The basic goal of ASTM E2848, “Standard Test Method for Reporting Photovoltaic Non-Concentrator System Performance,” is to compare the ratio of a PV power plant’s actual in-field performance to its expected performance based on a system model. The most important things for beginners to understand about PV capacity testing relate to the model and schedule.
Here, I take a deeper look at some more subtle considerations related to weather file and shade model; commissioning and instrumentation; seasonality and location; and technology and design. By paying close attention to these criteria, you can expedite the process and improve the probability of passing a PV system capacity test.
Weather File & Shade Model
While there are many inputs to a PVsyst model, the weather file and shade model have the greatest influence on the long-term energy yield. Since a PV system capacity test is normalized for weather, the shade model and its influence requires special attention. Most importantly, a shade model is always a prerequisite for a capacity test. If you omit a shade model, your PVsyst results will overestimate the modeled results, which sets performance expectations too high.
Shade models are also used to help identify filtering points for the test, which ensures that you do not incorporate data in these sub-optimal regions. More importantly, 3D terrain has an influence on the average POA. This outsized influence carries through to the capacity tests used for project acceptance. When actual PV system performance in the field fails to correlate to the expected performance based on the PVsyst model, the root cause is often due to this issue with the shade model.
Courtesy Hukseflux
Commissioning & Instrumentation
Proper commissioning and instrumentation are mission critical to capacity testing. In terms of general project commissioning, capacity tests need to be preceded by pre-qualification activities that confirm that sites are operating normally. Acceptance testing takes place after system commissioning. With that in mind, a PV system is not fully commissioned if there are outages in the system and things that are not fully operational. In addition, it is recommended to have a minimum of a 48-hour period to qualify system performance prior to commencement of performance testing.
For best results, this commissioning process needs to include the instrumentation used to verify plant performance against the model. As a rule, instrumentation is rarely well commissioned ahead of acceptance testing. Prequalification checks are often limited to verifying that sensors are installed and reporting data. Once we start running tests and using this data, we often find that sensors are not aligned or calibrated. These types of issues should be addressed prior to acceptance testing.
In this context, expediting the process of passing a capacity test is often as simple as verifying sensor placement and calibration. Commissioning activities include confirming POA gain and alignment versus the transposed global horizontal irradiance (GHI); verifying that sensors are clear of obstructions at all points of the day; and correlating ambient temperature to module temperature.
With POA sensor placement, ensure that the sensor is placed in a location most indicative of the irradiance across the entire site. If the site has a rolling terrain, the actual average POA is not the exact figure targeted by the racking design, the sensor must be adjusted to match the peak irradiance of the site. This angle should align with the shade model average azimuth and tilt angle rather than to an exact figure from the racking design.
With temperature sensors, ensure that ambient temperature sensors are protected from the influence of direct sun and that module temperature sensors are not influenced by cold spots due to the racking or module frames.
Seasonality & Location
The time of year and project location are critical to capacity testing success. ASTM E2848 defines a range of acceptable testing conditions, as well as certain disqualifiers. These grounds for disqualification relate to low or high irradiance levels and the presence of shade or inverter power limiting. Note that these are often interrelated. When irradiance is low, you tend to have shade based on the time of day or year. Similarly, when irradiance is high, the system is most likely to experience inverter power limiting and power curve clipping.
Note that the time required to conduct a passing test is highly variable based on the project location. This is especially true in winter when you might have only eight hours of sunlight per day. If you have to throw out two hours in the morning and two hours in the evening due to low irradiance or shade, your window for conducting a capacity test is down to four hours. If you have to throw out two hours in the middle of the day because of clipping, that leaves you with two hours—or eight 15-minute interval data points—per day.
Given that you need 50 valid data points, you may need a week of perfect weather to pass a capacity test. While that might be possible in California, Colorado or Nevada, winter commissioning is likely impractical in the Midwest or the Northeast. If the weather is unlikely to cooperate, it is better to postpone capacity testing to a later date. It is best to plan and contract for this contingency in the event that construction schedule delays push performance testing into the November to January timeframe.
Courtesy The Coloradoan
Technology & Design
Last but not least, you may need to adapt your capacity test methods to account for technology and design details, such as bifacial modules or high dc-to-ac ratios. While it is possible to run capacity tests under these circumstances, you will need to plan for it and implement a different strategy.
While the ASTM E2848 standard does not specifically address bifacial modules, for example, there are ways to work around it. Specifically, you have to add a back-of-module POA sensor in order to model these new capacity contributions. In a bifacial system, in other words, the total POA irradiance for capacity testing purposes is the sum of the front-side POA irradiance plus the rear-side POA irradiance. By accounting for the rear-side irradiance in the regression equation, you can adapt the test procedure to accommodate this technology.
To adapt to a high dc-to ac ratio, we want to modify system operation to fit the test criteria. Note that this requires a prequalification step. If the inverters are power limiting more than 50% of the day, the system probably qualifies as having a high dc-to-ac ratio. In this case, you should consider some means of modifying system operation. If it is a tracker system, you could defeat the trackers and run the system flat. Alternatively, you could remove half of the dc capacity, effectively shutting off half or the dc strings and conducting capacity tests on one half of the array at a time. Note that whichever option you pick, you will need to adapt your model to reflect this temporary performance condition.