A BEM-based genetic algorithm for identification of polarization curves in cathodic protection systems

The purpose of this work is to apply an inverse boundary element formulation in order to develop efficient algorithms for identification of polarization curves in a cathodic protection system. The problem is to minimize an objective function measuring the difference between observed and BEM-predicted surface potentials. The numerical formulation is based on the application of genetic algorithms, which are robust search techniques emulating the natural process of evolution as a means of progressing towards an optimum solution. Examples of application are included in the paper for different types of polarization curves in finite and infinite electrolytes. The accuracy and efficiency of the numerical results are verified by comparison with standard conjugate gradient techniques. As a result of this research, the genetic algorithm approach is shown to be more robust, independent of the position of the sensors and of initial guesses, and will be further developed for three-dimensional applications. Copyright (C) 2002 John Wiley Sons, Ltd.