The economics of public EV charging

We have liftoff

The rate of electric vehicle adoption in Australia has finally started to ramp up, and plenty of businesses and brands are jostling to charge all those new EVs.

The exam question we’re attempting to answer is, what’s the breakeven sell price in c/kWh for a site owner investing in public EV charging infrastructure?

And, how does that breakeven price change with different sizes of charger and when those chargers are located in different Australian cities, on different networks and exposed to different energy markets?

For the purpose of this exercise, we’re going to a model a shopping centre where the charging infrastructure is to be installed behind the existing connection point.

What makes a viable EV charger investment?

The main drivers of commercial viability for EV chargers are:

1. Infrastructure costs
2. Energy supply costs
3. Network service costs
4. Physical constraints
5. Charger utilisation
6. Margin/markup

We’ll have three different sets of infrastructure costs, corresponding to three different sizes (power ratings) of EV charging infrastructure.

We’re ignoring margin for now to just focus on the breakeven price.

Here are the assumptions we’re making for these drivers.

1. Infrastructure costs

We’re assuming the shopping centre owner is leasing a single EV charging of each type over 10 years, and that the lease cost is inclusive of supply, installation, operation, and maintenance.

2. Energy supply costs

To keep this model simple, we’re assuming the site owner is taking the wholesale energy price in each respective NEM spot market for Brisbane, Sydney, Melbourne and Adelaide, and the WEM’s balancing market for Perth.

We’re also including the associated environmental certificate costs for these markets, and the capacity market costs for the WEM.

The forward wholesale energy curves we’re using in this model have been provided courtesy of our friends at Cornwall Insight Australia.

3. Network service costs

The chargers are located behind-the-meter in an existing shopping centre that’s connected to the low voltage network, and so exposed to the applicable distribution network tariff.

For Brisbane that’s Energex 8100, for Sydney Ausgrid EA310, for Melbourne CitiPower CLLV1, Adelaide SAPN’s LBAD (outside of Adelaide CBD) and for Perth it’s Western Power’s RT6 metered demand tariff.

4. Network and other physical constraints

EV chargers are high power assets that can stress the capacity of existing electrical infrastructure at a site, in particular the capacity of a site’s connection to the network.

For this exercise we’ve assumed there is no physical constraint to be managed, we’re only adding a single charger to what is already a fairly hefty shopping centre load, but in other situations physical constraints will impact the economics, either by restricting charging use or requiring investment in additional infrastructure, like battery storage.

5. Charger utilisation

An EV charging business case is very sensitive to assumptions about utilisation. There are hundreds of billions of dollars of liquid fuel sales each year that are up for grabs, but how quickly will Australian’s adopt electric vehicles and how much charging will be done at public place chargers?

For this modelling, we’re assuming each charging event supplies 30kWh of energy, and we’re assuming utilisation increasing by around 350% over the 10-year modelling period in response to increasing EV adoption, from a starting point of 5 charging events per day.

It is worth noting that the chargers are only available for use during the operating hours of the shopping centre, so 7am to 8pm Monday to Sunday. That is, the maximum possible utilisation is about 54%.

Modelling approach

For our baseline, or ‘do nothing’ scenario, we’ve simply taken our shopping centre and transported it to each of our target cities where it is exposed to the relevant wholesale energy costs, environmental certificate costs, and network costs, and in the case of Perth the capacity market.

We then created an alternative scenario with three different sized EV chargers in each site in each city, a 50kW, 175kW and 250kW asset, and we applied our EV utilisation schedule to simulate usage of the charges.

Our utilisation model will stochastically schedule the forecast number of charging events each day within the specified charging window. This includes queuing the vehicles if there isn’t enough capacity to keep up with the forecast number of charging events due to each vehicle’s energy requirements, charging times, EV charger capacity, and power constraints at the site.

The animated gif below shows the process of building up the simulation in Gridcognition.

In all cases the chargers are required to serve the same number of EVs with the same energy requirement, so total energy provided is the same. Of course, the rate at which it’s provided varies between the chargers depending on their size.

The chart below shows total supplied energy (red) and charger utilisation (blue) over the modelling term.

The EV chargers will generate additional energy and demand requirements at the site, which you can see in the chart below as the grey trace, that are then re-costed against the wholesale (orange trace), network, certificate and capacity market prices for the site owner.

The difference between these costs and the baseline cost represents the additional cost incurred to provide the charging.

Modelling results

The big picture is illustrated in the chart below, where each row represents a different sized charger and each column represents a different location, with the value stack for each.

For the 50kW size there are broadly similar costs across all our cities, with Brisbane, Melbourne, Adelaide and Perth all having a breakeven cost of about 25c/kWh. Sydney is the most expensive city at 28c, driven by higher wholesale energy costs (coloured orange).

Despite the overall costs being similar across each location, the value stack can be quite different, particularly in relation to the share of costs from energy and network.

However, as the charger size increases so the variations between locations becomes much more pronounced.

Adelaide is the cheapest city for all charger sizes with Sydney and Perth the most expensive.

It’s clear that network costs (coloured red in the chart) are the main driver of cost for higher powered chargers, a reflection of the demand-heavy network service costs imposed by some of the distribution networks, particularly Western Power (Perth) and Ausgrid (Sydney).

If we drill down we can see that there’s significant variability in costs from year to year, a reflection of both the changing wholesale market costs, as well as the rather stochastic nature of public EV charging events interacting with site load and network tariff structures.

Again, Perth and Sydney are particularly volatile in this respect.

It’s also worth noting that charging costs are far more sensitive to charger size (power requirements), in kW, than charger throughput (energy requirements, in kWh)

There is a 3.5 fold increase in charger utilisation over the simulation period, with the total delivered energy increasing from about 55 MWh in the first year to 175 MWh in the final year of the model.

This is reflected in the annual costs per kWh of delivered energy which factors in increasing charger utilisation plotted on the chart below.

It’s one of the reasons that management of peak demand from EV chargers is a very important consideration, and will inevitably become a key driver of battery investment along with co-optimisation of site load with vehicle charging (demand flex), and potentially a shift towards more complex time-of-use or demand-based pricing of EV charging.

Wrap up

One of the great things about the electrification of passenger vehicles is that it opens up a whole world of choice for vehicle owners, and a lot of new commercial opportunities for businesses and brands.

Any business whose customer’s drive, park and spend some time at a particular location is already thinking about how they capture some of this new market from incumbent fuel retailers.

But modelling costs and revenues for EV charging is challenging. There are many factors to consider. As we’ve shown in this modelling exercise, location and charger capacity are both very significant factors.

All of these factors need to be taken into account when new EV charger asset owners are making investment and pricing decisions.‍