Accurately Predicting Spring-Clip Stress Using AI Data Analytics｜physicsAI

 Software used : physicsAI 

【 Background 】

       Inside an electronic connector slot, signals are transmitted through the contact of spring pins. The contact force provided by the spring pins directly affects the connector’s ability to transmit electrical signals. If the contact force is too small, contact resistance increases; if it is too large, it may cause excessive mating/unmating force or other issues. Therefore, spring-pin design in electronic connectors often requires repeated refinement, consuming a great deal of time and cost.
       Altair provides a complete AI big data solution, including: AI data analytics Rapid Miner, the integration of AI and CAE analysis with physicsAI, and ExpertAI. In this case, we use physicsAI to build a spring-pin stress prediction model, allowing rapid evaluation of different spring-pin designs.
 

【 Results 】

Building an AI prediction model
​​​       Using historical CAE analysis results from different spring-pin designs, we apply physicsAI to perform AI big data training and build an AI prediction model to estimate the stress of newly designed spring pins.

Fast prediction of analysis results
        By replacing repeated CAE re-modeling and analysis, we reduce 1–2 hours of work time down to just a few seconds, significantly shortening development lead time.

High accuracy
       The stress distribution predicted by the AI model and the CAE analysis results show a difference of less than 3%.

【 Technical Features 】 

• Using physicsAI deep geometric learning technology, past design and analysis results are used as training data to train the AI model.

• For new design configurations, providing only the CAD or FEM model is sufficient to directly obtain predictions using the already well-trained AI model.

 

Further reading:  AI big data-driven CAE result prediction for bicycle forks｜physicsAI