For the long-pulse high-confinement discharges in tokamaks, the equilibrium of plasma requires a contact with the first wall materials. The heat flux resulting from this interaction is of the order of 10 MW/m 2 for steady state conditions and up to 20 MW/m 2 for transient phases. The monitoring on surface temperatures of the plasma facing components (PFCs) is a major concern to ensure safe operation and to optimize performances of experimental operations on large fusion facilities. Furthermore, this measurement is also required to study the physics associated to the plasma material interactions and the heat flux deposition process. In tokamaks, infrared (IR) thermography systems are routinely used to monitor the surface temperature of the PFCs. This measurement requires an accurate knowledge of the surface emissivity. However, and particularly for metallic materials such as tungsten, this emissivity value can vary over a wide range with both the surface condition and the temperature itself, which makes instantaneous measurement challenging. In this context, the multi-spectral infrared method appears as a very promising alternative solution. Indeed, the system has the advantage to carry out a non-intrusive measurement on thermal radiation while evaluating surface temperature without requiring a mandatory surface emissivity measurement. In this paper, a conceptual design for the multi-spectral infrared thermography is proposed. The numerical study of the multi-channel system based on the Levenberg-Marquardt (LM) nonlinear curve fitting is applied. The numerical results presented in this paper demonstrate the design allows for measurements over a large temperature range with a relative error of less than 10%. Furthermore, laboratory experiments have been performed from 200 °C to 740 °C to confirm the feasibility for temperature measurements on stainless steel and tungsten. In these experiments, the unfolding results from the multi-channel detection provide good performance on temperature measurement, which supports our numerical evaluation and demonstrates the potential feasibility for metallic surface high temperature measurement with this method.
