This master thesis research investigates the possibility to repair carbon fiber thermoplastic aircraft structures using induction welding. Carbon fiber is conductive and heats up when placed inside an alternating magnetic field. A generator, coil, and pressure frame are needed to perform induction welding. The support plates needed to pressurize the carbon fiber parts need to be non-conductive, non-magnetic, temperature resistant, stiff at high temperatures, and thermally insulating. The material used is five harness satin weave carbon fiber PPS supplied in pre-consolidated plates and unconsolidated semi-preg. Three different joint geometries are used in this investigation: a conventional scarf, a continuous scarf, and a stepped lap joint. The stepped lap joint and the conventional scarf are milled using a CNC machine. The continuous scarf is produced in a press using specialized tooling and a press program prescribed by TenCate. Of each type two specimens are produced: an induction welded specimen and a press joined specimen. The specimens are tested in tension to determine the tensile strength and stiffness. This gives the performance of the induction welded joints with respect to the press joined specimens. Additionally, both of the welded specimens are compared to the pristine specimens. By measuring the temperature along the weld-line of multiple test specimens, an induction welding program is obtained for each joint type. Achieving a consistent temperature along the weld-line is challenging due to the thermal conductivity of carbon fibers and other effects inside the laminate. Similar to the continuous scarfed specimens, the press joined specimens are created using specialized tooling. Before testing the specimens in tension, they are scanned using a C-scan. The press bonded specimens show no flaws, whereas the induction welded specimens do not return the signal to the transducer. Micrography of the specimens shows similar results as the C-scan. The press joined specimens show little to no flaws and the induction welded specimens show a significant amount of voids and small unjoined sections near the tips for the continuous scarf and stepped lap specimen. The tensile tests show that all induction welded joints perform less than the press joined specimens with a percentage of about 73-78%. The conventional scarf and the continuous scarf result in a similar failure strength recovery of about 44% of the pristine failure strength for the press joined specimens. Both the conventional scarf and the continuous scarf show similar behavior in welding and the tensile testing, but the continuous scarf is more difficult to produce. Therefore the conventional scarf joint is preferred over the continuous scarf joint. The stepped lap joint has the best performance, recovering about 59% of the pristine stiffness. The stiffness of all tested specimens is between 92%-99% of the pristine stiffness. Inspection using digital image correlation and finite element modeling indicates the failure initiates inside the PPS matrix in between the parts at the location of the first or last step. Finite element modeling also suggests that a smaller angle for the scarf joint or a longer overlap for the stepped lap joint would increase failure strength.