Feasibility study of a semi floating spar buoy wind turbine anchored with a spherical joint to the sea floor María Sanz Martínez DTU Wind Energy DK-4000 Roskilde, Denmark msma@dtu.dk Anand Natarajan DTU Wind Energy DK-4000 Roskilde, Denmark anat@dtu.dk Lars Christian Henriksen DTU Wind Energy DK-4000 Roskilde, Denmark larh@dtu.dk Abstract: The feasibility of a semi floating platform offshore wind turbine system is investigated at 120m water depth. The semi floating system consists of a 5MW wind turbine on a floater with mooring lines similar to a spar buoy and strongly anchored with a spherical joint to the sea soil. The stability of the newly designed floater and mooring assembly are analyzed from static and dynamic simulations. The design loads on the universal joint on the sea floor are tuned with the needs for a ballast chamber. Using load simulation in the HAWC2 software, ultimate and equivalent fatigue loads are obtained and compared with the loads from the same wind turbine mounted on a spar buoy system and a land based wind turbine at the same points. The results show a decrement in the ultimate and equivalent fatigue loads for the new system. Keywords: Design loads, spherical joint, semi floating platform, mooring system 1. Introduction Fixed offshore wind turbine structures may be cost effective till 60 metres water depth. Monopile fixed sub structures are usually used in wind turbine installations till 30m water depths and frame structures till 60m. Floating wind turbine prototypes are designed for water depths near 150m-200m. Therefore, the wind industry may require other designs of cost effective sub structures in the range of 60-120 metres. A floating spar-buoy substructure which is anchored with a universal spherical joint to the sea soil is analyzed at 120m water depth as potential solution for moderate water depths (60m to120m). Oil industry has used similar substructures [1]. However, the wind loads experienced by the oil rig structures are negligible compared to the rotor generated dynamic loads experienced by wind turbines so a further study is required. In this study, the wind turbine used is the NREL 5 MW [2]. The simulations were performed in the aero-hydro-servo-elastic code HAWC2 [3]. For a better analogy, the new system was compared with the NREL spar buoy [4] and a land based wind turbine. Hence, the same simulations were run for all systems. 2. Design of the platform and static analysis of the system The new sub structure configuration should achieve static and dynamic stability. Therefore the submerged volume needs to be large enough to equilibrate the system with its restoring force (the higher the angle of tilt, the bigger the submerged volume). To match these requirements, the final geometry was chosen as two cylinders of different radii linked by an interface with the shape of a truncated cone (see Figure 1). The design of the new platform was obtained through an iterative optimization process between the total weight of the system and the submerged volume under normal conditions. In each iteration, the weight of the system was determined and then, the submerged volume needed to stabilize that weight was computed. We considered the weight for the initial iteration as the weight of the spar buoy system. The final mass of the platform was seven times lower than the mass of the spar buoy floater because there is no 1
ballast requirement in the semi floating system. The final geometry of the floater consists in: a bottom cylindrical part of 80 m length and 2 m radius, a transition conical part of 40 m length, and a top cylindrical part of 10 m length and 3 m radius. The total draft of the platform is 130 m; the floater is designed for a water depth of 120 m. As the floater was redesigned, the Morison equation applicability was verified with 8 different significant wave heights. A stability analysis using two tools was performed: an approximate Matlab script that considers the weight, buoyancy and thrust loads and calculates the pitch angle of equilibrium of the platform; and detailed HAWC2 simulations that includes also the mooring line dynamics and hydrodynamics. The results of both analyses are listed in Table 1. Table 1 Platform equilibrium angle of both systems depending on wind speed Pitch angle Pitch angle Mean Wind (Matlab) (HAWC2) speed [m/s] [deg] [deg] 5 +21.3 +1.2 8 +23.5 +2.8 11 +25.9 +5.2 13 +24.6 +3.9 16 +23.7 +2.9 20 +23.0 +2.3 24 +22.6 +2.0 Table 2 Natural frequencies of the semi floating system Frequency [Hz] Description 0.02977 Platform roll 0.03004 Platform pitch 0.03518 Platform yaw These results are within the same range of the natural frequencies calculated for the spar buoy system in other studies as [5]. The floater is depicted in Figure 1. The largest pitch angle is obtained for mean wind speeds close to the rated wind speed because the thrust force is the main input of the system. The mooring line forces cause the different equilibrium angles of both analyses. This shows the system is statically stable without mooring system, but the resultant equilibrium angle is too high to be operative. Mooring lines should be considered, and designed for improved performance. A modal analysis was then performed to obtain the lowest modes of the static system. The results are listed in Table 2: Figure 1 Sketch of the floater with a universal joint at the soil 2
3. Specification of new mooring system We defined a new mooring system based on [6]. Each mooring line is divided into a number of elements or Timoshenko bodies that are analyzed individually together with the rest of the system in HAWC2. The equation of motion for each of the mooring bodies is: JONSWAP (Joint North Sea Wave Project) Spectrum [7] with an irregular airy pattern. The peak enhancement factor, γ was assumed to have a constant value of 3.3. The significant height and the period of the waves are defined in Figure 3. ( ) ( ) ( ) ( ) The main parameters of the mooring system are presented in Table 3. The mooring system uses catenary lines because it has simple and cheap anchors, it is easy to install and it is suitable for shallower waters. Its sketch can be seen in Figure 2. Table 3 Summary of properties of the mooring system Number of mooring lines 3 Depth to fairleads, anchors Radius to fairleads, anchors Un stretched line length Line diameter Line mass density (in air) Line mass density (in water) 30, 120 m 9.4, 850 m 853 m 0.1 m 113.09 kg/m 86.19 kg/m Figure 3 Wave parameters depending on the mean wind speed. We define semi floating system as the 5 MW NREL wind turbine defined in [2], mounted on the platform described in section 2 and the mooring system defined in section 3. To confirm the good operation of the semi floating system the mean power and thrust are compared with the corresponding figures for the floating spar buoy mounted wind turbine in Figure 4 to Figure 7. Line extensional stiffness 381700 kn Figure 2 Mooring system scheme. 4. Wind Turbine System and Environment Environment definition The wave conditions for the aeroelastic simulations were defined following the Figure 4 Power curve comparison. 3
wind turbine and for a land based wind turbine. The response at the tower top for a single degree of freedom system is ( ) { ( ) ) ( ) ( ( ) Figure 5 Thrust curve comparison. The tower top displacement was fitted to the previous equation and the system parameters were calculated: The mass used in each case was the whole system mass (tower, nacelle, rotor and floater). The obtained results were: Table 4 Comparison of damping values Semi floating system Spar buoy system Land based system Figure 6 Pitch curve comparison. ξ 0.214 0.147 0.578 c [kns/m] 7.955 10 2 1.756 10 3 1.063 10 3 k [kn/m] 3.265 10 3 4.748 10 3 1.210 10 3 The design load cases 1.1 (NTM, normal turbulence model) and 1.3 (ETM, extreme turbulence model) from the standard IEC 64100:3 [8] are simulated. Both design load cases use 6 different wind turbulence seeds for each mean wind speed. 5. Fatigue analysis Figure 7 Rotational speed curve comparison. The differences are minimal. Estimated damping for the pitch of the platform We consider the semi floating system as a second order system to estimate its damping. In the steady state, the response of the system to a step force was simulated. The same test was completed for the spar buoy For the site we assumed the mean wind speed has a Rayleigh distribution with a scale parameter σ = 6.80 (data from [9]). A yaw misalignment angle of ± 10 degrees is included 50% of the simulated time (25% each) in the normal turbulence model simulations. In this study, the fatigue analysis was used as a methodology to compare the dynamic loads on both systems (spar buoy and semi floating) in the bottom of the tower. For a better comparison, a third system, a land based 4
wind turbine, was simulated and compared together with the semi floating and the spar buoy system. In the analysis of the semi floating system, the joint was also studied as a point of interest, whose lifetime loads should be used as guidelines in the design of the joint. Figure 8 shows the fatigue damage equivalent loads for the fore-aft bending moment (M x ) at the bottom of the tower (lowest node of the tower). The semi floating system obtained results within the same range of the land based wind turbine, while the spar buoy reached higher equivalent loads. These results were due to higher platform pitch oscillations for the spar buoy system compared with the oscillations of the semi floating system for the same angle. The reason for the high differences is the semi floating system has only positive bending foreaft moments and this has a lower contribution on the fatigue loads (see Figure 9). At high wind speeds, the semi floating system reached lower negative moments in comparison to the land turbine. Figure 10 shows the fatigue loads for the side to side bending moment (M y ) in the bottom of the tower. The semi floating system obtained lower loads for high wind speeds than both the land based and the spar buoy systems due to the lower variations in the amplitude of the loads than for the other systems (in Figure 11 it is noticeable the lower values of the standard deviation for the semi floating and land based systems). For low wind speeds the spar buoy system reaches high equivalent fatigue loads while the semi floating and the land based systems have loads in the same range. For the torsional moment at the bottom of the tower (M z ), the semi floating system achieved lower equivalent loads compared with the other two systems, which was again due to the lower standard deviation of loads on the semi floating system (Figure 12 and Figure 13). The equivalent lifetime loads for the three systems are listed in Table 5 (the lifetime of the turbine is assumed as 20 years of operation 10 7 cycles). Table 5 Equivalent fatigue loads in the tower bottom [knm] Semi Float Spar Buoy Land Based M x 57360 137720 46110 M y 22950 50970 21920 M z 6210 11210 11330 The ratios between the results obtained for the sea turbines and the land turbine have been compared with previous studies in this field ( [10] and [11]) obtaining similar results to previous publications. Figure 8 Equivalent for aft fatigue loads M x at the tower bottom (continuous line: no yaw misalignment; dotted line: -10⁰ yaw misalignment; dashed line: +10⁰ yaw misalignment). Figure 9 Statistical data of the NTM simulations for the M x loads at the tower bottom (Max maximum value; Min - minimum value; Mean - averaged value; SF 5
semi floating system, SB spar buoy system, Land- land based system). yaw misalignment; dotted line: -10⁰ yaw misalignment; dashed line: +10⁰ yaw misalignment). Figure 10 Equivalent side to side fatigue loads M y at the tower bottom (continuous line: no yaw misalignment; dotted line: -10⁰ yaw misalignment; dashed line: +10⁰ yaw misalignment). Figure 13 Statistical data of the NTM simulations for the M z loads at the tower bottom (Max maximum value; Min - minimum value; Mean - averaged value; SF semi floating system, SB spar buoy system, Land- land based system). At the joint (Figure 14 to Figure 16), the maximum equivalent loads were obtained in the wind and waves direction (y). In the other axis, the changes in the forces were minor, so the equivalent fatigue loads were smaller. The equivalent fatigue loads at the joint for the semi floating system are listed in Table 6: Figure 11 - Statistical data of the NTM simulations for the M y loads at the tower bottom (Max maximum value; Min - minimum value; Mean - averaged value; SF semi floating system, SB spar buoy system, Land- land based system). Table 6 Equivalent fatigue lifetime loads in the joint (semi floating system) Load Direction Equivalent fatigue lifetime load (kn) F x 664 F y 1128 F z 665 Figure 12 Equivalent torsional fatigue loads M z at the tower bottom (continuous line: no 6
from the ETM simulations (DLC 1.3). For the second method utilized, the NTM simulations were used. The threshold was defined as Where µ and σ are the mean and the standard deviation of the NTM simulations. All the peaks above the threshold were considered. Their average value was calculated and multiplied by 1.35 to obtain the ultimate load expected for each wind speed. Figure 14 Equivalent lateral fatigue loads at the joint (F x ). Figure 15 Equivalent longitudinal fatigue loads at the joint (F y ). Again, for a better analysis, the three systems -semi floating, spar buoy and land basedwere compared. Figure 17 shows the ultimate loads at the tower bottom fore aft bending moment (M x ) for the semi floating system. They reached their maximum around rated wind speed (due to the maximum thrust). The spar buoy system obtained higher ultimate loads than the other systems because of isolated extremes. Even though the mean values obtained for the three systems were quite similar, their standard deviations differed for high wind speeds (see Figure 18). Figure 19 shows the ultimate loads of the three systems for the case of the side to side tower bottom side to side moment (M y ). The semi floating system obtained lower loads than the other two systems. For the case of the torsional bending moment at the bottom of the tower (M z ), the semi floating system obtained the lowest ultimate loads (see Figure 20). Figure 16 Equivalent vertical fatigue loads at the joint (F z ). 6. Ultimate load analysis The ultimate loads analysis is done according to two different methods described in the IEC standards [8] and [12]. The first method used was simply extracting the maximum values Figure 17 Ultimate loads for fore aft bending moment (M x ) at the tower bottom (SF semi floating system, SB spar buoy system, Land- land based system). 7
to the high thrust force obtained at rated wind speed. To decrease the values achieved in Figure 23 (vertical force at the joint), ballast could be added to the floater. Figure 18 Standard deviation of the fore aft bending moment (M x ) at the tower bottom (SF semi floating system, SB spar buoy system, Land- land based system). Figure 21 Ultimate lateral loads at the joint (F x ). Figure 19 Ultimate loads for side to side bending moment (M y ) at the tower bottom (SF semi floating system, SB spar buoy system, Land- land based system). Figure 22 Ultimate longitudinal loads at the joint (F y ). Figure 20 Ultimate loads for torsional bending moment (M z ) at the tower bottom (SF semi floating system, SB spar buoy system, Land- land based system). At the joint of the semi floating system (Figure 21 to Figure 23) the maximum loads were reached around rated wind speed. This is due Figure 23 Ultimate vertical loads at the joint (F z ). 7. Conclusions The semi floating system is a promising solution for moderate water depths. The 8
oscillations of the semi floating system under turbulent wind excitation are much lower than the spar buoy. The pitch angle of the platform is steadier and has lower oscillations against possible excitations and that reduces the fatigue loads considerably. According to the results of the fatigue analysis (Table 5) the semi floating system expects significantly lower damage equivalent loads than the spar buoy system. This means that the structural requirements are less demanding and, more likely, cheaper than for the spar buoy system. The design of the joint should be carefully evaluated and consider the impact of the fatigue loads depicted in Table 6 and the ultimate loads displayed in Figure 21 to Figure 23. The ultimate loads analysis shows that at the tower bottom, the maximum loads obtained for the semi floating system are lower at almost all wind speeds than for the spar buoy wind turbine. Further analysis should be done to verify the feasibility of this new semi floating system. Acknowledgement The work presented in this paper is part of the Collaborative Project "INNWIND.EU" supported by the EU Seventh Framework Program (FP7), grant no. 308974. The financial support is greatly appreciated. References 1. Glanville RS, Paulling JR, Halkyard JE, Lehtinen TJ. Analysis of the Spar Floating Drilling Production and Storage Structure. Offshore Technology Conference (OTC-6701) May 1991, pp. 57 68. 2. Jonkman J, Butterfield S, Scott G. Definition of a 5-MW reference wind turbine for offshore system development. NREL. February 2009.NREL/TP-500-38060. 3. Larsen TJ, Hansen AM. How 2 HAWC2, the user's manual. Risø National Laboratory 2012 Technical Report R-1597, (ver.4-4). 4. Jonkman J. Definition of the floating system for phase IV of OC3. NREL May 2010 NREL/TP-500-47535. 5. Bir G and Jonkman J. Aeroelastic Instabilities of Large Offshore and Onshore Wind Turbines. Journal of Physics: Conference Series, The Second Conference on The Science of Making Torque From Wind, Copenhagen, Denmark, 28 31 August 2007, [online journal], Vol. 75, 2007, 012069. 6. Hansen AM and Kallesøe BS. Detailed and reduced models of dynamic mooring system. Aeroelastic Optimization of MW Wind Turbines. Risø National Laboratory Risø-R- 1803(EN) 2011; 20-34. 7. Carter D. Prediction of wave height and period for a constant wind velocity using JONSWAP results. Ocean Engineering vol. 9 1982; 17-33. 8. International Electrotechnical Commission. IEC 61400 3 Ed. 1. Wind Turbines Part 3: Design Requirements for Offshore Wind Turbines. Geneva, Switzerland: International Electrotechnical Commission, 2008. 9. Bezrukovs V and Bezrukovs V. Wind speed and energy at different heights on the Latvian coast of the Baltic Sea in proceedings of the Conference World Renewable Energy Forum, WREF 2012 May 2012, Denver. 10. Matha D, Jonkman J, Fischer T. Model Development and Loads Analysis of an Offshore Wind Turbine on a Tension Leg Platform, with a Comparison to Other Floating Turbine Concepts. NREL February 2010, NREL/SR-500-45891. 11. Jonkman JM. Dynamics Modeling and Loads Analysis of an Offshore Floating Wind Turbine. NREL November 2007. NREL/TP- 500-41958. 12. International Electrotechnical Commission. TC88-MT1 (ed.). IEC 61400 1 Ed.3 CD. 2. Revision. Wind Turbines. Part 1: Design Requirements. International Electrotechnical Commission: Geneva, Switzerland, 2005. 9