Application of Adaptive Neuro-Fuzzy Inference System for Estimating the Global Ductility of EBF steel frames under Pulse-type Near-fault Earthquakes

The need to solve the complex, nonlinear, and variable problems grows with time. Conventional mathematical models perform linear and constant analysis effectively. Although techniques that work on a particular model, capable of analyzing complex nonlinear and time-varying problems, however, they also face some limitations. Combining these with other issues such as decision making, etc., has inspired the development of intelligent techniques such as fuzzy logic, neural networks, genetic algorithms, and expert systems. Intelligent systems mainly employ a combination of these techniques to solve very complex problems. Although both fuzzy logic and artificial neural networks have been very successful in solving time-varying nonlinear problems, each has its own limitations which reduces their use in solve of many of these problems. The roof global ductility, is a comprehensive reflection of various engineering demand parameters (EDP), such as story-drift, plastic rotation at member ends, roof displacement, etc. Careful estimation of this parameter will certainly lead to greater accuracy in the design of structural members. One of the methods which establish a good estimate of the nonlinear seismic response is the using of EDP parameters and measuring the seismic intensity index. The main purpose of this paper is to establish an accurate intelligent model related to the geometrical characteristics of the structure, performance level, the behavior factor and global ductility in eccentrically steel frames, under earthquakes near-fault. For this purpose, genetic algorithm is used. Initially a wide database consisting of 12960 data with 3-, 6-, 9-, 12-, 15- and 20- stories, 3 column stiffness types, and 3 brace slenderness types were designed, and analyzed under 20 pulse-type near-fault earthquakes for 4 different performance levels. To generate the proposed model, 6769 training data were used in the form of adaptive-neural fuzzy inference system(ANFIS). Subtractive clustering and FCM methods have been used to generate the purposed model. The results showed that Subtractive clustering provides more accurate results than the other FIS. To validate the proposed model, 2257 test data were used to calculate the mean squared error of the model. The proposed model is an intelligent model in the range of data used, and can be used to estimate the global roof ductility of EBFs. To evaluate the efficiency and performance of the model, correlation coefficient and common error calculation criteria including RMSE and MARE were used. The correlation coefficient for the Subtractive clustering method was 0.888, based on intelligent model in the test data. In the other hand, the developed intelligent model can be used as a precise alternative to prediction of (μR) for EBFs under near-earthquakes. To evaluate the model’s efficiently and accuracy, various error criteria including Error, Mean Error, RMSE, MARE% and R were used between model values and real values, in the test data. From the results of this study, it can be pointed out that, the developed intelligent model can be used as an accurate substitute method to predict the (μR) for EBF structures, under near-fault earthquakes. The results of correlation analysis of the proposed model show that the proposed intelligent model has high accuracy.