Grasp and Prediction of Joint Resistivity in Sonic-Welding Process of REBCO Coated Conductors Based on Limited Numbers of Experimental Data

A data-efficient analytical approach was developed to quantitatively predict and control the joint resistivity in sonic-welded REBCO coated conductors. By employing the proposed Acoustic Energy Activation Bonding (AEAB) model, which describes an exponential dependence of bonding probability on the input acoustic energy, the AEAB model successfully reproduced the experimental data within approximately ±22% relative deviation. Even when only five distinct process conditions were used, the model achieved reasonably accurate prediction of joint resistivity (12.1 ± 2.1 nΩcm<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>) and identified optimal conditions realizing 9.8 nΩcm2 without degradation. This suggests that, unlike conventional empirical methods requiring extensive trials, the present approach provides a data-efficient framework for controlling joint resistivity from a limited dataset. The model was further applied to MOCVD-processed tapes, confirming its consistency and general applicability across different types.