| Popis: |
As an alternative energy source to fossil fuels, marine renewable energy is expected to have a huge potential. Wave energy is one of the marine renewable energy sources. The feature of a PA-WEC, which is one of the Wave Energy Converter (WEC), can generate electricity directly from the change of magnetic field caused by the relative motion of the coil and the moving part of the floating body with permanent magnet. It is also possible to control the stiffness relation by applying a force to the moving part of the floating body by changing the current in the coil. By taking advantage of this characteristic, it is possible to generate electricity efficiently in a wide periodicity band. As for this efficient power generation, the authors have shown the analytical optimum solutions for the regular and irregular wave problems, including arrayed PA-WEC problems, for maximizing the power generation considering the copper loss in the generator. On the other hand, it is not easy to decompose the unknown irregular wave components into multiple regular wave components with high accuracy from time to time in the actual ocean. Therefore, the authors have been investigating the possibility of estimating the optimal control force at each moment in an unknown irregular wave by using an AI model. Then we showed that the AI model trained on regular waves can estimate the control force close to the theoretical solution for irregular waves. In this study, the AI model is incorporated into the equations of motion, and the time series response of the power generation is obtained in an unknown irregular wave. From the comparison with the analytical optimal solution of the power generation, the possibility of applying the AI model was investigated. As a result, it was shown that, depending on the combination of the training model and input conditions, it is possible to obtain about 70% of the power generation compared to the theoretical solution. |
