,
ABSTRACT
The use of distributed generation has witnessed a dramatic increase over the past decade. The two most popular of these – solar and wind – provide a unique set of regulation problems given the periodic and semi-unpredictable nature of their primary sources: the sun and wind. This white paper touches on some of the voltage regulation and stability issues that can arise as a result of the introduction of PV (photovoltaic) systems into the grid and how PMI’s Canvass web-based power quality portal can be used to quantify and troubleshoot the issues presented by this type of generation.
WHERE TO BEGIN
To begin, it is important to understand just what types of issues can occur when PV systems are introduced as a source of distributed power generation. Some whitepapers that provide background and analysis are Introduction to Residential Photovoltaic Systems, New Graph Templates for Distributed Generation, and Photovoltaic Monitoring Setup Recommendations. The latter two papers are focused on using power quality recorders and PC-based ProVision. In this paper, PV monitoring is discussed using web-based Canvass with the Boomerang and/or Revolution products.
Additionally, PMI offers a wide range of products that can be used to generate Canvass data – this white paper will go over these devices and provide a general overview.
PMI CANVASS-ENABLED ANALYZERS
PMI currently offers two product families that are capable of generating Canvass data: the Revolution and the Boomerang. These devices range in capabilities from single-phase, voltage-only measurement to the full gamut provided by the Revolution (three phases plus neutral of voltage; current; real, reactive and apparent power). See Figure 1 for an example of PMI’s Canvass-enabled Boomerang devices.
Figure 1. PMI’s Boomerang family of devices (from left to right: Pole Mount, 2S Meter Socket, Plugin, and 3-phase Boomerangs)
HOW DO THEY WORK?
PMI’s Canvass-enabled devices take a series of one-second average readings for the different measures that they are capable of collecting (such as RMS Voltage, RMS Current and Real Power). After a predetermined period of time, these devices will “flush” their data into the Canvass system (usually anywhere between five and 30 minutes. As soon as the data has been sent from the device, it’s stored by the Canvass system and is immediately available for analysis in a web browser.
Unlike conventional recorders, Canvass-enabled device don’t have on-board memory limitations or recording limitations – all data is streamed from the device to Canvass, and stored online. All available measurements are always sent, so there’s no need to compromise on what’s measured or the trend data interval to save memory. The 1-second data stored in Canvass can be processed to calculate up to 15-minute averaging periods for different steady-state analysis needs.
The bulk of PMI’s Canvass-enabled devices send their data to Canvass using a secure cellular connection through a Virtual Private Network (VPN) and a cell carrier private network that is not accessible from the public internet. The other devices work through a standard Ethernet connection through a LAN and back through to the PMI Canvass network.
WHY USE CANVASS?
As mentioned above in “HOW DO THEY WORK” the Canvass-enabled analyzers store all the 1 second readings on the web. It is reported to Canvass in a continuous stream and is always available to the user (so long as the user has an internet connection and a modern web browser). What this allows for is serious long-term trending analysis. This provides tremendous value when the task consists of identifying the effects of PV systems over a period of time (hours, days, weeks, months or – in some cases – even years). As in the example below (Figure 2), the user can see a 1-month graph for three single-phase voltage and current Boomerangs all attached to different residential PV installations geographically distributed throughout a single town.
Figure 2. One-month graph for three single-phase Boomerangs attached to PV units
SOME USAGE EXAMPLES
The weather has a very direct and often times severe impact on PV generation, with temperature and cloud cover making the biggest difference. Some aspects of the weather, of course, tend to be seasonally predictable (i.e. monsoon season in the Desert Southwest, snowy/overcast winters in the Northeast, overcast/rainy winters in the Pacific Northwest, etc.). In addition to the seasonal weather patterns, there are seasonal variations in the amount of daylight hours one experiences (as you travel farther from the equator).
The following examples (Figures 3 & 4) come from a series of PV panels installed throughout a region in California. The data was collected from a series of single-phase Form 2S Boomerangs over the course of several months as part of a pilot project with a California-based utility. One of the things that Canvass provided was a very good overview of what effects the weather had on distributed PV installations which, ultimately, can help a utility with load and generation prediction. The examples and graphs that follow will show some of the seasonal and weather impacts on PV distributed generation. Take special note of the power generation (negative on the graph) versus consumption over these two periods.
Figure 3. Impact on PV generation during sunny weather
The reader will most likely notice the difference in the shapes of the graphs. On the cloudy day (Figure 4), the customer seems to have “broken even” at best, whereas on the sunny day the customer was able to net produce more than they were consuming during the period from 12:00 through 18:00. Of special interest is the 18:00 hour – starting from about 17:00-17:30 and running through about 18:30 on both the sunny and cloudy days, this customer’s consumption went up quite dramatically. The difference was the net consumption / production (i.e. “net metering”). In this particular case, on the cloudy day the owner of this panel was consuming between 6.8kW and 7.2kW, whereas on the sunny day, the owner was only consuming 1.2kW (1/7th the load) at their peak.
Figure 4. Impact on PV generation during overcast weather
Figure 5 shows a graph for one week’s output for this same unit (covering both the cloudy and sunny days above).
Figure 5. PV system’s power output for one week
Graphs like those above in Figures 3 and 4 are very useful in helping a utility identify trends in distributed generation as well as general consumption.
Finally, the season itself has an impact in that the number of daylight hours is diminished during the winter months and is augmented during the summer months. See Figure 6 for a comparison of a winter day versus a summer day with the same unit from above. The maximum PV output in January is around 1 kW, while in the summer the same PV system can produce 1.5 kW – a 50% increase.
Figure 6. Output of a PV system during a summer day (top) and a winter day (bottom)
Notice how during the winter day, generation begins at around 09:00 and end at around 12:00. Contrast that with a summer day, where generation begins at around 08:00 and ends at around 18:00.
Voltage regulation can be a challenge when faced with PV systems producing variable amounts of power. Keeping track of tap changers and other voltage regulation devices is important in the face of PV arrays that can increase the voltage downstream of the substation. Figure 7 shows a single day, with RMS voltage on the top graph, and net power at one residential location on the bottom plot. In the early morning, the voltage starts to sag due to regular, steady load increases. A local regulator kicks up the voltage around 06:30, and again a short time later. The voltage continues to sag, bottoming out around 07:00. By 09:45, the PV generation (throughout this neighborhood) is causing the voltage to raise enough to require a drop in voltage. By 16:00 the PV output is falling, and the system voltage also falls. The regulator kicks in again around 18:00 to bump the voltage back up.
Figure 7. RMS voltage of PV system (top) with net power (bottom)
Continuous trend graphs of system line voltage in conjunction with PV output provide the information necessary to find and quantify regulation problems. In some cases, cap banks or simple tap changers may need to be reprogrammed, or replaced with voltage-sensing regulators or SCADA-controlled devices. Long term monitoring before and after distributed generation is in place is essential for understanding what changes are needed, and ensuring that the service voltage is within limits at all times and under all weather conditions.
CONCLUSION
Canvass is a very versatile tool. One of its greatest strengths is that it provides a near-limitless repository of data for engineers who deploy Canvass-enabled devices in the field which, in turn, allows for long term analysis. The two brief examples in this whitepaper are a simple yet clear example of the value in this. Long-term voltage, current and power trends, seasonal and meteorological correlations and much more can be observed now with a few clicks of a mouse from the user’s web browser.
Caleb Payne
Software Development Manager
cpayne@powermonitors.com
http://www.powermonitors.com
(800) 296-4120