下载文档原格式
[4]SUN J Y,ZHENG E.CHEN J J.Full-scale experiment of longitudinal ventilation smoke control system and central smoke exhaust system in city underwater iunncl[J].Procedia Engineering, 2013,52:330-335.⑸(；B50016-2014,建筑设计防火规范[S].[6]徐志胜，姜学鹏.防排烟工程[M].北京:机械工业出版社,2011.[7]U Q Q,YI L,XU Z S,刃al.Preliminary study on exhaust efficiency of smoke management system in tunnel firesfJ].Procedia Engineering,2013,52:514-519.[8]徐琳，张旭.集中排烟水平隧道排风诱导风速('FD分析[J].地下空间与T•程学报.2007,3(3):555-558.⑼丁•军，张旭.多向风流耦合作用对隧道火灾温度场的影响[CJ//2008年铁路暖通空调学术年会,2008.[10]吴德兴，乍伟平.郑国平•国内外公路隧道火灾排烟设汁理念比较[J].公路交通技术,2008,10(5):113-117.[11]XU Z.ZHAC)D,ZHANCi X,et al.Numerical study on effecls ofinduced velocity on central extraction system in large tunnel fire [J].Procedia Engineering,2012,45(2):678—684.[12]YI L,WEI R,PENG J,el al.F3xpcrimental study on heat exhaust coefficient of transversal smoke extraction system in lunncl under fire[J].Procedia Engineering,2013,52:268-278.[13]徐志胜，于年澈，张新，等.集中排烟公路隧道入口段火灾下诱导风速研究[J].灾害学,2012,27(3):97-101.[14]XU Z S.SHAN W.Influence analysis of effects of longitudinalinduced velocity on the smoke control under central smoke ex traction system in tunnel fire[J].Procedia Engineering,2013,52：500-507.[15]F•刚,赵冬.点式排烟诱导风速对大功率隧道火灾烟气影响数值模拟[J].安全与环境学报,2013,13(5):165-169.[16]TANG F F,MEI Z,WAN(J Q、e〔al.Maximum temperature beneath the ceiling in tunnel fires with combination of ceiling m(「chanical smoke extraction and longitudinal veniiIation[J].Tunnel ling and Underground Space Technology,2017,68:231—237. [17]JIANG X,LIU M,WANG J,er al.Study on induced airflow velocity of point smoke extraction in road tunnel fires[J].Tunnel-ling and Underground Space Technology,2018,71:637—643. [18]IJN('J,CHUAN YEW K.A study on long tunnel smoke extraction slrategies by numerical simuiation[J].Tunnelling and Underground Space Technology,2008,23:522—530.[19]湖南省质址技术监督局.公路隧道消防技术规范[S].The effect of induced ventilation on the efficiency under tunnel central smokeexhaust modeZHAO Jia-ming1,XU Zhi-sheng1,XU Ran2,NIU Zhan-wei2,LI Zhi\XIE Bao-chao1,HF：Lu1(1.Institute of Disaster Prevention Science and Safely Tech nology,Central Soulh University,Hunan Changsha410075,Chi na, 2.China Railway Tunnel(iroup Co.,Ltd.,(iuangdong Guang「............in,.,"in,...................on“”"mi”.•“hi......................mi,.............uh,........ ...............Xu,..•科技信息•超高性能混凝土的火灾爆裂研究法国研究人员开展了一种超高性能混凝土的火灾爆裂性能的试验研究。
局域共振声子晶体板的减振降噪研究郭旭1，2，崔洪宇1，洪明1（1.大连理工大学船舶工程学院，辽宁大连116024；2.东风日产乘用车公司技术中心，广州510800）摘要：针对船舶设备及结构的振动噪声控制问题，提出一种附加圆柱型振子的声子晶体板结构。
基于周期结构的Bloch 定理，采用有限元方法对声子晶体板的带隙特性进行计算，并通过振动传递率和隔声损失来描述声子晶体板的隔振及隔声效果，在此基础上分析了散射体几何参数及布置形式对隔振及隔声效果的影响。
研究表明，声子晶体板在特定频率范围内具有很好的隔振及隔声效果，通过合理设计散射体几何参数及布置形式，可以实现结构的低频宽带隔振及宽带隔声，在船舶振动噪声领域中具有很好的应用前景。
关键词：声子晶体；带隙特性；船舶结构；减振降噪中图分类号：O48O734文献标识码：A doi:10.3969/j.issn.1007-7294.2021.04.014Research on vibration and noise reduction of local resonant phononic crystal plateGUO Xu 1,2,CUI Hong-yu 1,HONG Ming 1(1.School of Naval Architecture Engineering,Dalian University of Technology,Dalian 116024,China;2.Technical Center,Dong Feng Nissan,Guangzhou 510800,China)Abstract:To suppress the propagation of vibration and noise in the ship equipment and structure,a phonon⁃ic crystal plate structure was designed with additional cylindrical vibrators.Based on the Bloch theorem of pe⁃riodic structure,the band gap of the phononic crystal plate was obtained with finite element method,the ef⁃fects of vibration isolation and sound insulation were described by vibration transmission rate and sound insu⁃lation loss.Moreover,the effects of the geometric parameters and arrangement of the scatterers on the vibra⁃tion isolation and sound insulation effects were analyzed.It is shown that the phononic crystal plate has a good vibration isolation and sound insulation effect in a specific frequency range.By rationally designing the scatterer geometric parameters and arrangement form,the low-frequency broadband vibration isolation and broadband sound insulation of the structure can be realized,which supplies more application opportunities in the vibration and noise reduction.Key words:phononic crystal;band gap characteristic;ship structure;vibration and noise reduction.0引言随着船舶向轻量化、快速化及重载化方向发展，船舶的振动噪声问题日益突出。
大型低温高雷诺数风洞及其关键技术综述廖达雄;黄知龙;陈振华;汤更生【摘要】With the development of air transportations,detail-optimized designs of advanced aircrafts demand aerodynamic data under the flight Reynolds number rge-scale cry-ogenic wind tunnels,such as ETW and NTF,are the best ground testing facilities to obtain air-craft flow characteristics in the real flight conditions.To facilitate the development of large-scale high Reynolds number wind tunnels,the achieving means and types are summarized,the current developing status is discussed,the key technologies and solutions are analyzed in depth for design methodologies and construction concerns.Finally,the future designs and constructions of large-scale continuous cryogenic wind tunnel in China are prospected.%随着航空运输业的发展，先进飞行器的精细化设计要求有飞行雷诺数下的气动数据为支撑。
大型低温高雷诺数风洞（如ETW、NTF）是真实再现飞行器飞行状态流动特性的最佳地面试验设备。
doi: 10.11978/2023026舟山海域夏季上升流的年际变化及其与ENSO 的关系全梦媛, 王慧, 李文善, 王爱梅, 骆敬新国家海洋信息中心, 天津 300171摘要: 本文利用1968—2021年的海表温度和风场数据, 分析舟山海域夏季上升流强度的年际变化, 并结合同期的Niño 3.4指数分析ENSO (El Niño-Southern Oscillation)对上升流的影响。
温度和风上升流指数表明, 1982—2021年夏季舟山海域上升流均呈下降趋势, 下降速率分别为0.062℃·10a −1和0.35m 3·s −1·(100m·a) −1。
近年来, 沿岸风应力的减弱是影响温度上升流指数减弱的一个重要因素。
统计更长时间段内(1968—2021年)El Niño 和La Niña 年风上升流指数的强度发现, El Niño 年平均风上升流指数较小, 仅为−10.33m 3·s −1·(100m) −1。
La Niña 年平均风上升流指数较大为7.60m 3·s −1·(100m) −1, 高于El Niño 和气候态, 且多达4级(比例为75%)。
进一步分析ENSO 与舟山海域风上升流指数的关系发现, ENSO 主要通过影响风的变化进而影响上升流的强度。
El Niño 年, 舟山海域东南风减弱, 导致上升流强度较弱, 甚至发生下降流。
La Niña 年主要为偏南风且风速较大, 更有利于上升流的发展。
关键词: 舟山; 上升流; 年际变化; ENSO; 风场中图分类号: P731.21 文献标识码: A 文章编号: 1009-5470(2024)01-0048-08The interannual variation of summer upwelling in Zhoushan Islands and its relationship with ENSOQUAN Mengyuan, WANG Hui, LI Wenshan, WANG Aimei, LUO JingxinNational Marine Data Information Center, Tianjin 300171, ChinaAbstract: Based on the sea surface temperature and wind data from 1968 to 2021, this paper analyzes the interannual variation of upwelling intensity in Zhoushan in summer, and the impact of El Niño-Southern Oscillation (ENSO) on upwelling. The temperature and wind upwelling indices both show that the upwelling in Zhoushan sea decreased in summer during 1982—2021, with the decreasing rates of 0.062℃·10a −1and 0.35 m 3·s −1·(100m·a)−1, respectively. Recently, the weakened coastal wind stress causes the temperature upwelling index to decrease. According to the results, the wind upwelling index during La Niña events is larger than that during El Niño events and climatology. Further analysis of the relationship between ENSO and the wind upwelling index shows that ENSO affects the intensity of upwelling mainly by influencing the wind. In El Niño events, the southeast wind dominated Zhoushan sea weakens, leading to a decreasing upwelling intensity. While in La Niña events, the enhanced south wind benefits the development of upwelling.Key words: Zhoushan; upwelling; interannual variation; ENSO; wind收稿日期：2023-03-03; 修订日期：2023-03-30。
German-Dutch Wind Tunnels DNWThe foundation DNW is an independent non-profit organisation that has been established by NLR and DLR.In 1971, NLR was developing an atmospheric low speed wind tunnel with an 8x6 m2 test section. At the same time, in Germany, the Deutsches Zentrum für Luft- und Raumfahrt (DLR) defined a similar wind tunnel with tandem test sections of 9.5x9.5m2 and 6x6 m2. Both developments had reached the status of building pilot facilities. From 1974, NLR and DLR started a co-operation that combined the best of both worlds. The German-Dutch Wind Tunnel (singular) [Deutsch-Niederländisches Windkanal DNW, Duits-Nederlandse Windtunnel DNW] was built at NLR NOP site and offers all three test sections mentioned above.In 1993, the low-speed wind tunnels NWB (DLR Braunschweig) and LST (NLR NOP) were added to the foundation that is since then called "German-Dutch Wind Tunnels" (plural) and the original DNW wind tunnel was re-baptized "Large Lowspeed Facility LLF". Around 1998, the remaining industrial wind tunnels of DLR and NLR were added to the DNW foundation, including facilities in Göttingen and Amsterdam.Presently, DNW operates 12 facilities covering a Mach number range from 0 to almost 7.For more information on DNW: DNW web site .NLR used to operate its own wind tunnels until those were placed with the foundation German-Dutch Wind Tunnels DNW. Although NLR and DNW are separate and independently operating foundations, NLR and DNW can (but must not) operate together with transparency to the customer, synergizing their skills in terms of wind tunnel operation and model design.The Dutch National Aerospace LaboratoryThe Dutch National Aerospace Laboratory (NLR) carries out applied research on behalf of the aviation and space sectors. NLR is an independent technological institute.NLR performs research to develop new technologies for aviation and space travel, not only from a scientific perspective, but also for the application of this research in industrial and governmental sectors.NLR has two locations, one in Amsterdam and another about 100 kilometers to the northeast . Approximately 700 people - from aircraft engineers to psychologists, to mathematicians, to materials experts - work for NLR. All these people work constantly to make aviation safer and more environmentally friendly. They support Dutch government policy, they assist the Dutch military and use their own research to enhance the innovative capacity of private businesses. In this way, NLR contributes to more responsive authorities and competitive industries.NLR's clients include governmental authorities, large and small industries, and aerospace organizations - both in the Netherlands and abroad. NLR has a number of specialized research facilities such as wind tunnels, which it operates together with its German sister organization, DLR.NLR is a non-profit organization that carries out market-oriented andsocially-relevant studies. Three-quarters of the research it performs is commissioned by clients. It also receives subsidies to perform basic research. It is one of the Netherlands' major technological institutions.NLR Organisation ChartThe large low-speed facility (LLF) in Marknesse is an industrial wind tunnel for the low-speed domain.Type of wind tunnelClosed circuit, atmospheric, continuous low-speed wind tunnel with one closed wall and one configurable (slotted) wall test section and an open jet.Main featuresClosed wall test sectionsFixed section- 9.5 m x 9.5 m: 0 < V < 62 m/sConfigurable section with the following two configurations- 8 m x 6 m: 0 < V < 116 m/s- 6 m x 6 m: 0 < V < 152 m/sOpen jet- 8 m x 6 m: 0 < V < 80 m/sModel support- Remotely controlled sting support system with four degrees of freedom for models with internal balance- External six-component balance- Floor-based model support system for open jet testing with three degrees of freedomAuxiliary systems- Compressed air supply with a capacity of 5 kg/s continuously at 80 bar- Vacuum system- Moving belt ground plane for ground simulation- Microphone traversing system- Microphone wall arraysTypical tests- Configuration studies, database creation (civil and military transport aircraft, fighters, helicopters, spacecraft, cars and trucks)- Engine integration studies with air-powered simulators- turbofan-powered aircraft by means of TPS- propeller-driven aircraft- Air exhaust simulation with compressed air- Air intake surveys for fighters and helicopters- Aeroacoustic and performance testing on rotorcraft models- Aeroacoustic testing on full-scale aircraft components (landing gears, wings) - Aeroacoustic investigations on scaled turbofans- Full-scale cars and trucks (drag and aeroacoustics)。
CFD predictions of NREL Phase VI RotorExperiments in NASA/AMES Wind tunnelMukesh M. Yelmule*‡, EswaraRao Anjuri VSJ**Engineering Competence Center, Assistant Lead Engineer, Vestas Technology R&D Chennai Pvt. Ltd.yemuk@, esran@‡Corresponding Author; Mukesh M. Yelmule, Vestas Technology R&D Chennai Pvt. Ltd. Block A, 8th Floor, TECCI Park, 173, Rajiv Gandhi Salai, Sholinganallur, Chennai-600119, India., +91 9600012903, yemuk@Received: 17.01.2013 Accepted: 13.02.2013Abstract-This article presents the computational predictions of NREL Phase VI rotor, a stall-regulated two bladed wind turbine with full-span pitch control and a power rating of 20 kW, in the NASA/AMES 80 ft. X 120 ft. wind tunnel. A 3D CFD-RANS approach is used, modeling single blade of the rotor utilizing periodicity, in a rotating frame of reference; over several upwind cases. All the simulations are performed using the commercial multi-purpose CFD solver ANSYS CFX 12.1. The blade is modeled with simplified spherical hub excluding nacelle and tower, at stationary wind conditions neglecting wall shear effects due to tunnel blockage. The comparisons are done for the blade with 0° yaw angle and 3° tip pitch angle. Reasonably good agreement is obtained when comparing modeled mechanical effects Viz. power, thrust, and span wise force components with measurements over wind speeds ranging from 5m/s to 25m/s. The capability of CFD in predicting complex 3D wind turbine aerodynamics is demonstrated in this paper with NREL Phase VI data campaign as a case study.Keywords-NREL-VI Rotor, Navier–Stokes equations, Computational fluid dynamics, Wind turbine aerodynamics, Simulation.1.IntroductionAccurate aerodynamic predictions are required in the design of new rotor blades and additional passive/active performance improvement devices. This requires continued validation of new and existing design tools & methods, increased accuracy and efficiency of the results. CFD is one such design tool & extensive research has been done in developing the CFD tools and methods for predicting aerodynamics of wind turbines during the last few years.During the last decade, CFD modeling of wind turbines has evolved from scientific work performed at research institutions and investigations performed at wind turbine manufacturers with the application of commercial codes. Traditionally the wind turbine blades are designed using first principles (BEM theory) utilizing 2D airfoil tables from wind tunnel. Empirical corrections are used to account for 3D effects Viz. tip losses, root losses, rotational effects, and dynamic stall effects. High fidelity CFD naturally includes these phenomena, but has more difficulty in modeling and other wind turbine phenomena such as variable turbulent inflow and boundary layer transition [7]. CFD has been used to improve the aerodynamic design of wind turbines including tip shapes, winglets and hub modeling [8, 9, 10, and 11] where it captures flow physics better at which BEM models are no longer applicable. High fidelity Navier-Stokes based computational fluid dynamics is currently making inroads into many phases of industrial wind energy design [1, 2]. CFD is used for the analysis of both 2D airfoils and also 3D blades [1, 2, 3, 4, and 5].The NREL Phase VI Unsteady Aerodynamic Experiment [1, 2, and 18] provides an excellent validation test case for 3D CFD Rotor analyses. The Phase VI test campaign performed in the NASA Ames National Full- Scale Aerodynamic Complex (NFAC) was completed in the year 2000. The 2-bladed, 10.058m diameter, stall regulated turbine has a power rating of 20kW. The blades are twisted and mildly tapered. Multidisciplinary measurements were obtained over a wide range of operating conditions. Experimental measurements included blade pressures and resulting integrated air loads, shaft torque, sectional inflow conditions, blade root strain, tip acceleration and wakevisualization. Both upwind and downwind configurations with rigid and teetering blades were run for speeds from 5 m/s to 25 m/s. Yawed and unsteady pitch configurations are also available. Free and fixed transition results were measured. The blade uses specially designed S809 airfoil for which experimental aerodynamic performance parameters are available. Blade structural properties are well documented [21]. Various researchers [3, 12, 13, 14, 15, 16, 17, 18, 19, and 20] have investigated this configuration numerically using a range of CFD methods and grid topologies.Researchers at Risø computed the isolated rotor with and without wind tunnel walls using a multi-block, structured mesh, incompressible solver EllipSys3D with a RANS turbulence model [18] and a detached eddy simulation [19]. Performance was generally well captured although stall initiation at 10 m/s wind speed was missed.The objective of the present work is to validate the CFD method utilizing commercial multi-purpose CFD solver ANSYS CFX 12.1 & Multi block structured mesh generator ICEM-CFD, with the NREL Phase VI rotor (Test sequence S) wind turbine experiments. This report mainly consists of ∙ CFD Modeling of the NREL Phase VI rotor ∙ Comparison of integrated quantities ∙ Flow visualization∙ Comparison of span-wise sectional details ∙ Comparison of 2D airfoil characteristics ∙ Comparison with other CFD predictions ∙Conclusions2. CFD Modelling of NREL Phase VI RotorIn the present work a compressible Navier-Stokes solver (CFX) is utilized to predict the aerodynamics of the Phase VI rotor from the National Renewable Energy Laboratory. The two-bladed 10.058m diameter rotor geometry is based on the S809 airfoil. The details about the blade & measurement convent ions can be found in [21]. The rotor cone angle is 0° and the tip pitch angle is set to 3°. In this investigation, only Sequence S upwind configuration is examined, and the operational conditions for the cases computed can be found in Table 1.Table 1. Sequence S Operating conditions.In the current work, a single blade is modeled in CFD considering the periodic boundary conditions that isequivalent to 180o periodic sector of the rotor, to save computational resource. Only the wind speed, RPM and density are used as input variables for CFD simulations without any empirical tuning of the existing models. ANSYS CFX 12.1 uses a finite-volume based unstructured parallelized coupled algebraic multi-grid solver with a second order advection scheme and second order overall accuracy [22]. The computations have been performed with compressible Reynolds Averaged Navier-Stokes (RANS) equations and the SST [23] turbulence model. The transition from laminar to turbulent flow is modeled using Langtry and Menter correlation based Gamma-Theta transition model. The default correlations in the model are proprietary of ANSYS and therefore not known in detail by the user. In general the default correlation for R eθt is based on the free stream turbulence intensity and the pressure gradient outside the boundary layer. The value of R eθt determined outside of the boundary layer is diffused into the boundary layer by a standard diffusion term. The physics of the transition process is not directly modeled by the two additional transport equations. Instead, the physics of the transition process is entirely contained in the underlying experimental correlations.As the turbine is upwind type, exclusion of tower in the CFD model has negligible effect on rotor aerodynamics & is a sound choice. The theoretical definition of the S809 airfoil has a very sharp trailing edge; whereas the geometry used for CFD simulations has trailing edge thickness of 1mm along the entire span of blade that resembles the actual blade used for experimentation. Rotor computations are stationary, at constant uniform wind speed, constant pitch and RPM neglecting the unsteady inflow, which is a fair choice considering that the experimental data set is arrived statistically from a large number of repeated measurements. Uniform velocity normal to the inlet is used at inlet boundary and atmospheric static pressure is used at outlet and far-field boundaries. Blade and hub surfaces are defined as no-slip walls with specified rotation. Figure 1 shows different boundaries and the blade.Figure 2 shows the mesh on different boundaries including the blade. All computations are run in parallel on the computing cluster.Fig. 1. Computational domain, boundaries and blade.Fig. 2. Computational mesh on boundaries and blade.The steady state simulations are performed for approximately 900 iterations ensuring convergence (residuals <=10e-4 & imbalances <1 %).It took approximately 24 hours of computing time with 32 CPU’s.parison of Integrated QuantitiesMechanical Power (P) is calculated by monitoring the torque T about the flow axis and multiplying with the angular velocity Ω (as shown in equation 1).P = TΩ (1) Fig. parison of measured and CFD Integrated quantities.It is observed from Fig.3 that integrated quantities Viz. Power & Thrust, from CFD compare well with experimental results at all wind velocities, except at 10 m/s, where an over-prediction up to 20 % is observed.4.Flow VisualisationsFigure 4 shows surface streamlines and Fig.5 shows turbulence intermittency contours on the suction side of blade indicating transition at 7 m/s, 10 m/s and 20 m/s wind speed. The vertical lines show span-wise sections Viz. 30%, 46.7%, 63.3%, 80%, 95%, where pressure measurements are available .It is observed that at low wind speed up to 7 m/s, the flow is attached except up to 30 % span & the transition line is clearly visible. At 10 m/s the flow is separated over the entire span except close to 95 % span, where transition can be clearly seen , and close to mid-span (47.7% span) the separation line has moved to the leading edge. At 20 m/s the separation has spread over the entire blade and the flow is fully turbulent across the span.parison of Span-Wise Sectional DetailsFrom the experiment pressure measurements are available at five span-wise sections Viz. 30 %, 46.7%, 63.3%, 80%, 95% r/R. The stagnation point dynamic pressure is used to non-dimensionalise the pressure.Fig. 4. Surface streamlines at 7 m/s, 10 m/s & 20 m/s wind speed.Fig. 5. Turbulence intermittency on suction side of the bladeat 7m/s, 10 m/s & 20 m/s wind speed.Fig. 6. Comparison of CFD and measured pressure distributions at 5m/s wind speed. Fig. 7. Comparison of CFD and measured pressure distributions at 7m/s wind speed.Figure6 and Fig.7 show pressure distribution at 5m/s and 7m/s wind speed, which is categorized as low wind speed region. Referring Fig.4 to Fig.7, the flow is mostly attached and is in good agreement with the measured pressure distribution except up to 30% span, where flow is separated and we observe some deviation that is due to a known difficulty of RANS turbulence models in solving separated flow.Figure 8 shows pressure distribution & force coefficients at 10 m/s wind speed which is categorized as onset of stall. A discrepancy in pressure near the leading edge of suction surface (Peak suction pressure) is observed at 46.7 % span due to flow separation close to mid-span and resulting localized transient stall effects making the peak suction pressure practically difficult to capture in the experiments. The same can be observed from sudden dip of tangential force coefficient CT at 46.7% span.Figure 9 & Fig.10 show pressure distribution and force coefficients at 13 m/s & 15 m/s respectively, which are categorized as stall region, where the flow is separated over the entire blade except close to the tip. The separation that started at mid span for 10 m/s wind speed moves progressively towards the outer span of blade with increase in wind speed, the same is observed from widening ofCTdip. Figure 11 & Fig.12 show pressure distribution and force coefficients at 20 m/s & 25 m/s respectively, which are categorized as deep stall region, where the flow is separated over entire span and the blade is completely stalled. Deviation is observed in the suction side pressure distribution in stall and deep stall region, due to difficulty of RANS models in solving separated flows. This difference however does not cause substantial differences to integrated quantities, which is a characteristic specific to S809 airfoil at higher angle of attack and is in agreement with observations from [18].Although quantitative differences exist in the normal & tangential force coefficients between computed andexperimental data, the trends agree well for all the wind velocities. The fact that the number of pressure probes placed along the airfoil to reconstruct the pressure distribution from experiments has a physical limitation compared to no limitations in CFD simulations reflects in deviations. For e.g. in experiment data at 22 pressure probes is available along the airfoil, while CFD simulations have 700 grid/data points along the airfoil. Thus CN, CT calculated with experimental pressure distribution assumes linear pressure variation between any two consecutive measurements, whereas CFD has a much finely resolved pressure profile.Fig. 8. Comparison of CFD and measured pressure distributions at 10m/s wind speed.Fig. 9. Comparison of CFD and measured pressure distributions at 13m/s wind speed.Fig. 10. Comparison of CFD and measured pressure distributions at 15m/s wind speed.Fig. 11. Comparison of CFD and measured pressure distributions at 20m/s wind speed.Fig. 12. Comparison of CFD and measured pressure distributions at 25m/s wind speed.parison of 2D Airfoil CharacteristicsFigure 13 shows the 2d characteristics extracted from 3D CFD at 5 span-wise locations Viz. 29%, 48%, 66%, 79%& 93% using the method described in [25].Fig. 13. 2D characteristics extracted from 3D CFDIt is compared with 2D OSU WT (wind tunnel) data of S809 airfoil with natural transition. It is observed that the lift is comparatively higher in the inboard section for high angle of attack that is in good agreement with theoretical himmelskamp effect. It is also observed that the drag is comparatively higher, thus the higher lift in inboard sections is at the cost of higher drag. For the outer span-wise locations Viz. 66% and 79 % span Cl & Cd compare well with WT data , these locations thus are least affected by inboard and outboard radial flow. At 93 % span the tip effects influence the flow and lowers Cl & Cd that again is in good qualitative agreement with theory.parison with Other CFD PredictionsFigure 14 shows the comparison of our CFD results (CFX) with other CFD predictions Viz. BEM, Risoe (Ellipsys 3D) & GRI (Acusolve) for NREL Phase VI rotor test sequence S. Note that the Risoe (Ellipsys 3D) predictions are for fully turbulent flow conditions.Fig. 14.Comparison of CFD results with other researchersA good comparison is observed for CFX predictions with experiments and other CFD researchers.8.ConclusionComputational fluid dynamics calculations have been executed for NREL Phase VI rotor at upwind conditions with 0° yaw and 3° tip pitch. A single blade has been analyzed with 180° periodicit y and structured mesh using commercial multi-purpose CFD solver ANSYS CFX 12.1. The steady state CFD data is compared with the measured wind tunnel data. CFD rotor computations show good qualitative and quantitative comparison with measurements except at the onset of stall (10 m/s).At low wind speed (5 m/s & 7 m/s) flow is attached except up to 30 % span and transition occurs close to mid-chord. CFD predictionss are accurate & within the measurement range.At the onset of stall (10 m/s) separation occurs close to mid span at leading edge. Steady state CFD fails to predict the power and thrust within measurement range and is well known for its difficulty in capturing highly transient effects at the onset of stall. This reflects in over-prediction as high as 20 % in power.In stall (13 m/s and 15 m/s) and deep stall (20 m/s & 25 m/s), the separation that initiated at mid-span (at 10 m/s) progressively moves over the entire span with increase in wind speed. Steady state CFD predicts integrated quantities within measurement range, although quantitative differences are observed in pressure distribution on suction side, due to the specific stall behavior of S809 airfoil.Additionally flow visualization and comparison of 2D airfoil characteristics extracted from 3D CFD with OSU WT airfoil data gives more insight and understanding of complex 3D effects making CFD computations more competitive & generic compared to design methods based on 2D theoretical models that rely on empirical tuning & corrections. AcknowledgementsWe acknowledge Vestas Technology R&D Chennai Private Limited, ECC (Engineering Competence Centre), Simulation competence for the technical and computational resources availed during these CFD computations. We are also grateful to Vestas Americas, GRI (Global Research and Innovation) for collaborating with NREL & making authentic NREL data available for comparison. References[1]Larsen, J., “ANSYS CFD Applied to Wind Turbines atSiemens Wind Power,” ANSYS Conference & 26th CADFEM Users’ Meeting, Darmstadt, Germany, October 2008.[2]Standish, K., “CFD at GE”, Wind Power 2007, L osAngeles, CA, June 2007. [3]Le Pape, A., and Gleize, V., “Improved Navier-StokesComputations of a Stall-Regulated Wind Turbine Using Low Mach Number Preconditioning,” 44th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, January 2006, AIAA 2006-1502.[4]Standish, K. J., and van Dam, C. P., “AerodynamicAnalysis of Blunt Trailing Edge Airfoils,” Journal of Solar Energy Engineering ,Vol. 125, No. 4, Nov. 2003, pp. 479-487.[5]Fuglsang, P., and Bak, C., “Development of the RisøWind Turbine Airfoils,” Wind Ene rgy, Vol. 7. No. 2, May 2004, p. 145-162.[6]Xu, G. and Sankar, L., “Development of EngineeringAerodynamics Models Using a Viscous Flow methodology on the NREL Phase VI Rotor”, Wind Energy, vol. 5, no. 2-3, pp. 171-183, 2002.[7]Langtry, R. B, Gola, J, and Ment er, F. R, “Predicting 2DAirfoil and 3D Wind Turbine Rotor Performance using a Transition Model for General CFD Codes,” 44th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, January 2006, AIAA-2006-395.[8]Johansen J., Madsen, H. A., Sørensen, N. N., an d Bak C.,“Numerical Investigation of a Wind Turbine Rotor with an Aerodynamically Redesigned Hub-Region,” 2006 European Wind Energy Conference and Exhibition, Athens, Greece, 2006.[9]Johansen J., and Sørensen N. N.: “Aerodynamicinvestigation of winglets on wind turbine blades using CFD”, Risø-R- 1543(EN) report 2006.[10]Hansen, M. O. L., and Johansen, J., “Tip StudiesUsing CFD and Comparison with Tip Loss Models,”Wind Energy, 2004, p. 343 -356.[11]Hjort, S., Laursen, J., and Enevoldsen, P.,“Aerodynamic Winglet Optimization,” Sandia National Lab Blade Workshop, May 2008.[12]Duque, E. P. N., Burklund, M. D., and Johnson, W.,“Navier-Stokes and comprehensive analysis performance predictions of the NREL phase VI experim ent,” Journal of Solar Energy Engineering 2003; 125: 457-467. [13]Chao, D. D., and van Dam, C. P., “ComputationalAerodynamic Analysis of a Blunt Trailing-edge Airfoil Modification to the NREL Phase VI Rotor,” Wind Energy, July 2007, 10:529-550.[14]Gonzalez A., a nd Munduate, X., “Three-dimensional and Rotational Aerodynamics on the NREL Phase VI Wind Turbine Blade,” 45th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, January 2007, AIAA 2007-0628.[15]Schmitz, S., and Chattot, J-J., “Application of a‘Paralleli zed Coupled Navier-Stokes/Vortex Panel Solver’ to the NREL Phase VI Rotor,” 43rd AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, January 2003, AIAA 2003-0593.[16]Zahle, F., Johansen, J., Sorenson, N., and Graham,J., “Wind Turbine Rotor-Tower Interaction Using an Incompressible Overset Grid Method,” AIAA 45th Aerospace Sciences Meeting and Exhibit, Reno, NV, January 2007, AIAA 2007-0425.[17]Simms, D. A., Schreck, S., Hand, M., and Fingersh,L., J., “NREL Unsteady Aerodynamics Experiment in the NASA Ames Wind Tunnel: A Comparison of Predictions to Measurements” NREL/TP-500-29494, June 2001. [18]Sørensen, N. N., Michelsen, J. A., and Schreck, S.,“Navier-Stokes Predictions of the NREL Phase VI rotor in the NASA Ames 80ft x120 ft wind tunnel,” Wind Energy, Vol. 5, No. 2-3, 2002, pp. 151-169.[19]Johansen, J., Sørensen, N. N., Michelsen, J. A., andSchreck, S., “Detached-Eddy Simulation of Flow around the NREL Phase-VI Rotor,” Wind Energy, Vol. 5, No. 2-3, 2002, pp. 185-197.[20]Mark A. Potsdam, Dimitri J. Mavriplis,“Unstructured Mesh CFD Aerodynamic Analysis of the NREL Phase VI Rotor”, AIAA 2009-1221, 47th AIAA Aerospace Sciences Meeting Including The NewHorizons Forum and Aerospace Exposition, 5 - 8 January 2009, Orlando, Florida.[21]M.M. Hand, D.A. Simms, L.J. Fingersh, D.W.Jager, J.R. Cotrell, S. Schreck, and S.M. Larwood, “Unsteady Aerodynamics Experiment Phase VI: Wind Tunnel Test Configurations and Available Data Campaigns”, NREL/TP-500-29955, December 2001. [22]Technical information regarding the ANSYS-CFXsolver: /products/cfxadvanced-solver.asp.[23]Menter FR. Two-equation eddy-viscosity model forengineering applications. AIAA-Journal 1994;32(8):1598-1605.[24]Robin Langtry, Florian Menter, “Overview ofIndustrial Transition Modeling in CFX”, ANSYS Germany GmbH, ANSYS CFX.[25]Jeppe Johansen and Niels N. Sørensen, “AerofoilCharacteristics from 3D CFD Rotor Computations”, Risø National Laboratory, PO Box 49, DK-4000 Roskilde, Denmark.。
Alan G. Davenport Wind Engineering GroupThe Boundary Layer Wind Tunnel LaboratoryThe University of Western Ontario, Faculty of Engineering ScienceLondon, Ontario, Canada N6A 5B9; Tel: (519) 661-3338; Fax: (519) 661-3339Internet: www.blwtl.uwo.ca; E-mail: info@blwtl.uwo.caTABLE OF CONTENTS1INTRODUCTION 1 2THE MODELLING OF THE SITE AND THE WIND 22.1General 22.2Scaling 2 3THE CLADDING LOADS TESTS 33.1Pressures and Suctions on Exterior Surfaces 33.2Scaling 33.3Internal Pressures and Differential Pressures 3 4THE DETERMINATION OF OVERALL STRUCTURAL LOADS AND RESPONSES 44.1Introduction 44.2The Force Balance Test 44.3The Two-degree-of-freedom Aeroelastic Test 54.4The Multi-degree-of-freedom Aeroelastic Test 54.5Overall Loads from Local Pressure Measurements 64.6Effective Static Force Distribution 64.7Load Combination Factors 7 5THE PEDESTRIAN LEVEL WIND SPEED TEST 8 6THE TESTING OF LONG SPAN BRIDGES 9 7OTHER TESTS 10 REFERENCES 11 APPENDIX A A-1 THE DEFINITION OF WIND CLIMATE A-1A.1Introduction A-1A.2Natural Wind A-1A.3Availability of Wind Records A-2A.4Probability Distribution of Mean Wind Speed and Direction A-3A.5Applicability of the Wind Climate Model A-4 APPENDIX B B-1 THE DEFINITION OF A HURRICANE WIND CLIMATE B-1B.1Introduction B-1B.2The Approach Used B-1B.3Verifying the Approach B-2B.4The Wind Climate for a Particular Site B-2 APPENDIX C C-1 THE MEASUREMENT AND PREDICTION OF SURFACE PRESSURE C-1C.1Experimental Technique C-1C.2Experimental Time Scale C-1C.3Choice of Sampling Period C-2C.4Definition of the Pressure Coefficients C-2C.5General Characteristics of the Pressure Response C-3C.6Predictions of Peak Pressures and Suctions C-3C.7Limitations on the Predicted Peak Pressures and Suctions C-3 APPENDIX D D-1 PREDICTING PEAK RESPONSES FOR VARIOUS RETURN PERIODS D-1D.1Introduction D-1D.2The Prediction Process D-1D.3The Rate of Up-crossing of Peak (Maximum or Minimum) Response Values D-2 APPENDIX E E-1 STORM PASSAGE PREDICTIONS OF WIND LOADS AND RESPONSES E-1E.1Overview E-1E.2Extreme-value Predictions from Time-domain Analysis E-1E.3Examples of Predictions of Wind Loads and Effects E-2E.4Wind Directionality E-3 APPENDIX F F-1 DETERMINATION OF INTERNAL PRESSURE COEFFICIENTS AND THE FORMATION OF DIFFERENTIAL PRESSURE COEFFICIENTS F-1F.1Summary F-1F.2Introduction F-1F.3Mean Internal Pressures: Distributed Leakage F-1F.4Mean Internal Pressures: Large Openings F-2F.5Fluctuating Internal Pressures F-3F.6Forming Differential Pressure Coefficients at BLWTL F-4F.7Limitations of Predicted Differential Pressures F-5 APPENDIX G G-1 DETERMINATION OF TOTAL DYNAMIC LOADS USING A RIGID MODEL/FORCE BALANCE TECHNIQUE G-1G.1Summary G-1G.2Introduction G-1G.3Concepts of the Force Method G-1G.4Concept of the Balance G-2G.5Linear Elastic Response Calculations with the BLWT Balance G-2G.6Torsional Response G-4 APPENDIX H H-1 AEROELASTIC SIMULATIONS OF BUILDINGS USING TWO-DEGREE-OF-FREEDOM MODELS H-1H.1Introduction H-1H.2Aeroelastic Modelling H-1H.3Details of the Aeroelastic Model H-2H.4Experimental Procedure and Preliminaries H-2H.5Predictions of Peak Wind-Induced Response H-3 APPENDIX I I-1 AEROELASTIC SIMULATIONS OF BUILDINGS USING MULTI-DEGREE-OF-FREEDOM MODELS I-1I.1Introduction I-1I.2Aeroelastic Modelling I-1I.3Details of the Aeroelastic Model I-2I.4Experimental Procedure and Preliminaries I-2I.5Predictions of Peak Wind-induced Response I-3APPENDIX J J-1DETERMINATION OF TOTAL DYNAMIC LOADS FROM THE INTEGRATION OF SIMULTANEOUSLY MEASURED PRESSURES J-1 J.1Summary J-1 J.2The Integration Procedure J-1 J.3Response Calculations J-1APPENDIX K K-1THE EVALUATION AND USE OF EFFECTIVESTATIC FORCE DISTRIBUTIONS K-1 K.1Effective Static Force Distributions K-1 K.2Combined Load Cases K-2 APPENDIX L L-1THE PEDESTRIAN LEVEL WIND ENVIRONMENT L-1 L.1Introduction L-1 L.2Test Procedure L-1 L.3Statistical Predictions of Pedestrian Level Winds L-1 L.4Acceptance and Safety Criteria for Pedestrian Level Wind Conditions L-2 APPENDIX M M-1DYNAMIC WIND FORCES ON LONG SPAN BRIDGES USING EQUIVALENT STATIC LOADS M-1 M.1Introduction M-1 M.2The Description of Design Loads M-1M-1 M.3Evaluation of the Modal Load W and W12M.4Experimental Determination of Design Load Components M-4 M.5Determination of Design Wind Loads M-4 M.6Conclusions M-5TABLE L.1CRITERIA FOR PEDESTRIAN COMFORT AND SAFETY............................................L-3 TABLE L.2EXTRACTS FROM THE BEAUFORT SCALE...............................................................L-4FIGURE B.1COMPARISON OF TYPHOON WIND SPEEDS AT WAGLAN ISLAND TOMEASURED DATA CORRECTED FOR TOPOGRAPHIC EFFECTS..........................B-5 FIGURE B.2PREDICTED PEAK ACCELERATIONS FOR THE ALLIED BANK DURINGHURRICANE ALICIA......................................................................................................B-6 FIGURE D.1 ILLUSTRATION OF THE PREDICTION PROCESS....................................................D-4 FIGURE E.1OBSERVED 10-MINUTE AVERAGE SURFACE WIND SPEED AND WINDDIRECTION AT HONG KONG DURING TYPHOON YORK (SEPTEMBER 16,1999)...............................................................................................................................E-5 FIGURE E.2 TYPICAL WIND INDUCED RESPONSE SHAPES.......................................................E-5 FIGURE E.3COMPARISON OF GENERIC WIND LOADS AND EFFECTS PREDICTEDFOR DIFFERENT WIND DIRECTIONS USING CONVENTIONALSTATISTICAL METHODS AND TRACKING THE EFFECTS OF INDIVIDUALSTORMS.........................................................................................................................E-6 FIGURE E.4COMPARISON OF PEAK STRUCTURAL WIND LOAD EFFECTS FORBUILDINGS A AND B HYPOTHETICALLY LOCATED IN DIFFERENT WINDREGIONS PREDICTED BY CONVENTIONAL STATISTICAL METHODS ANDBY TRACKING THE EFFECTS OF INDIVIDUAL STORMS........................................E-6 FIGURE E.5COMPARISONS OF PREDICTED LOCAL PEAK PRESSURES ANDSUCTIONS FOR A SPECIFIC BUILDING AND FOR A GENERIC PRESSURECOEFFICIENT DATA SET IN DIFFERENT WIND REGIONS USINGCONVENTIONAL STATISTICAL METHODS AND BY TRACKING THEEFFECTS OF INDIVIDUAL STORMS............................................................................E-7 FIGURE E.6DIRECTIONALITY FACTORS FOR GENERIC PEAK STRUCTURAL LOADSAND RESPONSES USING CONVENTIONAL PREDICTIVE METHODS ANDSTORM PASSAGE TRACKING....................................................................................E-7 FIGURE G.1 DYNAMIC RESPONSE OF THE BALANCE-MODEL COMBINATION.......................G-6 FIGURE H.1 SCHEMATIC OF THE AEROELASTIC MODEL.........................................................H-4 FIGURE I.1SCHEMATIC OF THE AEROELASTIC MODEL.............................................................I-4 FIGURE M.1DISTRIBUTED WIND LOAD COMPONENTS...............................................................M-6 FIGURE M.2NOTATION.....................................................................................................................M-6 FIGURE M.3SPECTRUM OF MODAL LOAD AMPLITUDE..............................................................M-7 FIGURE M.4SUNSHINE SKYWAY BRIDGE......................................................................................M-7 FIGURE M.5SECTION MODEL RESPONSE (UNCORRECTED).....................................................M-8 FIGURE M.6VERTICAL VELOCITY SPECTRUM..............................................................................M-9 FIGURE M.7 AERODYNAMIC ADMITTANCE RESPONSE (UNCORRECTED).............................M-9FIGURE M.8 JOINT ACCEPTANCE FUNCTION.............................................................................M-10 FIGURE M.9DAMPING FUNCTIONS...............................................................................................M-10 FIGURE M.10WIND LOAD COMPONENTS ON COMPLETED BRIDGE.........................................M-111 INTRODUCTIONThis document provides a general outline of common wind tunnel tests performed at the Boundary Layer Wind Tunnel Laboratory (BLWTL) at the University of Western Ontario. It also details some of the techniques used to analyse the data from these tests. Since it is a general outline, it will cover some tests and analyses not performed for a particular project. Other than the wind climate modelling discussed in Section 2 and Appendices A and B, and the prediction methodology discussed in Appendices D and E, the various tests and analysis methodology are independent and the reader may skip sections that are not relevant. Also, this report does not, by any means, attempt to cover all of the types of tests and analyses performed at the Laboratory. Unusual tests are covered in separate reports for the projects employing them.In determining the effects of wind for a particular development, there are two main ingredients to consider. The first comprises the aerodynamic characteristics of the development. These are simply the effects of the wind when it blows from various directions. This information only has limited value, however, without knowing how likely it is that the wind will blow from those directions and how strongly it is likely to blow. This climatological information, in the form of a probability distribution of wind speed and direction, is the second main ingredient needed for determining wind effects for a particular development. The aerodynamic information is characteristic of the particular development and its immediate surroundings, while the wind climate information is characteristic of the geographical location of the development. Both are necessary to determine the wind effects for a particular development and, when combined, provide statistical predictions of the wind effects which are independent of wind direction.At the BLWTL, the aerodynamic characteristics of the development are commonly determined through model studies of the project. These studies may include measurements of various types of information of interest, such as cladding loads, structural loads and pedestrian level wind speeds, as detailed in the following sections of this report. The probability distribution of wind speed and direction, is determined from analyses of historical wind speed and direction records taken near the site of the development. Details of these analyses are included in Appendix A. Tropical cyclones, such as hurricane or typhoon winds present a special case and their associated statistical characteristics are handled separately using different analysis methods. These are detailed in Appendix B.In all cases, tests carried out at the BLWTL are in accordance with the state-of-the-art, and meet or exceed such test requirements as documented by the ASCE Manual of Practice (1).2 THE MODELLING OF THE SITE AND THE WIND2.1 GeneralThe basic tool used is the Laboratory's Boundary Layer Wind Tunnel. This wind tunnel is designed with a very long test section, which allows extended models of upwind terrain to be placed in front of the model of the development under test. The modelling is done in more detail close to the site. The wind tunnel flow then develops characteristics which are similar to the wind over the terrain approaching the actual site. This methodology has been highly developed and further details can be found in References 2, 3 and 4.The modelling is comprised of the following components:1. A detailed model of the development. Different types of model are used for the various types oftest. These are discussed below in the sections on the individual tests.2. A detailed proximity model of the surrounding area, built in block outline from wood andStyrofoam. Depending on the scale and size of the model, this may extend for a radius of approximately 500 to 600 metres.3. Coarsely modelled upstream terrains, chosen to represent the general roughness upstream ofthe site for particular wind directions. Typically, several models are chosen, each used for a range of wind directions.For project sites close to hilly terrain or with unusual topography, topographic study may be carried out to establish the wind characteristics at the site. This may be in the form of topographic model study at a small scale (~1:3000) or computational methods. The resulting target wind characteristics will be modelled in the larger scale used in the building or bridge tests.2.2 ScalingThe fundamental concept is that the model of the structure and of the wind should be at approximately the same scale. The natural scaling of the flow in the wind tunnel is in the range 1:400 to 1:600; however, in some cases, instrumentation or other requirements may demand a larger model. In these cases, additional flow modification devices may be used to approximate larger scale flows.In all cases, it is the mean wind speed profile and the turbulence characteristics over the structure that are most important to match with those expected in full scale. Guidance as to the latter is obtained through direct full scale measurements as compiled by ESDU (5, 6). Such data are also used to ensure that the test speeds near the top of the building are properly interfaced with full scale wind speeds predicted to occur at the full scale site.3 THE CLADDING LOADS TESTS3.1 Pressures and Suctions on Exterior SurfacesDetailed measurements of the pressures and suctions on exterior surfaces of the building or structure are made using a rigid model that accurately represents the detailed exterior geometry of the development. The model contains numerous (typically 300 to 800) holes or "taps" which are connected via tubing to pressure transducers. The transducers convert the pressure at the point where the tap is located to an electrical signal which is then measured by the Laboratory's computerized data acquisition system. The technology employed allows all pressures on the building to be measured essentially simultaneously for a particular wind direction. Measurements are usually made at 10° intervals for the full 360° azimuth range. A detailed description of the procedures followed and the definitions used are presented in Appendix C.These aerodynamic measurements made in the wind tunnel are subsequently combined with the statistics of the full scale wind climate at the site using the methodology outlined in Appendices D and E to provide predictions of pressures and suctions for various return periods.3.2 ScalingThe aerodynamic pressure coefficients can be converted to full scale pressure values based on consistent length, time and velocity scaling between full scale and model scale. This applies very well for sharp-edged structures. For structures with curved surfaces, additional care has to be taken to ensure that the flow regime is consistent in model and full scale, as well as in the interpretation of the results.For typical building tests, length scale is in the order of 1:300 to 1:500. Velocity scale is approximately 1:3 to 1:5. Time scale is in the order of 1:100. For example, 36 seconds in model scale represents about an hour in full scale and the data will be taken about 100 faster in the test than in full scale. Further details regarding scaling can be found in Appendix C.3.3 Internal Pressures and Differential PressuresThe net load on cladding is the difference between the external and internal pressures. Using the methodology described in Appendix F, mean internal pressures are determined at all wind angles. These are then subtracted from the appropriate external pressure coefficients to form differential pressure coefficients. Finally, the coefficients are combined with the statistics of the full scale wind climate at the site, using the methodology outlined in Appendices D and E, to provide predictions of differential pressures and suctions for various return periods.In the case of large opening due to operable windows or breach of the building envelope, large internal pressures may develop. Typically, the external pressure at the opening will be transmitted into the building interior volume. Building envelope at other locations within the building volume will experience both the external pressures at those external locations as well as the large internal pressure transmitted from the opening.For free standing elements with both sides exposed to air, such as parapets and canopies, the net differential pressures are the instantaneous difference in pressures on the opposite sides.4 THE DETERMINATION OF OVERALL STRUCTURAL LOADS ANDRESPONSES4.1 IntroductionThe dynamic response of most tall buildings to wind is primarily the results of building motions in the fundamental sway and torsion modes of vibration with relatively small contributions from higher modes. The mechanical transfer function, relating the load function to the response, is straightforward. On the other hand, the aerodynamic transfer function, relating the gust structure to the wind induced forces is difficult to establish without wind tunnel model tests. A further complication exists if body motion effects interact with the load function (aerodynamic damping).Multi-degree-of-freedom aeroelastic models have traditionally been used to study the action of wind on sensitive buildings and structures. While such simulations provide the most direct and reliable estimates, the required models are expensive and time consuming to design and construct. Two-degree-of-freedom aeroelastic models, which simulate the wind induced responses in the two fundamental sway modes of vibration, while less expensive, do not provide information on torsion effects, which may be significant for buildings of unusual shape and structural dynamic properties. In both of these cases, the model moves in the wind tunnel just as it would in full scale; its response in the wind tunnel can be scaled directly to full scale.A high-frequency balance/model system can measure the load function directly, provided that aerodynamic damping effects are negligible, which is usual for most buildings at practical wind speeds and practical structural damping values. (Note that if such effects are important, they can be accounted for by using a supplemental testing technique in which the model is oscillated). The now commonly used high frequency force balance technique was originally developed at the BLWTL.One other method for determining the load function is to integrate the point pressure measurements on an instant-by-instant basis to form time histories of the generalized forces.4.2 The Force Balance TestThis technique involves testing a lightweight, stiff, geometrical representation of the building on an ultra-sensitive force balance. The technique allows direct measurements of good approximations to the steady and unsteady modal forces acting in the fundamental sway and torsional modes of vibration of the building. The dynamic responses including resonant amplification at the natural frequencies of the building are derived analytically for each mode using random vibration analysis methods and are subsequently used to provide estimates of the full scale responses of the building. As a result, this method is very accommodating of changes to the structural properties after testing, since the analytical procedure can be simply repeated using the same experimental results, which remain applicable so long as the aerodynamic characteristics of the building remains the same. A detailed description of this method is presented in Appendix G.Time histories of the base shears and moments are taken during the tests. From these, the mean and rms (root-mean-square) base bending moments along orthogonal building axes, as well as mean and rms base torque are determined. The base bending moments represent a good approximation to the generalized modal forces in the fundamental modes. The spectra of the generalized modal forces are also determined from these time histories and used to determine the resonant component of the response. Measurements are usually taken at 10° intervals for the full 360° azimuth range.Scaling is essentially the same as for pressure tests, requiring scaling of the structure to the flow model only, with the non-dimensionalized data being independent of test speed; however, in this case, the time scaling must be chosen carefully to ensure that the prototype’s natural periods of interest fall within the accurate measuring range of the model/balance combination.Once the responses have been determined, they are combined with the statistics of the full scale wind climate at the site, using the methodology outlined in Appendix D, to provide predictions of loads and responses for various return periods.4.3 The Two-degree-of-freedom Aeroelastic TestThis technique requires scaling the dynamic properties (mass, stiffness and period) of the building in the fundamental sway modes and measuring the response to wind loads directly. The building is modelled as a rigid body, pivoted near the base, with the elasticity provided by appropriately selected springs. Implicit in this technique is the assumption that the sway modes do not include any coupling and can be approximated as linear, and that torsion is unimportant. These prove to be reasonable assumptions for a large range of buildings. A full discussion of this technique is contained in Appendix H.The advantage of this technique is that the measurements will include effects of aerodynamic damping that are not included when using the force balance technique. It is also a simpler, less expensive technique than a multi-degree-of-freedom aeroelastic test. The disadvantages of the technique are that it is limited by the assumptions noted above and it is more complicated and expensive than the force balance technique while being less accommodating of changes to the dynamic properties of the building after the test. Furthermore, its advantage over the force balance technique, namely the inclusion of aerodynamic damping effects, rarely proves to be necessary since the aerodynamic damping is usually small and positive (i.e. it reinforces the inherent structural damping), but can be negative if vortex shedding plays an important role in the dynamic response. For most buildings, vortex shedding occurs at speeds well above the range of speeds that the structure will be subjected to. Nevertheless, it is useful for confirming force balance results and confirming that aerodynamic damping effects are indeed negligible.As with other types of tests, once the aerodynamic data has been measured for a full range of wind directions, it is combined with the statistics of the full scale wind climate at the site, using the methodology outlined in Appendix D, to provide predictions of loads and responses for various return periods.4.4 The Multi-degree-of-freedom Aeroelastic TestThis technique requires scaling the dynamic properties (mass, stiffness, periods and mode shapes) of the building in the fundamental sway modes and the fundamental torsion mode, including any coupling within modes. Some higher modes of vibration are also modelled. The responses to wind loads are then measured directly. Typically, a building is modelled as a series of lumped masses joined by appropriately sized columns; for towers, other approaches to produce the elastic model can also be used. A full discussion of this technique is contained in Appendix I.The advantage of this technique is that the measurements will include effects of aerodynamic damping, vortex shedding, coupling within modes and some higher modes that are not fully dealt with when using the force balance technique. For most buildings, however, it can be argued that the aerodynamic damping effects are likely to be small, higher modes can be neglected and that the force balance adequately handles coupled modes analytically. Nevertheless, for more complicated structures, the additional reassurance of an aeroelastic test may be justified.The disadvantages of the technique are that the model is time consuming and expensive to build. The model is also designed for a single set of building dynamic properties and approximations must be made if these change. The force balance technique on the other hand, yields results equally applicable to any set of building dynamic properties.As with other types of tests, once the aerodynamic data has been measured for a full range of wind directions, it is combined with the statistics of the full scale wind climate at the site, using the methodology outlined in Appendix D, to provide predictions of loads and responses for various return periods.4.5 Overall Loads from Local Pressure MeasurementsThe development of solid state pressure scanners, which permit the simultaneous measurement of pressures at many points on the surface of a building, allows the determination of instantaneous overall wind forces from the local pressure measurements. The technique allows direct computation of the steady and unsteady modal forces acting in any number of modes of vibration of the building. In similar fashion to the force balance approach, the resonant amplification due to the building's structural dynamics is derived analytically for each mode using random vibration analysis methods and results are subsequently used to provide estimates of the full scale response of the building. A detailed description of this method is presented in Appendix J. (Note that integration can also yield true time histories of loads on building subcomponents of interest, such as canopies, large panels, or roofs).The advantages of this technique are that a single model used in a single testing session can produce both overall structural loads and cladding loads. The testing parameters would be extended to ensure that the local pressure data taken is also sufficient for the analysis of structural loads. The analysis required to determine structural loads from the local pressure information is nominally the same as that from the force balance tests. This technique also has the advantage of properly determining the generalized forces for three-dimensional non-linear mode shapes. Although these are also properly handled in multi-degree-of-freedom aeroelastic tests through the proper design of the aeroelastic model, such tests are expensive and time-consuming. The force balance technique can handle three-dimensional mode shapes but need to have corrections for non-linear modes. The pressure integration technique can also be used to examine higher modes with non-monotonic mode shapes.A disadvantage of this technique is that the cladding pressure test model typically includes more instrumentation and takes longer to construct than a force balance test model. Also, as in the force balance technique, it does not include any effects of the building's motion through the air, such as aerodynamic damping; however, neglecting these effects is usually slightly conservative. Proper integration of local pressures to obtain overall wind forces requires that all buildings surfaces are properly represented in the model instrumentation and subsequent calculations. This may not be possible for buildings or structures with complex geometry or small components.Once the responses have been determined, they are combined with the statistics of the full scale wind climate at the site, using the methodology outlined in Appendix D, to provide predictions of loads and responses for various return periods.4.6 Effective Static Force DistributionRepresentative effective static force distributions reflecting the combined static and dynamic response of the building are evaluated for the x, y and torque directions. The details of the procedure are included as Appendix K.In principle, for any azimuth, effective static force distributions can be determined which reproduce the measured dynamic peak base bending moment. The particular vertical distribution chosen must reflect the actual static and dynamic loading of the building.The loading is reasonably considered to be made up of the mean loading, the dynamic loading due to background or quasi-steady excitation and the dynamic inertial loading due to resonant oscillations. The mean loading distribution is determined from the integration of mean external local pressures from the cladding loads test or from an assumed distribution chosen to reflect the mean wind speed distribution. The inertial dynamic loading distribution is determined from the known vibration properties of the building. The measurement of the relative contribution of the unsteady and steady components at the base from one of the above structural loads tests is used to combine the mean and dynamic distributions into one total distribution. An overall effective loading diagram, independent of wind direction, is then constructed by averaging the effective force distributions, weighted by the relative importance of each wind direction.It should be appreciated that these effective static loading distributions are representative of the most likely severe wind loading conditions expected, but the detailed loading may change somewhat for。
姜超（中国广核新能源控股有限公司）摘要：本文研究的是风电机组功率曲线的优化方法。
首先分析功率曲线的影响因素，分别从空气密度（包括气压、环境温度和湿度）、扫风面积、风速、风能利用系数及机组效率这几个方面展开分析，研究各个影响因素对功率曲线的影响作用。
然后，针对风电机组功率曲线的影响因素，探索优化风机功率曲线的方法，包括：桨叶角度检查、风向标的检查、风速仪的检查、环境温度的检查、海拔高度参数的检查及功率控制参数K opt值的检查，并用实际风电机组整改案例去验证功率曲线优化方法的正确性。
关键词：风电机组；功率曲线；优化方法中图分类号：TM614文章编号：1006-8155-(2021)Z1-0076-05文献标志码：A DOI：10.16492/j.fjjs.2021.Z1.0014Research on Optimization Method ofWind Turbine Power CurveChao Jiang(China General Nuclear New Energy Holding Co.,Ltd.)风电机组功率曲线优化方法研究Abstract：The thesis studies the optimization method of wind turbine power curve.Firstly,the influence factors of power curve is analyzed,respectively from the air density(including air pressure,ambient temperature and humidity),swept area, wind speed,wind turbine power cofficient,unit efficiency are analyzed to study the influence of various factors on power curve.Then,explor the optimization method of wind turbine power curve through the influence factors of wind turbine power curve which include check blade angle,inspection of weathervane,inspection of anemometer,check ambient temperature, check elevation parameters,check the power control parameter K opt,the correctness of the power curve optimization method is verified by the actual wind turbine rectification case.Keywords：Wind Turbine;Power Curve;Optimization Method0引言随着风电产业的快速发展，越来越多的风电场出现机组功率曲线问题，包括：机组功率曲线不满足合同标准曲线的要求、满发风速严重滞后、中等风速段机组发电出力低等问题。
