Sunday, August 2, 2015

A gigawatt of solar panels

To become fully self-sufficient in our energy production, we need to more than ramp up our use of solar energy. Double isn't good enough, nor is 10x. Maybe not even 100x. No, we need to start thinking in terms of gigawatts. The capacity of commercial power plants, be they nuclear or fossil fuel or hydroelectric dams, is measured in GW.

I am not here to argue climate change or the relative pros and cons of various generating mechanisms, concern over low sun angles in the winter, or the rather obvious problem of the sun not in the sky half of the day. Rather, how big an array would you really need to start thinking in terms of GW, in a metro area the size of Pittsburgh?

Start with a real product, a solar panel made by Sharp, model ND-F4Q300. Each panel can put out 300 watts. It's about 1x2 meters in size and weighs 50 pounds. I won't discuss price, as any major generating plant would cost in the billions of dollars to design and construct. If we end up spending a billion here, too, that's reasonable, but irrelevant for the current question. At least not yet.
My initial idea was to construct a large roof over an interstate highway, and mount the panels on that. Ignoring the problems inherent in building a shed of that size, assume it's possible. I am trying to visualize the scale of a structure big enough to generate 1GW of power. It does not need to be a roof over a highway, but that is the model I will use here.

Start small, let's say a kilowatt. At 300 watts a panel, that would require four panels -- 1200 watts total, building in some buffer to account for clouds and low sun angles. For purposes of estimation, this is close enough. A panel array 4m wide and 2m high would thus give us a kW. Single house residential roof installations usually consist of a few of these, some multiple of 1kW, depending on the size of the roof and the depth of the homeowner's pocket. We should have 10,000 of these, but in addition to, not in place of, the GW-scale installations, as there will certainly be tens of thousands of homes and businesses without their own installations. The power they use has to come from somewhere.

Back to our highway roof idea. A typical interstate highway is two 12-foot lanes and a 12-foot shoulder, both directions, with a 36-foot median, and 50 feet of "clear recovery zone" on either side -- space free of trees, light poles, large road signs, and supports for big solar roof structures, so you won't get killed if you fall asleep at the wheel and drift off the road. Total, 208 feet of space between roof supports. Again, not that we would actually build this thing, but how much space does that give us to mount some solar panels? 208 feet is about 63 meters. Assuming a simple trapezoidal cross section, let's say that would give us about 40 meters horizontal space on the roof to work with. Rearrange our array to be 1m wide and 40m long, that's 20 panels. So, 1m of expressway distance would let us generate (300 watts x 20 panels) 6,000 watts, or 6kW, about what's needed to power one house.

Now let's play with numbers. Ten meters of expressway distance, 40m across, is 60kW. That's roughly the size of a gas station or small suburban strip mall parking lot, or the building's roof. We have lots of those around. That's a perfectly manageable size. A city the size of Pittsburgh probably has 1,000 such locations, be they roofs of existing structures, or adjacent parking lots that could accommodate a 10x40m panel array. Do the math; that's now 60 megawatts. Together with the previously mentioned private residences, we are now only one order of magnitude removed from that GW.

Meanwhile, back at the interstate, let's lengthen that roof to 100 meters. Imagine now an airport hangar, not that you would put a solar array on one, but a building big enough to house a full-size jet is going to be on the scale of 40x100m. That size of an array would generate 600kW. That's enough juice to support a small neighborhood's summertime air conditioners, typically 3kW apiece, and changing that sunshine to go juice exactly when those 200 A/C units are running. This is roughly the size of a school building or small industrial building in a suburban office park, or its adjoining parking lot. Pittsburgh has hundreds of these types of spaces.

Not big enough yet, though. Make the array a full 1km long and 40m wide. Six megawatts. That can power a small town. If it was square, it would be 200 meters on a side. In terms of scale, think of shopping mall, hospital, or stadium parking lots. We have dozens of these in Pittsburgh. Building a solar roof over them is a lot easier to manage than building one over an interstate highway.

Let's change the multiplier here. How big an array to get to 10MW of generating capacity? At 40m wide, you would need 1,550 meters of highway roof, or a square 380m on a side, still in the scale of a stadium or regional shopping mall. A human can still stand in one spot and see it all, and moreover, nothing below it needs to change. Ross Park Mall's parking lot would do exactly what it does now, sit 80% empty 95% of the time, but require a lot less wintertime snow removal, and shoppers would not have to worry about being caught in a summer thunderstorm on their way into or out of the mall.

The next order of magnitude gets more difficult to envision. To generate 100 megawatts, our interstate would be covered for 15.5 km of its distance. For Pittsburgh, that would be as if I-279 was covered from Allegheny General Hospital all the way up to Camp Horne Road, or the Parkway East from the Squirrel Hill Tunnels to the Turnpike at Monroeville. A square would be 780 meters on a side. The Mall at Robinson takes up about that much real estate, considering both building and parking lot.

To get to that GW, though, we go up one more order of magniude. The highway roof idea is now 155 km, or roughly I-79 from Cranberry to Erie. A square is 2.5 km on a side, covering an entire village the size of Crafton. Even to me, that sounds challenging beyond comprehension.
But we don't have to do that. Add up all those other things I mentioned:
* 10,000 residential rooftops at 5kW each
* 1,000 gas stations at 50kW each
* 100 schools and office parks at 500kW each
* 50 big parking lots at malls, stadiums, hospitals at 1MW each
* 10 entire malls or similar huge, coverable structures at 10MW each

Each of them is generating a piece of the puzzle: 60 MW here, 100 MW there, it adds up. Rough numbers, the above is at about 300 MW, not quite 1GW but getting closer in magnitude. We could be building them all, bit by bit, but at orders of magnitude greater speed. We're not. What we're doing is incremental, at best, and at worst, we are making it difficult for ourselves. That residential solar tax credit expires 31 Oct 2015.

Here is where things get interesting, though. We've been here before, three times, at least. About 1830, enough little demonstration projects for steam locomotives and railroad trackage had been constructed to allow some visionaries to dream, and put those dreams into action. Hardly a generation later, we had built railroads to connect every city, and were well on our way to connect oceans. A generation after that, that was where all the money was -- railroads, and the steelmaking to support it all. The second and third biggies happened at once -- automobiles and electric power, both just after 1900. The amount of infrastructure to be built, and the support industries that grew up to feed them, and be fed by them, drove our economy like gangbusters for 100 years. Can we please learn from our own past success?

Getting to GW-scale power generation will not be easy. Putting roofs over entire shopping malls will require lots of everything -- engineering, construction, manufacturing, legal. But we can do this. Every major metropolitan area in the country will face a similar need for the same reason. Pittsburgh will need its own GW generator, so will St. Louis, Salt Lake City, Pensacola, etc. It might be easier to generate farther south, but it's also warmer, so more constant draw from A/C systems, too.

However, I have understated the need. Since the sun doesn't shine at night, we really need to double what I stated above. We need the first GW to run the A/C in the daytime and a second GW to run a set of pumps to refill the reservoirs to run the turbines to generate the juice to run the lights at night. It's all a mere matter of how many GW capacity can we build, and how soon?

We will need lots more electricity than we do now, as car fuel moves from fossil to electric. Recharging your Tesla while you are at work is a great idea, but the juice has to come from somewhere. May as well be solar.


  1. It's difficult to imagine how big a solar collector would have to be to generate a gigawatt of electric power, but in this blog post, I will try.

  2. Some idea of cost from this article: Roughly $5/watt. So, five billion dollars. For what a gigawatt nuke costs, that's reasonable, particularly when you consider the power starts coming online with each site installed, not waiting five years before anything happens.

  3. Well, whaddyaknow... South Korea has solar panels along a 20- mile highway.