A thermal efficiency analysis of a Gas Tungsten Arch Welding process using a temperature moving sensor

Abstract This works proposes an Inverse methodology to estimate the thermal efficiency of a Gas Tungsten Arch welding (GTAW) process as a time-dependent function. The direct model consists in solving the non-linear three-dimensional heat diffusion equation through the Finite Difference Method in a C++ code. As the inverse methodology, the Golden Section optimization technique was used together with the Time Travelling Regularization to estimate the applied heat rate in the GTAW process. The Time Travelling Regularization was used to reduce the noise on the estimated heat input. Furthermore, a numerical Temperature moving sensor is proposed to determine the heat input as a time-dependent function. The methodology was applied in lab-controlled experiments based on a Robust Project matrix. The numerical analysis revealed that the thermal efficiency decreased as the arch torch moves. The proposed methodology also presented a higher sensitivity than the usual methods for the heat rate estimation and is a cheaper way to determine the melting efficiency in welding processes.