Premium
The study of base isolation on the precise machinery system for regional ground motion records with modified back propagation neural network approach
Author(s) -
Wan S.,
Yen J. Y.
Publication year - 2007
Publication title -
structural control and health monitoring
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.587
H-Index - 62
eISSN - 1545-2263
pISSN - 1545-2255
DOI - 10.1002/stc.177
Subject(s) - base isolation , artificial neural network , ground motion , isolation (microbiology) , computation , base (topology) , motion (physics) , engineering , bearing (navigation) , strong ground motion , bar (unit) , backpropagation , vibration , computer science , structural engineering , artificial intelligence , algorithm , mechanical engineering , mathematics , geology , mathematical analysis , microbiology and biotechnology , frame (networking) , biology , oceanography , physics , quantum mechanics
The machinery industry of central part of Taiwan suffered a major catastrophe earthquake on 21 September 1999. This strong ground motion introduced a great deal of economic loss for Taiwan. To reduce the impact of earthquake to the precise machinery systems, the related research of the developments on the device for decreasing earthquake excitation loads should be highly emphasized. Limited by solution techniques, numerous papers have been written on the solution for reducing structural vibration which is only considered on single structural equipment with base isolation system subjected to a given ground motion. Little has been done to the parallel study of overall regional records to find and implement influence factors on the behaviour of various structures and the design of new buildings. In the past, the structural response under a given earthquake was obtained by step‐by‐step integration which was very time consuming. A learning machine, based on neuron cell concept was developed to investigate the relationship of massive given inputs and certain outputs in this study. Back Propagation Neural Network (BPN), a well‐known learning machine was used as a predictor for the simulation. The BPN takes the advantage of extended delta bar delta (EDBD) algorithm by fast‐convergency and low‐error characteristics for computation. The influence factors for the lead‐rubber bearing (LRB) devices are solved based on the regional ground motions. Further, a better understanding on the general response to the base isolation design problems may even bring about brilliant concepts on the design in the future. Copyright © 2006 John Wiley & Sons, Ltd.