Abstract:To solve the problem of uncertainty in the estimation of vertical ground motion parameters, we propose a prediction model based on random adaptive neuro fuzzy inference. Firstly, based on PEER NGA strong motion database from Pacific Earthquake Engineering Research Center, we take as input parameters the earthquake magnitude, fault distance and average shear wave velocity, and take the PGA and PGV as estimation targets to establish training data sets and test data sets. Secondly, according to the prediction equation of ground motion parameters, we use random adaptive neuro fuzzy inference technology to construct the prediction model of vertical ground motion parameters to predict PGA and PGV, and give comprehensive parameter study and reliability test. The vertical ground motion attenuation of ANFIS model shows the near-field strong earthquake saturation effect, site amplification effect and soft soil damping effect of PGA. The average absolute percentage error of ANFIS model is about 0.15. Compared with Campbell-Bozorgnia ground motion attenuation relationship, the accuracy of PGA and PGV prediction is improved by about 77.4% and 62.7% respectively, thus has better estimation accuracy.
YOU Shan,HU Qizhi,ZHANG Jie et al. A Method Based on Adaptive Neuro Fuzzy Inference System to Predict Vertical Ground Motion Parameters[J]. jgg, 2023, 43(5): 517-522.