Neurofuzzy prediction model for aseismic ability of buildings

Depending on human judgement, the current approach carried out to assess earthquake-damaged buildings' aseismic ability is time-consuming and vulnerable. The paper emphasizes on the development of an automatic model to assess the aseismic ability of buildings using neurofuzzy. The methodology is to collect data of damaged buildings destruction in both categorized earthquake regions of Taiwan. Statistically analyzing collected data to identify significant factors by analysis of variance and Chi-square test is conducted, resulting that the factors are earthquake zone, coefficient of usage, strength of concrete, number of building floor(s), and residual life. Using a hybrid model of fuzzy logical control and artificial neural network methods for building aseismic ability is subsequently completed to moderate the impacts by human effects and to facilitate the inspection procedure. With the help of the neurofuzzy prediction model (NPM), destruction degree of buildings can be rapidly assessed.