Determining model parameter from self-potential data using quantum-behaved particle swarm optimization

Abstract

A new approach for quantitative analysis of self-potential (SP) data is introduced. In this paper, anomaly of SP is associated with simple geometric models such as a vertical cylinder, a horizontal cylinder and a sphere object. Then, in order to estimate the depth, the electric dipole moment, the anomaly body’s centre, the geometrical form factor and polarization of the anomaly, the method was developed and implemented. The development and implementation of the method is based on the global optimization concept. This method uses Quantum-behaved Particle Swarm Optimization (QPSO) algorithm to overcome the inversion problem on SP anomaly modelling. The QPSO algorithm was randomly tested on synthetic data which consist of different random noise levels. The result shows a close agreement between the assumed and the measured parameters. At last, the validity of the method was tested on real SP anomaly data and compared to the results given by other advanced inversion approaches.