Should I Clean My Solar PV Panels?

We've had solar PV panels for over 17 years. Some long-term performance data and other details have been discussed here, here, here and here.

While solar PV panels are largely maintenance-free during their expected lifetime (25- 40 years), regular cleaning is often recommended. My local installer claims a 10% - 30% improvement in performance, on average, with regular cleaning. A 10% better performance seems feasible but I am more than a little skeptical about the 30% claim.

Most, if not all, solar PV panels are designed to be self-cleaning - a hydrophobic (water-repellant) coating encourages the panels to shed rain droplets taking any dirt and dust with it. To work efficiently you want regular rainfall (preferably at night!) and the panels tilted to allow the water to run off quickly. Roof-mounted panels will usually be pitched at the same angle as the roof (Photo 1).

We have two solar PV arrays - one on the extension (Photo 1) facing South-East (135⁰) and the other (Photo 2) on a lean-to shed at the rear of the house facing South-West (225⁰).

Photo 1: Main Solar PV Array facing SE

Photo 2: Secondary (near horizontal) PV Array facing SW (with solar thermal above)

The secondary PV array was installed horizontally because of planning restrictions - a more upright position and the reflective glare would have been a nuisance for the houses adjoining our rear garden (see Photo 1).

In Photo 1, you can clearly see the flight path of seagulls as they take off from the main house? Seagulls produce copious amounts of guano but they are not the only culprits.

Our solar PV panels are, fortunately, on single-storey roofs (or rooves?) so all I needed was a small ladder (Photo 3) and a long-handled mop (Photo 4) for the cleaning operation. I already had the telescopic handle used with the lopper tool to trim the wild cherry tree and our small hedge. So my outlay for the whole cleaning operation was less than £30 with subsequent procedures costing less than £5 a time.

Photo 3: Easy Ladder Access to the Main Solar PV Array

Photo 4: Extendable (3 metres) Mop

I half-filled the yellow trug (Photo 4) with hot water (50 ℃) - heated by solar thermal - and added some household detergent (Ecover Washing Up Liquid). Using the large trug meant I could easily access the cleaning fluid from the top of the ladder thus saving both effort and time. After gentle washing with detergent, I used the hosepipe to rinse down the panels before leaving them to dry overnight; note, all the water used for cleaning was collected in our water butts, via the gutters and downpipes, for re-use in the garden.

On a safety note, I need to point out that I carried out the cleaning operation in the evening between 9 pm and 10pm (civil twilight, Photo 3) when the solar PV was inactive. If you want to do this cleaning job during the day then you will need to isolate the system from the mains before commencing.

The big question, of course, is whether the cleaning has improved the efficiency of the solar PV panels. This is not a trivial task to determine. You cannot just measure the generated electricity before and after cleaning because the kWh produced will vary from day to day depending on sunshine strength and the number of sunshine hours in the day. Furthermore, the two arrays are differently-sized and face different directions so will contribute different proportions to the overall electricity production at different times of the day.

My Davis Weather Station has a solar radiation sensor that can be used to estimate daily sunshine hours. Solar radiation (W/㎡) values also show an excellent correlation with the amount of electricity (kWh) generated by our solar panels.

Figure 1 shows the daily solar radiation profiles measured by the Davis Weather Station for the period 2nd to 8th June 2023. Each daily profile is different - sometimes subtly and sometimes drastically - due to changes in sunshine strength, cloud cover, length of day, etc. The profiles for the 3rd and 4th June are most similar due to cloudless skies all day on both days. Data for these days, therefore, represent near-optimum solar PV generation before the panels were cleaned on the evening of the 4th. The profiles on the days immediately following cleaning were too dissimilar so I had to wait until June 13th & 14th (Figure 2) for comparable daily solar radiation profiles (again thanks to cloudless blue skies). Data for the 13th & 14th June represent near-optimum solar PV generation after the panels were cleaned.

Figure 1: Daily Solar Radiation Profiles between June 2nd - 8th 2023

Figure 2: Daily Solar Radiation Profiles between June 9th - 15th 2023

Note: Day length on 13/14th June was some 12-13 minutes longer than on 3/4th June. The solar radiation detector detected the first 'light' at 5.15 am on all four days; the last 'light' was detected at 9.15 pm (3/4th) or 9.30 pm (13/14th).

Over the two-day period, 3rd - 4th June (before cleaning), the solar panels generated 54.83 kWh of electricity.

Over the two-day period, 13th - 14th June (after cleaning), the solar panels generated 56.27 kWh of electricity.

At first glance, this might suggest that cleaning the PV panels improved output by 2.6 %. However, although the solar radiation profiles for all four days looked similar (Figures 1 & 2), there were in fact small differences. The average solar radiation was 314.5 W/㎡ for the period of 3/4th June whereas it was 319 W/㎡ for the period of 13/14th June - an increase of 1.4 % as we approached the summer solstice.

Once we take the small increase in solar radiation into account, we are left with an improvement of just over 1 % in the efficiency of the solar panels after cleaning.

My attempt at cleaning was far from perfect as the photo below (taken 11 days after cleaning) shows ...

Photo 5: Persistent Grime on the Solar PV Panels even after cleaning

... but I was wary of causing damage to the panels with a too vigorous cleaning regime.

I will give cleaning another go if only to collect another set of data to see if it backs up the results seen here.