COMPARISON OF SEISMIC STRUCTURAL RESPONSES IN DIFFERENT STORIES VIA FUZZY CLUSTERING

Determining differences and similarities of earthquake records is one of the challenges ahead for dynamic analysis. Here, clustering is used as a branch of machine learning as it can carry out the task without supervision. On the other hand, in addition to earthquakes, properties of the seismic response of the structure are also important for such a choice. Fuzzy membership functions are used to compare the clustering of earthquakes and structural responses. To this end, with the implementation of the fuzzy means algorithm, earthquake records have been compared before and after structural analysis. Seismic responses are obtained at the center of mass roof 2-story structure, center of mass third floor and roof 6-story structure, center of mass fifth floor and roof 10-story structure. Time history analysis is performed by OpenSees software, and then clustering is done by MATLAB functions. For interpretation of the results, a comparison is made between the degrees of membership the earthquake records and the resulting responses to each cluster

Keywords:

Data mining , Earthquake records , Structural response , Fuzzy logic