tidal hydrokinetic energy and site characterization

Tidal Current Energy and Site Characterization

Kristen Thyng 1

Jim Riley 2

1Texas A&M University

2University of Washington

November 22, 2013

Kristen Thyng (TAMU) Tidal Energy November 22, 2013 1 / 37

Outline

1 Tidal Hydrokinetic Energy

2 Admiralty Inlet

3 Siting Metrics

4 Conclusions and Future Work


Outline


2 Admiralty Inlet

3 Siting Metrics



Tidal Hydrokinetic Energy

Like wind energy

http://windeis.anl.gov/guide/photos/photo7.html



Like wind energy... but under water

http://verdantpower.com


Some differences between Wind and Tidal Energy

P = 12s3Ac , density, s speed, Ac area

Wind: s 7 m/s; tidal: s 1 m/sbut 1 kg/m3 in air and 1000 kg/m3 in water

Limited space underwater in constricted channels- also vertically due to ship traffic

Wind turbine has larger cross-sectional area

Very different environments

Underwater make access more difficult



Pros:

Renewable resource

Geographic location

No carbon output

Relatively predictable

Potential for lowenvironmental impacts

No visual impact

Cons:

Not constant resource

Possible effects on marinemammals and fish

Physical flow effects

Additional stress oncoastal oceanenvironments


Resource

image: Atlantis Resources Corporation, opportunity:energy


Resource

image: Aquaret


Resource (wave)

image: Aquaret


Resource

US: 4,000 TWh of electricity used; tidal could produce 250 TWh

image: http://www.tidalstreampower.gatech.edu,

http://www1.eere.energy.gov/water//marine assessment characterization.html


Resource

Alaska

image: http://www.tidalstreampower.gatech.edu


Resource

Washington. Seattle: 1200MW used; tidal could produce e.g. 210MW

image: http://www.tidalstreampower.gatech.edu, Polagye et al 2009


Resource

Bay area, CA



Resource

Texas



Renewables in Texas

Currently:

On-shore wind energy

Solar energy

Biofuels and biomass

Future:

Off-shore wind?

Wave?

Texas Renewable Energy Industry Report, Electric Reliability Council of Texas (ERCOT)


Renewables in Texas: On-shore Wind

Texas #1 in the U.S. for wind energy capacity and biodiesel production- about 9% of power consumed in Texas; 12000MW installed capacityimage: http://commons.wikimedia.org/wiki/File:Desert-Sky-Wind-Farm.jpg Texas Renewable Energy Industry Report, Electric

Reliability Council of Texas (ERCOT)


Renewable Energy Mandate

Many communities have passed renewable energy initiatives:

United Kingdom: 20% renewable energy by 2020

The Department of Defense: 25% renewable energy by 2025

In Washington State: 15% renewable energy by 2020

Local public utility districts in Washington are looking for local renewableenergy sources


Northwest National Marine Renewable Energy Center, UW

Materials survivability

http://depts.washington.edu/nnmrec



Velocity field around a turbine in 3 turbine models




Marine mammal monitoring




Background acoustics measurements



Turbine Siting Considerations



Turbine



Turbine

HighSpeeds



Turbine

HighSpeeds

Directionality



Turbine

HighSpeeds

DirectionalityShear



Turbine

HighSpeeds

DirectionalityShear

UpwardFlow



Turbine

HighSpeeds

DirectionalityShear

UpwardFlow

Turbulence


Site Characterization

Maximize in-stream power atturbine locations to maximizeprofit

Micrositing turbines in areasof highest speeds (power s3)Limited space

Minimize fatiguing effects onturbines to minimize cost

Turbine survivabilityTurbine efficiency

Figure : Large vertical velocities inAdmiralty Inlet

Understanding the flow field underlies all considerations


Basic turbine properties

Cut-in and rated speeds

Ability to yaw or fixed axis

Typically 5 to 20 meterdiameters

Mounting system

Design

8

Wind turbines have a fourth operating regime defined by a cut-out speed, above which the turbine

blades are feathered to avoid damage during periods of extremely high winds. Since tidal currents

are largely predictable, there is no tidal analogue to extreme weather. The possible exception to

this is a tsunami event, which is not generally considered in the design of a tidal turbine.

0.00.51.01.52.02.53.03.54.04.5

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

Water speed (m/s)

Ext

ract

ed p

ower

(kW

/m2)

I

II

III

Figure 1.3 Representative turbine power curve. Region I is below the cut-in speed and the turbine extracts no power. In Region II, power is extracted in proportion to the kinetic power incident on the rotor swept area. Region III is above the rated speed and power extraction is constant.

Device utilization is quantified by the capacity factor, defined as the ratio of average power

extracted to power extracted at rated speed. Feasibility studies indicate that the lowest cost of

energy for in-stream tidal turbines would be achieved with capacity factors between 30 and 40%

[6] depending on the particulars of the tidal regime. Therefore, the selection of the rated speed is

an economic decision.

1.4. Available Resource While the similarities between tidal and wind energy are obvious and striking, there are also

important differences. Even the largest wind turbines extend no more than a few hundred meters

into the air, while the characteristic length scale for the atmosphere is measured in kilometers. For

most tidal energy sites, the characteristic length scales of the device and resource are comparable

(e.g., 20 m rotor in 40 m water) and the extracted power may constitute an appreciable fraction of

the total power in the system. As will be discussed in this dissertation, kinetic power extraction

from tidal streams has the effect of increasing the frictional resistance to flow. Since tidal streams

are generally subcritical, the effect of increasing friction is felt estuary-wide. While small

increases in friction due to extraction may be indistinguishable from natural friction, large-scale


Tidal Turbine Designs

Verdant: 35.9 kW (5 m diameter)http://verdantpower.com

http://www.aquaret.com/images/stories/aquaret/pdf/chapter3.pdf



Open Hydro: 1.52 MW (15 m diameter)http://www.openhydro.com




Marine Current Turbines: 0.75 to 1.5 MW (15-20 m diameter x 2)http://www.marineturbines.com




ORPC turbine: 25kW (each unit)http://www.oceanrenewablepower.com/



Outline


2 Admiralty Inlet

3 Siting Metrics



Area of Interest

Strait of Juan de Fuca

Pacific Ocean

Puget Sound

Columbia River


Area of Interest

Admiralty Inlet

SeattleHood Canal

Deception Pass

Tacoma Narrows

Pacific Ocean

Canada


Realistic Model Domain

Surface salinity of regional modelhttp://faculty.washington.edu/pmacc/MoSSea

D. Sutherland, J. Phys. Ocean, 2011

Bathymetry of nested model ofAdmiralty Inlet.


Realistic Model Domain

Run in ROMS: hydrostatic, 3D,parallelized, large usercommunity

Horizontal resolution of 65meters

20 vertical layers

k- turbulence closure scheme

Boundary and initial conditionsfrom regional model Bathymetry of nested model of

Admiralty Inlet.


Tidal Projects in the Puget Sound: Admiralty Inlet


Flow Features in Admiralty Inlet: Eddies

Google Earth


Flow Features in Admiralty Inlet: Fronts

Google Earth


Flow Features in Admiralty Inlet: Hydraulic Control

Harvey Seims thesis at UW, 1993


Surface Speed


Lavf52.64.2

aissurface.mp4Media File (video/mp4)

Density behavior

fresher

denser


Density behavior

Ebb tide brings fresher water

northward


Density behavior

Flood tide brings denser water south


Density behavior

front


Free Surface

Free surface comparison with NOAA tide gauge station

09/03/06 09/10/06 09/17/06 09/24/062

1.5

1

0.5

0

0.5

1

data

model

Similar behavior as in regional model


Outline


2 Admiralty Inlet

3 Siting Metrics



Site Characterization Metrics

Summarize flow field in relation to turbine siting

Local fluctuations in speed important

Kinetic power density = 12s3

Small increase in s large increase in KPDQuantitative and qualitative metrics to address size and quality ofresource


Quantitative Metrics

Quantifying the size of the resource available

Mean speed

Mean power

Timing of currents

Vertical profiles

Bias of currents to ebb or flood tide


Kinetic Power Density

Value of 0.5 or 1 kW/m2 economically viable (Bedard et al., 2006, EPRI;http://www.aquaret.com/images/stories/aquaret/pdf/chapter3.pdf)


Operation Timing

Turbine operation increases environmental stressors (Polagye andThomson, 2011).


Power production bias

Even power in time

Biased power

Timing of power production throughout day


Qualitative Metrics

Measuring the extractability of the resource and potential effects onturbines

Shear

Directionality

Vertical velocity

Turbulence


Flow asymmetry example

(a) Large asymmetry, large directionaldeviation

(b) Small asymmetry, small directionaldeviation


Flow asymmetry

More bi-directional flow

Less bi-directional

flow

Flood Ebb

(c) Asymmetry

Less directional deviation

More directional deviation

(d) Directional deviation


Flow asymmetry

60 40 20 0 20 40 60

0.2

0.4

0.6

0.8

1

Yaw angle (degrees)

Po

wer

red

uct

ion

fac

tor

Data

cos2

cos3

15 reduces power production to 87% of full power40 reduces power production to 59% of full powerMadsen 2000


Shear and Turbulence

u

z

5

4 J

3

2 L I t I , t i I i I L I I 101 1 0 `2 1 0 3 1 0 4 1 0 5 1 0 r 1 0 7 1 0 8

T ( s )

Figure 5 Comparison of extreme predictions. Z(t) =Zma x (dotted line), time scaling (27) (dashed line), exact vy(~) and (Ox/Oz)2 = 1.0 (upper full line), (oxloz)= = 0.2 (lower full line)

model, x2 The model implies that the damage during N time segments each with n i sinusoidal cycles having a stress range AS i becomes:

N (ASitm D = E ni \ - ~ [ / (29)

i = 1

in which S~ and m are the material constants for the S-N curve. For the complex signal Y(t) the first question is how to define a cycle. For a complex signal with a known time history the rainflow counting procedure by Matsuishi and Endo, ~a which counts cycles and ranges, gives good results. It is, therefore, the goal for an analytical approach to yield similar results.

As Y(t) is a stochastic signal, it follows that statistics for a number of cycles and stress ranges must be determined to calculate the expected damage rate, i.e. the damage per unit time. For a narrow-band Gaussian process with zero mean X(t) , the realizations resemble sinusoids with slowly varying amplitudes. In this case the rises and falls are distributed as an envelope, i.e. Rayleigh distributed, n The expected damage rate can then be written:

e { D } = .0 (30)

where the number of cycles are represented by Vo, which is the characteristic frequency of the process in Hz. In the case of a pure stochastic signal, Vo is:

f Vo = ~.fxx(0, .~) d.~ = -- ]~-2/xo (31) 27r o

where .fx is the joint probability function of X(t) and its time derivative. The associated stress range AS is:

AS = 2 X/~Ox p(1 + m/2) 1/m (32) where P( ) is a gamma function. For a combined sinusoid and a narrow-band stochastic process with a centre frequency equal to the frequency of the sine wave, Rice m has determined the mth moment of the distribution of the rises and fails. Using these results the stress range AS becomes:

A S = 2V~Ox [P(1 + m / 2 ) M ( - - m / 2 , 1, _~2)]~/m (33) in which M( . . . . . ) is the confluent hypergeometric function and:

R = 2 X/~ o ~ (34)

Wind-induced failure o f wind turbines: P. H. Madsen and S. Frandsen

in terms of the range of the sinusoid R and the standard deviation of the stochastic part.

For a Gaussian stress signal which is not narrow-banded, Wirsching and Light 14 proposed an equivalent stress range of the form:

AS = g(a , m) 2 X/2o x P(1 + m/2) '/m (35)

where g(a , m) is an empirically determined correction function, which depends on the material parameter m and a:

a = po/V= (36)

v m is the expected number of maxima given by: lo o

/ 1 ~m = E l 2 2 (0, 2) dx = - - ~ (37)

2~

In independent simulation studies, is a better agreement was found for multi-modal spectra of the stress signals, when the damage computed with the rainflow algorithm was correlated to m and the parameter:

),1 8 - - - (38)

Based on a regression analysis a correction function was proposed of the form:

g(~i, m) = 1.0 + (0.66 -- 0.045m) (8 -- 1) (39)

For a combined periodic and stochastic wide-band signal a similar model based on a correction to the narrow-band expression, equation (33), was suggested) s Basing the central frequency Uo and the bandwidth parameter ~ on spectral moments of iche combined signal:

X~ = 2 cokSx(w) +- ~ (JIo~n126(~o--molO)d~ 2 ~=1

o (40)

20 10 0

E -10 Z v -20 ~ -3O @ E -40 0 IE -50

-60 -70

0 I 1 I I I i I I I I

i 2 3 4 5 Time, (s)

1000000 000

I00 000 000

io ooo ooo x/ - k j\

1000000 3

10000O

I000C ' t , t , i , , ,~,,L ,,lJl J i , I,L= O. 1 1 10 I(](]

W, (rod/s)

Figure 6 Time series of periodic part and power spectrum of stochastic part of flapwise bending moment at r = 1.6 m

Eng. Struct . , 1984 , Vo l . 6, October 2 8 5

Shear as periodic forcing, turbulence stochastic

Can affect both turbine survivability and resource extraction



High shear areas

(e) Mean Shear (f) Bathymetry

Shear correlates to bathymetry gradients



Turbulent kinetic energy


Outline


2 Admiralty Inlet

3 Siting Metrics



Turbine Placement?

Areas with high mean KPD and bi-directionality

Also small area near headland

Resource and high bi-directionality


Turbine Placement?

Areas with high MKPD and low mean TKE

Resource and low turbulence


Effect of turbines on flow

Roc et al, Renewable Energy, 2013; Thyng et al, EWTEC 2013.


Optimization of turbine layout

Funke et al, Renewable Energy, 2014.Kristen Thyng (TAMU) Tidal Energy November 22, 2013 36 / 37

Thank you!

This work was done as part of the Northwest National Renewable EnergyCenter at the University of Washingtonhttp://depts.washington.edu/nnmrec/

Partial funding for this project was provided by the US Department ofEnergy.

Additional support came from the PACCAR chair.


Tidal Hydrokinetic EnergyAdmiralty InletSiting MetricsConclusions and Future Work

tidal hydrokinetic energy and site characterization

Documents