Can 100% renewables work?
Can “renewables + storage” suffice as the basis for a fully decarbonized energy economy? The assertion from renewable energy advocates has long been an insistent “yes”. The issue, however, appears increasingly in doubt. Resilience.org, for example, recently posted The Rising Chorus of Renewable Energy Skeptics. The article focuses on the very high materials and mining cost of renewable energy resources and battery-powered electric vehicles. Excerpts:
<..> the world needs a better plan to avoid collapse other than replacing one unsustainable fossil fuel system with another intensive mining system powered by even more extreme energies.
<..> For largely ideological reasons many greens and “transitionists” have presented the transition to renewables as a smooth road with no potholes. In so doing they have ignored much basic geology, energy physics and even geopolitics. As a consequence many imagine the construction of millions of batteries, wind mills, solar panels, transmission lines and associated technologies, but they downplay the required intensification of mining for copper, nickel, cobalt and rare minerals you’ve probably never heard of such as dysprosium and neodymium.
Resilience isn’t alone in challenging the feasibility of the common vision of a renewable energy economy. Dr. Simon Michaux, a mining expert working with the Geological Survey of Finland (GTK), has laid out what appears to be a rather compelling case against it. In a 1000 page report packed with charts and data, he compiles the materials requirements of a green energy transition based on wind and solar plus storage, with electrification of transport. He calculates the expansion of mining and processing required to produce those materials, based on present global reserves and estimated resources. Mining and minerals processing are energy intensive, and becoming more so as the grade of available ores declines. From his results, it appears that the required expansion of mining is so great that just the expansion itself could consume 100% of the increased energy production for decades to come.
I’ve downloaded but haven’t reviewed Dr. Michaux’s full report yet. I can’t offer any firm opinion on its validity. A summary of findings available on the same Finnish Geological Survey website looks plausible, as are various presentations that Dr. Michaux has given that can be found on Youtube. Of the latter, the one he gave at a seminar at the Sustainable Materials Institute at the University of Queensland (here) is probably the most informative. It’s long – an hour and 12 minutes – but worth it for anyone who seriously wants to understand the materials issue.
I may devote a future post to Dr. Michaux’s report. In the meantime, there’s another interesting article I came across that focuses on the sufficiency of “renewables + storage”. It doesn’t deal directly with the materials issues that Dr. Michaux and others have raised. Rather, it centers on the economics of coping with the intermittency of weather-dependent energy resources. I found the article on the website for Low Tech Magazine. It was published in 2017, but there’s a free archive of all articles that Low Tech Magazine has ever published. The article is How (Not) to Run a Modern Society on Solar and Wind Power Alone.
The cost of intermittency
The author of the article is Belgian writer Kris De Decker. He feels that it might be possible to run a modern society on solar and wind power alone – but not in the way we’re trying to go about it. A society running on solar and wind power alone would be profoundly different than what we have now.
De Decker’s article, IMO, is an outstanding piece of technical journalism. It’s well researched, peppered with facts, figures, and references, without coming across as stuffy or academic. It includes helpful charts and images, plus illuminating bits of history. So what’s the bottom line?
De Decker doesn’t put it this way, but I’ll say it: intermittency is a real bitch! It isn’t that there’s nothing we can do about it; it’s just that none of the strategies currently available for dealing with intermittency appear good enough to displace the use of fossil fuels. They’re not presently and perhaps never can be economically competitive.
Matching supply with demand
De Decker reviews four strategies for “how to match supply with demand”. All have problems. The first strategy is backup power plants – i.e, what we employ today. We give priority to clean renewables, but when their production is too low to meet demand, we fall back upon fossil fuel resources. The renewables, when they’re delivering, serve to reduce net carbon emissions. But they do so at the cost of maintaining a system of backup generation with the capacity to meet virtually the entire worst-case demand load on its own. That’s because production from wind and solar resources can fall to nearly zero for extended periods of time. Thus, it means maintaining two full-capacity systems in parallel. It increases the net cost of electricity and reduces the already marginal net energy return on investment (EROI) of wind and solar.
Strategy 2 is oversizing renewable energy production. That reduces the fraction of time that backup power plants must be called upon. In doing so, it further reduces carbon emissions. But it doesn’t significantly reduce the capacity of backup power plants required to meet worst-case demand load. It just reduces the fraction of time that they’ll be operating. There will still be extended periods of time when output from wind and solar resources is a small fraction of its yearly average. “Wind droughts” and prolonged periods of overcast skies will still arise from time to time. The system must be able to cope. Hence, oversizing of renewables ends up further increasing the cost of electricity while reducing the system’s net EROI.
Strategy 3 is supergrids: interconnecting RE power plants over wide geographical regions. The regions must be large enough that production from different subregions will not be strongly correlated. The presumption is that low production in one subregion can then be accommodated by importing power from one or more subregions where production is high. As a bonus, robust long distance transmission would enable geographically restricted RE resources (e.g. wave and tidal power, offshore wind, and geothermal) to be exported more widely to areas of need. Plus, under scenarios that relax “only wind and solar” to “mostly wind and solar”, it permits more efficient utilization of a smaller number of backup power plants.
Unfortunately, implementing such supergrids is problematic. Weather systems impacting wind and solar production can sometimes cover tens to hundreds of thousands or even millions of square miles. Though connections between grids over areas that large do exist, the existing connections are far too weak for the power levels that high renewable energy content would sometimes require them to deliver. De Decker cites European studies concluding that a 10-fold increase in long distance transmission capacity would be needed for a robust renewable energy grid across the whole of the EU. For most of the time, the added transmission capacity would be idle or utilized at a fraction of its capacity. Thus it’s economically inefficient and hard to finance. Not to mention that new high voltage transmission lines are generally unpopular. Outside of China, the permitting process is typically slow, arduous, and chancy.
Strategy 4 is energy storage. “Renewables + storage” has become a mantra for activists promoting RE solutions. But how much storage? 4 MWh of energy storage capacity for every MW of RE power is currently an informal rule of thumb. That’s sufficient to maintain grid stability under high RE penetration scenarios and to buffer short term variations in supply vs. demand. It also allows time for smooth ramping of fossil fueled backup power plants. But it’s not remotely sufficient to eliminate the need for those backup power plants altogether.
To be able to eliminate the need for backup plants, the system’s energy storage capacity must be large enough to deliver as much electrical energy as the backup power plants would be called upon to deliver over the worst case supply shortfall periods. In regions subject to large seasonal variations in RE supply, that could easily amount to a several hundred-fold increase over the 4-hour rule of thumb for systems that avail themselves of backup from fossil-fueled generation. There is no energy storage technology currently deployed or under development that would be economical at that scale. Until something radically better comes along, no purely storage-based solution for intermittency can be viable.
Combined strategies
De Decker’s article isn’t perfect. Its treatment of the strategies in isolation is arguably a serious flaw. The feasibility of matching supply with demand is evaluated for each individual strategy in turn. None of the strategies is found to be an economically viable alternative to the existing reliance on backup power plants. But the strategies aren’t mutually exclusive; they can and would be combined. Combinations could prove more cost-effective than individual strategies on their own.
As an example, adding long distance transmission capacity reduces the amount of energy storage capacity needed. Conversely, adding energy storage reduces the amount of long distance transmission capacity needed. Both are reduced by oversizing wind and solar resources to the extent that some fraction is routinely curtailed. But none of these options are without cost. They trade off. Somewhere in the space of tradeoffs there should be an optimum mix that will minimize the overall amount of construction and capital expenditure needed for a workable system solution.
There’s no guarantee that even an optimum mix of transmission, storage, and oversizing will yield a solution that’s good enough to displace the current reliance on fossil-fueled backup generation. It’s likely that with the current state of the technologies involved, it won’t. But that can’t be taken as proven. The studies that De Decker cites don’t include rigorous tradeoff analyses on the balance of combined coping strategies.
I can appreciate why De Decker didn’t attempt such a multi-variable tradeoff analysis himself. It would have turned the article into something like a PhD thesis that readers would have to wade through. Lost in the details would be the most interesting point that De Decker wanted to get to. As it happens, there’s an entirely different approach to running a society on wind and solar plus (limited) storage. It’s an approach we should perhaps be taking more seriously.
Reversing the approach
In reviewing the four strategies summarized above, De Decker was careful to state the goal as matching supply with demand. None of the three alternative strategies to backup power plants that he reviewed looked promising. Possibly no combination would either. But what if we were to change the game and go at it the other way ‘round? What if we focused on matching demand with supply?
In a separate article, How to Run the Economy on the Weather, De Decker writes:
Before the Industrial Revolution, people adjusted their energy demand to a variable energy supply. Our global trade and transport system -- which relied on sail boats -- operated only when the wind blew, as did the mills that supplied our food and powered many manufacturing processes.
The same approach could be very useful today, especially when improved by modern technology. In particular, factories and cargo transportation -- such as ships and even trains -- could be operated only when renewable energy is available. Adjusting energy demand to supply would make switching to renewable energy much more realistic than it is today.
What De Decker suggests is effectively Demand Side Management (DSM) taken to a higher level than most clean energy scenarios envision. Most scenarios do call for some level of DSM to be added to the mix of oversizing, transmission, and storage to help reduce the net cost of implementing the scenario. But so far, those DSM measures have been modest.
Advanced DSM
To date, DSM measures have mostly been limited to voluntary agreements between grid operators and selected commercial and industrial consumers. In exchange for reduced rates, the participating customers agree to cut power consumption during power emergencies. In the past, cutbacks to consumption couldn’t be commanded directly by the grid operator’s scheduling and dispatch SCADA computers (Systems Control and Data Acquisition). They were managed by telephone calls in emergency situations. That’s perhaps starting to change, with the advent of support for “virtual battery” systems.
Implementations for virtual batteries have varied. One of the earliest and largest involved a smelter that was built in 2016 by German aluminum producer Trimet. (GTM article here.) The plant’s smelter lines were designed to operate over a range of power levels. The design range was roughly +/- 50% around a nominal 1.2 GW. Swings over that range allow the plant to free up as much as 600 MW for use on the grid during periods of low RE generation, or productively utilize up to 600 MW of otherwise surplus power during periods of high RE generation.
Tesla Solar has recently begun implementing a different type of virtual battery system for the grid. It’s built around residential solar systems that include Tesla Powerwall batteries. Tesla Solar serves as a point of contact for the grid operator. On the wholesale electricity market, it will ultimately bid as both a consumer and as a supplier for power and ancillary services. It will be a service aggregator, communicating with the Powerwall units of its customer base. The idea is that it will exploit rate arbitrage to optimize customers’ return on their Tesla battery systems. Presently, however, the VPP program is in early trials with simplified protocols.
Virtual battery systems of the sort that Trimet or Tesla Solar have implemented are certainly steps in the right direction. They are not complete solutions, however. Systems that ramp consumption by heavy users according to power availability have no ability to actually supply power to the grid. That’s OK, as long as the grid maintains enough firm generation capacity to meet minimum demand when production from wind and solar resources is down. It reduces the peak load that backup generation must be sized to support, but the requirement for firm backup capacity remains. The type of virtual battery that Tesla Solar aims to support is better; it can provide real generation capacity. However, the capacity is necessarily limited. It can and will run out, if low production from wind and solar resources continues. Then it’s back to the requirement for firm backup generation.
To reach the goal of a 100% renewable energy economy and avoid the need for fossil-fueled backup generation, something more potent than virtual battery systems is needed. That would be something resembling the weather-based energy economy that De Decker writes about. What might that look like, in practice?
Soft Outage Capability
To shift the economy to run more on as-available energy (AAE), grid operators must be able to signal the real-time state of generation capacity to customers and be guaranteed an appropriate real-time response. When generation is in short supply, service to customer premises must be at least partially curtailed. Currently, the only way to do that is with rolling blackouts. But better options are possible. Rather than unconditional cutoff of power to an entire distribution loop, partial cutoffs could be implemented behind the meter at customer premises. Firmware controlling the partial cutoffs would let critical loads remain connected. Thus communications, emergency lighting, and critical medical equipment would always get power. In all but the most severe generation shortfall situations, refrigerators and freezers used for food preservation would as well.
There are various ways this “soft outage” capability could be implemented. It’s largely a matter of agreement on new standards, protocols, and regulatory framework. Some new hardware is needed as well, but the requirements are modest. In fact, the key hardware functionality to support this new operational paradigm is already in the field. Service panels are now being produced with the ability to monitor power consumption and to enable or disable individual circuits under software control. Examples are the smart electrical panels from Span and Lumin, and the Smart Breakers available for Leviton’s Load Centers.
The purpose of these products isn’t specifically to support AAE or advanced DSM. Rather, it’s for smarter and more flexible management of limited backup power capacity when the main grid goes down. But that’s the same hardware functionality needed to support AAE and advanced DSM. What’s lacking is a standard signaling protocol and the software to manage partial curtailments in real time in response to generation shortfalls. When appropriate signaling protocols and software are put in place, the result is a power distribution system that’s able to smoothly adapt to dynamically varying levels of power availability. The grid, in a sense, will be able to function as its own backup system for soft power outages.
Social and Economic Impact
The difference between a grid system with soft outage capability and one with only conventional DSM is mostly a matter of degree. Conventional DSM is intended for short term load shifting or load shaving in a manner that will not be disruptive to businesses and consumers connected to the grid. It might manage to reduce peak demand by up to a third, but the avoided demand will mostly be made up by increased demand during off-peak hours. Conventional DSM therefore can’t help much in getting through extended periods of low wind and solar production without fossil-fueled backup generation. By contrast, a soft outage, in severe cases, might cut active demand by 95%. At that greatly reduced level of demand, battery storage sufficient to maintain service to critical applications for up to a week wouldn’t be prohibitively expensive. After that, emergency backup generators might be necessary, but carbon emissions would be minor.
Soft outages would be disruptive to businesses and consumers connected to the grid. That’s assuming those businesses and customers hadn’t made other arrangements for backup power and hadn’t structured their operations to accommodate intermittency. But those options would exist; the soft outage capability would leave it to individual customers to decide. Meanwhile utility companies, relieved from a statutory obligation to maintain enough firm reserves to meet worst case demand from all customers at all times, would incur lower costs. Prevailing utility regulations would require the cost savings to be passed on to ratepayers in the form of lower rates. Some ratepayers might happily opt for the disruption of occasional soft outages in exchange for lower power bills; others might give it a try to see how it worked out, while still others would choose to pay for their own alternative backup arrangements.
Theoretically, in a world where energy supply was dominated by intermittent renewables, a reworking of the utility model to incorporate the soft outage concept would make sense. The utility would guarantee customers a minimal level of power for critical functions that they could nearly always count on; beyond that, it would be up to customers to decide individually how much backup capacity to purchase. That would be economically efficient, as it would not arbitrarily impose the high cost of full backup capacity on customers who didn’t need it. It wouldn’t magically create an economic order able to operate purely on renewables plus limited storage, but it would facilitate movement in that direction. Taking power from the main grid as available would nearly always be less costly than taking it from a backup source. The economic incentives would align with using energy, as much as possible, on what was available from renewable and carbon-neutral energy resources.