The project developed a vision algorithm to detect anomalies in processing in real time, preventing fire risks and supporting timely and targeted interventions.
Meccanica del Sarca
Cefriel
Meccanica del Sarca specializes in the mechanical processing of walnut wood for the production of rifle stocks, forends, and magazines.
One of the processes typically carried out by Meccanica del Sarca is woodworking using machine tools. These tools are enclosed and cause accumulations of shavings and dust, generating a high fire risk within the plant. Inside these machines, the so-called ‘fire triangle’ is always present, i.e. the three conditions that can trigger a fire are always met: the simultaneous presence of fuel, an oxidiser and an ignition source. This is typically due to an increase in cutting effort caused by contact between the tool and the wood, non-wood inclusions within the matrix or high-density areas within the wood. This leads to a localised increase in heat, which can potentially lead to a devastating fire.
For this reason, the company is at high risk of fire during the mechanical processing, also because the fire and smoke/heat detection sensors typically available on the market have not proven to be effective.
The solution lies in developing an algorithm that can analyse images from a standard HD camera installed inside machine tools in real time, identifying processing anomalies that could potentially lead to a fire. Such an algorithm would guarantee the immediate reporting of critical issues, allowing for an intervention strategy that can vary depending on the level of risk and the specific organisational structure of the company.
Artificial intelligence applied to computer vision techniques used for automatic, real-time identification of sparks that could trigger a fire.
Intelligent continuous monitoring system that uses standard optical cameras positioned on the machines to automatically detect fire hazards in all operating conditions and generate an alarm signal.
The creation of an artificial intelligence algorithm allows a traditional camera to be equipped with additional features: identification and classification of hazardous and risky conditions, appropriately catalogued for Meccanica del Sarca’s operations.
The integration between the connectivity features offered by the camera and those offered by the AI algorithm allows for real-time detection, identification, and recognition.
The integration and connection of the described solution with the alarm system allows for the creation of an intelligent real-time monitoring system for the working conditions of the woodworking process, enabling the company to carry out unattended operations.
Improved risk management: based on hazard identification, the most suitable solution to mitigate the detected hazard can be identified.
Reduction of damage caused by fires inside the plant and related machine downtime: it is estimated that this system can reduce the economic impact of potential fire damage by 90%.
Implementation of an intelligent solution that goes beyond the functionality offered by existing market solutions. There is also the possibility of protecting the results obtained from the prototype solution through industrial property rights, which is the subject of the innovation project.
