Variable Renewables & the Options for Large-Scale “Flexibility”

Because they involve buying no fuel, wind and solar PV have always been lowest marginal cost electricity generating options. In many markets they have now become – or are fast becoming – the lowest cost options on a lifetime, levelized cost basis too. So it makes sense for future, renewable-dominated energy mixes to be built around these lowest cost options. However wind and solar are inherently inflexible, generating when nature allows rather than when consumers demand. So what is built around them needs to be flexible, with system mixes moving away from concepts such as “baseload” generation (as I’ve written about previously).

Scaling up wind and solar to very high percentages will not be possible without scaling up system solutions to provide for when the wind doesn’t blow or the sun doesn’t shine.

This will be problematic and requires that we pay more attention to full-system costs rather than “plant gate” generation costs in future. However, if you regard problems as opportunities awaiting solutions, then the growth of wind and solar must also create big growth markets for those who can create cost-effective system flexibility.

In this article I’ll outline the fundamental requirement of systems with large-scale wind and PV capacity and briefly indicate how this points towards potential categories of solution. This article aims to step back and “see the wood, not just the trees”: I’ll be delving into the details of specific solutions in future.

Power and Timing

Power (electricity) systems exist to perform one function: delivering the right amounts of electrical power to individual end users at the exact moments they demand it.

It’s important to focus on the term power here. Power delivered over time multiplies up to energy, which usually determines most of an end-user’s electricity bill. But it’s the power supplied at a specific moment that allows them to do what they want: they can’t drill a hole in the wall if the network only supplies energy at a rate (power) sufficient to charge their mobile phone.

The fine detail of how systems do this, connecting multiple sources of generation through transmission and distribution networks to end-users, becomes mind-bendingly complex. But all this complexity is fundamentally about one thing: timing – the right amount of power at the right time.

So if we want to transition towards power systems which are largely supplied from renewable sources (eventually 100%?), these systems need to continue to meet that same timing requirement.

There are a variety of scales and timeframes we can consider from bulk power and energy supply over hours, days and months, to fine-scale, second-to-minute timeframes requiring “ancillary services” (frequency regulation, voltage control and so on). The latter are important, and I’ll return to them in future. Here I want to consider the big picture, making some fundamental observations about bulk power balancing in the transition to large-scale renewable power.

Least Cost, Least Flexible

Some sources, hydro and biopower for example, are inherently dispatchable and slot neatly into existing power systems. Indeed hydropower is very much a “conventional” form of supply: the first plant started operation back in 1882, in Wisconsin, US. Biopower comes in a variety of guises, but biomass is basically a non-fossil hydrocarbon: so biomass and derived biofuels effectively provide direct alternatives to fossil fuels of various types at various scales: solid (replacing coal), gaseous (natural gas) or liquid (oil/diesel). As with fossil fuels, biomass has a cost but you can choose to burn it when there is demand for the electricity produced.

Wind and solar are different.

They have been growing much more quickly because they have become cheap, quick and easy to build. However they are inherently inflexible, generating only when nature provides the resource. If that isn’t when our user wants to drill a hole in the wall, then our system design requires some new thinking. With “plant gate” generating cost ceasing to be a barrier to growing wind and solar capacity, overcoming this basic supply/demand conundrum becomes the key issue to address. 

If supply doesn’t match demand at a point in time, there are basically two things you can do: change the supply, or change the demand. To help make the required changes each way smaller, you would ideally adjust both at the same time.

Supply > Demand

We are already familiar with situations where this occurs. It’s been most common with wind power, resulting in negative prices (see below) or curtailment. Although those are two current “solutions” to the problem, I’d suggest that neither cutting off zero marginal cost production (but still paying for it) nor considering it worse than valueless (by paying people to take it) are desirable in the long-term.  


Figure 1. Two very windy days at the end of 2016 in Germany


Instead, we can either turn down supply or create additional demand.

If we don’t want to be forced to curtail our cheap, zero marginal-cost resource, then turning down supply requires:

  • Designing our system such that the available capacity of wind (or solar) will never exceed demand


  • Designing our system such that the remaining capacity includes sufficient that is flexible enough to turn off whenever and at whatever rate is required.

The problem with this is that it may mean capping the market for what could be (already or in future) our cheapest sources of electricity. If large-scale (100%?) renewable power is the goal, it also relies on sufficient resource (hydro and/or biomass) that is both renewable and flexible, to make up the remainder. In many markets, this may be neither realistic nor – for various reasons – desirable.

We have more options for creating additional demand to eat up our “excess” supply:

  1. Extend the system to reach more electricity customers by removing local, intra-grid bottlenecks (by grid extension or reinforcement). This relates to the specific case whereby demand does exist, but simply can’t be connected with supply because of grid constraints.
  2. Build more inter-system connections (interconnectors): tap into demand within adjacent systems.
  3. Put electricity into storage: use it to charge batteries, pump water up hills etc.
  4. Encourage a bigger group of users to use electricity now rather than later, through multiple means: e.g. smart meters and price signalling, demand-response and aggregation.
  5. Extend the system to serve “non-electricity” customers: in particular, use electricity to generate and store heat (or cold), or convert electrical energy to fuels (power-to-fuels: hydrogen or hydrocarbons). Both are a form of energy storage, but I’m separating from option 3 on the basis that the stored energy isn’t going to later be converted back to electricity. This is often termed substitution

Some of these solutions are at earlier stages than others, but they will ultimately all be required in some combination. They represent potential growth areas for innovators (and their investors) who can offer solutions.

Between them, they need to scale sufficiently to create enough extra demand, such that our supply not only has somewhere to go but also has a market value. (For the last three options, negative prices – i.e. payments – certainly encourage increased usage, but these are really just a symptom of an inflexible system). 

Some of the same growth opportunities will crop up again as we consider our other potential power mismatch.

Demand > Supply

This is the classic criticism of “variable” renewable power sources: how do we keep the lights on when the wind doesn’t blow or the sun doesn’t shine? Currently, the answer is: burn something else (see below – which also shows price spikes at times of “tight” supply).


Figure 2. Two calm days in Germany at the start of 2017


To balance a situation of insufficient supply (relative to demand) we obviously we need solutions which can do the reverse of those discussed previously. We need to be able to either reduce demand or increase supply.

Increasing supply can be done by:

  1. Turning on some other flexible generating capacity that is available within the system.
  2. Generating electricity from energy that was previously stored: discharging a battery or letting water drop down a hill.
  3. Accessing some other flexible supply outside the system: through an interconnector.

In future, “burning something else” is an option we want to phase out (unless we have a sustainable and scalable source of biomass). So the first of these options requires that we already have (or build) and maintain, sufficient capacity of other, dispatchable supply – just to cover these periods. A market with substantial hydro resource could use this, if available at the right time. However whatever the source, it might be difficult to build a business case for building new flexible capacity: who will invest in an asset that will be called on infrequently; increasingly so as alternative solutions grow? Guaranteed capacity payments provide one business case and in a transitional phase they may prove essential. However in the long-term, is it economically efficient to operate a system reliant on such payments to keep alive often-idle capacity at very large scale?

We’ve already encountered storage: as a solution to increasing demand. Here it is acting as a source of generation. This highlights two very important points when comparing storage to “conventional” flexible generation.

Firstly, storage can act as both a source of demand and a source of supply (at different times). Whereas a conventional generator can supply when needed but lies idle when not, storage can supply when needed to and, at other times, can be busy charging itself up. If these other times are ones where additional demand is helpful to the system – as discussed previously – then storage has provided two advantageous “services”, not just one.

Secondly, storage is not a net addition to power/energy supply. In fact, because there are round-trip losses, it is likely a net subtraction (unless locally sited where grid losses are offset). It isn’t adding new electricity into the system; it’s just moving it about. That’s important. It means we still need enough primary energy generation (installed capacity) somewhere in the system – it isn’t saving us from investing in that. But if the latter is inherently inflexible (e.g. cheap wind and solar), additional storage investments are providing flexibility. Electricity consumers will ultimately need to cover the costs of both.

Our third option for increasing supply was to import it. Like storage, the building of interconnectors can provide a dual benefit, moving electricity both in and out of the system. Unlike storage, which can be large or small, centralised or distributed, interconnectors are large-scale, slow-to-build projects. More problematic is that their usage depends not simply on events within one system, but on events outside it too: you can only shed/gain supply if the adjacent system wants to buy/sell it. To be efficient, connected systems really need single trading markets. Also, individual energy mixes would ideally develop to acknowledge an interdependence on those in adjacent systems, creating one sensible overall mix (but potentially ceding some degree of self-contained energy security in the process). That last issue would present major political challenges.

Reducing demand can be done by:

  1. Encouraging more users that they will be better off if they use electricity later rather than now, through multiple means already mentioned: e.g. smart meters and price signalling, demand-response and aggregation

“Better off” means making it worthwhile financially, by reflecting the avoided cost of peak demand in significant savings on electricity bills. They could avoid peak costs by not drilling that hole in the wall just now. This demand reduction/shifting is something well-established for large commercial, industrial or other business consumers. Any strategy towards large-scale renewable power integration must seek to extend this to residential consumers too: smart meters are essential and I expect smart aggregation at the residential level to become key too, in order to achieve the management, monetisation and delivery of demand shifting at meaningful scale.

As with storage – and unlike flexible supply (generators) – demand-shifting and aggregation solutions are able to contribute to both sides of the balance equation. As a result, and because storage can be sited directly at sources of demand (consumer premises), the two will naturally combine as elements of the same mechanism.

For example, consumers with local storage will contribute to demand reduction not by stopping drilling that hole, but by taking the required power from their own storage instead (because price signals tell them it is cheaper to do so), thus reducing the load on the wider system. Similarly, at other times, when the system is short on supply, consumers will sell power into it from their own storage. 

This integration and optimised management of supply, storage and demand between both distributed (grid edge) and systemic (grid-centre) levels will be an increasingly crucial element of the renewable power integration story. It will be transformed by innovators in software and data analysis; creating the “smart” intelligence and algorithms that glue together and optimise the behaviour of the multiple hardware nodes within future smart grid systems.


Concluding Remarks

It’s a consequence of natural resources – both renewable and fossil – that different markets will follow different routes towards more renewable-dominated power systems. They will start from and evolve towards different mixes at different rates of progress, depending on which resources are available, at what level of abundance and at what cost. However, wind and solar are likely to remain or become the lowest-cost “plant gate” generating options in many (most?) locations. Our ability to include them in ever-growing proportions depends on overcoming the challenge of matching supply from these variable sources with demand from electricity consumers.

This is a genuine challenge, and one which will come under increased scrutiny. It creates strong long-term growth opportunities for solutions which can scale and which minimise full-system costs (which are the ones which ultimately matter to end-users). The most successful innovators will complement other aspects of the energy transition: notably moves towards a greater proportion of small-scale and distributed power generation (and the transition of end-users from consumers to “prosumers“)

I’ve mentioned various solutions in this article. It’s important to note that during any transition, flexible conventional sources will continue to play a role, perhaps requiring regulatory mechanisms such as capacity payments to keep them in business. Where available, flexible renewable sources such as hydro will help markets achieve the renewable power transition more quickly. Increased connectivity between power systems will always offer benefits, by extending the balancing area across which sources of supply and demand can be matched.

However here are three areas where I think investors and innovators would be wise to focus. They are small contributors now, but if we are to integrate solar and wind at large scale, they offer huge growth potential.

Electricity storage – particularly battery storage – is already growing rapidly. Although much focus just now is on business models which create value from ancillary services, batteries are becoming cheaper and projects are becoming bigger. That’s good, because storage will need to contribute a big share of our large-scale, bulk power/energy time-shifting challenge, through both distributed and centralised solutions. It’s an industry still ripe for innovation: not just cheaper batteries, but also entirely new batteries, mobile batteries (i.e. EVs) – plus a variety of non-battery solutions.

Demand response (both demand-up and demand-down) is a natural complement to storage and becoming established with large commercial and industrial users. The aggregation of multiple users can help optimise and monetise it. Extending demand response and aggregation to a wider base and ultimately to all consumers, residential included, is a natural progression and a huge opportunity for a variety of new market players. Even with more time-of-use pricing, there will be limits to how much we can really time-shift final electricity end-use (i.e. waiting until later to drill that hole in the wall). However in combination with storage as a source of both demand reduction and demand creation, the “smart” manipulation of system demand will be the other big contributor to system flexibility.

In my view, Power-to-fuel technologies will also come to play a major role. They are at an early stage (and noncompetitive) just now, but the energy density of fuels, the ability to store them over long periods (to balance seasonal variations for example) and their multiple end-uses (substitution) can only increase the flexibility of both the electricity and the wider energy system. Rather than in cars, as once assumed, a “hydrogen economy” is more likely to emerge as a way to create new demand (and value) from very large capacities of installed solar and wind supply.

For similar reasons of long-period and bulk energy shifting potential (particularly to suck up excess supply), Power-to-thermal (both heat and cold) is very much on my watch-list too.

Naturally, when it comes to the evolution and deployment of any of the solutions discussed here, the devil will be in the details. All these solutions will cost money, so while the debate about generating (wholesale) costs has tilted in favour of solar and wind, focus needs to move towards full-system, retail costs instead. These, after all, are what matter to end-users. They won’t be impressed if we claim renewable power is cheap, while at the same time raising their bills. Rather than generating costs, it is the combined cost of generation and integration that needs to start guiding the direction of travel. Allocating these costs is going to be a key challenge for regulators.

Whatever the scale of the challenge, stepping back and focusing on the fundamentals tells us that change within power systems is essential. That’s good news, opening large market growth opportunities for those who provide the enablers for this change.


Like this? Then Share it!

Leave a comment