Solar power in space: Vanguard 1

Today is an interesting anniversary: the sixtieth anniversary of solar power in space. On 17 March 1958, the American satellite Vanguard 1 entered orbit, becoming the fourth satellite ever to do so, and the first to use solar power. Vanguard 1 was small enough to be held by a person in one hand – 1.5 kg (3 lbs) in mass and 16 cm (6”) in diameter, or 76 cm (30”) wide including the antenna aerials. Altogether, Vanguard 1 had 6 silicon solar cells which generated about 1 watt in total. For comparison, the power produced by a typical rooftop solar PV system in 2018 is several thousand times greater.

Why did the first satellites use electrical power? Primarily, to send radio signals back to Earth. The first satellite ever to reach orbit, the Soviet satellite Sputnik 1, broadcast radio signals at 20 MHz and 40 MHz which could be detected on Earth, even by amateur radio users. The radio signals provided additional proof that the satellite was really in orbit around the Earth (it could also be seen with large telescopes), and allowed observers to measure its position and trajectory. The power to provide this radio signal came from batteries, which allowed it to broadcast radio signals for 21 days before the batteries went flat.

This radio communication was one-way only. Early satellites could not receive communications from Earth, and were essentially a battery, radio transmitter, and an antenna. The second and third satellites, the Soviet Sputnik 2 and US Explorer 1 also used batteries to power radio signals, which lasted somewhat longer (Explorer 1 had a lower power mode that lasted 105 days). Vanguard 1 had two radio transmitters. One transmitter (at 108 MHz) was powered by a battery, which lasted for 20 days. The second transmitter (at a slightly higher radio frequency, 108.3 MHz) was powered by the (as yet unproven in space) solar cells, and that broadcast for over six years, or over 23,000 orbits of the earth, before going silent.

Vanguard 1 schematic (Image credit: NASA)

Vanguard trajectory (Image credit: NASA)

Satellites and spaceflight were one the first uses of solar power in which – environmental reasons aside – it was clearly cheaper, more reliable, and more practical than other sources of electrical power. Generation of electricity on Earth, then and today, is most commonly done with a combustion engine. On Earth, oxygen is in the air everywhere and fuel can usually be found, whether it is wood, coal, natural gas, or some other fuel. Photovoltaic cells at that time were incredibly expensive, and their intermittent output was difficult to manage since rechargeable batteries were also very expensive.

In space, solar power had clear advantages. Combustion in space requires that both fuel and oxygen (or another oxidant) be brought up from the Earth in tanks. For example, the Saturn V rocket engines that sent humans to the moon had tanks of liquid hydrogen and liquid oxygen. (Those engines were obviously used to make the rocket move, not to generate electricity). Non-rechargeable batteries could not provide power for long without a lot of “chemical fuel”. In space, above the clouds, solar power was more or less constant the whole time that the panels were not behind the Earth. No fuel needed to be carried up. Satellites typically orbited the earth every few hours, so that when rechargeable batteries did start to be used with solar panels, they were small compared to those needed to last the entire long night on earth. What happened to Vanguard 1? Although its radio no longer transmits, Vanguard 1 continues to orbit the Earth today, at an altitude between 600 and 4,000 km, going around the Earth approximately every 2 hours. It is now the oldest object made by humans to still be in space. Sputnik 1 and Sputnik 2 reentered the atmosphere after around one year after launch, and Explorer 1 reentered in 1970. Vanguard 1 is in a very stable orbit, and is predicted by NASA to remain in orbit for another 180 years. Several websites track its location (Satflare, n2yo). You can see where Vanguard 1 is right now on this map provided by Satflare:

3D globe current position of Vanguard 1, by Satflare

As a final note, in researching this article I came across some amazing audio recordings of Vanguard 1 transmissions made by amateur radio enthusiast Roy Welch in 1958 and 59. (The signals have been modified to bring them into audible frequencies). Roy Welch wrote that he believed that the variations in frequency were due to the satellite spinning around, varying the light exposure to the solar cells. The one from 1959 I find quite eerie:

Vanguard 1 signal recorded by Roy Welch (link).

It is quite something to listen to the recording and imagine the tiny Vanguard satellite, moving high above the Earth, broadcasting its radio signal, and slowing turning around in the sunlight.

Update: this article is also published on Medium:

Electricity demand response at the Supreme Court (Demand response part 2)

In part 1 of this article, I wrote about the importance of demand response in the electricity industry – i.e., why power companies might choose to pay customers to use less electricity at certain times. Here, I discuss what the right price for demand response would be, and a recent US Supreme Court case over the fundamental economic principles of that price – including the question of whether reducing consumption was equivalent to generation.

When companies offer demand response programs, the offer needs to be simple, compelling and easy to understand for customers. Demand response is typically used only a few times per year, so it is not worth customers spending the time and effort to learn a complex price scheme. You might also ask, what is the point of demand response as a separate scheme? We already have a wholesale market. If the price gets very high, then people will reduce their consumption. There is more that can be said on this, but one point of separate demand response schemes is to target customers who are not able, or not interested, to respond to wholesale market prices in real time. So demand response schemes are offered in different ways that make sense for different customers. In part 1, I described a scheme to pay people cash up front in return giving the distributor Energex the ability to remotely turn down their air conditioner at certain times. Another example will be trialled by the retailer Powershop this summer, supported by The Australian Energy Market Operator (AEMO) and Australian Renewable Energy Agency (ARENA). Powershop will use a smart phone app to tell people to reduce their demand at certain times, with the reward being a certain about of free electricity at other times.

With these indirect schemes, it is often difficult to calculate the exact price in $/MWh being offered to customers. That’s probably fine if both buyer and seller are happy with the deal. But is there are a “correct” price? Given that demand response is something of an alternative to the wholesale market, the correct price probably has to do with the wholesale market price, right?

Let us give a specific example. If a person normally pays $50/MWh for their electricity use, and prices rise to $500/MWh … how much should they be paid if they use less electricity than they normally would?

  • $500/MWh for each MWh of reduced demand (the current price), or
  • $450/MWh (the current price minus the customer’s usual retail price)?

This was argued in a 2016 Supreme court case, where the two sides disagreed on the fundamental economic theory of the question. The case was significant because this was not an example of a buyer and seller mutually agreeing on price, but an example of when the government was forcing power system operators to pay demand response providers a particular price, even if the operators did not agree with that price.

In the case, FERC (the US regulator) had created a demand response program that required energy market operators to pay the current price ($500/MWh in the above example). The EPSA (a generator industry group) was opposed to the demand response scheme, and argued that:

  1. In the EPSA’s view, FERC did not have the legal authority to create a demand response scheme, and
  2. That if FERC did have the authority, then the compensation should be the current price minus the normal price for the customer ($450/MWh in the above example).

What I think was amazing about the case was that all these famous economic professors then wrote submissions to the court, arguing the economic theory of what the price should be. Harvard economics professor Bill Hogan contributed a “friend of the court” brief in support of the EPSA, arguing that FERC was forcing grid operators to overpay for demand response by, in the example above, forcing them to pay $500/MWh.

Here is an extract from Hogan’s brief:

“To offer an analogy, consider a manufacturer that produces an automobile it can sell to a dealer for $20,000; the dealer has agreed to then sell the automobile to a customer at cost ($20,000), but cars are in high demand and another customer wants to buy the car for $30,000. No one would say that the first customer should be paid $30,000 for not buying the car, just because another customer wants it or cars are in short supply. If one customer has a right to buy the car at $20,000, while another is willing to pay $30,000–and lack of supply means that both cannot purchase cars–the dealer could, in theory, sell the car to the second customer and give the first customer the $10,000 difference between the market price and the price at which she has the right to purchase. That would allocate the car to the customer who values it more, while giving the first customer an incentive to allow the second customer to have it. We would never, however, say that the dealer must: (1) pay the manufacturer $30,000; (2) pay the first customer $30,000 (the car’s current value) for not buying the car; and (3) sell the car at $30,000 (again its current value) for a loss. But that is what FERC effectively has done: It provides the first customer with a windfall while requiring [grid operators] to pay twice (to the electricity producer and the non-buyer) for a unit of electricity that they may only sell once for less than the total price paid.”

However, in an earlier affidavit, Professor Kahn of Cornell University argued (and FERC agreed), that reducing demand was equivalent to increasing generation, and should receive the same compensation, meaning that the $500/MWh would be the correct price above. From Kahn’s affidavit:

“Demand response is in all essential respects economically equivalent to supply response; and that economic efficiency requires, that it should be rewarded with the same [price] that clears the market. Since DR is actually–and not merely metaphorically–equivalent to supply response, economic efficiency requires that it be regarded and rewarded, equivalently, as a resource proffered to system operators, and be treated equivalently to generation in competitive power markets.”

What do you think? It is a subtle difference. In the above example, the $450/MWh case is equivalent to the customer buying electricity at their regular price of $50/MWh, and immediately selling it to the market at the market price of $500/MWh, for a profit of $450/MWh, instead of using the electricity themselves.

In the end the Supreme court did not endorse any particular price method. By a 6–2 verdict, the court decided that:

  1. FERC did have the legal authority to run a demand response scheme, and
  2. That FERC also had the authority to set the price in that scheme, and therefore (even though both sides had made good points) FERC’s price should be used.

What do I think? I admit that I am not an expert in these matters – nothing like the experts quoted above – but for what is worth, my view is that Hogan’s argument is correct, and that $450/MWh would be the correct price. From my perspective, demand response by a customer is equivalent to selling their right to use electricity to someone else. To do that, they need to have the right to use electricity, and if they have to pay $50/MWh to obtain that right (and hence receive $450/MWh overall), then that is what they need to do.

However, I acknowledge that there are other benefits to demand response schemes than avoiding wholesale market costs. For example, demand response schemes at times of high network demand also reduce the costs of “poles and wires”, and that the financial benefits of reducing network costs could even exceed those of wholesale costs. (Note that in the air-conditioner example above, the scheme is offered by the distributor Energex (a network operator) rather than the wholesale market operator AEMO). Network costs are recovered very inefficiently in general, mostly through simple charges. A demand response scheme with non-ideal prices could still benefit the public overall. Another submission to the court, from Charles Kolstad, an economist at Stanford University, argued in favour of FERC because the public was better off with demand response than without it, even if the non-ideal FERC method was used. Note that there were also implementation challenges with the EPSA method that I haven’t described, such as the need for the market operator to know the normal price paid the customer. (There are other details which I have omitted, and if you’re interested I recommend reading the references below).


Supreme court dockets 14-840 and 14-841

Hogan brief to the court

FERC quotation of Alfred Kahn

Kolstad brief to the court

Supreme Court decision (January 25, 2016)

Scientific American article on the case

Paying people to use less electricity (Demand response part 1)

In the electricity industry, where I work, it is important to keep supply (the energy being provided by the generators) balanced with demand (the energy being used by customers) at all times. If this is not done, the entire system will quickly collapse to and fail to what is called “system black” or “blackout”. This is different to the market for almost every other product or service. Consider, for example, what happens when Apple creates a new model of iPhone, but does not make enough so that everyone who wants to buy one on the day they are released is able to do so. What happens in this case? Well, all the iPhones that are made will be bought by someone, and those people will get to use them. Everyone else who wants to buy one will either have to wait until more phones are made, or buy another kind of phone. If the phone market were similar to the electricity market, an analogous situation would be that if even one person went to the store and could not purchase a new iPhone, then all other iPhones stopped working within a few seconds.

For electricity grid operators, this fact makes it hard to run the power system securely when electricity demand is very high (which may happen only a few times per year, typically on hot days). There is nothing to stop people from connecting additional appliances to the grid. Even though wholesale prices rise very high at these times (over 300 times the average price), most people are either unaware of such prices or any not exposed to them (i.e. the price paid by their retailer rises but the customers themselves have a fixed price). For many kinds of loads, the demand rises automatically without any person even being involved. For example, many air conditioners are trying to cool a building down to a set temperature. If the outside temperature increases, the air conditioners automatically start working harder (therefore using more electricity) in response.

One way that this problem is managed is called demand response, which is essentially paying customers to use less electricity at certain times than they normally would. Sometimes this is done by paying customers directly. Sometimes there is simply an appeal to customers, “Let’s all pitch in and use less power tomorrow afternoon, it will help keep the lights on, and we’ll all share the benefits and reduced costs later.” Here is an example from 2017 in NSW.

When demand response is financially compensated, it is often paid for in indirect or simple ways. As with rooftop PV under the renewable energy target, demand response programs for the masses often pay the entire subsidy upfront in cash, at the time of purchase. For many customers, this is easier to understand, and more compelling, and it means that you don’t have to manage an ongoing financial relationship to pay thousands of people small amounts of money on a regular basis. For example, in Queensland, there is an interesting program called PeakSmart, which is managed by the distributor Energex. When you buy a new air conditioner at the shop, if you opt in to the PeakSmart program, you get up to $400 cash back which helps make the system cheaper. The PeakSmart system allows the distributor to remotely turn down the air conditioner to about half the output. They do this a few times per year, for an hour or so each time. From their annual report:

“We activated our PeakSmart airconditioner technology on 1 and 2 February 2016. On these two days South East Queensland experienced 40 degree temperatures. More than 50,000 air-conditioners were signalled to reduce their demand by approximately 25 per cent between 4.30pm and 5.30pm. More than 25,000 air-conditioners were active at the time and reduced peak demand on our network by 11.2 MW on 1 February and 16.4MW on 2 February. These load reductions are the equivalent of more than 7,200 homes on our network. Surveys completed with participants after the event indicated no impact on comfort.”

From what I can tell, this is a successful and well run demand response program, and is well designed for its target customers.

In the next post, I’ll discuss what is the optimal price to pay for demand response, which is a dispute that went all the way to the US Supreme Court in 2016, with a disagreement about the fundamental economic principles that should be used.