Active vibration control of transverse vibrating segmented marine riser

Abstract

Vortex induced vibration (VIV) could be regarded as a fluid-structure interaction vibration type where the bluff structure vibrates due to fluid flowing around the body. The separation of boundary layer has created vortex layer that staggers the structure in cross-flow direction. VIV suppression work has attracted numerous researchers to build a passive device that could reduce the vibration. However, such device requires an intricate design which incurs high expense and indirectly contributes to higher chance of VIV occurrence due to the additional mass to the system. This research proposed a method to overcome those shortcomings by introducing an active flow control concept to the system. Since the vibration originates from unhindered flowing fluid, the approach is to avoid the development of the vortex by attaching a single control rod to the system as an actuator. The actuator injects momentum to the boundary layer thus preventing the VIV phenomenon. Both simulation and experimental works were implemented in this study. The input-output data of the system were measured directly from the experimental rig. For system identification, three methods were employed which were least square (LS), recursive least square (RLS) and differential evolutionary (DE) algorithms. It was found that the DE methods were stable, had considerably lower mean squared error (MSE) and the transfer function itself represented the natural frequency of the system. The study was continued by tuning the proportionalintegral- derivative (PID) based controllers to the simulated system plant in offline mode. The PID based controllers were tuned using heuristic and Ziegler-Nichols (ZN) methods. The best performance was recorded. However, it was observed that once the disturbance of the system changed, the performance of the PID tuned using heuristic and ZN were deteriorated. To overcome this drawback, adaptive tuning algorithms were introduced, namely ZN-Fuzzy-PID and ZN-Fuzzy-Iterative Learning Algorithm-PID (ZN-Fuzzy-ILA-PID) based controllers. In simulation, it was found that the ZN-Fuzzy-ILA-PD controller outperformed other controllers with 57.82 dB of attenuation level. In experimental works, dynamic response comparison was made between the bare pipe, fixed single and double control rods. It was observed that the fixed single and double control rods could not effectively attenuate the system, but amplified the vibration instead. Further experimental work was conducted by varying the rotating speed of the actuator at various disturbances. The result shows that at 100 % actuator rotating speed with 33 Hz disturbance flow to the system, the vibration was successfully reduced with attenuation level of 20.71 dB. However, by changing the disturbance, the actuator performance was reduced. Therefore, the controller was adaptively tuned using the fuzzy and iterative learning (ILA) schemes. It was observed that the maximum vibration attenuation was achieved by ZN-Fuzzy-ILA-PD controller with 13.8 dB of attenuation level at changing disturbance. Overall results show that by adopting the single rotating control rod, the vibration of VIV could be successfully attenuated.