Fuzzy modeling of holding capacity of offshore suction caisson anchors

Summary A fuzzy algorithm, the Takagi–Sugeno model, is implemented to develop a fuzzy inference system for predicting the holding capacity of suction caisson foundations for offshore platforms. The premise parameters of the fuzzy model are optimized by using a subtractive clustering algorithm. The consequent parameters are optimally determined via a weighted least square estimation. The input variables used for training the fuzzy model include the aspect ratio of the caisson, the undrained shear strength of the clay, and the angle that the chain force forms with the horizontal. The output of the proposed fuzzy model is the capacity of the suction caisson anchor. To demonstrate the effectiveness of the fuzzy modeling framework, the results of extensive finite element analyses are investigated. Comparisons of the trained fuzzy model with the data demonstrate that the proposed modeling framework is an effective method to estimate the holding capacity of offshore suction caisson systems. Moreover, the performance of the fuzzy model is robust against higher levels of input data uncertainties. Copyright © 2017 John Wiley & Sons, Ltd.