Research on geometric inverse problem based on the firefly conjugate gradient method

In this paper the two-dimensional steady-state heat transfer geometric inverse problem is solved using Finite Element Method (FEM), Conjugate Gradient (CGM), and Firefly Algorithm (FA). Based on finite element method for solving forward model, based continuous iterative optimisation of conjugate gradient method, accuracy error function measured estimated values kept within a certain range, so that geometry object under test can be calculated in an way. study problem, temperature field distribution circular pipes as research objects, feasibility applying to problems verified. The takes tube considers two different corrosion defects inner wall tube. Simultaneously, firefly algorithm introduced stochastically optimize data suppress fluctuation inversion result. Numerical experiment results indicate perform more accurate shape recognition when there measurement or information relatively complete.