yet the speed they reported was basically the same speed that Google reports as normal for said route. Consequently, these – while “redlined” – should probably not be considered slowdowns.
In all of these cases, allowing highway travel for just part of the trip (sometimes just a dozen miles or so) makes the red-line go away.
While one may expect that trying to avoid all highways would lead to frequent detours and backtracking to get to charging stations, this turned out not to be the case – indeed, in only three situations did this happen, 6-10 miles each time. One was one of the North Dakota trips, while the other two were:
This honestly should not be unfamiliar to people driving gasoline cars in remote areas.
In every single other case – regardless of how awful the driving conditions were or where the trip was – there was neither red-lining slowdowns nor any backtracking detours at all. All remaining valid cases follow (read: prepare to scroll down! Summary follows)
What can we take away from this? In summary:
Excepting the “North Dakota Hole“, which will surely be closed very early in the lifespan of any EV purchased today, a “good road tripping EV” can easily manage backwoods trips between arbitrary points in the continental United States – regardless of the road and weather conditions, payload, and even accounting for battery degradation in older vehicles. In rare cases one might need to take a leg or two on a highway, but these are edge cases, and far from the general rule.
It's almost impossible to drive for long distances without crossing a major highway unless you're driving in circles.
EVs frequently travel further on backroads than they do on highways, with less frequent stops required.
Electrify America really needs to lower their prices. The stark difference between Supercharger and Electrify America pricing really stood out in this comparison.
Extreme-cold temperatures, and particularly gale-force headwinds, will ruin your energy consumption – but even that had little impact on the overall ability of EVs to make comfortable road trips on backwoods roads with today's infrastructure – let alone tomorrow's.