Energy supply, not just electrical power, is what I’m considering – the goal is total replacement of fossil fuels, aside from niche applications. For some data to work with I’ll use the overall energy usage statistics for Australia. In many ways my homeland is an extreme example, due to the vastness of the Great South Land and the sparsity of human habitation. Thus we use a lot of energy for transport of all kinds – freight, fuel, water and people.

For 2017 the usage was as follows (to two figures):

Oil: 2,400 petajoules

Coal: 1,800 petajoules

Gas: 1,600 petajoules

Renewables: 400 petajoules

Total was 6,200 petajoules or 6.2 exajoules (6.2E+18 J), which is equivalent to 200 gigawatts power. That’s about 8 kW per capita.

Due to the inefficiencies of turning thermal energy into useful work, the actual electrical energy required is quite different:

Oil: 2,400 petajoules (PJ) = 600 PJ Work (average car useful work is 25% of thermal power)

Coal: 1,800 PJ = 540 PJ Work (assuming 30% thermal efficiency)

Gas: 1,600 PJ = 560 PJ Work (assuming 38% thermal efficiency. NB: A combined cycle with heat recovery has achieved 53%)

Renewables: 400 PJ = 360 PJ Work

Roughly ~2,400 PJ of non-thermal renewables could supply the 2,100 PJ of Work that Australians use.

That last efficiency assumes 95% distribution and 95% end use. Of course the intermittent nature of wind and solar mean their utilisation is sporadic – thus the need for storage. Pumped hydroelectrical energy storage is proving a cost effective energy storage approach, but it has a finite capacity in this country – and others.

Solar is the more predictable of the two chief renewables. It has a hard availability of at most 50% of the day, but in reality, thanks to the atmosphere and the seasons, the maximum is about 25%. To supply continuous power therefore, 4 times the power is needed to produce the excess you need to store. Battery efficiency is roughly 90% (charge/discharge), similar to pumped hydro. Thus 4.5 times the average power required needs to be installed to keep the storage topped up.

Power from a Solar Power Satellite is available for 8,764 hours of an 8,766 hour Julian year. Its storage requirements are a lot lower and as the eclipse season is predictable and sequential, a series of SPS can cover each other’s drop out.

For comparison we have two options:

(1) Solar Power Satellites, with minimal storage, but a large transportation cost.

(2) Renewables that require much higher installed capacity with storage to ensure continuity of supply.

If we assume the total electrical power needed by Australia is 40% of total thermally supplied energy, then the total power needed is 2.48 PJ per annum or 80 GW. Using the extreme, that means 360 GW of renewables and 280 GW* storage is needed.

Solar Power Satellites do require a substantial area for the microwave downlink, but a rectenna array doesn’t need to shade the ground – it can be situated over cropland, for example. To minimise the possible biological effects, the power density of the downlink is kept to ~240 W/m^{2}, which is much less than daylight.

For a first guess estimate of pricing let’s look at current solar power and battery pricing. The biggest Solar-with-storage project that’s in progress is Sun Cable’s Australia-Asia PowerLink, which promises 3 GW of dispatchable power (power available anytime, on demand), 14 GW installed (Variable Renewable, the peak power) and 33 GW.h storage (11 hours of dispatchable), for $23 billion. A 5,000 kilometre undersea cable will supply power to Singapore from Australia’s Northern Territory.

Thus 3 GW of PowerSat power costing $23 billion would be competitive. As a guesstimate, let’s say 1/3 the cost of the Australia-Asia PowerLink is PV’s (all 14 GW worth), 1/3 is storage (33 GW.h) and 1/3 is power cable, some 5,000 km worth. 3 GW of PowerSat power on the ground is about 5 GW in space, but designs like CASSIOPEIA use concentrators, minimising the raw photovoltaic array cost. The bulk of the mass is structure and the microwave phased-array, as well as the space-rated concentrator lenses. The other cost is the ground rectenna, to convert microwaves into mains power and the cables to the power-grid. Unlikely to be as big as the Australia-Asia PowerLink!

The final cost is the launch to GSO. We’re talking the delivery of 3,000 tonnes to GSO via Starship. Previously we determined the total number of launches required for 600 tonnes to GSO is 26 i.e. 20 Tankers and 6 Payload flights. Thus 26 x 5 = 130 launches of Payload and Tanker Starships, will orbit 3 GW of CASSIOPEIA’s. Elon Musk has stated the long term cost of a Starship launch will approach $2 million per launch. That’s $20 per kg, but that’s a long term goal. If so, then 3 GW of PowerSats will approach ~$260 million to launch. Of course we need to convert to $AUD, so that’s ~$ 350 million AUD.

Therefore if we have a launch budget of $7.7 billion, assuming launch is 1/3 the cost and the total cost matches the AA PowerLink, then that’s a budget of $59 million AUD per launch. If SpaceX achieves anything like its ambitions, then PowerSats, at least like CASSIOPEIA, will be competitive with ground-based solar utility-level power projects being pursued today. Potentially there’s a BIG saving and thus a solid motivation to pursue PowerSats.