Estimation of earthquake performances of reinforced concrete residential buildings by artificial neural network rapid assesment method

Depending on the geological features and geographical location, Turkey stands in an area where there are experienced earthquakes frequently. Considering a possible earthquake, one of the most important points to focus on to minimize the loss of life and property is the evaluation of existing building stock. The rules which are the basis for the evaluation of the existing buildings in the building stock have entered into the Turkish Earthquake Regulation (TDY) published in 2007 as a new concept. According to these rules, it is required to carry out detailed investigations such as; surveying in reinforced concrete structures, coring for determination of material strengths and structural analysis in the existing structure evaluations. If the high amount of existing structures in building stock is taken into account, it is obvious that too much labor, time and funds are required for the analyses stated in the earthquake regulation. Considering the possibility of a devastating earthquake at any moment, it is important that the existing building stock needs to be evaluated in a way that quickly and cost-efficiently. Many rapid evaluation methods have been developed with many studies, however, the rapid assessment of the existing building stock still remains a necessity. With this study, it is aimed to develop a more practical and efficient method of rapid assessment than the current evaluation methods with minimum number of data input and a new method has been developed for residential buildings up to 8 stories with Artificial Neural Network (ANN) models. To determine the building performance classes precisely, the structural analysis of 402 different structure models were performed by the method of non-linear mode combination and the ANN models were trained according to these results. Verification was made with existing buildings and it has been seen that; the earthquake performance classes of the buildings can satisfactorily be determined with the method developed in this study. Keywords: (RC Residential Buildings, ANN, Earthquake Performance, Rapid Assessment Method)