Hydrodynamic response analysis of a semi-submersible platform using a hybrid rans-bem approach

Abstract

Over the past 50 years, climate change has been observed all over the world. In the last 30 years, it has ceased to be a simple issue among developed countries and has become a significant concern and topic of discussion under the term climate crisis. Authorities have begun seeking solutions to prevent future catastrophe scenarios by taking measures against this crisis. Scientific studies have repeatedly highlighted the role of greenhouse gases in the climate crisis, especially carbon dioxide emissions. Experts have determined that the largest share of carbon emissions results from the combustion of fossil fuels for the generation of electrical energy. Since global energy demand is a fundamental pillar of modern civilisation, finding sustainable and low-carbon energy sources has become inevitable for humanity. Renewable energy sources have come to the forefront, and offshore wind energy has gained significant momentum. The offshore wind concept is a logical solution for the future of the world, considering that today, nearly two billion people live along coastlines, and it is well-known that nearly 80% of offshore wind potential lies in areas deeper than 60 meters. Since fixed platforms can only be used in shallow waters, the concept of floating wind turbines emerged to harness wind energy in deeper waters. Countries aware of the importance of energy independence and their potential for offshore wind energy have started working on the implementation of these structures within their territorial waters. In recent years, a semi-submersible platform with three conical pontoons has been proposed for use in the Black Sea, as part of ongoing research into offshore wind platform designs to deploy within Türkiye's territorial waters. Conical geometries are uncommon in semi-submersible platforms. The few studies conducted so far have shown that they perform quite well and are worthy of attention. This study aimed to create a baseline for future research on conical semi-submersible platforms by examining the proposed platform in detail using numerical methods, while supporting Türkiye's green energy transformation under the Twelfth Development Plan (2024-2028). The analyses to be conducted in this study focused on heave, pitch responses, which can be compared with experimental data, and also addressed roll responses that were not considered in previous experimental or numerical studies. Since prototype or model experiments for floating wind turbines are costly, creating a baseline through numerical analyses during the early design process has a great importance. To reveal the hydrodynamic characteristics in both frequency and time domains, a commercial boundary element method (BEM) code, AQWA, was primarily employed, and the analyses were supported with Reynolds-averaged Navier-Stokes equation (RANS) based computational fluid dynamics (CFD) code. AQWA is based on linear potential flow theory. To present the details of the analyses conducted, the fundamentals of fluid mechanics, key problems in this field, and their solution methods were covered, along with the theoretical foundation of AQWA. Following the theoretical background, validation studies based on various works in the literature were conducted to evaluate AQWA's capabilities. These studies primarily used shapes commonly found in ocean engineering designs, such as a sphere and a cylinder. Numerical analyses were conducted, and the obtained results using AQWA were compared with experimental data of chosen studies through the heave motion. While successful results were achieved for the sphere in frequency- and time-domain analyses with differences lower than 2%, significant amplitude differences, up to 10 times stronger in frequency-domain, were observed in the cylinder case. The cause of these differences was that potential flow theory assumes no viscous effects in the solutions. While this assumption was not very significant for the sphere, it made a noticeable difference in the cylinder. To determine whether the threshold of viscous damping cannot be ignored for accurate motion predictions, the existing cylinder model was analysed under different draught-to-diameter ratios, 0.5, 1.0, 2.0, 3.0 and 4.0. The analyses were conducted using RANS-based CFD code to capture viscous effects and BEM-based AQWA to observe only radiation damping. During the process, the threshold was determined as a range for the uncertain region and a critical point. The critical threshold was accepted as the point where the ratio is 2.0. Subsequently, the calculated viscous correction was applied to the cylinder as frequency-independent additional damping in AQWA, and the numerical results showed strong agreement with experimental results. To support future work, it was decided to represent the conical semi-submersible platform with two different models in AQWA. One of the models excludes structure elements such as the tower (semi-simplified model), while the other excludes smaller submerged elements such as beams, retaining only the main conical pontoons and the central cylinder (simplified model). To validate the redefined physical parameters, initial numerical analyses were conducted using a RANS-based CFD code through free decay tests. Free decay tests were conducted for heave, roll, and pitch motions in the investigation. The computational domain was constructed similarly to the previous RANS-based analysis in terms of dimensions. Solution parameters for the analysis were determined by reviewing the literature and considering the platform's experimental results. Both heave and pitch motions were based on experimental results, while the time step for roll motion was selected after a preliminary diffraction analysis due to the lack of data. Different mesh structures in terms of quality were created systematically, and their results were evaluated under the same solution parameters. The convergence of the results for the finest mesh structure was determined, and the uncertainties remained below 3%, with the highest uncertainties mostly observed in the damping ratios. The RANS results were compared with the results of the experimental study's results. After it was seen that the differences were within an acceptable range, viscous damping was introduced as frequency-independent additional damping to AQWA. The diffraction analyses were repeated for both models, and hydrodynamic coefficients were obtained for both cases, with and without damping. Hydrodynamic coefficients were presented with comparisons for different models and the viscosity effects. After the diffraction analysis was conducted in AQWA, the dynamic behaviour of the structure in the time domain was examined, first via free decay tests. Initial conditions applied in the RANS-based analysis were applied here as well. The free decay tests were conducted for heave, roll, and pitch motions using both models. The results of the experiment, the RANS-based simulations and AQWA were compared using graphs and tables. As in the RANS-based analysis, notable agreements were also found in AQWA, particularly between the semi-simplified model and the experimental results. In the time-domain analyses, the AQWA semi-simplified model exhibited less than 3% deviation in natural periods, whereas discrepancies in damping ratios reached up to 10%. These similarities between the results were considered to indicate that the newly defined physical parameters are consistent with those of the experimental model. The performances of AQWA models were compared based on the free decay test results, and the most suitable model was determined as the semi-simplified one to use in the regular wave analyses. The wave conditions were simulated based on the deployment region's sea conditions, selected considering an average 20-year service life, and within the limits of AQWA's capabilities. To minimise time step errors, translational movements along the x and y axes and rotational movements about the z-axis were constrained. The results were presented from the moment the system began regular motion. During the regular wave analysis, it was observed that the structure was significantly affected by coupled dynamics, which was important in evaluating the platform. For example, in waves characterised by a height of 0.15 meters, it was expected that the dynamic responses in pitch and roll motions would decrease between the period of 1.24 seconds and 1.55 seconds; however, an increase was observed because of amplified heave response. Similarly, under wave conditions characterised by a period of 2 seconds and a height of 0.15 meters, it was observed that the heave motion was influenced by pitch motion. Although coupled movements and some unexpected behaviours were observed, no excessive responses or risk of capsizing were recorded under the design conditions. Based on the results, it can be stated that the behaviour of the structure is quite stable in its current configuration, safe under the design conditions for an average 20-year service life, and ready for future studies.