Dunkelflaute

I was surprised to get email and an SMS from my electricity provider* this week warning me about price spikes (I am on a wholesale pass-through arrangement). This is normally a summer phenomenon. Fortunately, the spikes were not as bad as predicted, but I thought I’d better see what was happening. It turns out that both Tuesday and Wednesday evenings saw a combination of high demand (due to the cold), no solar (no surprise after 5pm in winter!) and low wind. Dispatchable plant had to step in to make up the gap, mostly gas and hydro, and of course this led to a price spike, although not all the way up to the price ceiling of $14,700.

Source: OpenNEM

As the snapshot above shows, variable renewables were only providing about 1.5 per cent of total generation as the evening peak approached. Note also the negligible contribution of batteries discharging. With the current energy mix, the NEM handles this kind of situation just fine. But as this illustrates, as coal exits, there are going to be times when having more renewables doesn’t mean more supply and the gap has to be filled by other resources.

The Germans have been thinking about this for a while. They even have a word for it: Dunkelflaute, which loosely translates as a “dark lull” or “dark doldrums”. The solutions depend on how long it lasts for. As can be seen from the snapshot, we’ve had sunny winter days recently so even though there has not been much wind, solar has been able to make a material contribution during the day. So the issue is more a diurnal one that could be addressed through a range of flexible resources, from batteries to demand response to pumped hydro. The analysis in Germany found that there could be a period of up to two weeks with low wind and little sun during the day, which would require some longer-term storage options. Pumped hydro would be helpful here as could hydrogen or compressed air storage. Depending on the carbon constraints, gas plants would be effective in both scenarios.

Will this be expensive? This approach – of renewables backed up by dispatchable resources – to meeting our energy needs requires more “steel in the ground” or capital investment than the traditional use of coal, gas and hydro plants. However, it avoids fuel costs (except to the extent gas is still a contributor) and cost projections are for renewables and batteries to continue to get cheaper. The German modelling assumes that there will be a large fleet of electric vehicles in any case, so there’s relatively little incremental cost in setting up the grid so that they can discharge into it through the evening peak and then recharge as necessary before the morning commute. Here in Australia, we will have Snowy 2.0 and perhaps the Tasmanian equivalent: “Battery of the Nation“.

The key to containing the cost and making sure we don’t pay more than necessary, is as always, to use the incentive properties of price signals and appropriate risk allocation. This flows from choices made in the design of the electricity market and of support policies for renewables batteries and other new technologies. The reason my retailer alerted me to the expected price spikes was because I am on a wholesale pass through and so I can save money if I can minimise my consumption during the price spikes. At the moment that’s hard to do, but at least I’ve got the incentive to engage with service providers who can help me set up automated load-shifting, for example. Even if I wasn’t on that style of contract (most people aren’t) that just shifts the risk to my retailer, so the key is to make sure they have the right incentives to look for ways to manage their customers’ load as a cost-effective alternative to protecting themselves through hedging contracts. On the supply side, what’s important is to ensure the price cap is high enough to incentivise investment in some resources   that might only be used a few hours a year, if that.

Thanks to Paul McArdle at WattClarity for reminding me about Dunkelflaute.

*NEM Risk Bulletin is not in the business of promoting one company over another, but I thought it appropriate to provide a link for those interested in checking out this business model.