Dynamic response of wind turbine towers in warm permafrost

Benjamin Still ,ZhaoHui (Joey) Yang ,2*,Simon Evans ,FuJun Niu

1.Dept.of Civil Engineering,University of Alaska Anchorage,Anchorage,AK 99508,USA

2.Visiting Professor,State Key Laboratory of Frozen Soil Engineering,Chinese Academy of Science,Lanzhou,Gansu 730000,China

3.Alaska Native Tribal Health Center,Anchorage,AK 99508,USA

4.State Key Laboratory of Frozen Soil Engineering,Cold and Arid Regions Environmental and Engineering Research Institute,Chinese Academy of Sciences,Lanzhou,Gansu 730000,China

1 Introduction

"Wind generated electrical energy is the world’s fastest growing energy source," and Alaska has an abundant wind energy resource (AEA,2011).Western Alaska has nearly continuous winds which sweep across the barren tundra,providing an ideal situation for wind energy harvest.Over the last decade wind turbines have been installed in several villages throughout western Alaska,with foundations constructed in warm,ice-rich permafrost.

Vast areas throughout western Alaska are underlain by warm ice-rich permafrost that requires extra precaution in foundation design.Standard gravity foundation for most wind turbines consists of a large mass of reinforced concrete,which is generally cost prohibitive in western Alaska due to the cost of construction materials.In addition,this mass of concrete will retain heat and facilitate degradation of the underlying permafrost,which will in turn compromise the stability of such a foundation.Current practice favors a deep foundation consisting of steel piles and a mechanism for keeping the frozen status of the permafrost.

The dynamics of wind turbines are important and often are a controlling aspect in foundation design.Resonance between the foundation-structure system and the rotating parts must be avoided and an adequate factor of safety must be applied.Research on these deep foundations in warm permafrost has been sparse.Satari (2008)developed a model of wind turbine structure-foundation-soil interaction in SAP2000 software(Computers &Structures,Inc.,Walnut Creek,CA) for two towers in western Alaska and found the frequency of the tower nearly resonated with the operational frequencies of the wind turbine for certain soil conditions;proper separation of these frequencies is often controlled by design.Dilley (2010) investigated several foundation designs for wind turbines in Alaska,and concluded that a standard foundation was not cost-effective for all turbines as the soil conditions vary greatly from warm permafrost to bedrock.

2 Foundation design

Three 100-kW Northwind wind turbines (Northern Power Systems,Inc.,Barre,VT) were installed near Quinhagak,Alaska for power generation.Quinhagak is a town in western Alaska located on the eastern shore of Kuskokwim Bay and is underlain with warm permafrost between-0.6 °C and-0.1 °C (Golder,2009).The wind turbine site consists of flat terrain with gently sloping hummocks with less than 1.5 meters of vertical relief.Frozen soil was found to the depth of boring at 12 meters and consisted of an organic mat at the surface followed by organic silts between 0.5–2 meters,mineral silts from 2–4 meters,and well-graded gravelly sand to 9 meters deep.Below 9 meters in depth is frozen silt.The foundation consists of six 40.6 cm diameter steel piles with 1.9 cm wall thickness embedded 9 meters into the permafrost.The piles are equally spaced at the corners of a hexagon,as shown in figure 1.

Figure 1 The wind turbine foundation consisting of thermal syphons,piles,and pile cap in Quinhagak,Alaska

Temperatures below-0.6 °C are needed to ensure stability of the foundation.This is achieved with a passive cooling system.The installed two-phase liquid-to-vapor thermal syphons are designed to cool the subgrade to below-0.6 °C and maintain this temperature throughout the life of the structure.A thermal syphon is placed vertically next to each pile for the total length of the embedment.This system relies on the cold temperatures to drive the heat exchange process.The piles and thermal syphons are generally installed at the same time when the ground is frozen.Oversized holes are augured out and the piles are inserted plumb with enough space on the outside perimeter for each thermal syphon.Water and sand slurry is poured into each hole and allowed to freeze back before continuation of construction.A gravel pad is placed on top of the tundra organic mat for access to the wind turbines for maintenance.

3 Field instrumentation

The wind turbine is located 1.6 km south of Quinhagak and is accessed via gravel roads raised above the surrounding tundra surface.The westernmost wind turbine in the group of three was chosen for instrumentation.The south side of the wind turbine gravel pad had recently sloughed into a small pond that formed along the south edge of the gravel pad,as shown in figure 2.When the instruments were installed in September 2013,a small ribbon of water was observed to reach partially around the wind turbine gravel pad.

Figure 2 Wind turbine gravel pad sloughing into a newly formed pond

Figure 3 Accelerometer,strain gage,and data logger locations on the Quinhagak wind turbine

Accelerometers:Low-noise,single-axis accelerometers were used with an operating temperature down to-40 °C.A total of six accelerometers were installed on the inside surface of the wind turbine tower (Figure 3).Two accelerometers were placed at three locations inside the wind turbine tower:just below the nacelle at the top of the tower,the midpoint of the tower,and near the base of the tower.At each location,an east-west and a north-south direction accelerometer were mounted on a custom mounting bracket.The mounting brackets were glued to the side of the tower using a quick-setting high-strength epoxy.

Strain Gages:Five strain gage rosettes were placed equally spaced around the interior of the wind turbine tower near the base.The rosettes were placed such that the horizontal strain,the vertical strain,and the strain at a 45° angle were measured at each location.Each of these strain gages was connected to a dummy gage for temperature compensation.The dummy gage was mounted onto a steel coupon which was placed near the strain gage rosette with a thermally conductive deformable epoxy.

Data Logger:A Campbell Scientific CR5000 data logger (Campbell Scientific,Logan UT) was connected to a NL200 network interface device providing communication in real time.The data logger is housed in a custom enclosure,and is powered by a deep-cycle marine battery with a trickle charger connected to a power source.In the event of loss of power,the battery is capable of providing continuous power supply to the data logger for two to three weeks.The sampling frequency of the acceleration and strain data is 20 Hz.

4 Results and discussions

The acceleration data were analyzed by fast Fourier transform (FFT) to examine the operational and structural frequencies.The tower vibration data recorded by the accelerometer in the east-west and north-south direction at the top of the tower during operational and non-operational periods are shown in figures 4 and 5,respectively.The wind turbine operation period varied between 1.024 s and 1.069 s while the structural period varied from 1.205 s to 1.280 s,as shown in table 1.The maximum operation rotation is specified by Northwind as 59 rpm,equivalent to a period of 1.017 s.The smallest difference between the structural and the operation periods was 6%,as shown in figure 6.This value was smaller than the value of 15% recommended by Satari(2008).In particular,during starting and braking of the wind turbine operation,the rotational period will overlap with the structural period for a short time period and cause resonance between the spinning blades and the tower,possibly leading to shortened lifespan of the tower and foundation system.This is typically not considered acceptable.

Table 1 Wind turbine periods

Figure 4 Acceleration time history (a) and typical operation and structural periods (b)

Figure 5 Acceleration time history (a) and typical wind turbine structural period while not operating (b)

Figure 6 The percent difference between the structural period and the operation period

The structural period decreased over the winter of 2013–2014,as shown in figure 7.This decrease in period was most likely due to refreezing of the active layer and cooling of the warm permafrost by the thermal syphons,which stiffens the soil around the base of the turbine and the piles.Also,the stiffness of the structure itself increased with decreasing temperature.The structural period of the wind turbine decreased in colder temperatures,as shown in figure 8.

Climate data from the National Oceanic and Atmospheric Administration (NOAA,2014) for the nearby town of Bethel,Alaska shows that all months from October 2013 to February 2014 experienced above normal temperatures as seen in table 2.The month of January 2014 was the warmest on record with a temperature of 10.8 °C above normal.Eight record highs were recorded during this month resulting in a loss of snow coverage.This warm winter helped lessen the stiffening of the soil and hence the structure in that season,but an average season could see more of a decrease in the tower period.

Table 2 Climate data from the 2013-2014 winter compared to normal for Bethel,Alaska

Figure 7 Change in structural period over the winter of 2013–2014

Figure 8 Structure period vs.air temperature inside the tower

5 Conclusion

This paper presents a vibration and load assessment for a wind turbine-foundation system constructed in warm permafrost at Quinhagak,Alaska.It presents,analyzes,and discusses the data collected during the winter of 2013-2014.Field-installed accelerometers and strain gages are able to perform real-time monitoring of the turbine-foundation and collect vibration as well as strain data at the site.Acceleration time histories,the natural period of the turbine structure,the wind turbine blade flapping period,and the rotational period of the wind turbine itself were found using a fast Fourier transform based on the collected data.Climate data for the nearby town of Bethel,Alaska were used to help understand the current freezing regime of the soil.The following conclusions can be made from the results of this study:

1) The structural natural period ranged from 1.205–1.280 s;it decreased as winter progressed and the ambient temperature dropped.

2) The operating period of the turbine structure ranged from 1.024–1.069 s;resonance between the spinning blades and the tower structure is likely to occur during starting and braking.This period was also at times below the 15% safety margin (sometimes as low as 6%)specified during design.

It is recommended that more data during the coming spring and summer months be collected for further study of the dynamic properties,further examination of the resonance issue,and wind load assessment.

We would like to acknowledge that the Alaska Energy Authority (AEA) provided financial support to this project and acknowledge the support from the Western Project Program of the Chinese Academy of Sciences(KZCX2-XB3-19),the State Key Development Program of Basic Research of China (973 Plan,2012CB026101),the Foundation for Innovative Research Groups of the National Natural Science Foundation of China (Grant No.41121061),and the National Sci-Tech Support Plan(2014BAG05B05).We are thankful to Mr.Richard Stromberg,Wind Energy Coordinator of AEA,for his strong support for this study,and Mr.Bill Thompson and others at the Alaska Village Electric Coop for their continued support in accessing the wind turbines.

Alaska Energy Authority,2011.Community Wind Toolkit:A Guide to Developing Wind Energy Projects in Alaska.Renewable Energy Alaska Project (REAP),March.Anchorage,Alaska.

Dilley L,Thornley J,2010.Design and Testing of Deep Foundations for Wind Turbines in Western Alaska.Hattenburg Dilley &Linnell,Anchorage,Alaska.

Golder Associates,Inc.,2009.Geotechnical Reports.Anchorage,Alaska.

National Oceanic and Atmospheric Administration,2014.Climate Data from Bethel,Alaska.Available at:www.nws.noaa.gov/climate/index.php?wfo=pafc,accessed on March 20,2014.

Satari M,Hussain S,2008.Vibration-based wind turbine tower foundation design utilizing soil-foundation-structure interaction.Proceedings of 14th World Conference on Earthquake Engineering,Beijing,China.International Association for Earthquake Engineering,Tokyo,Japan.