Saturday, June 27, 2009

So, you want to buy a solar plant, part 2.

I thought I'd extend my calculations from So, you want to buy a solar plant, part 1, to make them a bit more realistic. You'll find the basic logic for this post at that location. Here's a recap:


(1) Outgoing$Monthly = TotalPeakPower (kW) * Cost/Wp * C1

(2) Incoming$Monthly = TotalPeakPower (kW) * Rate ($/kWh) * C2

(3) C1 = i / (1 - (1+ i)^-n)

Note: i = .0042, n = 300. Previously, I had set n to 240 for a 20 year loan, but in this case, I'll be setting n to 300 for a 25 year payoff period.

(4) C2 = Insolation Ratio * 1year * 365days/year * 24hours/day / 12months

Note: Insolation Ratio = .1875. Previously I had used .20 for the Solar Plant's local Insolation Ratio. The particular location that's being considered is in the area of Bangalore. I see from this Insolation Map that that area seems to be recieving between 4kWh/day/m2 and 5kWh/day/m2 in Solar Energy. For this estimation's sake, I'll pick the middle, or 4.5kWh/day/m2. Using previous calculations, this works out to an Insolation Ratio of 18.75%.


(5) Setting Outgoing$Monthly = Incoming$Monthly gives the break-even long term average energy price per kWh with respect to the cost per Watt of the Installation:

(6) Cost/kWp ($/Wp) = Rate ($/kWh) * C2/C1

Or conversely:

(7) Cost/kWp ($/Wp) * C1/C2 = Rate ($/kWh)



I want to take this one step further, and take into account degradation of the system over time, as well as various losses like Inverter / Substation losses. I'll try to be aggressive on this correction, to err towards the worse case. So, over the 25 years over which I'm considering, I'll say that the modules lose 1% per year, and Inverter losses are 8%. Averaging the degradation over 25 years gives a loss of 12.5%, plus the 8% Inverter loss, gives a total of a 20.5% loss.

I'll reflect this in my equation by going back to another equation from Part 1 (one of the roots of the (6), above):

(8) TotalPeakPower (kW) * Cost/kWp * C1 = TotalPeakPower (kW) * Rate ($/kWh) * C2 (Hours/Year)

Just prior to (6), I had canceled TotalPeakPower out of my equation. Essentially, these equations should give a rough estimate of the Levelized Cost of Energy irrespective of the total size of the Installation. Of course, there are alot of variables involved when changing scale, and these aren't reflected here. Consider this to be more along the lines of an ideal solar farm. It costs the same per Watt to add a thousand Watts of capacity, or a hundred thousand Watts.

Remember that on the lefthand side of the equation is the initial cost per Watt of the system, while on the righthand side of (8) is the income per unit energy. So, essentially, the "TotalPeakPower" on the left is the ideal TotalPeakPower, while the TotalPeakPower on the right is the effective TotalPeakPower as reflected in the Energy produced over the lifetime of the system, by which income is derived.

So, I'll throw in some subscripts.

(9) TotalPeakPowerideal * Cost/kWp * C1 = TotalPeakPowereffective * Rate ($/kWh) * C2 (Hours/Year)

In the ideal case, TotalPeakPowereffective = TotalPeakPowerideal, but in this case, there's a 20.5% loss, so:

(10) TotalPeakPowereffective = .795 * TotalPeakPowerideal

Plugging (10) back into (9) gives:

(11) TotalPeakPowerideal * Cost/kWp * C1 = TotalPeakPowerideal * .795 * Rate ($/kWh) * C2 (Hours/Year)

So, I can now still cancel out the TotalPeakPowerideal, and I'm left with a correction to the income side of the equation, which reflects system degredation losses and Inverter losses.

The final result is:

(12) Cost/kWp * C1 = (Rate ($/kWh) * C2 (Hours/Year)) * .795

Taking a case, let's say your installation is going to cost $5.25/W, or $5250/kW.

Solving (12) for Rate gives:

(13) Rate ($/kWh) = Cost/kWp * C1/.795C2

C1 = i / (1 - (1+ i)^-n) = .0058 (i = .0042, n = 300 (25 years))

C2 = .1875 * 8760 Hours/Year / 12 Months/Year = 137 Hours/Month

C1/.795C2 = .00005325

Rate ($/kWh) = $5250/kWp * .00005325 = $.28/kWh

Monday, June 15, 2009

Wondering about Intel.

The New york Times told us the other day that Taiwan Semiconductor Manufacturing Company is planning to enter the Solar Cell market.

What I want to know is, where is Intel? These guys know the Silicon Wafer as well as anyone else. Are they passing up an opportunity?

Wait, looking up "Intel Solar" brings up SpectraWatt. SpectraWatt was formed by Intel in June '08. Their plan was to build a manufacturing plant in Oregon to start deliveries in mid-2009. In January of '09, SpecraWatt halted construction of their Oregon plant, and on April 9th of this year they announced their intention to build a headquarters and manufacturing plant in New York.

Ok, so deals went sour in Oregon, I don't know the details, can't say much about that, but still, I gotta say WTF? Come on, INTC. Do you REALLY want into this market? After a year with pretty much nothing to show for it, you are now planning to manufacture 60MW by 2010, and 120MW annually within a couple of years after that. Even if the Implementation had gone flawlessly, this tiny output demonstrates a lack of vision, at the very least. SpectraWatt will have to do better than this if they want to compete in this market.

So, it looks then like Intel has actually moved to enter the Solar Market. They've just done so in a half-assed noncommital sort of way.

Of course, they've got cash. Maybe, like HP, they've discovered that the bottom line doesn't necessarily support building new manufacturing plants in the US. They could buy a heck of alot of Asian production, all in one fell swoop. There won't be too many opportunities to buy into the Asian industry on the cheap, though.

My guess then is that they're not actually idiots; they know what's up, and they're just being sneaky. It's too bad, though, that this Leading US Company is taking so long to get into the game.

Presentation - GT Solar

Here's a newish presentation by GT Solar (dated 6/5/09). Thanks to Parabequ of Yahoo for the find.

Also, here's a nice brief on the process of manufacturing Polysilicon.

It's tough stuff to make. The technology has been kept pretty carefully under wraps for years, but it's breaking out with help from the folk's at GT Solar. Note page 13 of the presentation and the estimated price per Kilogram of $25. If this works out to be true, then it will be a significant step in taking Silicon-based Solar Energy to Grid Parity and below in the mid-term.

Sunday, June 14, 2009

Innovalight - Silicon-based Thin Film.

News came up this last week on Innovalight installing "the industry's first high-throughput silicon-ink inkjet printing system."

If there's any thin film competition for wafers in the large-scale installation niche, I suspect that it must be silicon-based, simply because of the whole supply-constraint problem in the rarer materials like Tellurium, and to a lesser extent, Indium. Innovalight looks to be a good potential contender in Silicon-based Thin Film.

It doesn't look like Innovalight is giving out its cell's efficiency, and they also apparently haven't mentioned the capacity of their new production line, but they are talking about $.50/Watt in the long term (cost vs price, unknown). It'll be interesting to see how far they'll have to scale in order to approach $.50/Watt, and how much money it's going to take them to get there.

Wednesday, June 10, 2009

Global Oil Reserves Fell in 2008 on Russia, Norway, Says BP

Global Oil Reserves Fell in 2008 on Russia, Norway, Says BP.

Don't forget Peak Oil.

Take the Crash Course.

Crash Course Chapter 17a: Peak Oil



Crash Course Chapter 17b: Energy Budgeting



Crash Course Chapter 17c: Energy and the Economy

Wednesday, June 3, 2009

Thoughts on Solar Materials - Thin Film - 6/3/09

I hear the arguments on future domination of the Solar Industry by Thin Film Technologies, but I would suggest that this is far from certain.

What particularly gets me is when people talk about how they're going to come up with solar paint or some such thing, and all of our problems will be solved, just like that; like a snap of the fingers. The argument goes that there's little point in spending all the time and effort on the massive industrialization of Silicon, because some futuristic technology will simply come along to make it obsolete.

Ok, so maybe it's true that some revolution will come along that will completely change how we see Solar Energy. Maybe one day you'll be wearing Solar Clothing to charge your remote devices, and cars and homes will wear coats of Solar Paint to provide for their Energy needs.

Even if this Solar future is to be the case, though, we know that it will have to meet certain requirements, particularly in terms of Scalability / Material Availability, and Cost Efficiency.

Remember that only 1000 Watts of Power strike the surface of the earth per Square Meter on average in the middle of a clear day. That's it. No matter the wonderful technology that you develop, you can't generate more energy than what's available. To generate the incredible amounts of Energy that will be required of the future Solar niche will require a mind-dizzyingly vast array of "panels" distributed around the Planet. That's miles upon miles of glass and aluminum frames housing some kind of protected PV material, whether Wafer-based or Thin Film; or else unhoused, or lightly-housed thin films of various types, even potentially including "painted" Solar surfaces.

The main point that I want to mention at this point is on the value of Cell Longevity.

Note that when you invest in a Solar Panel, you are actually paying upfront for the entire future energy production of that panel. Normally, Solar Panels are rated in Cost per Watt Peak, or Peak Power, which is an indication of the amount of Energy that the Panel would produce at an instantaneous moment of time in ideal midday conditions. Peak Power, however, is no indication of how much Energy that the Panel will actually produce over its lifetime. Two different kinds of panels may cost the same number of Dollars per Watt, but if one lasts only half as long as the other, then ultimately it is twice as costly in terms of its total Energy Production over its lifetime.

This is where the Levelized Cost of Energy (LCOE) comes in. When you Calculate the Levelized Cost of Energy of a Solar System, you are basically determining the overall cost of the System per unit Energy over the Entire expected Life of the System. I did a rough version of this kind of calculation here. Sunpower Corp provides this nice description of the factors involved.

There are a several reasons why a Solar Cell might stop working. One reason that a cell could fail would be from molecular damage to the PV material simply by the bombardment of Solar Energy (including various cosmic rays). This could slowly degrade any kind of Solar Cell. Other types of Solar Cell may be chemically susceptible to degradation, such as today's Organic and Plastic Cells. These materials degrade quickly under common exposed conditions, and at this stage of the game, a lifespan of five years or so seems to be the cutting edge. Finally, of course, smashing a Solar Cell by way of storm debris or a baseball can destroy a panel, and dirt and grime can cover the surface and degrade its performance.

Solidly encasing the PV material in an aluminum and glass (or possibly plastic) module will, in most cases, help to protect the cells from physical damage, but that housing will certainly add to the cost of manufacturing the module. For a thin film product aiming to compete on very low manufacturing cost, this added expense is going to be a killer. In fact, during First Solar's Q1 '09 Conference Call, Jesse Pichel of PJC suggested that Glass was actually FSLR's largest cost. First Solar didn't disagree, and nobody mentioned Tellurium.

Now, if glass is actually even a significant portion of the cost per watt for a thin film, then it sets a kind of a lower limit on the potential cost to manufacture Thin Film Cells housed in glass (adjustable by efficiency). So, to some extent, the decision is whether to go for extreme affordability (or flexibility) and avoid a robust enclosure, but lower the operating lifetime of the cells; or else go for a longer lifespan, but adding significantly to the total cost of the module. First Solar, for example, is targeting a production cost of $.65 per Watt.

Though I'm certain that nanotech of various sorts will be able to make headway in durable exposed thin film cells, I can't help but think that it's going to have its limits. For comparison's sake, a tarp is made of very tough stuff, yet I've seen my share of tarps shredded by fall winds, and a tarp doesn't depend on the same kind of exacting chemical structure that a PV cell does. You can beat the crap out of a tarp, and it will still keep the rain off of your stuff. I'll be very impressed if I see a thin film material that you can roll into a ball, peat with a stick, and still use to generate electricity. I can't wait to see the infomercial.

There are numerous conclusions that I could follow with, but for now, I'm going to leave this with a simple idea for the consumer. Don't just buy solely based on Cost per Watt, or one day you're going to be led astray. Know what you're buying, and make sure that it has a solid warrantee over a time period to assure your expected payback. If you're offered a deal too good to be true on a cost per watt basis, it could simply be that the product you're buying is going to crap out long before it pays itself off.