Page 22 - 2019

temperature. The inverse analysis for solving heat transfer problems to obtain accurate
prediction of thermophysical quantities is one of the efficient methods. The boundary
conditions are customized to resemble to the real life problem of interest. The numerical
solution to the two-dimensional heat convection problem with turbulence is obtained
using inverse heat transfer concept. Prandtl Mixing length theory turbulence model and
K- turbulence model incorporated in the solution and compared. In this optimization
problem a squared residue functional is minimized with the conjugate gradient method. A
sensitivity problem is solved to determine the step size in the direction of descent, and an
adjoint problem is solved to determine the gradient of the functional. Tikhonov
regularization approach is second-hand to promote smoothness to the solution of
problem.
The proposed methodology makes use of inverse heat convection approach. In heat
transfer through convection boundary, the surface temperature is not known. The
simulated temperature values at the measurement locations are used to minimize the
cost function iteratively. The unsteady heat source function is estimated accurately in an
iterative manner. The accuracy of the inverse analysis is examined by using simulated
exact and inexact temperature measurements. The solution to the convection problem
provides temperature profile at the thermal boundary of DPHE. The final results showed
low noise and the algorithm is robust in multiple trials. The estimation of the unsteady
time varying function is agreeable. The temperature contours authenticate the heat
convection process reliability in the annulus domain.

Keywords: CFD, Insert, Conjugate Gradient Method, Heat Recovery, Inverse Analysis,
Double Pipe Heat Exchanger.

03