Utilities and investors have had very little help in making technology and financial decisions when developing new solar power generation systems.

Every choice has risks and uncertainties and all have financial implications. Do you build your new solar installation using wafers of crystalline silicon or thin film cadmium telluride? Do you start now or wait another year or two for a more favorable discount rate? Do you locate it outside of Boston or Seattle?

To help model costs and their uncertainties, Argonne National Laboratory and analyst firm Gartner, Inc. have developed a methodology for evaluating the lifetime cost, cost uncertainties, and associated risks of building and operating a commercial-scale solar power generation system.

The researchers believe their approach can give utilities and investors a better understanding of a system’s total costs and help them identify areas where costs and risks can be minimized.

Today, the common metric to assess a solar facility is the dollar per watt. The price of the panels, installation, wiring, subsystems, and other components are totaled up and divided by the maximum total output of the system. Argonne and Gartner claim a better metric is the levelized cost of energy (LCOE), which is derived by dividing the lifetime cost of the system by the lifetime energy produced.

LCOE can be thought of as the price at which energy must be sold to break even over the system’s lifetime. It is measured in center per kilowatt-hour and takes into account the project costs, operating costs. LCOE calculations also can and should factor in other parameters like the amount of sunlight in a given location, conversion efficiency of the solar technology, solar panel degradation rate, and financial considerations including the discount rate and taxes.

However, while LCOE has been around for many years, it has been misused when applied to solar. “In typical LCOE projections for solar energy, many assumptions are swept under the rug, and we wanted to make a small step toward lifting up that rug and showing how you can truly get a handle on those assumptions to develop a more accurate picture of the potential costs,” said one of the researchers.

For example, many of parameters are characterized by a single number, rather than being represented by a more realistic range of values. And in many cases, the true value of a parameter is simply not known, so a rough estimate is made assigning that parameter a single value. In fact, there is great uncertainty in many of the parameters used in typical LCOE calculations. By using a single number for any of the parameters (the degradation rate, for example), the calculated LCOE does not capture or reflect that uncertainty or the potential financial risks of a particular project.

The Argonne and Gartner model tries to address these variations and uncertainties by factoring in the probability associated with each parameter value instead of a single value. For example, the researchers studied the 30 year history of solar insolation for Boston, Chicago, and Sacramento and created a probability distribution for each city.

They also developed probability distributions for degradation rates, discount rates, and operational costs. Essentially, these distributions specify a range of values for each parameter and the probability that each value will occur. (The researchers noted that the distributions may not be the best; people should focus on the methodology, not the specific values used in their model.)

Having these distributions, the LCOE is calculated by picking values from each distribution randomly based on the probabilities. A simulation might produce an average LCOE of $.15 per kilowatt-hour, with a 90 percent chance that the rate could be between $.10 and $.18. An investor might not like this large a range in possible prices. The beauty of this model is that the investor can now look at the sources of the price volatility and look for ways to reduce financial risks.

For instance, the choice of a lower cost panel certainly cuts the initial project cost, but a more expensive product might offer a better degradation rate. The discount rate might be offer great uncertainties, but a utility might seek other funding or lock in a slightly higher fixed rate early on to reduce the chance of greater variation over time.