INTEGRAL METHOD FROM INITIAL VALUES TO OBTAIN THE BEST FIT OF THE GAROFALOS CREEP EQUATION

An integral computational method has been developed to provide initial values to a subsequent fitting of creep data based on non-linear and iterative methods. The fitting of the Garofalo's equation, which descr ibes creep data, utilizes a strong non-linear objective function. Vari ous algorithms are developed which provide the solution for the fittin g. An analysis is made of the significance of this solution in relatio n to the problem studied. The dispersion of the solution values is rel ated to the final values provided by other non-linear and iterative me thods. The method is applied to three types of steels and screens the experimental creep data with a statistical analysis that improves the fitting results. (C) 1998 Elsevier Science S.A. All rights reserved.