Open Collections

Atkinson, Stephen Hamilton

University of British Columbia

Thermoplastic composites are increasingly being used in aerospace applications where a significant benefit is the ability to implement fusion bonding techniques (welding). Continuous resistance welding (CRW), a promising form of fusion bonding, uses a movable apparatus to incrementally heat sections of a resistive implant placed between parts by imposing electrical current and pressure to melt and fuse the interface. Precise temperature control is necessary to melt the polymer matrix while avoiding degradation however, direct in-situ measurement is invasive. Additionally, parameters such as boundary conditions, substructure properties, or part geometry may vary along the length of the weld. Therefore, a physics-based simulation framework using a systems definition was developed to determine appropriate pressure, power, and speed inputs required to achieve optimal results. Previously validated melt/crystallization kinetic models for the PEEK thermoplastic matrix were used to determine a quality metric for weld simulations based on prediction of melting and degradation. Analyzing the model outputs revealed a high degree of non-uniform heating both across the width (edge effects) and length of the weld with significant pre-heating providing insight into complications for potential control schemes. A parametric study was conducted and found the electrical parameters to be highly sensitive to thermal behaviour, especially the contact resistance of the electrode blocks. However, the thermal materials and boundary conditions were generally less sensitive or insensitive, including latent heat of melting allowing for decoupling of melting kinetics to simplify model implementation. Demonstration of the digital twin using the melt prediction quality metric was shown to successfully mitigate typical processing complications including boundary conditions, material changes, and a process stoppage. Future work is recommended to mitigate the non-uniform heating and uncertainty of the electrical parameters as well as utilize the framework to develop reduced order models to achieve accurate, fast simulation predictions for real-time control of the welding process.

Text

2024-03-11

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/87549

Civil Engineering

University of British Columbia

UBCV

http://creativecommons.org/licenses/by-nc-nd/4.0/