Lodz University of Technology and FABE to Build AI That Predicts Solar Panel Fires

Lodz University of Technology and the Lodz-based company Fabryka Bezpieczeństwa FABE (FABE Safety Factory) have announced a joint research project aimed at teaching artificial intelligence systems to recognize fire risk in photovoltaic installations and energy storage systems before the first visible signs of a fault appear. The venture is expected to be funded with around PLN 6.8 million, with the goal of producing technology ready for commercial deployment by FABE.

What Exactly Is Being Built

The university and the company want to move beyond the current approach to photovoltaic safety, which mostly relies on detecting failures only after they occur, for instance through smoke, temperature, or leakage current sensors. The new system is meant to continuously analyze operational data from panels, inverters, and energy storage units, picking up on subtle parameter deviations that precede damage before a device actually fails.

In practice, this means building predictive models trained on data from real installations, capable of recognizing patterns characteristic of degrading cells, connectors, or energy storage components. Lodz University of Technology is responsible for the scientific side and algorithm modeling, while FABE contributes engineering experience gained from building its own fire-prevention systems for power infrastructure.

The Scale of the Problem in Poland

The number of photovoltaic installations in Poland has been growing for years, and so has the number of fire incidents at buildings fitted with panels. According to data cited by the trade outlet Gramwzielone.pl, more than 300 fires at buildings with solar panels were recorded in 2025 alone, with over 100 more in the first quarter of 2026 alone. These statistics cover all fires at properties with PV installations, not only those where the installation was the direct cause of the blaze.

Root-cause analyses suggest that the biggest problem isn't the panels or inverters themselves but errors made during installation and component defects. The most commonly cited causes include careless or rushed installation, wiring short circuits, lightning strikes, and improper system disconnection during a fault, which makes safe intervention harder for firefighters.

FABE's Role

FABE is no newcomer to this field. The Lodz-based company previously carried out an EU-funded project in which it developed a smart fire-prevention system for electric vehicle charging stations, sold under the name estcharge. The solution continuously assesses a system's fire risk and generates preventive recommendations before a failure occurs.

The new project with Lodz University of Technology aims to carry the same logic over to photovoltaic installations and their accompanying energy storage systems, a market segment that is growing much faster than EV charging infrastructure and generating proportionally more insurance claims.

What This Means for the Market

For installation companies and solar farm operators, early risk detection could mean lower insurance costs and faster service response before a minor fault turns into a costly failure or fire. The rising number of fire incidents at installations with energy storage systems is already prompting insurers to tighten policy terms, so predictive systems could become a bargaining chip in rate negotiations.

For Lodz University of Technology, the project fits into a broader trend of commercializing AI research in partnership with local business, alongside other university initiatives in photovoltaics and materials for solar energy.

The parties have not yet disclosed a precise timeline for rolling out the finished product or a planned completion date for the research work. The next step will likely be building a prototype and testing it on real installations before the system is added to FABE's offering as a commercial tool for PV farm and energy storage operators.