Freedom for the megawatts
Power to the, err, people!
Almost weekly I meet someone moving to Cape Town from Joburg, and this seems like a very natural response when I look around at the growing number of potholes I need to avoid daily. Well, at least a very natural response for a middle class family1.
One of the topics that comes up regularly in these conversations is the reduced loadshedding in CT, which is clearly an awesome reason to move. The odd thing about it though, is the vague libertarian undertone with which its couched. It seems to say, "in CT we can be free of these national shackles and therefore free from loadshedding."
Every time I hear this, I think "except that won't work." This drive for an independently operating electrical grid feels instinctively to be misplaced and misguided (my brain goes "techno-economically too expensive, socio-politically infeasible.," and "the-last-10% problem is surely compounded at this scale"). But hey, where are the MODELS to tell us, amiright2.
To be fair to the City of Cape Town themselves, they seem to be much more circumspect in their ambitions and promise only to "protect" residences from Eskom loadshedding (and not to supplant Eskom entirely).
I made the rookie error of saying "energy independence for CT is a pipe dream" on a group chat, and then made an even bigger error of defending my position poorly. The upshot of all of it is I have been seriously thinking about the question for a bit since then.
So in the absence of a real model (the best way to approach this problem), can we draw any conclusions on the feasibility on energy independence for an SA city/metro?
Let's make some working assumptions briefly. Let's assume a copper plate model (power can flow unconstrained from any generation site in the country to the city), no market design constraints (the market miraculously facilitates the purchase of energy and ancillary services etc.) , no regulatory constraints (gov doesn't have anything to say!) and that the metro/city enters into contracts to buy power directly from IPPs (doesn't have to go through gov etc.)
That's as unrealistic a set of assumptions as they come, and also provides a scenario very similiar to solving for the entire country. But there are important differences even under these circumstances.
In this scenario, I think it's the nature of the changes on the demand side that actually make the biggest difference. Compared with procuring power for the entire country, your demand is now geographically concentrated (that's not the case on the supply side, given the copper plate assumption). For example, this means when the metro is experiencing a heat wave, that's a major upward shift in demand that's not smoothed out by normal weather in another part of the "demand centre."
Your demand shape is also different. That is, compared with a country scale you now have concentration in terms of demand type. So in the metro you have less agriculture and mining but more heavy industry possibly. Overall you have a much less flexible demand, and therefore the system has less "inherent" ability to absorb changes. You therefore also have less room to employ "demand side management."
All of this means that, when compared to country scale, a metro/city will have to contract larger "reserve margins" to ensure demand is always met. This naturally pushes up the cost of the system, and my instinct tells me (we need models!) that the cost increase is dramatic.
One way that would be open to metros to manage this problem is to trade power - that is, buy power when needed and sell excess power when they have it. The trouble for South African metros is that most of the other SA small towns and smaller cities will probably not be in a position to trade with. It is unlikely a smaller municipality will independently have the organisation, skills and financial ability to be a reliable trading partner to a metro. So in reality there would be a ton of demand and supply around the metros, but no real reliable way to access this. Unless of course we get the independent system operator going, but that's been stuck for how long now?
Well, that now leaves SA metros with the choice of each other (but there are not many of them) and other countries (i.e. the Southern African Power Pool), the latter of which is the group that SA already trades with at country scale. Metros trading directly with other countries is pushing the "no market design constraints" assumption a little too far I think, but hey, these are strange times. Even if that does happen though, each SA metro landing in Harare (last I checked the Power Pool had their HQ there) with PPAs and LV Celini bags in hand, is going to prove logistically and politically tricky. But at least it will spur some renewed activity in Harare's bed and breakfast industry. The thing is, it's also technically tricky, because our neighbouring countries are by themselves small demand and supply centres. So if, say, 3 metros in SA all have excess capacity (say it's mid day and it's a great day in the Northern Cape where much of the contracted power sits), there's only so much the SAPP is going to be able to buy. I mean, here is January 2023 peak demand breakdown from SAPP3:
All of which to say, the metro are going to find trading power much more difficult than the country does, which pushes up the costs of going independent.
Then there is the question of economies of scale coupled with reliability. A 2018 report (I couldn't find more recent info) put the average peak demand in Joburg (City Power) somewhere between 2.5 and 3 GW. See graph below4.
Let's assume that's still about right5. A typical solar PV project is around 150MW now (nameplate) now, and so some very, very rough calcs would see a single PV project represent around 5% of total demand? I'm going to need to hear from my grid friends but I think that we have some concentration risk here (from a reliability stand-point). The metros could of course reduce the size of each project, but that pushes up costs by losing economies of scale, or they could increase the reserve margin (ceteris paribus), but that pushes up the costs by, well, increasing the power you're buying (and quiet possibly spilling).
So all in all, going independent seems like a costly affair. And I haven't even gotten into the practicalities of our assumption (copper plate? shew).
How costly? Maybe a 10% bump on current costs? Would metro citizens pay 10% more for energy independence? How about 20%? 200%?
I hereby request each metro to do and publish some modelling to answer this. And I hereby summon the god of democracy to encourage citizens to critically engage with "independence" rhetoric.
CT is a whole other experience for a lower income family, and let's for the moment ignore the inherent segregation built into the city.
I simply have to do a whole post/rant on needing models for these questions, but for now, here's a tweet:
COVID and growth should cancel each other out? But loadshedding has led to reductions I would assume given all the behind the meter generation and storage that has come online