What will Saskatchewan's electricity supply look like through 2030?

I asked SaskPower via the "Contact Us" form if they could share a copy of the "10-Year Generation Supply Plan" that was mentioned on p106 of the 2021-22 Annual Report. The reply was "it's on the website" (I don't think it is) so I've been piecing the path to 2030 together out of a few different sources.

The post will explore Saskatchewan's electrical supply mix up to 2030, as well as hit these themes:

• The difference between generating capacity (MW) and electricity supplied (GWh), because one looks good for ad campaigns and the other runs homes and industry,
• Saskatchewan's go-forward dependence on natural gas as our primary energy source, and
• A few thoughts about ownership and costs of renewables and who benefits.

It was fun to try and piece this together. There are definitely assumptions and missing pieces of information, and I've tried to flag those where they're obvious. Sources are linked in a spreadsheet at the bottom of this post. 

Scenario 1: 40% Renewables by Capacity (T&D Report) 

Here's what I've put together from SaskPower's annual reports, blog posts, found presentations, external reports, news articles, and more. Some key inputs to this model:

• 2022 is from the System Map, 2027 is derived from the 5-year forecast in the 2021-22 Annual Report, and 2030 is based on a) a SaskPower Transmission & Distribution report found online dated October 2021 and b) a SaskPower blog post stating 7000 MW of generation capacity by 2030. 
• Project-based capacity is added where known, like the Bekevar Wind Energy Facility in 2023 and the Great Plains Power Station in 2024. 
• Most other values are linear interpolations between known/inferred data points (e.g. everything in 2028 and 2029 is an interpolation between 2027 and 2030). 

Insights from this model: 

• Coal is not completely shutting down in 2030: Boundary Dam Unit #3 (equipped with Carbon Capture and Storage) will continue to add ~115 MW of generation capacity to the grid. 
• Dispatchable power (available anytime: coal, gas, hydro) has a healthy buffer over the projected peak load. We need lights on calm, dark days! 
• The sum of hydro, wind, and solar "renewable" energy total about 40%, which is about the minimum SaskPower has committed to by 2030 (they've given a range of 40-50%). 
• Look at all that gas! Gas is the primary replacement energy source for coal.
• One invisible data point (and source of model error/uncertainty) is that in the late 2020s a 650 MW interconnect will be completed to the US, allowing SK to import/export power. I assume this is factored into existing 2030 generation capacity models that have been published to date, but I'm not showing a 650 MW increase in capacity in any one year. 

With the chart above, megawatts (MW) of generation capacity do not tell the whole story. What matters is GWh of power delivered to end-users (more commonly kWh on your power bill). Recall: MW is instantaneous; GWh is accumulated. 100 W is how much energy your laptop is using right now, 500 kWh is what your home consumed last month. MW is like the instant flow from your faucet; GWh is like the water collected in the sink. 

Anyway, when you convert from megawatts (MW) of generation capacity to GWh (electricity supplied) a slightly different picture emerges. Gas gets bigger, solar gets infinitesimally small. This is because solar's capacity factor - the ratio of actual power provided to theoretical power provided under ideal circumstances - is about 15%*. Natural gas has historically run at a capacity factor of 58% the last few years, coal about 74%. 

* (in fact, in 2021-22 solar's capacity factor was just 2.5% according to SaskPower's annual report, but I've fudged it up to 15% to align with typical expectations. I'm guessing there was a rounding error not in solar's favour in the report)

Note: I'm only showing 2022, 2027 and 2030 as those were the years with the most data available on the power supply mix in those years. 

We'll do one final transformation on Scenario 1, flipping the above chart into percentages:

The takeaway is the "40% renewables" narrative refers to capacity which is only available under the best possible conditions. In reality we expect the electricity supplied from renewables (based on historical data) to be more like ~32%. 

Scenario 2: 50% Renewables by Capacity 

SaskPower noted in their 2020-21 Sustainability Report that by 2030, renewables could make "up to" 50% of Saskatchewan's supply mix. Let's explore how that would look:

Assumptions and insights:

• I've assumed that most changes occur between 2027 and 2030. 
• I've also assumed a) we're still expecting 7000 MW of generating capacity in 2030, b) there's not much room for hydro growth over Scenario 1, and c) that wind/solar displaces new gas. 
• I've assumed, for lack of a published plan, that the ratio of wind-to-solar will be the same in 2030 as in 2027. 
• At first glance it looks very similar to Scenario 1, but the wind and solar bars are bigger at the end of the decade. Dispatchable sources are still safely above peak load. 

Let's transform to GWh and, based on historical utilization rates, see how much work can actually be done: 

There is a problem with this chart that may be hard to spot at first: in 2030 in Scenario 2, we're making less power than in 2030 in Scenario 1. Within the constant 7000 MW of generating capacity we have to play with, the lower capacity factors of wind and solar have lowered the total output of the grid. 

I've shown natural gas utilization on the above trend, and below I'll show what would need to happen to meet the energy output of Scenario 1: we'll need to increase utilization (capacity factor) from a historical 58% to about 64%. Unlike intermittent energy sources, increasing utilization is easy with dispatchable sources. 

Of course, this assumes that we want the same electricity supplied as in Scenario 1, and that Scenario 1 is somewhat accurate. Even if these assumptions turn out to be wrong, the point is that an increasing percentage of intermittent renewables within the same total generating capacity (i.e. 7000 MW) will deliver less electricity than a system with a higher percentage of non-intermittent sources, unless the dispatchable sources all run harder. 

Finally, let's look at Scenario 2, "50% renewables" in terms of actual renewable percentage on the basis of energy delivered/supplied:

In Scenario 1 we had about 40% renewables by capacity (1,470 MW of hydro, 1,050 MW wind, 241 MW solar) which would end up delivering 32% of electricity supplied from renewable sources. 

In Scenario 2, we add 10% renewables (+565 MW wind and +174 MW solar, replacing 739 MW of gas) to get 50% renewable by capacity. This works out to about 37% of electricity actually supplied. 

Again, it's marketing to talk about MW of capacity. I'll repeat this absurd example from a previous post: you could install 1000 MW of solar in a deep, dark mine and add that to your generation capacity, even if it's impossible for those panels to produce. 

Scenario 3: What if 50% of the electricity supplied was from renewables? 

You might be wondering what it would take to supply 50% of our electricity from renewable sources. I wondered about that too! 

We have to work backwards from the other two scenarios. We'll start with this mix of sources in 2030:

The basis of this scenario follows: I continue to assume hydro doesn't have much room to grow, so much of the growth will come from wind and solar. I used the same wind-to-solar ratio as 2027 (which was a credible projection in the 2021-22 annual report). We'll keep Boundary Dam coal, keep "Other" unchanged, and natural gas will take up the slack. 

The final assumption is that in 2030 we'll want the same total electricity supplied as in Scenario 1 and 2. This will mean we have to discard the 7000 MW generation capacity from the first two scenarios and figure out what it would take to supply ~31,300 GWh with this mix. Working backwards we get generation capacity mix: 

Takeaways and insights:

• 50% renewable energy could be supplied from 62% renewables by capacity. 
• The first major challenge is this model requires more generating capacity (8,035 MW) to supply 2030. The generation network would be overbuilt (capacity-wise) by 1,035 MW over Scenarios 1 and 2. All this extra capacity would increase costs to consumers. 
• The other major challenge is that in 2030, available dispatchable power intersects and nearly crosses below Peak Load. This means that if (or when) the highest-demand day of the year is super cold (wind turbines suffer losses in the cold and are often disabled in extreme cold) and super dark (it's nighttime), intermittent renewables like wind and solar provide no electricity, and our entire provincial dispatchable fleet (coal, gas, hydro) would have to be running at 100% capacity, with no slack in the system. That means no power plants could be down for maintenance, and they all must be completely reliable. 
• This model would require approximately 1.7 million more solar panels than in 2027 and 1,300 km^2 of land dedicated to wind farms (including turbine spacing). 

Someone on Twitter was not understanding the relationship between gas and intermittent sources. I don't know how to make it any clearer than p14 of SaskPower's 2021-22 Annual Report! 

It's unlikely the lights would actually go out if intermittent renewables failed to meet demand on that cold, calm, dark day. SaskPower could import power from neighbouring provinces and states (assuming they are also not running on razor-thin, renewables-heavy margins). That would also increase costs as imported electricity is 2-3x more expensive than electricity generated in-province (see image below).

Costs and Ownership

In Saskatchewan, 98.2% of wind power and 100% of solar power is privately owned, either by large Independent Power Producers (IPPs) or in the case of solar, customers of any size. SaskPower pays these producers to supply electricity. 

As a public utility, SaskPower is not allowed to make a profit or a loss over the long term. Private power producers, on the other hand, are all about profits. We can reason that wind power is even cheaper to operate than the above chart shows, because a fair chunk of that green bar SaskPower pays for wind electricity is profit for IPPs. 

Missing from this cost data: capital costs and OM&A (Operating, Maintenance, and Administration) costs for the grid operator (discussed a bit on a previous post here). 

These OM&A costs are important, and it's too bad SaskPower doesn't break them up by generating asset (it's one $700M chunk in the 2021-22 Annual Report) - they would help us understand the system costs it takes to support, integrate and work around intermittent renewable energy sources like wind and solar. These include backup costs (gas assets on standby for calm, dark days), balancing costs (juggling the high variability of weather-based supply with end-user demand), grid costs (all the infrastructure) and connection costs. 

In other words, for an IPP to throw up a field of wind turbines and crank out MWh (when it's windy) is easy (low cost), but it's hard (expensive) for a grid operator to back that capacity up, balance it with all the other variable sources, and build thousands of km of new transmission lines to connect that infrastructure to the grid. 

The Netherlands published a fascinating study where the feasibility of low- and zero-carbon energy sources were evaluated for their long-term energy supply mix. After accounting for system costs (backups, balancing, grid and connection costs), wind and solar's "levelized cost of electricity" (LCOE) came out as more costly than nuclear (note SMR and EPR both refer to nuclear). 

Note: VRE = variable renewable energy. I've been using the term "intermittent"

These hidden system costs of grids with high percentages of intermittent renewable sources will end up driving up costs to ratepayers. One author puts it this way: "What's more expensive? A grid with 100% controllable [dispatchable] energy, or a grid with a 100% controllable energy plus a bunch of intermittents mixed in?"

Here is a pair of charts from Goldman Sachs (source) illustrating the apparent contradiction between the "renewables are cheap" narrative, and the price of power in countries with high percentages of intermittent renewable energy. 

Conclusions / Wrap-up

• SaskPower's habit of talking about their generating assets and future renewables mix in terms of instantaneous supply (MW, and % MW) is somewhat misleading. "40% renewable" assets (Scenario 1) would end up supplying about 32% of electricity, "50% renewable" (Scenario 2) would supply about 37% of electricity. 
  
  
• In Scenarios 1 and 2 (which both seem possible), the percentage of electricity supplied by natural gas will increase by 13 to 20% over current levels. This will displace coal (unquestionably good) but may pose its own risks: gas prices are increasingly volatile, Saskatchewan is a net importer of gas so it may affect local energy security, and gas is potentially subject to future climate regulations and taxes. For more, my previous post about SaskEnergy and gas prices increasing by 22.7%.
  
  
• Scenario 3 is unlikely and highlights how dispatchable energy sources like natural gas are required for backing up intermittent sources like wind and solar, and how intermittent sources need to be overbuilt relative to dispatchable sources.  
  
  
• As I said at the top, there are likely assumptions and omissions in the data I've reviewed, and errors in my scenarios. So, this is more of a factually-grounded thought experiment than concrete fact. I am sure SaskPower has a whole team of smart people working in generation supply mix! 
  
  
• It is not easy to account for or show the costs of intermittent renewable energy sources. SaskPower's recent annual report is not granular enough to draw any firm conclusions. There is little dispute that renewables are cheap to build and operate at the plant level. 
  
  
• I am keen to see how the next few years of new energy projects will either enable Small Modular Reactors or work against the business case. By diversifying our energy supply mix with a few hundred MW of nuclear energy, Saskatchewan could improve energy security and stability of supply, and dramatically lower greenhouse gas emissions from the electrical sector. See my previous post about SMRs in SK and what we know about SMRs today. 

If you made it this far, thanks for reading! As always, I am keen to know about any corrections or constructive feedback. 

Spreadsheet with calculations, sources, and assumptions