Abstract
Extracting useful electricity from ocean waves is consistently receiving significant research attention. However, the technology of wave energy conversion is underdeveloped. There are many challenges that need to be addressed in order to improve the economic index of wave energy converters (WECs). One important challenge is the expenditures on developing the energy storage units to improve the quality of wave power for grid integration. It has been found in recent research that the wave power quality can be significantly improved when placing WECs in an array which can potentially save the cost of energy storage units considerably. Nevertheless, the investigation of the modeling of WECs array is not sufficient, specifically, the disturbed wave field (which is typically not considered). Although the disturbed wave field can be evaluated by using a Computational Fluid Dynamics (CFD) based solver, it is very cumbersome to be applied to large-scale arrays (e.g., in the level of kilometers). Therefore, two new models (mid-fidelity and high-fidelity models) of large-scale WECs array are proposed in this paper, which both consider the wave to wire dynamics (using ProteusDS) and the disturbed wave field. In addition, a low-fidelity model is also developed (which assumes all the devices in the array are isolated) for the sake of comparison. In which the mid-fidelity model only considers the wave attenuation in the leeward of the device (along the wave direction, 1-D). In contrast, the high-fidelity model takes all the wave diffraction, refraction, and directional spreading into consideration (using Simulating WAves Nearshore (SWAN), 2-D). The performance of the WECs array predicted in different model fidelities are compared/analyzed, as well as the computational cost aims at providing the WEC developers the most economic approach to assess array performance in different scenarios.
Numerical simulations are conducted for all the model fidelities for three different WECs (RM3, RM5, and Triton) across four representative sea states at the PacWave ocean site (Oregon, USA). These sea states have a significant height of 1.65, 4.33, 1.96, and 2.19m and a peak period of 8.81, 13.97, 16.42, and 11.92s respectively. The simulation results show that the wave power quality (assessed by the Coefficient of Variance (CoV)) predicted by the high-fidelity model is significantly better than that predicted by the low-fidelity model (considers same number of isolated WECs in the array). For instance, for RM3, the improvement of the power quality (reduction of CoV) is around 39%, 33%, 29%, 36%, for different sea states respectively. This again demonstrates the advantage of placing WECs in an array which is shown to improve the power quality. Moreover, it is also found that the wave field disturbance is weak (smaller than 2% of the incoming wave height) when the device is only controlled by a passive Power Take-Off (e.g., RM5). In which, the wave field disturbance (%) is computed based on the wave field differential due to the presence of WECs divided by the free wave field. In this scenario, a mid-fidelity model is more economic than a high-fidelity model (which has approximately three-folder higher computational cost). As far as the array energy production is concerned, the mid-fidelity model underpredicts this power production (general for all the devices) since it only takes the wave attenuation into consideration.