Electricity in a decentralization-ascendant world
A very broad theme I have been mulling over for the last 5–6 years is the process of decentralization and atomization of almost everything in the world — societies fragmenting by way of social media, centralized, dependable institutions like the US federal government or the European Union failing in moments of critical need, and (accelerated by the Covid-19 pandemic) the decentralization of work away from the office hubs in major cities to a remote work. In general, I hold the opinion that we are in a decentralization-ascendant world, wherein key institutions fail to keep up with the relentless pace of change and become irrelevant as distributed solutions to challenges naturally emerge. Great examples of such solutions are blockchain, smart contracts, open source, and remote-first companies.
Decentralization will touch the physical world as well, including everything from how we get our food (vertical farming), buy our clothes (small retail ecommerce instead of national department stores) to how we power our lives. This brings me to what I hope to cover in this piece: the proliferation of Distributed Energy Resources (DERs) and the emergence of a decentralized electric grid. If you haven’t been following along, last time I wrote all about the basics of the electric grid where I explained how electricity moves from generation to powering our homes and businesses, dove into electricity markets and how they work, and shared some areas I was interested in digging deeper on, including DERs. So, let’s get into it.
What is a DER?
A DER, or Distributed Energy Resource, is any asset, physical or virtual, deployed across the electricity grid that usually lives behind the meter (behind the meter means anything that occurs on the end user’s side of the electricity meter) and resides close to the ultimate endpoint of demand. A straightforward example of a DER is a residential solar panel: it’s a physical asset on someone’s private property that is connected up to the electric grid and can supply electricity to an end user. That end user could be the owner of the solar panel or it could be some other random customer on the grid that the utility supplies by paying the solar panel owner for his or her DER-generated electricity. Another slightly less intuitive example is a smart home thermostat. This example satisfies the behind the meter and near endpoint criteria, but it doesn’t really supply electricity in a traditional sense. Rather, the smart home thermostat is considered a DER because, as a grid-connected device, it can adjust the end user’s electricity demand by changing the home’s heating and cooling patterns, thereby smoothing the household’s demand curve and making it easier for the utility or retail energy provider to buy supply to satisfy aggregate demand. Other common examples of DERs include micro-turbines, car batteries, home battery storage, heat pumps, and connected appliances, and have been growing at a rapid pace (predicted to hit 387 GW by 2025 up from 264 GW in 2015) thanks to falling costs of the DERs themselves as well as federal, state and local policies that stimulate demand for the things.
DERs are a fundamentally different approach to supplying electricity to end users than what has traditionally been done through the electric grid. As I hinted at in my introduction, the fundamental difference between legacy electricity supply and DERs is the decentralized nature of DERs. Today, most people get their electricity from a utility, which is either vertically integrated (and thus generates its own electricity) or must purchase electricity from wholesale or retail electricity providers. Wholesale providers (who may also sell to retail providers) tend to operate large, centralized facilities like massive power plants, natural gas facilities, or utility-scale wind farms. For simplicity, I am going to refer to the entity selling the energy as the supplier (i.e. the owner and operator of the power plant) and the utility or retail provider as the buyer. For the buyer to make money selling electricity to end users like you and me, it has to have a good idea of how much demand for electricity there will be at a given time so it can purchase the right amount of supply ahead of time and lock in prices.
For those familiar with finance, it’s sort of like a forward contract: for example, buyer wants 4,000 kWh of electricity at $0.09 per kWh delivered next Tuesday to supply what it thinks will be demand next week Tuesday. Then the buyer can deliver that electricity to my house and I pay my monthly bill. If the buyer did a good job, it makes the spread on what it paid the supplier and what it made in revenue from all the households and businesses paying at the end of the month. But, if the buyer miscalculates and comes up short on supply, it is forced to go back to the supplier and purchase more electricity at whatever price is being asked, which could result in some sort of premium to what the buyer had locked in when purchasing supply in advance. That means the potential for a much smaller spreads or even a loss, and if this occurs systematically at scale, the buyer will eventually go out of business. The same logic holds if the buyer miscalculates and purchases too much supply — it’s then left with supply it paid for that doesn’t actually get used by end customers and is thus not monetizable (though important to call out that there is no physical product “left over”, rather it’s just electrons that never find a sink and just head out into the ether).
Value Propositions of Decentralization
How does this all relate to decentralization? DERs shift the power to generate and supply electricity to the grid away from centralized, vertically-integrated utilities, energy generators and distributors and move that power into the hands of individuals and businesses. In the future, no longer will a single household be reliant on the staid, old-school utility to ensure daily delivery of electricity — rather, households can take responsibility for their own electricity into their own hands and literally go off the grid (or, at the very least, can reserve viable options for electricity during an outage or other grid catastrophe). This is, of course, a very utopian vision for a DER endgame that I don’t think we are likely to see. Rather, it’s much more likely that the grid will continue to exist more or less as-is with legacy energy suppliers still participating but with far greater DER integration that will materially impact both demand and supply sides of the electricity market.
On the demand side, the proliferation of DERs will greatly improve the accuracy of demand forecasting, demand response, and other efficiency programs. Here’s the way I think about it: DERs are nodes scattered across a geographical region that provide a flow of information on household-level or business-level electricity usage back to some entity (could be a DER managing software company or a utility) which can be leveraged to better model electricity demand and thus improve the utility’s ability to plan supply purchases for various times. This is a huge increase in the amount of data inputs available for forecasting models and could substantially improve load forecasting abilities as a result. DERs are also good at demand response and can easily recommend to their owners / end-users actions or settings that will optimize electricity usage to minimize the amount the user is charged for their electricity. For example, a car battery charger could schedule the car for charging only in the very early hours of the morning, when electricity demand is lowest, thus saving the end user money and reducing stress on the grid at times of peak demand.
In another example, DERs could prompt the dishwasher to run in the middle of the night instead of at a time of peak demand to minimize grid stress and drive down the cost of electricity for the end user. In both these examples, the DER effectively shifts load from what may have been peak load times (when electricity is most expensive) to less in-demand times, thus lowering costs for the utility. A burning question I have on this area is the following: if the pattern of load shifting and demand response by DERs becomes sufficiently routinized (i.e. megatrends / broad patterns in shifting emerge), does a flywheel effect emerge between accuracy of load projection and load shifting / demand response patterns? The implication here being that getting a minimum viable number of customers with DERs participating in demand response could greatly improve load forecasting abilities and, in turn, could grow profits for the utility. If you have thoughts on this or know the answer to this, please reach out, otherwise I will continue my quest to find the answer.
The supply side is where things are much less fleshed out and, in my opinion, much more interesting. An important idea here is that energy generation for electricity is by and large a centralized affair by necessity. This is because hydrocarbon-based power like natural gas and coal as well as nuclear power generate electricity more efficiently when done so in a large-scale format; this is due to the concept of thermal efficiency, a corollary of which states that efficiency of energy generation scales with size of the combustion engine (up to a certain point). For other types of power, specifically solar, wind, storage devices like batteries, and modular energy generation, size of combustion engine is not a factor and energy from those sources does not require the large-scale aggregation we see in natural gas, coal and nuclear for maximum efficiency. In the most naïve interpretation of DERs supplying the grid, DERs like residential solar and batteries could take over responsibility for supplying electricity for the household or business in which the DER sits and work in an off-the-grid fashion, since they don’t need to be physically aggregated to operate efficiently. The two benefits of this naïve interpretation are the avoided costs incurred during central energy production and the avoided costs from the transporting the electricity to the end user (since a DER by definition produces energy closer to the end user) — it’s an overall more economically efficient way for a household to get energy. However, the truly impactful and large-scale way that DERs could supply electricity to the grid is more complicated.
In a DER-centric future, electricity supply could be provided by what I’ll call a net of DERs: a collection of connected nodes scattered across a geographical area that connect to the underlying grid system. Since they are DERs, these nodes know the demand for electricity coming from their owners and can relay that information to the grid. Scaled up to account for all potential sources of demand, this results in a demand curve that can be much more accurately drawn ahead of time and can produce the equivalent of perfectly discriminating pricing in economics: every end user’s demand at a given time is precisely understood, which helps suppliers figure out exactly how much and when to buy electricity while making more money doing so.
Of course, handling all this data will require a shift from deterministic demand modelling to probabilistic demand modelling, as well as a shift from making purchase decisions not based on worst-case scenarios but rather making them more efficiently based on the probabilities of different demand scenarios. These kinds of models require heavy computing power, since the number of possible permutations of DER to grid interaction is astronomical, and it’s likely utilities and retail energy providers will need to upgrade their computer / modelling systems to handle this as DERs proliferate.
Circling back to how the DERs supply the grid, since the demand curve is precisely understood, the entity managing all the DERs can actually just supply the exact amount of electricity needed to meet demand. This requires sufficient penetration in a geographical area, but basically the DER manager could aggregate all the supply potential from the DERs in the network and then supply that electricity to the grid at scale. Since the connected DERs know demand and supply across the network, the DER manager can pull on both supply and demand levers in the case of any mismatches. The result of such a system is something like a virtual power plant, in which distributed nodes supply electricity and the previously relied upon centralized energy generators become redundant. This is exciting because it functionally excludes hydrocarbon-based energy from the grid, since those energy sources typically only run cost-effectively in large-scale settings where they can benefit from thermal efficiency. Thus decentralization, the rise of DERs, and their integration on the supply side of the grid will accelerate the decarbonization of electricity and do so in a flexible and reliable way.
I don’t know about you, but I find this idea so interesting and exciting and I can’t wait for this future to be realized across the country.