As demand for clean energy heats up, robots are increasingly critical for the cleaning of large-scale utility solar panels, an industry leader says.

During a recent presentation at the Climate Smart Engineering Conference hosted by Engineers Australia in November, Ido Molad, Vice President of R&D at multi-national solar PV robotic cleaning provider Ecoppia, outlined the scale of demand for robots in cleaning solar panels along with challenges that need to be addressed in robot development.

According to Molad, demand for robotic cleaning is set to expand around the world amid exponential growth in use of solar PV panels.

According to the Global Power Intelligence Centre, solar PV power generation expanded from just 61.5 terawatt hours (TWH) in calendar 2011 to 1,064 TWH in 2021. By 2030, this is expected to reach nearly 4,000 TWH.

Much of this is generated through utility scale solar PV plants – massive power plants in which solar panels are spread over tens or hundreds of square kilometres.

These are often located in arid regions that offer plentiful sunshine and large areas of space for ground-mounted solar arrays. Some are ‘fixed tilt’ structures which are fixed in place at a particular tilt or angle. Others are single access tracking (SAT) structures where the panels rotate from east to west as the sun moves during the day.

In Australia, several projects are either planned, underway or in operation. These include the 720 MW New England Solar Farm in Uralla in northern NSW, a 460MW farm at the Western Down Green Power Hub in Chinchilla in south-east Queensland and the 336MW Darlington Point Solar Farm at Darlington Point in central NSW.

On a massive scale, the $30 billion Australia-Asia Power Link project being delivered by Sun Cable aims to provide up to 15 percent of Singapore’s power needs by harnessing and storing solar energy from the Northern Territory and transmitting this to Darwin and Singapore via a high voltage direct current transmission system. That project, however, has encountered challenges and the company delivering it entered voluntary administration on January 11 amid shareholder disagreement about project objectives and direction.

With all this activity going on, Molad says demand for cleaning of these panels will be massive.

This is especially the case as daily cleaning is needed in order to prevent ‘soiling’ – the accumulation of dust particles on the panels. Where panels are not cleaned, soiling can result in power generation losses of up to 50 percent. This is especially the case where the cells are located in arid, desert areas where water is scarce. Soiling can also lead to an overall degradation in panel performance over time.

For several reasons, robotic cleaning is emerging as the preferred option to either manual or vehicle-based cleaning.

In particular, robotic cleaning:

  • Is performed autonomously and avoids the need for large amounts of on-site manual labour (which is costly, time-consuming, hazardous and often not available in remote areas).
  • Is based on dry cleaning and thus avoids requirements for large volumes of water that is often not available in arid environments and needs to be distilled in order to avoid leaving residues.
  • Delivers a clean which is of consistent quality (unlike that which may occur in manual cleaning); and
  • Avoids the need for vehicle movement between the panels that is associated with vehicle cleaning. This is important as vehicle movement requires large gaps between the panels and creates the potential for module damage arising out of contact with vehicles.

Predominately, robots are deployed on large scale installations.

In Ecoppia’s case, the company has three offerings:

  • Its first E4 robot (pictured left below) that is designed for fixed-tilt installations in which the panels remain in a fixed position and do not rotate to follow the sun during the day. The robot travels on rails that are installed on the top and bottom of solar PV installations. Only the delicate microfibres touch the panels themselves.
  • A lighter weight robot (T4) (pictured middle below) which is designed for single access tracking (SAT) installations in which the solar panels follow the sun from east to west during the day. The robot sits at a docking station during the day from which it subsequently launches to perform the cleaning operation once solar production hours have finished.
  • A more recent hybrid solution (H4) (pictured right below) that can be applied on either fixed-tilt or SAT systems. This travels rapidly from one side to the other to perform the cleaning operation.

All three solutions are autonomous (no manual labour needed), use dry cleaning and are solar powered with their own self-charging panels. Dry cleaning is common for large-scale operations as ‘wet’ cleaning would require large volumes of water and is thus often not practical for many areas in which large-scale installations are located.

In addition, all solutions operate on a cloud-based management platform which uses artificial intelligence and data collected from sensors that are located on top of the panels.

(From left: Ecoppia’s E4, T4 and H4 solutions – see above for descriptions).

When selecting a cleaning solution, Molad says decisions can be influenced by several factors. These include site location (i.e. remoteness), the scale of the utilities, the harshness of the environment (including a potential lack of water resources) and the complexity or otherwise of the site. On the last point, sites can adopt different structures (i.e. fixed/rotating), feature different types of panels, be located on sloped terrain (increasingly common as the vast numbers of installations sees sites with flat terrain become more scarce) and be subject to movement over time.

When developing robots for solar panel cleaning, Molad says six things are needed.

These include:

  • The ability to operate reliably and autonomously with minimal on-site intervention and to communicate (data collection etc.) with the remote operating control systems.
  • Effectiveness. This includes the ability to operate each day for 25 years, to use dry-cleaning and to remove all soiling (99.5 percent) daily. It also includes the ability to adapt and be maneuver to overcome obstacles such as ground movement or change in structures over time.
  • Be compatible to cater for different structures and configurations which may be present on sites. This includes catering for different configurations from tracker manufacturers, different variations of installations (with fixed structures, for example, the width of structures can vary from 2 meters up to 11 meters) and increasingly larger, longer and more powerful modules with a supporting structure that will contain progressively less metal in order to minimise cost. Robots need to be adaptable and to cater for all of these variations whilst continuing to work flawlessly.
  • Environmental robustness. This includes the ability to withstand high/low day/night temperatures, high humidity, lots of dust, high winds (that can impact not only the robots but the installations on which they travel), monsoons and high levels of UV radiation. This requires careful selection of components for both mechanics and electronics.
  • Being safe and reliable. Whilst people are not often present on these sites, the robots must still be safe for the modules and structures on which they travel. They must not degrade panels during operation or harm the anti-reflective coating. To ensure that robot operations do not jeopardise long-term panel warranties, certification for their use should be obtained from panel manufacturers.
  • Be long lasting and work each night for 35 years on end. This involves selecting components which are both durable and which will be available as spare parts if needed, designing robots to enable future upgrades and/or to use replacement parts, and considering life-cycle operational and capital costs when selecting components.

Speaking of Ecoppia, Molad says the company has adopted four strategies to overcome these challenges.

First, it engages in widespread stakeholder collaboration.

This includes collaboration with panel manufacturers and single axis breaker manufactures to ensure that robots will meet demands for current and future offerings; with site owners including engineering/procurement/construction contractors to understand different site layouts and to adapt robot development to meet these; and internally with production, sales and marketing, R&D departments and field technicians. The last of these (technicians) are particularly important as products are sold with 25 years of maintenance and service. As the people who will perform this, technicians can provide input about requirements in this area.

Next, an agile approach is adopted to hardware development. This enables feedback and design modifications across the development lifecycle from early stage though to final delivery.

Third, there is testing – verification and valuation. Early last year, Ecoppia moved into a new R&D facility in Israel. Stretching across more than 2,000 square meters, the facility is filled with simulators of leading trackers and structures that the robot may encounter in the field along with modules from different manufactures. These enable the robots to be tested against different trackers, structures and modules at all stages of development.

A key part of this is life-cycle testing, through which the robots are run 24 hours per day in the facility to ensure that they are able to work reliably and autonomously in the field. This includes testing not only of entire robots but also individual components. For example, one machine tests the lifecycle of the wheels and tyres. The wheel travels on another wheel that resembles the edge of a solar panel on which the wheel will travel in the field. When doing this, the firm looks at how different materials perform against certain trade-offs. In the tyre example, this may include the level of durability that a material may offer as opposed to the level of traction that it may deliver when cleaning the panel. Another trade-off could include the likely cost of maintenance as against that of production.

Finally, it is important to define information and data requirements. As with any device that is operated remotely, data is critical with solar panel cleaning robots. Informational requirements include operational performance (which must be correlated with weather conditions) as well as maintenance data such as battery health and charging status.

These data requirements need to be defined early in development as they influence the sensors, electronics and mechanics which are put into the robot.

Overall, Molad says opportunities and challenges are significant.

“Solar PV growth is huge,” he said.

“And it requires automated solutions to enable soiling control on these huge sites.

“Developing a 3D robotic cleaning solution is really tough. The market is changing fast. We have to adapt and to make sure that the product we develop fits this market.

“And of course, as we are in the era of data, data collection is critical to enable really large-scale operation.”


Enjoying Sourceable articles? Subscribe for Free and receive daily updates of all articles which are published on our site


Want to grow your sales, reach more new clients and expand your client base across Australia’s design and construction sector?


Advertise on Sourceable and have your business seen by the thousands of architects, engineers, builders/construction contractors, subcontractors/trade contractors, property developers and building industry suppliers who read our stories across the civil, commercial and residential construction sector