PID controller for NARX and ANFIS models of marine pipe cylinder undergoes vortex induced vibration

Abstract

PID controller discrete time on marine cylinder pipe risers which represented based on system identification models under vortex induced vibration (VIV) has been verified in this work. Input-output data have been provided from the experimental setup of previous paper. Two System identification methods used to create models which are: Neural Network based on Nonlinear Auto-Regressive External (Exogenous) Input (NARX) and Adaptive Neuro-Fuzzy Inference System (ANFIS). NARX and ANFIS models have selected based on Mean Square Error (MSE) technique. While, PID controller has been applied to overcome on the pipe cylinder oscillation for all models. Also, the controller performance has been compared on each model during from tuning the controller parameter (KP, KI and KD) depending on the heuristic method and validation the gain values based on Mean Square Error (MSE) technique. Finally, the consequences demonstrated that the ANFIS model better than the NARX model to forecast the dynamic behavior of the system. By contrast, PID controller has been managed to decrease the pipe cylinder fluctuation for all models specially the NARX model system identification