Accurate Estimation of the Well Test Parameters By Using a Hybrid Algorithm and Comparing It With Conventional Industrial Software

Permeability, skin factor, and wellbore storage coefficient are three essential parameters for well and reservoir. Moreover, well testing is the conventional method for determining these parameters. In this method, pressure data are plotted versus time in semilogarithmic and logarithmic scales, and specified relations determine well test parameters. In this study, a new approach is introduced to obtain well test parameters using artificial intelligence. First, using endpoints related to two real wells, pressure data are produced by reservoir simulator software. Because the data have a lot of noises, the additional data are removed, and analyzing the data is more comfortable. Then, with the combination of a genetic algorithm and the Levenberg Marquardt algorithm, the basic parameters of the reservoir have been calculated. In the last step, the pressure data are entered into standard software, and their values are calculated. Finally, the combined algorithm with excellent accuracy has been able to calculate these parameters in comparison with the conventional Saphir 4.10 well testing software.