Data-driven 3c assembly process mechanism correlation mining and parameter real-time tuning method

Abstract

This invention relates to a data-driven method for mining the correlation between the mechanism and parameters of 3C assembly processes, belonging to the field of intelligent manufacturing technology. The method includes the following steps: online acquisition and storage of clamping torque values ​​using sensors; preprocessing the data to obtain torque-rotation time-series data with the same feature dimensions; expanding the sample size of existing clamping failure experimental data using a denoising probability diffusion model; capturing hidden information in the failure torque experimental data using a Mamba model to build a torque threshold prediction model; and optimizing the torque threshold for different objects based on the prediction model. This invention effectively captures the hidden process information in the torque data, ensuring that the 3C assembly process can optimize the target torque data in real time according to different individuals, thus guaranteeing product quality during the clamping process in 3C production.

Images