Wyniki wyszukiwania - BazTech

Ograniczanie wyników

1 Acta Geophysica Polonica

1 2005

Znaleziono wyników: 1

Liczba wyników na stronie

Wyniki wyszukiwania

Sortuj według:

Ogranicz wyniki do:

Prediction of seismicity cycles in the Himalayas using artificial neural network

Sharma M. L., Arora M. K.

Acta Geophysica Polonica

2005

Vol. 53, nr 3

299-309

In this paper, cyclic behaviour of seismicity cycles in the Himalayas has been exploited to predict the future earthquake activity using Artificial Neural Network (ANN). The Himalayan region has been divided into six seismogenic zones. A feed forward multi-layer ANN has been used to evaluate the seismicity fluctuation in the time series containing data from historical times to 1998 for each zone. The most widely used Back Propagation Algorithm (BPA) is applied to train the neural network. BPA iteratively minimises an error function over the network outputs and a set of target outputs taken from the training data set. The results show that the probability of occurrence of moderate to great earthquake in next 50 years is relatively lower in the Hindukush-Pamirs zone. Since the intense release of energy will take place in the Kashmir-Himachal Pradesh zone, between 2030 to 2055, the probability of occurrence of moderate to great earth-quake is higher. The accumulation of energy stage is still going on in the India -Western Nepal Border zone, and there will be an increase in seismic activity after 2030 for the next 50 years. The hazard parameters could not be estimated for the Nepal-India-Sikkim Border zone because of lesser number of data to capture cyclic behaviour. In NE India, intense release and remnant release will take place up to 2030 due to which there will be an increase in the probability of occurrence of moderate to great earthquake in this zone. In Burma-Andaman Nicobar, the energy accumulation stage for the next cycle has started in 1990 and will continue till 2020.