ARTIFICIAL NEURAL NETWORK APPLICATIONS IN ARTHQUAKE PREDICTION
IEEE Student Member,
Department of electrical engineering
Faculty of Engineering and Technology,
Jamia University, New Delhi,.INDIA.
Earthquakes are undoubtedly most disastrous, unavoidable natural calamity on Earth. Seismological map of the Earth reveals that the epicenters of earthquakes form seismological belts that surround the whole planet. Most of the seismo-active belts lie in diversely populated areas within favorable climatic zones which shows that half of the population of Earth live under the permanent threat of potentially devastating earthquakes. The precarious condition arising due to earthquakes could be avoided only by making a reliable diagnostic to predict the location, magnitude and time of eminent earthquakes.
This paper presents Artificial neural network applications in the area of earthquake predictions. An electromagnetic model of fault is also proposed which make use of the piezoelectric effect and the elastic dislocation theory to investigate theoretically the spatial distribution of the stress induced charges associated with faulting. The concerned electric field associated with these induced charges can be estimated quantitatively. The amazing effect of this electric field variation in the atmosphere can be very well seen on animals, which start showing anomalous behavior at the time of earthquake because of their sensitive nature towards minor electric field exposure.
This paper is an effort to establish a relationship between the anomalous electromagnetic emission parameters and the earthquake parameters to understand the mechanism governing this process. One of the main goal of this paper is to systematize the enormous accumulated data and another is to draw the attention of the scientists towards the result obtained through the correlated parametric observations and calculations. To draw solid conclusions for a complex geophysical phenomenon like earthquakes, there is a need to systematize the results of the field experiments and to agree on a general approach to interpreting the observed behavior. Neural network classifiers are used to solve the purpose of finding a proper place for the parameters and effects under consideration along other geophysical phenomena. Trial networks were developed herein that act upon weather and atmospheric condition data to generate an opinion on earthquake activity historical earthquake data. . The consistency and reliability with which these assessments can be further improved because of neural network applications.
This paper presents a diagnosis of different parameters through the application of artificial neural networks. Although the volume of research involving non-traditional methods is constantly growing, we have to separately inter-link and observe the hydro-chemical parameters, atmospheric electrical disturbances, electromagnetic emission, anomalous disturbances in ionosphere and the magnetosphere to be recorded by satellites, seismic noise, and acoustic fluctuations. This paper not only presents a new model but also incorporates the most significant results, concepts and phenomenological model related to field research in chronological order wherever possible, cause many scientific articles published in this field even contradict common sense, wrong implementation or interpretation of the data may lead to a serious handicap to scientific outlook i.e., What is going to happen if we say an earthquake is going to come in N.Delhi in year 2000!!! , 20 million soles from N.Delhi can't be evacuated for one year. Its absolutely useless, the reaction time should be less and that is what on which the emphasis is laid in this paper.
(This is a student paper)