Solar plant construction evolution: robots are what comes next.
Solar plant construction evolution: robots are what comes next.
Few construction projects are as repetitive as a solar farm: the same gestures, hundreds of thousands of times, across hundreds of hectares. Benoît Froidurot, Director of Innovation & Development at Equans Solar & Storage, has been doing it since 2009. He explains why robotics is now entering the picture, and what it will take for it to deliver at scale.
Introduction: Benoît, you've been in solar construction since 2009. What's the one thing that's changed most?
The scale: when I started, a large project might cover a few dozen hectares. Today we're building on hundreds of hectares, installing hundreds of thousands of panels per site. On one recent project, we put in 700,000 modules, each accompanied by supporting metallic structures and connected with cables. Every single one had to be carried across the site, lifted into position, fixed to a structure and connected.
A solar panel nowadays weighs around 35 kilograms (roughly the size of a large ceiling tile) and you need at least two people to carry it. Multiply that by 700,000, and you start to understand the productivity challenge. You also start to understand the human one: lifting heavy loads repeatedly, outdoors, for months, is physically demanding work. Musculoskeletal disorders are a real occupational health issue in this industry. Robotics helps address both: not to remove people from the equation, but to support teams in working more effectively and in better conditions.
Question 1: Why hasn't robotics arrived sooner on solar sites?
It's coming, just not all at once. Solar construction involves a chain of very different tasks: driving steel piles into the ground, assembling the racking structures, placing the modules, running kilometres of cabling. Each one requires a different set of skills and, potentially, a different type of robot, while rationalization of machines is part of the equation as well.
The work is also highly repetitive, which is exactly where automation delivers the most value, both in terms of productivity and in reducing strain on workers. But repetitive doesn't mean simple, as these sites cover hundreds of hectares, often on uneven terrain, in extreme weather. Getting a machine to perform reliably in that environment is genuinely hard.
What's changed recently is that the technology has caught up with the ambition. We've moved from talking about it to actually testing it on real sites, in real conditions. Two years ago, we deployed a GPS robot to mark the position of each pile before installation. It sounds modest, but precision on this scale matters. That was the start.
Question 2: What does a robot actually do on a solar site today, and what can't it do yet?
The most advanced work is in pile driving. We tested a system at a site in Australia: two fully autonomous machines that locate the GPS position and ram each pile, with no operator in the cab. The piling operation performance is increasing while testing new versions of machines and starting to reach competitive costs vs conventional ways to ram the piles.
For module installation, we've tested robots that pick up a panel, use sensor-based recognition to position it precisely on the racking structure, and fix it in place. On Goorambat East project in Australia, we ran tests several robotics companies for comparison and testing purposes. The level of autonomy is already impressive and starts to facilitate installation tasks for humans.
What robots can't do yet is work in a fully integrated way. Right now, robotic tasks and manual tasks happen separately. The real step change will come when we will be able to orchestrate them together, when a connected layer sits above the whole site, collecting data from each machine, optimizing their routes and sequencing, and coordinating with the human teams in real time. That's where we're heading.
Question 3: Once a solar farm is running, do robots still have a role to play?
A significant one, and in some ways it's more mature than the construction side.
Module cleaning is a good example: dust accumulation obviously reduces panel efficiency. On a large site, manual cleaning requests people to spend hours cleaning modules with all related drawbacks. Robotic brushing systems, like those proposed by numerous of manufacturers, move across the panels automatically, with adapted processes according to the kind of soiling and the local climate.
Vegetation management works the same way: robotic mowers keep the ground around the panels clear, preventing shadows that would impact PV module energy production.
For inspection, we combine aerial drones with ground-based robots for a more accurate understanding of the plant behavior. The flying drones map the site from the top and flag anomalies, while the ground robot dogs or rovers can detect specific anomalies by getting into spaces flying drones can't reach, with the same goal to detect underperforming panels.
About a month ago, we ran our first test of a robot dog on one of our sites. The results were promising! The mobility of these machines has improved remarkably fast.
Question 4: Why is most of this development happening in Australia rather than Europe?
Several things come together there.
Australia has a strong will to decarbonize its energy production to get rid of fossil fuel solutions and solar is the most accessible path, knowing the advantageous local conditions for PV solar technology deployment, basically space and sun irradiance. On top of that, regulatory environment is favorable to move quickly.
We are also closely working with ARENA (the Australian Renewable Energy Agency) to develop robotics specifically for solar construction. That kind of institutional backing makes it possible to invest in technology that is still maturing.
There's a labour dimension too. The solar market is growing faster than the available workforce, particularly for physically demanding site roles. Robotics helps bridge that gap, taking on the most repetitive and strenuous tasks, while allowing human teams to focus on work that requires judgement, coordination and expertise.
Europe is developing fast too the solar technology with other particularities in terms of land types, installations, reglementations, ressources and geopolitical situation. Robotic will have a role to play as well using different kinds of construction methods. But robot market will not be able to adapt to multiple philosophy of construction. That is the reason why we have chosen the scenario of machine deployment running on the site for all areas, and benefit from what we experienced in Australia.
Question 5: Ten years from now, will solar farms build themselves?
Views in the industry vary enormously: from five to twenty years to have autonomous construction of solar plants. What is the reasonable timeline? I would say ten to stay at the middle but with the innovation works we are working on, we expect to minimize the timeline for site implementation.
What I'm confident about is the direction: the robotics technology available today is already far beyond what existed five years ago, and it's improving fast. Five years ago, what we're doing in Australia wasn't possible. Humanoid robots, machines that can navigate complex environments and handle unfamiliar objects, are advancing at a pace that's hard to fully appreciate from the outside.
My expectation is that within five years, we'll run our first tests of autonomous solar farm construction. From there, the shift to standard practice will depend on cost parity, on safety protocols for robots and human teams working together, and on clients willing to be early adopters.
One thing I think about is the feedback loop this creates. Today's panels are sized for human handling: around 35 kilograms, two people to lift. If robots become the standard installation method, some restrictions can be released with a probable impact on size and weight of components. They could be larger, heavier, more powerful: designed around what machines can carry, not what people can. That could change the economics of solar significantly, still having in mind efficiency and LCOE reduction.
Human expertise will remain at the heart of every project to drive the whole construction, ensure logistic of components, maintain and operate the machines, coordinate operations, handle the unexpected. What changes is the nature of the work: less physical strain, more skilled oversight. New roles will emerge around robotics management and data analysis. In my view, it's a positive shift for the people in this industry.
Conclusion: What would you say to solar developers or operators who haven't yet thought about robotics?
Solar technology has experienced massive cost reduction for the last 15 years and this trend still occurs today while constraints are becoming stronger on materials (silver, steel, copper…), geopolitics (taxes, industry protection), network integration and land scarcity.
However, some promising major technology shifts will raise bringing down the LCOE: String voltage increase, module efficiency increase (tandem), hybridization, etc… Robotics will be among the technologies accompanying the solar development to minimize construction costs and will also probably have an impact on component landscape.
Consequently, an advice would be to start paying attention now, even if you're not ready to act. The economics will shift, and the companies that engage with the technology early will be better positioned when it does become standard practice.
At Equans, we believe in testing on real projects, with real constraints. That's how we've built the knowledge we have today with partners who are pushing what's possible, with a significant impact on the sector. The productivity gains are promising. So are the benefits for the teams on the field: less physical strain, safer working conditions, and new skills to develop.
