Artificial intelligence can accelerate the planning, construction and operation of photovoltaic systems, as evidenced by the recent Solar Quality Summit in Barcelona, Spain. But its effectiveness depends on data quality, as poor data, unclear responsibilities and weak documentation can become major cost drivers.
“Are you ready for artificial intelligence?” With this question, David Moser, head of the Becquerel Institute Italy, opened the recent Solar Quality Summit in Barcelona.
The two-day conference, attended by 250 representatives from project development, operations, financing and monitoring, focused more intensively than ever on AI. The event, organized by SolarPower Europe and Intersolar Europe, made AI the central theme of almost every session, highlighting its role in faster evaluation of business models, accelerated planning, improved forecasting and more accurate verification of construction quality.
At the start, organizers showed a video in which participants described how they are already integrating AI into their daily work. The examples focused on productivity tools, analogous to the shift from letters to email. Users can work faster and make decisions based on a broader data base, while those who ignore AI risk being left behind. A key challenge: Software solutions developed at high cost to differentiate businesses can be quickly replicated by competitors using AI at a fraction of the cost, potentially turning some risky investments into wasted investments.
The most repeated phrase at the summit was “Garbage in, garbage out.” As decisions become more data-driven, poor data quality can quickly become costly. AI not only makes mistakes – which all participants acknowledged – but can also reproduce them quickly and on a large scale. “You can’t leave your brain at the door,” one participant noted. The conclusion: AI results need to be validated; without human responsibility, AI does not eliminate risk.
In a panel on AI in quality management and operations, participants discussed its application in photovoltaic projects. The public associated AI with automation, predictions and data analysis, but the benefits depend on high-quality structured data.
Sensors alone are not enough; responsibility is just as important. A public survey illustrated this: five days before handover of a ground-mounted PV system, a defect is discovered that could cause problems within three to five years. Should construction be halted, formally reported, ignored, or informally passed on to O&M? Most participants considered stopping construction unlikely. One panelist wryly noted that he had never seen such a scenario treated in that way. Formally passing on responsibility to operational management seemed more plausible. Time and budget pressures, along with fragmented responsibilities, often allow defects to trickle down.
AI could help alleviate these problems. Systematically recording construction and quality data – including location, evidence, severity and status – makes undetected problems harder to ignore. Drone images compared to digital twins can reveal abnormalities early. Yet the reality remains challenging: checklists are often filled with desired values instead of actual measurements to save time, multimeter readings are sent via WhatsApp instead of properly documented, and daily reports are compiled weekly, causing details to be lost. Humans, not just AI, make mistakes.
Most participants agreed that AI has the potential to create value for the industry, although measuring it is not easy. In the applications discussed, AI prevents losses instead of generating revenue, making prevented errors difficult to quantify. It can reduce rework, increase standardization, and shift corrections to the planning stage, where costs are lower. One panelist described AI as a tool to identify deficiencies in the discipline early, eliminating long-term costs rather than ensuring quality.
Cyber security concerns
Cybersecurity was the second major topic of the summit. Recent incidents, such as the suspected Russian hack of power stations in Poland, shaped the discussion. As photovoltaic systems become more interconnected – from inverters, SCADA, park controllers and cloud monitoring – the attack surface expands. New European regulations – NIS2, the Cybersecurity Act, the Cyber Resilience Act and the Network Code on Cybersecurity – require compliance, but are highly technical and inconsistent between countries. In-depth cybersecurity expertise remains scarce.
Two points were emphasized: cybersecurity is an ongoing process and not a one-time checklist. Patching, updates, monitoring and incident response are continuous tasks. Geopolitics adds risk: Networks involve manufacturers, software and cloud infrastructure with different vulnerabilities. Simply avoiding hardware from a particular manufacturer or country does not guarantee security; In the Poland incident, hackers gained access to European and Japanese controllers with weak passwords and missing updates. Yet excessive strategic dependencies should be avoided, and early integration of cybersecurity controls is essential.
The Solar Quality Summit showed that the industry is entering a new phase. AI can speed up processes, prioritize risks and detect errors earlier, but it also increases weaknesses when data, documentation and accountability are not clearly defined. As the summit’s unofficial motto put it: “Garbage in, garbage out.”
This content is copyrighted and may not be reused. If you would like to collaborate with us and reuse some of our content, please contact: editors@pv-magazine.com.
