Skip to content

Renewables Dis-integration?

This post is co-authored with Duncan Callaway.

Ahhhh Hawai’i, where the waves are big, the beaches are long, and the renewable energy ambitions are large:

surfSource                                                                           Source

As you may have heard, Hawaii has set a goal of 100% renewable energy penetration by 2045. This is pushing the Aloha state to the bleeding edge of renewable energy integration.

We were lucky enough to spend our spring break at a conference hosted by our colleagues at the University of Hawaii. Although the focus of the conference was on the integration of renewables into the grid, there was lots of talk about possible dis-integration.

Living off-grid might seem like a pretty out-there idea. It’s true that, in the past, “grid defection” has been limited to survivalists and bohemian-types who can get excited about living like this:grid


But with solar PV and storage costs falling relative to conventional grid-supply systems, the economics of grid defection are changing fast. And off-grid living is not what it used to be:solarhouse

A Hawaii developer is planning a 410-home project that would become the first off-grid community of this magnitude.

If grid defection takes hold, this would take concerns about stranded assets and utility death spirals to a whole new level. Grid dis-integration is a development we should be paying attention to.

The economics of grid defection

Let’s start by distinguishing a grid defector – the subject of this blog-  from a load defector (that’s you if you have PV panels on your roof).

“Grid defectors” are consumers who fully disconnect from the grid and supply their electricity needs with their own power generation. “Load defectors” remain grid-connected, but get some fraction of their electricity from a source other than their incumbent utility. Importantly, load-defecting solar PV customers continue to rely on grid services. They draw power when demand exceeds their solar supply, and inject surplus generation when there’s excess supply.

The economics of solar PV load defection already pencil out in many places (even the Kentucky Coal Museum!). For many consumers (the US passed the 1 million solar PV installation mark last year), net bill savings appear to exceed private investment costs. This is partly due to falling PV technology costs. But it’s also thanks to retail prices that can significantly exceed the variable costs of supplying electricity, net metering policies, and other subsidies. For example, the graph below shows how Hawaii’s residential prices per kilowatt hour (kWh) reflects not only the variable costs of generation (green line), but also a substantial amount of fixed cost recovery (black line). When a net metered solar customer generates her own power, she avoids paying the sizeable fixed cost component that was being collected through her volumetric payments for electricity.

robertsgraphSource: This graph shows the average residential electricity price in Oahu. The green line measures the generation component (fuel costs and costs of buying energy from independent power producers). The black line measures non-fuel and fixed costs.

The economics of grid defection are more complicated. If you want to unplug from the grid AND maintain the same level of reliability and power quality, you’ll need to make investments in a battery (or a backup generator) in addition to PV. In 2014, analysts at the Rocky Mountain Institute estimated that an off-grid PV-battery system would average about 80 cents/kWh in Hawaii. The graph above shows how 80 cents/kWh is still a long way from grid parity.

We’ve made some updated calculations with an eye towards rapidly declining solar PV and storage costs (with the help of Berkeley graduate student Jonathan Lee). The graphs below show just how fast PV and battery prices are falling.

pricegraphSolar PV source: Tracking the Sun IX                                                                 Battery source: While the battery figure shows costs for EV batteries, these cost reductions carry over.  It’s been estimated that over 18 months in 2015 and 2016, grid-connected battery system costs fell 70 percent.

Without getting too far into the weeds, we use a moderately aggressive scenario for solar and storage costs ($0.40/W for solar panels, $100/kWh for batteries), plus a host of other assumptions for system installation costs. We estimate that a stand-alone system would cost about 30 cents per kWh (assuming less than an hour a year of supply shortage). This is within the range of residential electricity prices in recent years (see above graph)… and that’s before accounting for state and federal incentives.

That grid defection could actually be cost-effective may seem inconsistent with everything you thought you knew about economies of scale in electricity generation. It’s true that generation costs per kWh are lower when electricity is generated at utility-scale, versus on your rooftop. The catch is that, as solar PV and storage costs fall, the additional cost associated with generating and storing electricity on a smaller scale could be more than offset by the transmission/distribution/retail service costs you can eliminate with a decentralized system.

If PV and storage costs get low enough, it could become more efficient to start grid dis-integrating (versus making additional investments in grid infrastructure). This is a mind-bending concept for those of us accustomed to thinking that the grid simply can’t be beat.

powerlinesStranded assets of the future?

Preparing for grid dis-integration?

Renewable energy developments in Hawaii can read like postcards from the future. Hawaii has been on the forefront of the distributed solar revolution, with PV penetration exceeding 15 percent on some islands. The state was the first to shut down net metering in response to costly solar load defection. Today, Hawaii is anticipating the next challenge that is grid defection.

Economists prescriptions for dealing with inefficient load defection emphasize cost causation: set real-time per-kWh prices at true variable costs. But there’s a hitch. When rates are set to recover real-time marginal costs, revenues can fall far short of total system costs. These residual fixed/sunk costs have to be recovered somehow.

If there’s no risk of grid defection, there’s room to be sloppy about recovering these residual fixed costs with some kind of fixed charge. But in a place like Hawaii, sloppy fixed cost recovery could lead to inefficient grid dis-integration in the not-so-distant future. So what’s the right way to recover sunk investment costs? Lumping them into customers’ fixed cost charges would lead to early/inefficient grid defection. Lumping them into per kWh costs has led to inefficient load defection. Exit fees could offer a solution. The socialization of cost recovery via general tax revenues has also been suggested. The jury in Hawaii is still out.

Grid defection may seem like a long way off for the rest of us. But the regulatory paths chosen in Hawaii’s renewable energy laboratory can inform decisions that the rest of the country will ultimately be confronted with. We’ll be watching this Hawaiian grid dis-integration story closely. And hoping to visit those long beaches and big waves again sometime soon.



33 thoughts on “Renewables Dis-integration? Leave a comment

  1. Forgive me if this issue was addressed somewhere in the long replies, but Hawaii is not a model for the continental U.S. simply because it is tropical. And I mean that in the formal sense. All of the Hawaiian Islands lie south of the Tropic of Cancer (which cuts through the southern tip of Baja and kisses the north coast of Cuba). Annual variations in the length of the day and available sunlight are relatively low, reducing the need for storage other than to get you through the night. Try the same exercise for Anchorage, AK, at 61 degrees N, where I used to live. The further north you go (or more precisely, the further toward the poles), the more storage you need to store excess summer production to get you through the winter.

    I’ve done some calculations for my solar PV system in the Bay Area, and how many Powerwall batteries I would need. If my residential PV system was sized to true up annually, the cost of batteries would be prohibitive. WAY prohibitive. It is much cheaper just to install excess PV capacity to drive down the need for batteries, although I don’t know what the cost-minimizing point would be.

    Hawaii is an interesting test case, if not exactly a model. UC Davis’s Andy Frank, the father of the plug-in hybrid, has written about this. Given the relatively short drives people take on the Islands, and the high costs of importing refined and unrefined fossil fuels, Hawaii should really benefit from PV/EV.

  2. Some of you know, I’ve lived part-time off-grid since 1990, with PV as my primary electricity source since 1994. Many would consider our house rustic, but we’re neither survivalists nor bohemians. Well, maybe just a little, occasionally. Take a look at our house.

    The array you see, plus 12 lead acid batteries and a 30-year-old gasoline generator are our island power system. While we don’t live here full-time, we definitely could, and sometimes spend multiple months off-grid. We’re also without mail delivery, city water, garbage collection, and various other unnecessary luxuries. We do have satellite internet, which is our lifeline and second biggest routine load. All this is just to say we know how to work around the limitations of renewable energy.
    Much of the time we’re not in the wilderness, we live in small Berkeley cottage, but we bring our frugal ways with us.
    While the CPUC promotes frugality to some extent, it has its limits. Our PG&E electricity use ranges from 70-100 kWh/mo, depending on how much we’re around. We’re always in the first tier, with an energy price around 10 ₵/kWh, but there’s a tricky customer charge of about 10 $/mo, which makes our usual average price around 20-25 ₵/kWh. Take a look.

    Meredith & Duncan’s post made me crunch some numbers I’ve been meaning to look into for a while. You won’t be surprised to hear that if I pencil a PV system for this tiny Berkeley cottage, I get a levelized cost around 15 ₵/kWh. Although this would be small system, about 500 W, so that may be too optimistic. The original system we had at our country home was 600 W, so we know what this lifestyle feels like. Bottom line, if we stay grid connected, we can’t (yet) beat PG&E’s 10 ₵/kWh.
    The more interesting question is what sort of battery system could we afford and come in under the 20-25 ₵/kWh, leaving us free to fly away from PG&E. Or alternatively, what hardship would disintegration impose?
    I like the promise of sodium-ion batteries for small systems, although they’ve had their problems. They come in at about 500 $/kWh, so I can afford a 2 kWh pack and get in around 23 ₵/kWh. Oh so close! (small print: that’s with 6 $/W, 5% interest for both PV and battery, 25-year PV life and 10-years for the pack, our original 1994 panels still work fine BTW). So close, but 2 kWh ain’t much, way less than we have in country. A small gasoline generator is cheap, but not a good way to stay popular with the neighbors. I already have a couple, actually, so sunk capital cost! Alternatively, if the battery cost erodes down to 250 $/kWh, a conservative target by Meredith & Duncan’s standards, I can have 3 kWh and still come in around 20 ₵/kWh. I think I gotta get on this before the Federal tax credit gets Trumped!

  3. A couple of things:
    1. “Grid defection” means that there is some responsiveness of grid connection to utility charges. This means that the fixed charge should be less and the variable charge more than in the case where everyone connects (see e.g. Auerbach and Pellechio, cited in Borenstein and Davis).
    2. The gap between residential prices and variable costs is not “fixed cost” (e.g. since prices exceed what is necessary to recover the cost of capital). It may also be useful to keep in mind that accounting (and actual) costs are inflated above the (minimum) cost function discussed in the regulation literature (e.g. Joskow).
    3. Hawaii did not entirely “shut down net metering.” Those with pre-existing PV systems were grandfathered in. Only new PV owners were put on a different system.
    4. Hawaii does not require that electricity generation will be 100% renewable by 2045. The RPS metric is total generation of renewables (including distributed solar) divided by MWs of utility sales. The range of this metric goes from zero to infinity. For example, suppose that generation is 45% utility coal, 15% IPP renewable, and 40% distributed solar. Net sales are 45% + 15% – 5%, assuming that line losses are 8% of utility and IPP generation. The metric is 55%/55% = 100% even though utility power is three quarters coal. California apparently uses a similar standard: Generation from “eligible renewable resources” must “be equal to or exceed 50% of retail sales.”

%d bloggers like this: