AI Slashes Product Testing Time from Seven Months to Five Days

In a significant leap for product development, Slovenian researchers have harnessed artificial intelligence (AI) and advanced modeling to slash testing times from an arduous seven months down to a remarkable five days. This innovation promises to revolutionize how products are brought to market.

Gregor Schoss, director of the Slovenian Institute for Quality and Metrology (SIQ Ljubljana), explained the core of the breakthrough. Manufacturers can now generate accurate digital models of products. These models are then used for simulations and testing before physical prototypes are even created. "This allows for the identification and elimination of errors in the early stages of development, significantly reducing testing time and costs," Schoss stated.

The "five-day project," initiated in November, saw a team from SIQ Ljubljana collaborate with Inpotron, a partner of 20 years. Utilizing AI, they successfully reduced the testing period for a power supply unit from approximately 240 days to just five. The team, including Primož Prijatelj, Rok Agnič, Primož Lesjak, and Mitja Rozman, used precise simulations to rapidly identify and fix flaws, enabling a faster market entry and lower development expenses.

"Such cases demonstrate how the use of advanced technologies can substantially improve the efficiency and competitiveness of companies in today's rapidly evolving market environment," Schoss noted. Shortening the development cycle allows for quicker product launches, potentially boosting company revenues significantly.

The process involved establishing a simulation model on a less complex power supply, allowing the AI to learn. This method was then applied to a more intricate power supply, yielding a usable model for the development phase. The digital model allows for precise checks of distances and critical areas, even those imperceptible to the naked eye. This meticulous approach, integrating electrical and mechanical component properties, helps identify potential failures and optimize testing sequences, thereby reducing the need for costly and scarce development samples.

“

To omogoča prepoznavanje in odpravljanje napak že v zgodnjih fazah razvoja, kar precej skrajša čas in stroške testiranja.