Response control of fixed offshore structure with wind turbine using MR damper

Due to the depletion of fossil fuels, offshore wind turbines have become significant contributors to energy production. The offshore wind turbines are subjected to severe environmental loads like wind, wave, and current. The combined effect of these loads induces significant vibrations in the structure, which may cause economic loss and reduce the power production capacity. Thus, the structure is designed to have a small response under environmental load, but it also increases the cost. Another alternative to reduce the structure’s response is implementing structural control devices. MR dampers are semi-active devices widely used in controlling onshore structures because of their adaptability to various loads. Since the environmental loads on the wind turbine have a wide range of spectral content, semi-active control is best suited due to their adaptability. The force generated by the MR damper is enormous for small input energy, and thus it is an attractive option for adaptive structural control in civil structures. Several semi-active control algorithms are available in the literature for nonlinear MR dampers. The semi-active algorithm obtains the optimal control force of the linearized system and smartly switches off the MR damper according to the system requirement. The offshore structure is nonlinear, so a nonlinear control algorithm is expected to have advantage over other techniques. One of the commonly used nonlinear control algorithms is the backstepping method. This method involves a systematic development of the Lyapunov function for a given nonlinear system. The nonlinearity in the MR damper arises due to the hysteretic property of the MR fluid. Many phenomenological models are studied in the literature to capture MR dampers' nonlinearities effectively. Bouc wen model is one such model that is widely used for modeling MR dampers.The backstepping controller requires full state feedback to produce the control signal but measuring many variables is difficult in real-time. Some controllers do not require an accurate structure model, and the control action is based on the measured response. This paper uses the feed-forward neural network controller (NNC) to control nonlinear dynamic systems. The controller is trained using the displacement time histories of the structure to predict the MR damper voltage. The NNC is trained with the output from the backstepping controller for various environmental conditions. In this study, a fixed offshore wind turbine in a water depth of 64.5m is controlled using NNC and backstepping controllers. Further, the control efficiency of the NNC is compared for different load cases.