2-DIMENSIONAL STEADY-STATE INVERSE HEAT-CONDUCTION PROBLEMS

In this work we estimate the surface temperature in two dimensional st eady-state in a rectangular region by two different methods, the singu lar value decomposition (SVD) with boundary element method (BEM) and t he least-squares approach with integral transform method (ITM). The BE M method is efficient for solving inverse heat conduction problems (IH CP) because only the boundary of the region needs to be discretized. F urthermore, both temperature and heat flux at the unknown boundary are estimated at the same time. The least-squares technique involves solv ing the equations constructed from the measured temperature and the ex act solution. The measured data are simulated by adding random errors to the exact solution of the direct problem. The effects of random err ors on the accuracy of the predictions are examined. The sensitivity c oefficients are also presented to illustrate the effect of sensor loca tion on the estimated surface conditions. Numerical experiments are gi ven to demonstrate the accuracy of the present approaches. (C) 1998 El sevier Science Ltd.