ENG

Introduction

Meccanica del Sarca S.p.A is an Italian metalworking company specialising in the production of walnut components for rifle stocks and sheet metal magazines for pistols.

The problem

Quality control of the production lines for loading tanks is currently carried out entirely manually, using specially designed gauges. This activity, performed by specialised operators, involves the aesthetic and dimensional verification of each individual piece, with any necessary corrective interventions carried out by means of plastic deformation. However, the process has several critical issues: it is highly repetitive, requires specific skills that are not always transferable, and is subject to margins of error linked to operator subjectivity. Furthermore, the time required for inspection has a negative impact on the efficiency of the production cycle, limiting productivity and the ability to respond quickly to market demands.

Technologies

Artificial intelligence and deep learning used for automatic identification of non-conformities in production (defects).

Sensoristica: sensori ottici analizzano il profilo del componente, per determinare se sia conforme alle specifiche di progetto.

Impacts

The introduction of an automated measuring station supported by artificial intelligence algorithms allows the company to significantly improve its quality control system. Through constant monitoring of defects and structured data collection, it is possible to identify and correct any deviations in the production processes in a timely manner. This approach reduces dependence on subjective operator assessments, increasing overall control reliability and promoting greater decision-making autonomy. The internal impact translates into a corporate culture that is more focused on quality, traceability and continuous improvement.

Benefits for the company

– Automatically identify the compliance/non-compliance of a tank.

– Automatically indicate the possibility of repair for non-compliant products.

– Indicate repair activities for repairable products.

– Have a system capable of automatically learning knowledge based on the operator’s alternative choices.