Artificial intelligence (AI) approaches that mimic human (or animal) cognitive functions have been developed over the past decades. Based on excellent learning and problem solving capability, AI has been applied extensively to a range of nonlinear problems for classification, recognition, prediction and optimization. Our research group utilize various AI methods to assist with the issues of offshore wind energy (e.g., operation/management of wind farm) as well as river and coastal disaster prevention (e.g., early warning for hazard mitigation). For the former issue, AI techniques could be used with the physically-based general circulation model (3D-Atom) to better predict the variation of wind speeds, ocean currents and waves which are essential for the design conditions of wind turbine, support structure, and foundation. Besides, safe ship-handling plays an important role in construction of offshore wind farm. The CubeSat with AIS system and AI techniques can be applied to obtain optimized planning for navigation. To achieve optimal benefits, the layout of wind farm can be determined by the genetic algorithm with ant colony optimization. Also, a smoothing control system that improves unstable electric power supply can be developed using recurrent fuzzy neural network. For the later one, the combined effects of strong wind, heavy rains and huge surge (and waves) during typhoon event often cause serious damages over upper catchment basin, metropolitan city, and coastal area. Particularly, the abnormal rise of sea level (storm surge) propagating from the estuary through the river (i.e., the backwater effect) could aggravate inundation and flooding. To strengthen disaster preparedness, early warning and response work, our research group also aims at developing long-lead-time (12hr) prediction models for both rainfall-runoff processes and storm surge by artificial neural networks with consideration of effective controlling parameters.
