The more efficient a distribution grid is, the less likely other potential supply bottlenecks (from lithium for batteries to trained electricians) will slow the transition to greater and greater electrification. State-of-the-art modelling of future grids is already happening, but the robust modelling of the distribution system is conspicuously missing because good data are notoriously hard to find, says Meredith Fowlie at UC Berkeley’s Energy Institute at Haas. Fortunately, in California investor-owned utilities are now required to collect and release these data, thus enabling an assessment of the distribution infrastructure’s capacity to absorb electrification-induced load increases. The resulting study can now come to some valuable conclusions. It shows that EVs will drive the pace of distribution system upgrades more than buildings electrification. Also, EV charging at commercial locations rather than at home will be more efficient. The data also allows for a decent estimate of the cost of upgrades to 2030 and 2050. Interestingly, rooftop solar has a smaller effect than expected, and smart-charging during the day (i.e. not just off-peak and overnight) has big benefits. Data access should be made universal and improved even further, says Fowlie, to ensure distribution systems play their full role in keeping generation capacity and costs to a minimum.
The pace of decarbonisation is starting to pick up. And “green bottlenecks” are cropping up. From lithium shortages to balsa wood shortages to electricians in short supply, supply-side curve balls are threatening to slow — or hamstring– the clean energy transition.
While working to transform our grand decarbonisation plans into a reality, we’re bound to stumble over costs and bottlenecks we failed to anticipate. The sooner we can identify and navigate around these complications, the better. In this spirit, a new Energy Institute working paper takes a deep dive into the electricity distribution system.
The local power lines and substations that deliver electricity to our homes and businesses may seem dull and pedestrian. But this infrastructure has a critical (albeit supporting) role to play in the clean energy transition. If electric vehicles (EVs) and building electrification drive electricity demand peaks into uncharted territory, system upgrades will be needed. How many, and at what cost, has been impossible to assess given a dearth of data.
An impressive research team (Salma Elmallah, Anna Brockway, and Duncan Callaway) brings great data and careful analysis to these questions. The key takeaway: if we’re serious about electrification, our distribution infrastructure needs a lot more love and attention.
The distribution system: a billion-dollar blind spot
Lots of important work has already been done to assess the costs of an accelerated clean energy transition. The distribution system is conspicuously missing from most state-of-the-art modelling (see, for example, here and here) because good data are notoriously hard to find. Modelling that does incorporate it uses simplified (and top secret) parameters to approximate how demand changes drive distribution system costs, so it is difficult to gauge credibility.
The paper uses new data from California to assess the capacity of substations and distribution lines to accommodate load increases. California’s investor-owned utilities are now required to collect and release these data (thank you, CPUC).
Elmallah, Brockway, and Callaway (EBC) focus on one large utility: Pacific Gas & Electric (PG&E). They combine circuit-level data on load hosting capacity with neighbourhood-specific estimates of the load increases we might expect to see as more electric cars and electric heating appliances (like those cool heat pumps everyone wants!) plug into the grid.
For each of PG&Es >3,000 feeders, EBC assess the capacity for local distribution infrastructure to absorb electrification-induced load increases. Given how hard it is to predict when/where people will charge their EVs, or how fast the residential electrification situation will actually unfold, they consider a range of scenarios.
There are too many interesting results to unpack in a single blog post. I’m going to focus on three:
1] EVs drive the pace of distribution system upgrades
The picture below summarises projected upgrade requirements across PG&E’s distribution circuits in terms of capacity and the number of circuits impacted. The graphs on the left show how EV adoption drives more upgrades as compared to building electrification on the right. Consistent with California’s EV targets, the scenarios on the left assume that PG&E territory reaches 3.1 million EVs by 2030 and 12.5 million by 2050.
It’s estimated that we will need between 95 and 260 feeder upgrades per year between now and 2030. That’s about triple the pace of projects that PG&E has planned for through 2025.
2] Distribution system upgrades cost real $$$
To map feeder and substation upgrades into dollars and cents, the authors use PG&E reported upgrade costs. The table below summarises these cost data. Let’s take a sentence to celebrate the transparency of these cost numbers. If readers have concerns about them, we can have a conversation about how these estimates could be refined and improved.
Taking these cost numbers as given, the authors estimate that upgrade requirements in PG&E territory will add up to approximately $1B between now and 2030 (closer to $5B by 2050).
3] Commercial EV charging holds promise
These graphs also show that dialling up the share of EV charging that happens at commercial locations does not increase distribution grid costs. You can see this by comparing the upgrade numbers across the “standard” scenario (67% of EV drivers have access to home charging) and the “more commercial” scenario (50% of drivers have access to at-home charging).
This surprised me because it’s assumed that charging at commercial locations will happen during the day, whereas residential charging happens at night. I had thought that daytime charging would be more constrained because it’s more likely to coincide with peak loads. But these data suggest there’s some excess capacity on commercial circuits.
Distribution cost work-arounds?
I’m fortunate enough to be married to one of the authors of this fine paper. So I’ve been able to pick his brain about some questions I had about possible work-arounds. Our kids impose a 5-minute time limit on how much time we get to spend on nerdy electricity talk. Here’s a lightning-round summary:
Rooftop solar to the rescue?..
Could more rooftop solar reduce the need for distribution cost upgrades?
More residential PV without storage has limited impact on system upgrade needs. This is not surprising given that much of the heating load and residential EV charging is assumed to happen at night. Distributed solar plus storage could reduce the need for distribution system upgrades. But could it really make sense to invest in distributed batteries to charge our EV batteries? Back-of-the-envelope calculations say no – probably better to bite the system upgrade bullet so that we can plug our EVs into the grid.
Smart EV charging to the rescue?..
These researchers consider a stylised demand response (DR) scenario that evenly distributes at-home vehicle charging between the evening hours of 10 pm and 5 am. They find that this kind of coordination reduces upgrade requirements and associated costs. Could costs be further reduced with more targeted demand response programs?
The answer is almost certainly yes. Remember that distribution system costs are no higher when EV’s plug into commercial circuits during the day (versus residential circuits at night). A big advantage of daytime charging is that it can be coordinated to soak up solar PV production (and low wholesale prices). Smart coordination of commercial/at-work charging could deliver bigger system-wide cost savings.
How to incentivise this smart charging behaviour? That’s a question for the economists in the room…
Electrification route planning without a map
This research provides some great insights. But there’s clearly more work to do. And more data to share. One issue that I’ve swept under the blog/rug is the uncertainty around these upgrade projections. The authors do the best they can with the data they have. But California utilities could provide better, more precise estimates if they updated their load integration modelling processes to capture and prepare for the most plausible electrification scenarios.
Finally, this is just one paper about one utility service territory. PG&E’s distribution system could look different from other parts of the country. But it’s hard to know given all that we don’t know about distribution system infrastructure and operations elsewhere! If more jurisdictions would make these data available, we’d be in a much better position to plan for the distribution-system-meets-electrification challenges ahead.
Meredith Fowlie is an Associate Professor in the Department of Agricultural and Resource Economics at UC Berkeley. She is also a research associate at UC Berkeley’s Energy Institute at Haas and the National Bureau of Economic Research.
This article is published with permission
Keep up with Energy Institute blogs, research, and events on Twitter @energyathaas