IDE 2015, Bremen, Germany, September 23 rd – 25 th , 2015 227 Determination of complex thermal boundary conditions using a Particle Swarm Optimization method Imre Felde 1 , Sándor Szénási 2 , Attila Kenéz 3 , Shi Wei 4 , Rafael Colas 5 1 Óbuda University, Bécsi út 96/B, H-1034, Budapest, Hungary, felde.imre@nik.uni-obuda.hu 2 Óbuda University, Bécsi út 96/B, H-1034, Budapest, Hungary, szenasi.sandor@nik.uni-obuda.hu 3 Hilti Werkzeug GmbH, 6000 Kecskemét, Hungary, attila.kenez@hilti.com 4 Tsinghua University, Beijing 100084, P.R. China, shiw@tsinghua.edu.cn 3 Universidad Autónoma de Nuevo León, San Nicolás de los Garza, Nuevo León, C.P. 66451, colas.rafael@gmail.com Abstract The methodology based on the Particle Swarm Optimization (PSO) method, as a recent stochastic optimization technique to solve complex inverse heat transfer problems is outlined. Temporal and spatial dependent Heat Transfer coefficient obtained on the surfaces of a cylindrical work piece is recovered by solving the inverse heat conduction problem. The fitness function to be minimized by the PSO approach is defined by the deviation of the measurements and the calculated temperatures is minimized. The PSO algorithm has been parallelized and implemented on a GPU architecture. Numerical results are demonstrated that the determination of Heat Transfer Coefficient functions can be performed by using the PSO method, as well as, the GPU implementation; provide a less time consuming and accurate estimation. Keywords Inverse Heat Conduction Problem, Heat Transfer Coefficients, Particle Swarm Optimization 1 Introduction Inverse heat conduction problems are known as “reverse engineering” problems, due to the reversal of a cause-effect sequence, in the field of heat transfer analysis. An inverse problem means that some of the initial, boundary conditions or material properties are not fully specified as determined from the measured temperature profiles at some specific locations. The inverse problems in most situations are likely to be ill-posed [Beck 1985]. Solutions of the inverse problem are very sensitive to measurement errors, i.e. small errors in the measured data values can produce very large errors in solutions. In general, the exclusivity and stability of an inverse problem solution is not guaranteed. In recent years, the inverse problems have been studied extensively due to their applications in various engineering disciplines. The most of the methods approach the inverse heat conduction problem, as an optimization problem, i.e. the problem is defined as the minimization of a cost function or a fitness function measuring the distance between measurements and predictions [Alifanov 1994, Özisik 2000]. With the improvement of computer capability, a variety of numerical techniques and computational methods have been developed to provide accurate solutions for inverse heat conduction problems (IHCP) in the last decade. Among these methods, stochastic optimization methods have become a popular means of solving inverse problems, due to their capability of finding the global optimal result without computing the complicated gradient of the objective function.