Article - Greenbyte

A wind perspective on solar

Solar panel rows.

With many of our customers moving into solar we are now expanding our product portfolio into solar photovoltaics with the launch of Bright. Last week I spent two days in San Francisco at the Solar Asset Management conference to learn more about asset management of solar photovoltaics and what it means to our customers and products from a data perspective.

With many of our customers moving into solar we are now expanding our product portfolio into solar photovoltaics with the launch of Bright.

Last week I spent two days in San Francisco at the Solar Asset Management conference to learn more about asset management of solar photovoltaics and what it means to our customers and products from a data perspective.

This post is a summary of our impression on the solar market with a wind perspective. The post goes through market segmentation, asset management and O&M, solar irradiance, availability and performance.

The purpose of the post is to share high level insights primarily with people who have a wind background.

Market Segmentation

The solar market is largely divided into four segments that are dependent on site or plant size (site and plant are used interchangeably in the same sense that wind farm is used in wind speak).

  • Residential:< 20 kW
  • Commercial: 20 kW - 100 kW
  • Industry: 1 - 5 MW
  • Utility: 5 MW+

Focus of the conference was USA. The US market is poised for growth. WOW!

Asset Management and O&M

The relationship between owners of solar PV sites for the segments commercial, industry and utility with asset managers is very similar to the approach taken for asset management in the wind industry. Utilities take care of their own asset management and investors outsource asset management to third parties.

O&M in the solar market is completely different from the wind market. In the wind market turbine manufacturers pretty much dominate the O&M market for time being. Especially on sites commissioned within the last couple of years. In the solar space there are no strong OEMs that take on O&M. Instead it seems EPC (Engineering, Procurement, Construction) contractors or specialized service companies are the ones taking on O&M. During the solar asset management conference a lot of time was spent talking about the lines between asset management and O&M becoming increasingly blurred. It will be interesting to see how this plays out in the US market that is set to grow from 20 GWs today to close to 100 GWs by 2020.

In terms of cost for asset management and O&M there seems to be a wide consensus that it should be between USD 15 – 20 / kW / year for large PV sites. Asset management in itself represents approximately 30% of the cost so in the range of USD 5 – 9 / kW / year. If we apply this figure to a 2 MW wind turbine it comes out to USD 30,000-40,000 / WTG / year for asset management and O&M where asset management represents USD 9,000 – 12,000 / WTG / year. So, the cost for asset management is very similar to wind and the cost for O&M is substantially lower on a per kW basis.

Approximate costs of asset management and O&M for solar and wind per year.

SolarWind
Asset ManagementUSD 5-6 kWUSD 5 kW
O&MUSD 10-14 kWUSD 20 kW
SumUSD 15-20 kWUSD 25 kW

Now, to the stuff we find most interesting for Breeze and Bright  Solar Irradiation, Availability and Performance.

Solar Irradiation

The resource that is converted into electricity in solar photovoltaics is solar irradiation. This is the equivalent to wind for wind power. Solar irradiation is measured in W / m2. During the night irradiation is zero. This has fundamental implications for everything in solar. During the day peak solar irradiation occurs during mid day, times when the solar irradiance is perpendicular (or close to perpendicular) to the solar modules. Peak solar irradiance varies from where you are on earth, but not by much. A normal peak solar irradiation is around 1,000 W/m2. What really differentiates a good project from a bad one is the number of hours with solar irradiation per year. In solar this is known as Yield – the wind equivalent of capacity factor. A normal Yield in solar is 15-22% compared to 30-35% capacity factor for wind.

For a viable project it is important to have a lot of days with sun. This is also highly dependent on region. Compared to wind though it is a lot easier to do resource assessment pre-construction. This is because the number of sunny days in a year is a lot less volatile than the number of windy days. Also, local phenomena that affect solar irradiation such as shadows and soiling are much easier to quantify in solar than in wind. Almost all solar sites are commissioned based on model data i.e. there is less need for pre-construction on site measurements. If measurements of solar irradiation were to be taken, an instrument known as a pyranometer would be used (picture above). This is the equivalent of an anemometer in wind industry.

Even though pyranometers may not commonly be used pre-construction they are very often used post construction. The data from pyranometers is absolutely necessary for availability and performance calculations.

Availability 

In the solar industry availability seems to be matter of philosophy rather than a contractual arrangement. “What does availability mean for you?” Is a question I heard a lot during the solar asset management conference.

But before I get into availability, let me digress with some words about the structure of a commercial, industry or utility scale solar PV site. The most basic component is the module. Here solar irradiation is converted into DC electricity. A module is normally around 260 W. That means a large site of 50 MW will have 192,000 modules. Modules are mounted onto a string. Strings then feed into a combiner box. A number of combiner boxes then feed into an inverter. The inverter converts energy from DC to AC current. The normal size for an inverter is approximately 500 kW – to 1,500 kW. For large solar sites the concepts of a block and field is used. A block or field contains multiple inverters and many strings. Further developments are also being made to install string inverters. These are smaller inverters installed at string level.

Even though there is no general consensus on availability in the industry it is common to measure availability at the inverter level and to then aggregate the average inverter availability to the site level for a site availability.

Since there will always be time when there is no solar irradiation, like during the night, availability contracts either treat these times as available or exclude them from the availability calculation. This is very similar to how availability contracts in wind power treat time periods when the wind is below cut-in wind speed or above cut-out wind speed. For solar PV sites the threshold value for solar irradiation is often set at 100 W / m2. Times with lower irradiation are either excluded or considered available.

All time when irradiation is above 100 W / m2 is categorized as either available or unavailable based on the condition of the inverter. The unavailable time can be the responsibility of the service provider or be carved out of the contract due to for example force majeure or grid issues depending on the contract.

By basing availability on inverter level insight into availability on the DC side is lost. Further, as it becomes more and more popular with single axis trackers for large solar PV sites the function of trackers is not taken into account. To capture failures in DC components or tracker failures time based availability on the inverter level is not enough. To go deeper in the availability calculation by basing it on for example combiner box availability or even string availability is often practically not possible as it, in most cases, is not possible to obtain the data.

Therefore it becomes necessary with robust performance metrics than can capture expected output compared to actual output at the inverter level. Performance metrics can give important and actionable insight into what may be wrong at a specific inverter and also deeper on the DC side or with trackers.

Performance

In wind the concept of power curve is widely established. It is impossible to talk more than two minutes about wind turbine performance before the warranted power curve is dropped into conversation. In solar, it is not so. One would expect graphs of solar irradiance on the X-axis as independent variable and power on the Y-axis as dependent variable. I have not seen any such graphs. In theoretical papers the “power curve” for solar sites is described as being dependent on irradiation, wind speed and temperature (both ambient and module). There exists a tremendous amount of research on the matter but in practice performance metrics seem to be based solely on irradiance.

In Bright we have chosen to apply the guidelines provided by the International Energy Agency in Task 13: Performance and Reliability of PV Systems. In essence it is a series of Yield, Performance Ratio Calculations and loss calculations that provide insight on the DC-side, AC-side and across the system.

A general observation is that performance is taken a lot more seriously in solar than in wind. 

Another method to assess performance is to perform a Module Level Thermal Audit - MLTA. This is done by flying over the solar site in an airplane equipped with a high resolution infrared camera. The camera captures the temperature of individual modules. The modules that are warmer than others are not performing well. In the hot modules solar irradiation is converted into heat instead of electricity. 

Above is an image from a company called Heliolytics that can detect faults using the MLTA technology.

Sum Up

I hope you found this post to be useful and we value your input as we continue to develop Breeze and Bright into the market leading renewable intelligence platforms. As always we value your feedback and thank you for reading through the blog post all the way here.

As a prize here is a video I found fascinating.