SolarOZ · PV soiling analysis and solar operations

Make better informed solar panel cleaning decisions.

PV soiling can reduce solar farm energy yield and operating value. Explore recorded power and weather examples, or compare cleaning strategies using your station assumptions. For site-calibrated analysis and advanced multimodal modelling, contact SolarOZ.

Decision support · modelled estimates

Recorded site inputs · 366 daily observations

Real site example: soiling and reward

Recorded power and weather from 8 Jan 2022 to 8 Jan 2023, evaluated under three cleaning strategies. The full period appears immediately; hover to inspect daily values.

Historical example

Panel condition factor

Higher is cleaner · 100% = clean

No cleaningRule-based cleaningModel-selected cleaning

Cumulative model reward

Unitless score · higher is better

No cleaningRule-based cleaningModel-selected cleaning

Hover over either chart, or focus it and use the arrow keys, to inspect values for a day.

Strategy definitions: No cleaning makes no manual visits; the rule-based strategy cleans every 90 days or after at least 60 days when modelled condition falls below 60%; the model-selected strategy uses the trained policy. Evidence boundary: power and weather are recorded inputs, while panel condition and reward are model outputs. Reward is unitless, not MWh or revenue; the model was trained and evaluated on this same period.

How SolarOZ analyses PV soiling and cleaning decisionsWhat the public examples show, and when a site-specific assessment may be useful

Public demo: The examples compare recorded site inputs and modelled strategy outputs. The station planner is an illustrative scenario tool. Its estimates depend on the information entered and should not be treated as measured savings or a site-calibrated forecast.

Inputs considered

Depending on the analysis, relevant information may include PV generation, weather, current soiling condition, cleaning history, energy value and operating costs.

Decision questions

Compare how cleaning choices may affect panel condition, energy loss, recovered generation, operating value and the timing of site visits.

Project-specific analysis

Sites with additional data, operational limits or more complex requirements may benefit from tailored calculations and advanced multimodal modelling. The suitable data and scope are discussed with each customer.

Scientific context: PV soiling and natural rain cleaning depend on site and environmental conditions. The following publications provide background on soiling measurement, rain effects and cleaning-schedule analysis. They do not validate the SolarOZ demo outputs.

Schill et al., Soiling Losses: Impact on the Performance of Photovoltaic Power Plants, IEA PVPS Task 13 (2022).
Norde Santos et al., Cleaning of Photovoltaic Modules through Rain: Experimental Study and Modeling Approaches, Solar RRL (2024).
González-Castillo et al., Cleaning scheduling in photovoltaic solar farms with deterministic and stochastic optimization, Sustainable Energy, Grids and Networks 36, 101147 (2023).
Micheli et al., Photovoltaic cleaning optimization through the analysis of historical time series of environmental parameters, Solar Energy 227, 645-654 (2021).

Your station · editable scenario

Investigate your station

Enter your site and operating assumptions, then compare an optimised schedule, a routine schedule and no cleaning. Results are modelled estimates for planning.

Tell us about your station

Use your measured values where available; otherwise enter a planning assumption.

Plant capacity MWInstalled AC capacity used for this scenario Clean-day yield MWh/MW/dayExpected daily output if panels were clean Energy value A$/MWhUse your expected realised value, not spot price by default Current soiling loss %Measured or estimated loss today Daily soiling rate %/dayRelative daily decline in panel cleanliness Cleaning cost A$/visitInclude labour, water, access and downtime costs

2 · Choose a strategy to inspect Optimised scheduleModel selects visit dates for the highest estimated net value Routine cleaningUse the interval selected below No cleaningUse as a baseline for comparison

Plan starts Planning period

Routine cleaning intervalUsed to calculate the routine strategy for comparison.

Weather and operational constraints optional

Expected rain dateDate is limited to this plan period; blank = no assumed rain eventRain removes soiling %Site-specific assumption, not a weather forecastMinimum gap between visitsDays

Use daily site projections optional CSV

For a variable solar/price outlook, upload one row per day with date,clean_mwh,value_aud_per_mwh. These projected inputs replace the three capacity/yield/value fields above. Your file stays in this browser.

No daily file selected. A constant clean-day yield and energy value will be used.

Ready to calculate

3Selected strategyOptimised schedule

Calculating...

Select a strategy and enter your site information to compare modelled outcomes.

Estimated net value vs no cleaning—Modelled A$ after cleaning costs

Energy recovered—Modelled MWh vs baseline

Cleaning visits—Over the selected period

Daily energy loss · MWh/day

Lower is better. Hover over the chart to inspect all strategy values for a day.

No cleaningRoutine cleaningOptimised schedule

Compare the choices

Projected values over the same planning period. No cleaning is the baseline.

Strategy	Energy recovered	Net value vs baseline	Visits

Selected strategy visit dates

Inspect the daily projection

Date	Clean-day MWh	No-clean loss MWh	Routine loss MWh	Selected strategy loss MWh	Action

Advanced analysis · by consultation

Need a more detailed, site-specific assessment?

SolarOZ can discuss customised calculations and advanced multimodal modelling for PV soiling, solar performance and cleaning decisions. Tell us about your site data and operational goals, and we can discuss a suitable analysis scope.

Discuss your project