Optimal Identification of Well Parameters Using Genetic Algorithm

Abstract

An effort has been made to optimally identify well parameters, viz. formation loss coefficient, entry loss coefficient, and turbulence coefficient in the framework of Genetic Algorithm (GA), which consequently resulted in the development of a computer model. To check validity of the model, results of the model are compared with the ones obtained through Jacob‟s graphical approach. Moreover, since minimization of the objective function „sum of square of errors (sse)‟ achieved through the GA approach have edge over the pertinent results obtained through Jacob‟s traditional method; further endorse validity of the model results. The model minimizes sse to 7.45 x 10-5 over 4.40 x 10-4 as computed by Jacob‟s method, signifying a percentage difference to the tune of 490. The model attains convergence smoothly except for minor level perturbations. Jacob‟s assumption for value of the turbulence loss coefficient is discarded, instead Rorabaugh and Todd‟s opinion is ascertained.