Chapter 5: Applications

Title: Neuro-fuzzy technology for the detection of weak seismo-electric signals

Author(s): Antony J. Konstantaras, Filippos Vallianatos, John P. Makris, Martin R. Varley, Gerry Collins, Phil Holifield

Address: Technological Educational Institute of Crete, Chania, Crete, 731 00, Greece | Technological Educational Institute of Crete, Chania, Crete, 731 00, Greece | Technological Educational Institute of Crete, Chania, Crete, 731 00, Greece | ADSIP Research Centre, University of Central Lancashire, Preston, PR1 2HE, UK | ADSIP Research Centre, University of Central Lancashire, Preston, PR1 2HE, UK

Reference: Atlantic Europe Conference on Remote Imaging and Spectroscopy pp. 151 - 157

Abstract/Summary: This paper presents the development and application of adaptive filters based upon prediction with neuro-fuzzy models (Konstantaras et. al., 2006) for the enhancement and detection of weak electro-telluric potential anomalies appearing upon recordings of the Earth’s electric field. These signals are most commonly known as electric earthquake precursors (EEPs) (Tzanis and Vallianatos, 2001) and they are believed to be related to forthcoming seismic events. In most cases EEPs are considerably weaker than the Earth’s electric field, and their detection can be further complicated by the probable appearance of severe additional distortions induced mainly by magnetic storms, and/or other physical and anthropogenic types of noise (Varotsos and and Alexopoulos, 1984). In this application, two neuro-fuzzy models are trained to learn and predict the recorded magnetic and electric fields, respectively. Minimization of the effect upon the electric field signal due to severe magnetotelluric distortions is achieved by the first neurofuzzy model which predicts the variations of the Earth’s permanent magnetic field, thereby ignoring any extrinsic influences. The processed magnetic field signal is then incorporated by the magnetotelluric method to generate the equivalent electric field. The main goals of this method are to present, enhance, and detect EEP signals upon the error signal, defined as the difference between the recorded and the predicted by the second neuro-fuzzy model electric field, because the latter have been extrinsically added on and are not a genuine part of the electric field.

Order a copy of this article