How can I monitor a PV system's performance?

Question

What are things I should monitor to track a photovoltaic system's performance?

I'm not looking for a recommendation for a particular piece of kit, because such a recommendation could quickly go out of date, and then the answers would be of no use to anyone else.

So I'm asking, what should be monitored, to track a typical home PV system's performance and health: roof-mounted PV peak capacity is 4 kW, and there's an inverter in the loft (attic) connecting it to the home power supply and to the grid.

Answer 1

Photovoltaic system's performance parameters belong to two categories: energy generation (electricity units like Volts, Amperes, Watts and Watthours) and supplemental (some other units). Although you don't mention it, to make the answer complete, I add energy storage category (if you had a battery storage attached).

A PVE system can be in one of these states:

1. Off - no energy is produced. No insolation, or the system is stopped manually.
2. Idle - panels are insolated but the DC power they produce is too low to output any AC power.
3. Active - panels generate DC electricity and the inverter produces and supplies AC power.
4. An error state - current break, under/overvoltage, frequency variations etc.

I. Energy generation parameters

1. Each string of panels has it's DC voltage and DC amperage in active state. Their product is current DC power. MPPT inverter always tries to gain the maximum power from panels. DC voltage is only monitored in idle state.
2. The most important performance parameter of the system is the inverter's current AC power. In case of a 3-phase inverter the total AC power can be split to more phases.
3. Because we generate electricity in time, the total energy produced in kWh is summing it up. We can aggregate through hours (hourly production), days (daily production), months, years etc. Beautiful graphs can be made from this (sorry for selfpromotion, but it's similar to the questioned system).
4. Current AC power supplying to the grid or total energy supplied to the grid throughout certain period of time are the final measurements.

II. Supplemental parameters

1. Temperature of panels. The higher temperature the lower maximum produced energy and shorter panel life.
2. Temperature of inverter. Inverters cool themselves down above certain temperature and consume (some) energy they produce.
3. Operating time - sum of the time when inverter operated in active state. We try to maximize the energy produced during active time.
4. Inverter should monitor AC voltage for under/overvoltage for safety purposes.

III. Battery storage parameters

1. Energy stored. Can be in kWh or in % of total capacity.
2. Battery (bank) voltage.
3. Time remaining to full charge. Approximate value based on current rate of charging.
4. Time remaining to discharge. Approximate value based on current rate of discharging.
5. Number of charge/discharge cycles.
6. Storage efficiency, i.e. energy stored/energy released.
7. Temperature of battery.

Answer 2

This is a bit specific to each solar PV installation, but I'll give it a shot.

Short Answer

For a 4kW array, the actual power output you see is likely to be around 3kW if your system is operating properly, assuming the array isn't too old. For more, read on ...

Proprietary Monitoring Software

First of all, many home solar installations now are done by companies that specialize in putting together packages of the solar panels, the inverter, and a control panel to monitor the system. If you're in that situation, then you basically get what the company who wrote the control panel's software decides to give you. It's unlikely that building your own monitoring circuitry or software would be a good use of time, although if you're a DIY type, I might suggest looking at LabView or OpenEnergyMonitor to create your own monitoring system.

It's also unlikely that the software features that come with your system are going to be (or should be) a primary factor in picking the system. I personally would rank the kind of panels, the inverter hardware, the experience of the installers, the price, and specific tax incentives as more important than the amount and kind of data you can see on the control panel, or web interface that the system supports.

Here are a few examples from some systems available where I live:

• https://enlighten.enphaseenergy.com/public/systems/ZHN8479
• http://www.greenraysolar.com/sites/default/files/GreenRay-SunTrakMonitoringSystem.pdf
• http://documents.locusenergy.com/SolarOS-Site_Owner_App_Product_Sheet.pdf
• http://www.civicsolar.com/resource/solar-energy-pv-monitoring-options

Metrics

In terms of what you should watch, I think solar panels are a bit like a car. If you only want to watch one thing to assess a car engine's health, watch your fuel consumption (e.g. miles per gallon, km/l, etc.). With solar panels, kWh produced per month is probably the equivalent metric. That number can be affected by a few different factors, but if the number is close to the output when your panels were first installed (adjusted for expected degradation ^‡), that's a good indication that the whole system is functioning reasonably well.

Some systems allow you to see the production per module, which can help identify one bad module in your array. Having electronics on each module helps both with fine-grained monitoring, but also allows micro-inverters or DC-DC converters to tune each's module's performance independently.

Another metric that may be available to you is panel temperature. Excessive temperature can be a sign of problems, and high temperatures can damage photovoltaics, like many electronics.

You may also be able to see the panel output voltage, or the frequency (Hz) of the power your inverter is producing. Familiarize yourself with the normal values for your system, and note any deviations from that baseline.

Finally, your system may have a separate status output that indicates that a particular module's health is Ok, Failed, Degraded, or Unknown.

Sunshine

Keep in mind that the solar panels can only convert energy from what the sun gives them. Variations in weather do make a difference, so if your panel output this month is 5% lower than the output from the same month last year, that doesn't necessarily mean there's a problem.

Take a look at the climate data aggregated at the US National Climatic Data Center (Note: this isn't just US data ... it's global). If you open up the PDF file for a given month (there's a couple month delay in the published data), find a city near you, and look at the far right column to see total hours of sunshine:

In the row I highlighted, the monthly sunshine was 26% higher than average for that month. So, in that case, you'd expect your monthly output to be higher than most years. This is a rough way to assess performance, but it's better than simply comparing to the same month last year, without adjusting for sunshine.

Some off-the-shelf software monitoring systems do interface with weather and climate data services, and show some measure of sunshine. The ones I've seen show instantaneous solar irradiance, which I don't think is as useful as longer term sunshine data.

(‡) Expected Performance Loss

It's important to know what to expect from a solar PV system. Panel performance will degrade slowly, and smoothly, over time, so that's to be anticipated. From the US National Renewable Energy Lab, 2010:

So, depending on the kind of photovoltaics you use, and when they were built, you might expect a 1/2% loss in output per year, or it might be 2% per year.

Inverter problems may show up more abruptly, and before the photovoltaics reach their end of life. Of course, physical damage to your panels can instantly affect their performance, too.

Finally

If you haven't already had your system installed, or have a maintenance visit coming up, ask this question of your installer, who will likely have important experience with the specific failure modes of your hardware.

Answer 3

As there are already two excellent answers I would just like to add what I found useful to monitor:

As Nate's answer suggests the energy produced per month is a good metric for keeping track of the PV systems status (supplemented by information provided by the inverter). And as the amount of energy is depending on multiple parameters and could not just be compared to the initial values of the system the comparison with other PV systems seems to be the most easy way. So if output data is made available by other PV owners and operators of the respective region this is a very simple yet pretty accurate estimate. The more PV systems contribute data to such a collection the more variations and degradation of single systems over time are averaged. (So this definitely calls for a citizen science project.)

If those values are not available for the region of interest I'd prefer to use global solar radiation (global irradiance) as a monthly aggregated value that's a more meaningful number than just the hours of sunshine per month. (In Germany such data is reported by the Deutscher Wetterdienst, but maybe they are also available from the US National Climatic Data Center or other national institutions.)