Recent Posts

Monday, February 22, 2021

Electricity Options

The recent power failures in Texas led me to think about a topic I had not thought much about since my electrical engineering undergraduate days -- electricity transmission and generation. In this article, I discuss big picture options for electricity, both from a technology and electricity markets/regulation point of view.

Although I have some relevant knowledge, it should be noted that I am not an expert on some these topics. For example, although I worked for a research centre that designed electrical generators and motors, my training was largely aimed at someone who was a customer of a utility. Only the relatively small cohort of engineers who went to work for utilities needed to worry about how utilities connected to each other. The other issue is that my technical knowledge is out of date.  Power electronics lagged behind electronics applications elsewhere, for a very good reason: high power levels melt silicon. This meant that power circuit technology was relatively stable when compared to other parts of electrical engineering (although DC transmission was developing). However, technology made impressive leaps since the early 1990s (when I graduated); a lot of what I am discussing herein would have only existed in research labs then.

I will first discuss the technology in broad terms, then swing towards regulatory aspects.

Decentralised and Intermittent Power

The rise of intermittent power -- solar and wind -- has been widely discussed. They may allow wider distribution of generation, putting the source of power closer to where the demand is. This decentralisation of generation helps  reduce the need for long-distance transmission of power.

(My views on power generation are shaped by location. In some provinces in Canada, we have ample hydro-power capacity, and so we do not face the same economic issues of countries that have to rely on thermal (natural gas, oil, coal) or nuclear generation as traditional sources. Provincial-owned crown corporations were dominant, although privatisation has happened. The centralisation of hydro is extreme, so it made no sense to create small generating firms using thermal plants. Canadian utilities sell into the American market, but we are otherwise bystanders to the American fights over the regulatory model.)

Dcentralisation reduces the strain on the grid. However, this is currently offset by the issue of electrical storage.


The weakness of wind and solar is the fact that it is intermittent. Meanwhile, electrical circuits store very little energy. (You can charge up capacitors and inductors enough to kill yourself with them, but the stored energy is small relative to the flow of power. Attempting to store a lot of energy in a capacitor would eventually result in the material exploding under the force of electrical repulsion.)

Solar and wind had to fight two battles to be economic: they needed to increase efficiency, and the ability to store energy was needed. Efficiency has made great leaps forwards, but the problem of storage remains. I do not know enough about the area to make any predictions.

However, even if storage is an issue, renewable energy can replace thermal sources. Utilities just need enough backup traditional power in case predicted wind/solar production drops. (Although there have been predictable attempts by some people to blame the Texas situation on wind/solar, initial accounts suggest that ERCOT would have had enough traditional power if the equipment had been cold-proofed.)

"Non-Electrical" Storage

I read "The Hydrogen Economy" by Jeremy Rifkin some time ago. I am not up to date on developments with hydrogen, but the concept does present one potential scenario that is radically different than the present.

The idea is that electrical power in excess of current grid demands is used to split water to create hydrogen. Hydrogen could then be used in transport, stored more easily, and shipped from one region to another. If we put aside the specifics of using hydrogen, the core idea is that we can store electrical energy in a fashion that can be used outside the grid.

This is different than "grid only" storage. If stored energy can only be used on the grid, it makes no sense to build capacity with expected output that is much greater than average expected consumption -- no point in storing it beyond what is needed to smooth bumps. Conversely, if we can dump the electricity into something like hydrogen, you can ramp up capacity and dump everything into "the hydrogen economy." This means that you end up with a huge excess of generation capacity over grid demand, and any hiccups are dealt with by shipping some hydrogen around. The resulting system is obviously more resilient.

It is possible that electric vehicle battery storage can fulfil this role. Shipping is more difficult, but it might allow excess electricity to absorbed by displacing gasoline/diesel use.

Smart Grids

Some genre of pop science has led to snarky comments about how power engineers were dinosaurs using fragile centralised technologies. This sentiment can be translated to: why did those rube power engineers not use technologies that did not exist?

There is a similar bit of hype about "smart grids." From my cynical perspective, this is just marketing. In the real world, power engineers are doing exactly what electrical engineers have been doing since the end of World War II: applying new technology to existing systems.

Advances in power electronics allows for connection to intermittent power sources, more complex connection patterns, and more advanced billing (which takes us to the economics discussion).


Although many people want to dump a lot of resources into upgrading the grid, I am unconvinced that attempting to force faster adoption of new technologies is the best macro policy approach. Sure, research funding and industrial policy would allow more options in the future, but in the short term, the equipment needs to be built by existing factories, and installed by a trained labour pool. 

To what extent those technologies are imported, one might start conversations with post-Keynesians about the external constraint. A small country might be able to snap up a lot of technology relative to its needs, but large economies will hit supply constraints quicker.

Alternatively, one could dial back to the 1970s, and think hard about energy conservation. The beauty of conservation is that is the policy of using less resources -- which can always be done without having to import specialised equipment. Furthermore, there are a great many ways to achieve it, creating many places to employ workers with different skill sets. For example, installing insulation requires much less specialised training than dealing with high voltage power lines.

Since conservation might be induced by economic incentives, this leads into the discussion of the economic side of electricity.

Consumers: Price Signals are Here to Stay

Post-Keynesians (and the left more generally) are not fans of arguments based on price signals. However, from the perspective of energy consumers, price signals are probably here to stay -- at least for industrial users of electricity. 

(As an aside, when I write consumers here, I mean all entities that consume electricity. This includes both industrial and household users of electricity; I write retail when I want to distinguish households.)

Barring some miracle (e.g., proponents of fusion power used to argue that electricity would be too cheap to meter) or a switch to a hydrogen economy where massive electrical capacity exists relative to demand, peak demand in the absence of consumption controls would always overwhelm capacity. There needs to be a means of forcing consumers of electricity to turn down their demand.

On the commercial side, this is well developed. Large buildings have backup generators, and industrial users can be induced to move some processes to off peak hours. (This is illustrated in more detail in the Daniel Breslau paper I will discuss later.) Putting prices to extremely high levels will eventually induce industrial users to cut optional consumption.

The alternative to using price is having the utility arbitrarily shutting off power. This could very easily cause discontent, and does not make a great deal of economic or engineering sense. To the extent the logic for cutting power is opaque, it leads to suspicions of skulduggery.

However, as events in Texas showed, it is crazy and cruel to expose retail customers to the same level of price volatility. That said, retail demand management can use price signals. My house has dual energy heating: an electrical heat pump, and a natural gas furnace. Once the temperature outside drops to -12 C, a LED switches from green to red. When it is green, my power cost is below the regular rate. But when it turns red, my heating switches to natural gas automatically -- and any electrical usage is at a multiple of the regular rate. You do not use a dryer, tell the kids to stop playing high performance video games, etc. It's an easily understood system, and unlikely to generate horror stories.

Oh Dear, Bitcoin

The rise of Bitcoin means that any system that offers cheap power at a fixed rate is likely to be used up by the pointless solution of cryptographic problems. Unless the activity is convincingly banned, any consumer-friendly electrical billing reform will end in disaster.

Generation Side: More Interesting

I highly recommend the article "Redistributing Agency: The Control Roots of Spot Pricing of Electricity" by Daniel Breslau. (Article link, non-paywalled link.) My comments here are based on my reading of the article, but the wording here is my own (as are any errors made).

(Breslau refers to control engineering throughout the text, but I am not greatly familiar with this topic despite a background in control engineering. This is because this work was in the domain of power systems, albeit the academics he discussed applied ideas and mathematics from optimal control to power systems. The mathematical structure of power systems is quite distinct from the models of mechanical and chemical plants that was my area of interest. I can only recall sitting through a few presentations by power systems engineers during my career in academia.)

As described by Breslau, the initial wave of power system development was characterised by utilities building large amounts of excess capacity in order to cope with the electrification of society. This excess capacity meant that meeting peak demand was not an issue. However, once the pace of electrification slowed, utilities had to adapt and only build capacity to meet near-run demand. (For hydro-power projects, long lead times meant that many Canadian utilities would build excess capacity and sell it to the United States.)

Since engineering margins were tighter, more thought had to go into how to schedule capacity. It is unsurprising that the engineers turned this into an optimisation problem, using optimal control theory (which so fascinates neoclassical economists).

Solving the optimisation problem was difficult, and the equipment was not nimble enough to react too quickly. This forced planners to work on a day ahead basis: forecast demand hourly, and then schedule which generators to run to meet this capacity.

The complexity came from the physics of the situation: transmission lines lose power (dissipate power to unusable forms if you want to be pedantic). It may be more efficient (in the engineering sense) to run an inefficient generator closer to expected demand than an efficient one further away. Hence, the optimal solution can vary widely depending upon the pattern of demand and supply.

Also, the problem arose of managing peak demand in a world of less excess baseline capacity. How to allocate electrical shutdowns among customers?

Homeostatic Control

The theoretical approach pioneered by a group of engineers and economists led by Fred Schweppe at MIT was to blend economic signals into the optimisation problem.

The optimal control problem led to a solution by Lagrange multipliers. These led to solution sensitivities that which in neoclassical economics are interpreted as prices. The group labelled this "homeostatic control" since human reactions were embedded into the mathematical system. (Having a cool name like this attached to pop science was a prominent feature of hippie-era academia.)

On the consumer side, the price interpretation makes a good deal of sense (as discussed above). You want to knock industrial usage at peak to meet available supply. The trickier part was the generation side. You needed to set up generation to meet projected demand (a day ahead). (Fine tuning during the day would be based on monitoring the status of the grid.)

You could set up the optimisation problem a number of ways. Unsurprisingly, each different specification leads to a different solution, and hence prices. The economic interpretation runs into two problems.
  1. If we can have different optimal solutions, we cannot visualise any mathematical solution as "the best" as normally conceived. You have to decide how to shape the problem, and the mathematics spits out a solution, with implied prices.
  2. There is nothing to guarantee that a traded market prices will correspond to optimal technical solutions. 

Pricing: Built Into the Engineering, But Structure Can Vary

Even if one is a fan of administered solutions, the technocrats managing the grid will tell you: you still have something akin to prices built into the problem, on both the generating and consuming side. (To repeat, you want to largely insulate retail from the massive price volatility industrial users need to face.)

The only point of potential divergence is ownership structure.
  • A section of the grid has to be managed by a single entity, with dictatorial powers. For example, the final say on what generators connect to the grid, and when to take down parts of the grid to protect the transmission equipment.
  • If generators are owned by the entity that controls the grid, the prices are technically driven internal administrative concepts.
  • If the generators are owned by other entities, they need to be paid for power, and the prices probably need to align with the solution of some optimisation problem. Prices paid need to match the costs of changing inputs, particularly for natural gas which features volatile prices and relatively small inventories at generators.
  • Revenues from consumers are split between the grid operator and the generator operators. Although there is smoothing due to caps on customers like households, consumer peak pricing surcharges would naturally tend to flow to generator operators. This suggests that there will be a de facto link between the generator side prices and consumer side load-sensitive prices, even though the two sets of prices can be conceptually de-linked.
  • Even if the local section of the grid is vertically integrated, the grid operator needs to deal with connected entities. Hydro-Quebec is not going to ship power to the United States on a charitable basis.

Which Way For the United States?

I am obviously not an expert on the power grid within the United States, but I would argue that there are no easy ways to get better outcomes. Although progressives will bristle at the notion that "there is no alternative," alternatives might only affect aspects of the electrical system.
  • The situation in Texas needs a proper engineering inquest. Despite red herrings floating around, the core problem was the lack of winterisation of equipment. Mandating winterisation is a technical engineering issue, and the question of which body should have mandated it is not necessarily obvious to me. Texas famously avoided being regulated by FERC, but FERC may not be the only regulator in play.
  • If Texas wants to remain a separate grid, the locals need to decide how to address reliability concerns.
  • An easy solution is for the Federal Government to throw money at problems in the form of tax incentives or subsidies. This can nudge utilities into investing more, but if engineers on the ground think an investment is a waste of money, it will not happen.
  • Technical regulation can be improved. However, any number of things can go wrong with a grid, and so this may always be reactive.
  • Market structures can be changed to create more incentive to build spare capacity. This would raise the cost of electricity, which is always politically awkward. (Texas lacked a capacity market, where operators were paid to have extra capacity ready for use. The problem is that it is a bit awkward to structure to pay entities to do nothing, and the payments need to be aligned with what consumers pay.) Meanwhile, given the diversity of systems, this is not a national policy.
  • Consumer pricing strategies need to be aligned with the wholesale electricity market somehow. There might be a number of ways an imaginative Federal government can subsidise preferred outcomes.
  • Advances in storage and wind/solar power might allow enough decentralisation to reduce grid loads to easily managed levels. This could be subsidised at the Federal level.
  • The most aggressive solution is to nationalise the electrical grid. I am not a fan of this option, but I imagine some readers are.
Nevertheless, unless the entire grid is nationalised, pricing of electricity will happen in some form of market -- even if the market is messages between vertically integrated utilities. (Even in the case of nationalisation, payments will have to be made to and from Canada and Mexico.)

Concluding Remarks

Form follows function. Although one can hope for improvements in technical management, electrical systems evolved the way they did because of technical necessity, and not just ideology.

(c) Brian Romanchuk 2021


  1. Increased storage would mitigate reliance on base load generation. We're not there yet, so coal and nuclear power plants remain necessary.

  2. re: using dams as batteries

    The Key to New York’s Green Dreams May Be Turning Quebec Into a Mega-Battery

    Quebec’s dams ship some power south across the Canadian border, supplying 15% of New England’s electricity right now. But the key to driving down emissions from both power grids lies in sending more electricity back and forth. Solar and wind facilities being built in the U.S. would help power Quebec on sunny or windy days, giving hydroelectric reservoirs time to recharge. Then, when the sun falls or the wind calms, the Canadian dams would take over.

    It’s similar to the system that has evolved between wind-rich Denmark and Norway, which boasts Europe’s biggest hydroelectric system. Denmark ships excess wind power to Norway, and that energy allows for Norway to refill its hydroelectric reservoirs. When Denmark needs more electricity, power flows the other way.

    source: The Key to New York’s Green Dreams May Be Turning Quebec Into a Mega-Battery

    1. Thanks. That’s a more efficient version of “pumped water storage”, where solar/wind is used to pump water back into a hydro reservoir. You still have generation constraints when the water flows, so it’s not as flexible as a battery, which can discharge quicker.

  3. "You still have generation constraints when the water flows, so it’s not as flexible as a battery, which can discharge quicker."

    Batteries might have high kW capacity but dams have much higher kWh capacity than batteries. They can be used to complement each other.

    Henry Rech


Note: Posts may be moderated, and there may be a considerable delay before they appear.

Although I welcome people who disagree with me, please be civil.

Please note that my spam comment filter appears to dislike long "anonymous" posts.

Note: if you want to post comments from Apple devices (iPhone, iPad), you apparently need to turn off "prevent cross-site tracking" in Safari privacy settings. (The reason presumably is that another URL handles comments, and so the user session needs to be preserved when redirected to that site. I don't like this, but this is not enough to make me switch my hosting service.)