Optimal Heat Transfer Design of an Ocean Turbine Pressure Vessel using Soft Computing

摘要

This paper presents a numerical optimization approach applicableto heat transfer design. A novel design procedure for theprediction of heat transfer inside a pressure vessel of an oceancurrent turbine using the finite element method of heat transferanalysis, artificial neural networks and genetic algorithms is presented.Numerical heat transfer analysis was done using commercialsoftware ANSYS for two-dimensional heat transfer in simplifieddomains. Computation was confined to heat conduction. TheANSYS simulations results were then used for training and approximatingthe unknown functional behavior of heat transfer byusing artificial neural networks (ANN). The trained ANN servesas the nonlinear objective function of the optimization procedure.Genetic algorithms (GA) were finally employed as the optimizationtool. The optimum results obtained from the GA were verifiedagainst both ANSYS and ANN results. Both the ANN andGA were implemented in MATLAB environment. The overallmethodology application was in effect validated by the results satisfactoryfor a specific ocean current turbine application.