Growth of Superchargers vs Teslas/EVs in Australia

[gazpachopolice]

Like many Tesla owners, I've been delighted by the huge spurt in new fast superchargers recently. Surely this holiday season will be much better than the last one? But to confirm my gut feel, I crunched the numbers (Google sheet), and I'm afraid it's not looking good at all.

There are 447 SCs in Australia today, vs 292 a year ago, so 53% growth, and in wattage terms we've gone up from 51 mW total to 90 mW total, 76% growth. However, there will be on the order of 92,000 Teslas by EOY vs 42,337 last year, for 117% (!) growth. It's the same story for EVs as whole: we've gone up from 83,000 to around 180,000, 116% growth, and unlike last year, they're now competing for some of those same Superchargers.

This means there are now 208 Teslas and 403 EVs for every supercharger, vs 145 and 284 last year:

And 0.97 kW capacity per Tesla (0.50 kW per EV), vs 1.2 kW/Tesla (0.62 kW per EV) last year:

These figures are based off supercharge.info and random news reports for Tesla/EV growth: if somebody has a solid data set for either of these, I'd love to see it. And there are a bunch of factors I'm not accounting for:

• Growth in non-Tesla fast chargers
• Tesla introducing idle fees
• Tesla opening some SCs up to all EVs

Am I missing anything else major? Or screw up my math somewhere? Feedback welcome.

 

[Jules22]

Seems right, only mitigating factor is that Tesla are not likely to be opening up the really congested chargers up anytime soon, but as you point out even just the growth in Tesla vehicle sales is outpacing the super charger rollout.

I expect this Christmas will be as bad or worse than last year, but the rate of rollouts of chargers does seem to be increasing, so I'm hoping if that trend continues we may start seeing the balance tip back over the next years.

Hopefully with some of the partnerships being formed with fuel companies, etc we may see more companies start contributing more to the infrastructure.

 

[Chuq]

The other thing to consider is that different parts of the network get different volumes of usage. It was mostly Sydney-Melbourne and Sydney-Brisbane that were overloaded. The growth of charging capacity on the Sydney-Melbourne route in particular has been huge this year, especially in the last month. Sydney-Brisbane not so much outside the 12-stall at Port Macquarie, which is better than nothing, but I think that route has the most potential for congestion (especially as reports in that thread suggest only 6 stalls will be active before Christmas).

 

[mukaibot]

I don't disagree with your numbers, nice job @gazpachopolice ! However, I think the approach to your analysis (which seems to be trying to understand if the holiday congestion is going to be better or worse than last year) has a couple of flaws.

While it's true there are more EVs on the road, and the proportion of Tesla Superchargers has not increased, other charging networks have also not been sitting idle (pun intended). Also there is a growing understanding from the EV community, particularly new owners, that they don't need to charge to 100% each and every time they use a fast charger. Sure, there is further education that can be provided, but levers like idle and congestion fees will continue to help.

So it summary, with more fast chargers overall (not just Tesla), especially in holiday destinations or routes to them, I'm not convinced that it's worse than last year, but of course am willing to be proved wrong!

 

[Jules22]

So it summary, with more fast chargers overall (not just Tesla), especially in holiday destinations or routes to them, I'm not convinced that it's worse than last year, but of course am willing to be proved wrong!

Valid point, but I haven't seen many non-Tesla installations at any decent scale (by that I mean more than 2-4 units at one location), we need more of the 8,12 and bigger sites especially on the major corridors.

 

[B3AU]

Valid point, but I haven't seen many non-Tesla installations at any decent scale (by that I mean more than 2-4 units at one location), we need more of the 8,12 and bigger sites especially on the major corridors.

Good point - AmpCharge at 6 bays (6x CCS2 and 3x CHAdeMO plugs) in either direction (12x if you include both sites) at Pheasants Nest, NSW is the biggest site I can think of from someone other than Tesla...

 

[Midnight Man]

A sideline observation, but unless I missed something, today at around 11.00am (first day of School Holidays in NSW), every single supercharger in Sydney and surrounds, except Campbelltown, had a queue. Whilst most were only 5 minutes (not unreasonable), I believe Broadway was 25 minutes - a bit rough). Central Coast was also full, and Heatherbrae had 1 stall available.

I know that was only a single point in time, but it did seem to indicate there's insufficient charging around Sydney and the northern side to handle peak demand without queues forming.

 

[moa999]

Sydney on weekends has been busy for most of this year (and the recent opening of Forest Lodge barely moves the needle), but we know at least there are plans and funding for a lot more sites.

 

[cafz]

Sydney-Brisbane not so much outside the 12-stall at Port Macquarie, which is better than nothing, but I think that route has the most potential for congestion (especially as reports in that thread suggest only 6 stalls will be active before Christmas).

This is true. Port Macquarie was always the epicentre of queueing on that route though (like Gundagai on the Sydney-Melbourne route) when the northbound and southbound traffic peaks meet in the middle at lunchtime.

 

[Vostok]

Interesting analysis @gazpachopolice (in fact it is exactly the sort of thing I would do ) so a few comments:

• As you noted, other DCFC networks have been growing as well, Evie in particular has doubled their sites this year - nearly 200 now I think. That will help a lot.
• DCFC throughput increases greatly the more of them they are, e.g. a site with 8 stalls can pump through 3.6 times as many cars in an hour that a 4 site stall can, at a 1% probability of cars needing to wait (this is called Erlang efficiency). A 16 stall site can push through 10.4 times as many cars as a 4-stall site. The effect of having a larger number of sites close to each other is similar to having a larger number of stalls at a single site (but not exactly the same).
• NRMA are now charging for charging™ - getting rid of a lot of the freeloaders and time-wasters which will improve throughput.

 

[Chuq]

this is called Erlang efficiency

I was trying to think of this name recently! There was a thread on here which explained it but I couldn't find it!

Capacity of Superchargers Using an Erlang-B Model

A capacity model developed over a hundred years ago for capacity of telephone trunks by Agner Krarup Erlang called the Erlang-B model should apply to Tesla Superchargers. See Erlang (unit) - Wikipedia, the free encyclopedia for background. The biggest error of the Erlang-B model is that it...

teslamotorsclub.com

 

[Quickst]

today at around 11.00am (first day of School Holidays in NSW)

When I drove into Pheasant Nest this morning on way to Exeter, there was only 1 MY charging. I didn't charge there

Exeter at 11:45am today 16Dec
I was the only one charging for my entire stay of 30min. The heatherbrae pie shop was very busy

Yass 16:30
Again I was the only one there

The only wait for a DCFC was the NRMA at Young NSW. The bloke had not set up the NRMA app etc etc

 

[Quickst]

(this is called Erlang efficiency

Which I think uses minimum 2 bays as a starting point??. Single charge bays are just extremely inefficient

 

[Hairyman]

A sideline observation, but unless I missed something, today at around 11.00am (first day of School Holidays in NSW), every single supercharger in Sydney and surrounds, except Campbelltown, had a queue. Whilst most were only 5 minutes (not unreasonable), I believe Broadway was 25 minutes - a bit rough). Central Coast was also full, and Heatherbrae had 1 stall available.

I know that was only a single point in time, but it did seem to indicate there's insufficient charging around Sydney and the northern side to handle peak demand without queues forming.

I‘m wondering if it will ever be the case that demand charges, land prices and cost of infrastructure will fall to the point of capacity meeting the extremes of demand.

Driving back from Sydney last weekend I saw very long queues to get into the big petrol stations on the M1 just north of Gosford.

That is a pretty mature market with the calculus well sorted out. I was charging at the Macquarie supercharger in Ryde that day and waited for 15 minutes - a first wait for me.

The market will eventually sort the matter out but I‘d be surprised if I didn’t end up waiting on the biggest driving days.

 

[Chuq]

The market will eventually sort the matter out but I‘d be surprised if I didn’t end up waiting on the biggest driving days.

There will always be waiting, but the impact of waiting will be less noticeable.

If a 2-stall site is full, and you turn up, it statistically likely that you'll need to wait 10-15 minutes.

If a 40-stall site is full, and you turn up (first one in the queue), it's statistically likely that a spot will open up in less than a minute. Or perhaps more realistically, that a car will be leaving a stall at the very time you arrive. Even a five car queue will be cleared in less than 5 minutes.

 

[Hairyman]

Completely agree

 

[SageBrush]

DCFC throughput increases greatly the more of them they are, e.g. a site with 8 stalls can pump through 3.6 times as many cars in an hour that a 4 site stall can, at a 1% probability of cars needing to wait (this is called Erlang efficiency). A 16 stall site can push through 10.4 times as many cars as a 4-stall site. The effect of having a larger number of sites close to each other is similar to having a larger number of stalls at a single site (but not exactly the same).

Interesting. Other throughput effects not considered in the OP analysis are Supercharger generational effects from improved power, and pre-conditioning. Tesla estimates these two effects as equal to 2x the throughput for the same # of cables. There are also very substantial non-linear effects from lessened intervals between Superchargers that allow owners to drive away with a lower SoC.

I would also point out that 1 kW per car is ridiculous riches overall, so I think the pertinent questions here are not trending averages but how well Tesla handles congestion points.

 

[SageBrush]

One other point: if holiday and festival driving is somewhere between 10x - 100x the usual location usage, the cost to fully mitigate waiting times everywhere will be exorbitant if all Superchargers are permanently installed. Tesla has experimented with relocatable Superchargers, and that strikes me as the smart strategy. Beyond that, drivers will have to learn to not bunch up at the same time

 

[Zinzilla]

Exeter at 11:45am today 16Dec
I was the only one charging for my entire stay of 30min. The heatherbrae pie shop was very busy

We were there yesterday (16 Dec) a bit after 2pm. The pie shop was about half full (we were there for the pies), and only 1 car charging. Goulburn a little bit later was about half full when we were charging there (stopped off to stretch our legs in the park).

 

[Vostok]

Which I think uses minimum 2 bays as a starting point??. Single charge bays are just extremely inefficient

You can use anything as a starting or reference point, including 1 stall. The Erlang formula does not have a singularity at 1.

But yes, a 1-stall site has very limited throughput. A 2-stall site has 5.5x the capacity of a single stall site at 10% chance of blocking, and 7x the capacity at 5% chance of blocking.