A 2-d inverse heat conduction formulation for determination of heat source characteristics from thermal images

The inverse heat conduction problems have shown to be ill-posed in the presence of measurement noise and this introduces the issue of regularization. A method of regularization for the two dimensional equation based on the maximum entropy method is suggested here. A novel, problem specific input parameterization scheme is introduced which leads to considerable reduction in computational burden. Preliminary results comparing the solutions both with un-regularized solutions are presented for various inputs and simulated measurement errors. Results comparing computational time with current existing maximum entropy methods, without the parameterization, are also presented.