Tuesday, December 6, 2022

Thought Experiment: Could we power Saskatchewan solely on renewables?

There's a kind of crystal-ball forecasting that makes claims about what could be possible with renewable energy. Namely that it is possible to run ALL cities, provinces, and countries on 100% Wind, Water, and Solar (abbreviated WWS). 

While possible in areas rich in hydro resources (check Electricity Maps for Quebec, Norway, and North-East Brazil) it is practically impossible where large-scale hydro isn't an option - like Saskatchewan. 

In Saskatchewan we recently had a calm, windless week (mid-November 2022) that can serve as a stress-test case study for determining requirements for 100% WWS. Here's a snapshot of wind capacity and wind energy delivered:

Figure 1. Source: https://skelectricity.info/

We'll focus on November 10-16, 2022, where wind utilization averaged 8.2% (meaning for every 100 MW of wind capacity installed, an average 8.2 MW of power was supplied). 

The question we're trying to answer is: what would it take to run Saskatchewan on Wind, Water, and Solar (and Batteries) during a winter week where the wind doesn't blow?

This post is hand-wavy and inexact, but the point should be the orders of magnitude required. My spreadsheet is here and I welcome corrections and feedback (leave a comment!).

First: let's look at today's power generation (in megawatts) and the Wind, Water, and Solar generating capabilities we'll have by 2027, based on information released by SaskPower. The second column excludes current fossil resources to show the generating gap we would need to make up: 

Figure 2. References in spreadsheet linked above.

Before getting to "what would it take," I'll list my assumptions and simplifications so that this simple model stays simple. 

  1. Our 100% WWS grid will need a minimum provincial power generation capacity of about 4,300 MW. This is 10% over the historical peak load of 3,910 MW. This is a slim, dangerous margin, but again: we'll keep it simple. (current SK generation capacity is about 5,400 MW). 

  2. We're not going to add any new solar in our models. Wind has a higher capacity factor (~37% vs. solar's annual 14% - source). This simplification should favour 100% WWS.

  3. Days and nights are each 12 hours. 

  4. Demand is an even 75 GWh of daily consumption, based on recent data from https://skelectricity.info/

  5. Demand is evenly distributed through the day (i.e. days and nights have the same, flat demand) 

  6. There is no new hydro capacity in Saskatchewan, and we can't get any more than we have from Manitoba. Assumption based on SaskPower comments in recent public engagement webinars, plus the intuition that if we go 100% WWS, our neighbours will as well. Their hydro will be a scare resource. 

  7. We'll generously assume the "Other / Imports" category can perform at 100% utilization, but we'll also assume we can't get MORE new power from our neighbours. Presumably, because they have also adopted 100% WWS grids, they have little extra to spare when the weather doesn't cooperate. 

  8. We assume existing hydro stays near its recent typical ~30% utilization, as Saskatchewan's hydro serves a secondary purpose of managing water levels in lakes and rivers. 

  9. We assume batteries are both simple (perfectly efficient), and possible on the scale of days and weeks (currently, to my knowledge, they are not). 


Capacity megawatts are deceptive...

As I've written before, megawatts from intermittent sources aren't always there when we need them. 100 MW of solar panels can never get above (approximately) 50% utilization because it's dark half the day (a City of Saskatoon pilot, linked above, came up with a 14% capacity factor). 

So we take the chart above and add more detail below. The second column highlights the "Gap to Fill" to achieve our minimum required generating capacity, based on the assumptions above. 

The third column, "Typical Capacity Factor Basis" shows the expected output in MW from the planned 2027 supply mix (megawatts are deceptive! It's MW + capacity factor + dispatchability that tells the real story, as we'll see). 

The final column, "November Wind Drought Effective Capacity" shows how our expected 2027 assets would have performed during the wind drought November 10-16, 2022, and the large supply gap we'll have to plan for in our 100% WWS scenario. 

Figure 3

That last column is concerning: ~3,460 MW of on-paper WWS generation capacity will perform like 1200 MW of generation in a wind/solar drought - and half of that 100% utilized imports. 

Broken up into a week of days and nights, the deficit looks like this: 

Figure 4

So where are we going to get ~3,150 MW of reliable capacity to run the province on 100% WWS through a calm, dark week? 


Overbuilding Wind 

One idea is overbuilding wind to an absurd degree. Simply dividing 3,150 MW by 8% utilization suggests we could add 38,580 MW of wind generating capacity to "guarantee" 3,150 MW of "reliable" output during a wind drought. 

38,580 MW of wind capacity is: 

  • 62x Saskatchewan's current wind capacity (626 MW)
  • 2.3x the capacity of Ontario's entire nuclear fleet (but direct comparisons fail, as nuclear MW are there when you need them, wind MW are not)
  • ~60% of the installed wind capacity of Germany, who have a horrible carbon footprint because they back up their (extensive) intermittent renewables with coal. Germany uses about 20x the total electricity of the province of Saskatchewan
  • ~9,600 new wind turbines to be built and installed (4MW each)  
  • ~26,500 km^2 of land that would be packed with turbines, with a turbine density of 2.75 km^2 per turbine. If centered around Saskatoon, this wind farm would encircle Blaine Lake, Wakaw, Viscount, Outlook, and Biggar:
    Figure 5

Our provincial power generation capacity (on-paper MW; not adjusted for utilization) would look like the third column: 

Figure 6

However, we know this entire scenario is ridiculous: 8% average utilization for a day could mean 16% utilization for one hour followed by 0% utilization the next, rendering the entire investment useless. Any purely wind-dependent scenario fails the reliability test, so "firm", dispatchable, reliable resources must be added. 

Can batteries help? What would it take to get through a week's wind drought?


Overbuilding Batteries

Batteries are the saviour of WWS scenarios because it's claimed they can fill in the (huge) gaps. They also don't (yet) exist at the scale WWS scenarios require, but theoretically could exist in the future.

The reality of batteries is they're "grid-scale" in that they exist on grids today, but they do not function as a long-duration grid backup. They are useful for providing fast, flexible, and dispatchable power when there is an excess of cheap, variable renewable energy. SaskPower is currently building one 20 MWh battery in Regina.

Batteries are not meant to provide bulk energy over long timeframes - so this whole section is extra made-up and hand-wavy - but we'll map it out anyway. For other useful functions of grid-scale batteries, see Table 1 in this paper from the US National Renewable Energy Laboratory (NREL). 

According to Wikipedia, the world's largest battery in service today is the Vistra Moss Landing Battery, with a power output of 400 MW for 4 hours (1600 MWh). The project was commissioned in 2021 at a cost of $400M USD. 

If batteries were capable of getting us through a week of low wind (again, they are not used for long-term bulk energy storage) Saskatchewan would need 523 GWh of energy storage capacity. This is the sum of the red bars of MW power needed in Figure 4, multiplied by hours, to get MWh (GWh).  

523 GWh of battery storage is:

  • 26,150 times more battery capacity than SaskPower has currently planned (20 MWh being deployed in Regina soon)
  • Equivalent to 336 Vistra Moss Landing batteries, currently the largest in the world
    • This would cost $134 Billion (CAD) although I'm sure some economy of scale would get that down... a bit. 
  • 23 times more battery storage than the entire United States of America has in 2022 (see interesting Twitter thread)

One of the risks of counting on batteries to survive a seven-day wind drought is... what if the drought is 8 days? 10 days? Two weeks? Once batteries are discharged, they require excess energy (over the base load of grid consumption) to be recharged, so that needs to be accounted for too. 

What about lots of small-scale batteries (and solar) installed at each home? This is implausible for two reasons. One, most of Saskatchewan's electricity is not supplied to homes. It's supplied to farms, businesses and industry (Figure 7), who depend on energy security for their businesses and operations to remain sustainable. Two, intuition says it would be cheaper and more impactful to do a handful of very large projects versus half a million very small projects. 

Figure 7. SaskPower 2021-22 Annual Report

What about pumped hydro, compressed air, or other storage technologies? Storing a week's worth of energy in any medium takes an enormous - and I daresay impractical - amount of resources, unless you're storing energy in the form of fossil fuels or uranium, which are far more energy-dense than the alternatives.  


Conclusions

Wind energy can be useful and plentiful: utilization averaged 85% for the five days following the seven-day wind drought. But due to its variability, we cannot solely rely on wind. 

Battery energy can be useful: the NREL paper discusses lots of interesting and useful functions batteries perform on the grid. But these are on the time scale of hours, not days. Batteries are not meant for long-term bulk energy storage, and storing a week's worth of energy in any medium other than fossil fuels or uranium ranges from impractical to impossible. 

Huge quantities of solar would not change these conclusions as it's useful far less than half the time in winter. 

It will take a mix of credible, viable, and economic energy sources to decarbonize Saskatchewan's grid. In a world where new hydro resources are constrained, only natural gas and nuclear energy can provide "firm", dispatchable (on-demand) energy at scale when intermittent renewables leave large power gaps. 

Building four 300 MW SMRs effectively replaces Saskatchewan's coal power plants, and will be less expensive and risky than building a Rube Goldberg machine of wind and battery (or other) energy storage projects to hopefully carry us through a calm, dark week. For safety, health, and prosperity we cannot gamble with energy security - this is not a service we can afford to outsource to our neighbours. Energy must be reliable and we must be responsible and accountable for it.  

This post was quite hand-wavy and imprecise. The exact numbers aren't the point, it's the orders of magnitude. Proponents of any energy solution must be able to show how the solution works here, a Northern, dark, cold, flat (hydro limitations) and often very calm environment. 

As I said at the top: I welcome constructive comments, corrections, and feedback: I am still learning all this stuff, too. 

Two books I cannot recommend enough are:

  • Sustainable Energy - Without The Hot Air, by David C. MacKay (my review). As MacKay says, "it all has to add up." That is the inspiration for many of my posts! While some of the cost data on renewables from the publishing date of 2008 is outdated, the premise of the book is evergreen: what are all of the credible options (that "add up") to building low-carbon grids? 

  • Shorting the Grid: The Hidden Fragility of Our Electric Grid, by Meredith Angwin (my review). An experienced deep dive into grid operations and governance, including challenges of integrating and operating renewable-heavy grids.