Earthquake Risk Assessment Using Fuzzy Inference Systems and its Application in Seismic Rehabilitation Studies

One of the important problems in seismic rehabilitation studies of existing structures is opportune decision making about ending or continuance of various stages rehabilitation in order to save time and cost. About that we can use decision maker systems to solve this problem and to give more rational assessment about that problem. This paper presents a procedure based on Fuzzy Logic that classifies structures into qualitative seismic hazard categories. The purpose of this study is to get a model that can speed existing structures seismic rehabilitation primary studies and also to prompt decision making about continuance of study process. In order to account real world data, in addition to expert’s knowledge, groups of school seismic rehabilitation data of different cities of Iran have been used for modeling. In order to reduce the input space and increase generalization ability of the system, a feature selection method has been applied to the data. Among available parameters of data, significant parameters have been selected by Decision Tree Learning method. Then, Fuzzy Membership Functions corresponding to these parameters have been defined. Appropriate defining of these functions, we can insinuate factors such as uncertainty on that parameter in computations also. Afterwards, the Fuzzy System has been designed by conditional regulations. It is worth to say that these regulations are optimizedcompletely. In order to ease the process of risk assessment based on this model, software named “Rapid Seismic Risk Evaluation” (RSRE) has been developed. Thus, we have a model that by inputting 7 entrance parameters of a structure (both structural and geotechnical parameters corresponding to existing structure), generates its seismic risk level. The proposed procedure has advantages among the rest we can recount the possibility of modeling uncertainties, inputting structural information qualitative and high speed of risk analysis process. It is clear that using Fuzzy Logic not only lead to more simple formation, but also speed the rate of risk analysis process intensely, that this case is one of the most important advantages of the proposed method. In order to scrutiny of the designed model, various controls have been done. These controls have been tested on different data. Outcome results are representative high accuracy of designed model. Finally, in order to survey the efficiency of proposed procedure, the designed model has been applied to some of Tehran and its suburb school structures and outcome results have been compared with main data real results. Outcome results are representative good efficiency of the method. We should notice that using Fuzzy Concluder Systems lead to speed structure risk analysis and so decision making about various stages of structure rehabilitation is performed with more rate than previous. Thus, use of procedure that proposed in this paper, can has suitable applications in rapid seismic risk evaluation of studied structures in first stage of rehabilitation process.