Layout Optimisation Brings Step Change in Wind Farm Yield by 6rlEDhL

VIEWS: 6 PAGES: 13

									Layout Optimisation Brings Step
     Change in Wind Farm Yield

         Dr Andrej Horvat, Intelligent Fluid Solutions
                      Dr Althea de Souza, dezineforce



Come and visit us on Stand V16
       Boyd Orr Hall
                          Acknowledgment

 Present work is a joint effort of
 •     David Hartwanger - Intelligent Fluid Solutions
 •     Richard Harding - dezineforce
 •     Althea de Souza1 - dezineforce
 •     Andrej Horvat2 - Intelligent Fluid Solutions


2Dr. Andrej Horvat                              1Dr. Althea de Souza
Principal Engineer                              Senior Design Engineer
Intelligent Fluid Solutions Ltd.                dezineforce
andrej.horvat@intelligentfluidsolutions.co.uk   althea.desouza@dezineforce.com
www.intelligentfluidsolutions.co.uk             www.dezineforce.com




                                           2
           Presentation Structure


•   Motivation and problem definition
•   Wind turbine modelling methodologies
•   Wind farm simulation
•   Maximizing investment yield
•   Optimisation of the wind farm layout
•   Conclusions & further work




                         3
     Motivation and problem definition
• Significant demand for renewable sources of energy, where
  wind power is (one of) the largest contributors

• Power output from a wind farm depends on wind availability
  (intermittency in strength and wind direction), local topology
  and turbine quantity

• Installation of wind farms is capital intensive

• In such an environment, accurate prediction of wind farm
  power output is crucial for planning installation capacity and
  to maximise return on the investment

• Interaction between turbine wakes means turbine layout
  affects total power output


                                 4
Wind turbine modelling methodologies
                     • Detailed simulations (CFD) of
                       entire wind farms are
                       computationally demanding
                     • Individual turbines can be
                       modelled using blade
                       element theory to reduce
                       overall computational
                       requirements
                     • Effects of a rotating turbine
                       on the flow are modelled
                       with momentum
                       sources/sinks
                     • Correct time-averaged
                       representation of axial and
                       tangential wake velocities
                 5
                 Wind farm simulation


                                                                           • Turbines are arranged in
                                                                             staggered (zig-zag) pattern
            ny
                                                                             to minimise wake influence
                 9
                     8
                         7 6

                                                                           • Covering fixed surface area
                                   5
                                           4
                                                   3

                                                                             of 2 x 3 km in streamwise
                                                           2
                                                                   1
                                                                   1

                                                       3
                                                               2
                                                                             (x) and spanwise (y)
                                                                             direction
                                                   4
                                               5
                                       6
                               7
                         nx
                                                                           • Steady-state simulations
                                                                             were performed for different
  wind                                                                       number of wind turbines in
direction                                                                    x and y direction



                                                                       6
Maximising Investment Yield

                 • Reduction of flow velocity in
                   the wake reduces the power
                   output of each subsequent
                   row of turbines

                 • Which wind farm
                   arrangement provides
                   maximum power generation
                   for a given investment?




             7
                  Maximising Investment Yield

   • To find maximum power output for given investment costs
        – Computational Fluid Dynamics (CFD) calculations of the
          different wind farm arrangements were performed and total
          power calculated as a sum of power output from each
          turbine
        – the investment costs were divided into fixed costs
          (construction, grid connection, development etc.) and
          variable costs (proportional to the number of installed
          turbines)
In this case study
 - 2MW Vestas V80 turbine was used as a representative example of a modern commercial turbine
 - single wind velocity of 15 m/s and single rotation speed of 16.7 rpm were considered
 - 28 mil EUR of fixed investment costs and 2.5 mil EUR per unit were assumed
 - The costs were estimated based on review studies prepared by Dept. of Trade and Industry
   (2001) and KEMA Nederland (2007)

                                                 8
Optimisation of the Wind Farm Layout
           •   A CFD based modelling methodology
               was developed to predict wind farm
nx             power output for a given investment
           •   For a set area and staggered layout, the
               number of turbines in each row (ny) and
     ny        the number of rows (nx) were varied
           •   Advanced design search and
               optimisation techniques were used to
               search for an optimal wind farm
               configuration
           •   This approach cost effectively assessed
               the range of design options available
           •   Additional variables can be considered,
               e.g. wind speed, direction, geographical
               site etc.
                     9
    Optimisation of the Wind Farm Layout
Based on a statistically significant but relatively small number of
simulations (~30) the entire design space (~120 designs) is
characterised

    13 in first row, 5 rows = 63 turbines
    Power/Cost metric = 0.55 W/€

    13 in first row, 2 rows = 25 turbines
    Power/Cost metric = ~0.44 W/€

     13 in first row, 8 rows = 100 turbines
     Power/Cost metric = 0.39 W/€

     11 in first row, 6 rows = 63 turbines
     Power/Cost metric = ~0.51 W/€

    9 in first row, 8 rows = 68 turbines
    Power/Cost metric = ~0.31 W/€


                                            10
           Conclusions & Further Work

• Blade element model was implemented in a commercial CFD
  package to simulate operation of wind turbines in a wind farm
  environment

• Different wind farm layouts were simulated to calculate power
  output of the wind farm

• The analysis shows that the same number of turbines in
  different layouts can result in significantly different yield

• With alternate offset rows, wide, shallow wind farms are most
  profitable

• The use of computational simulation methods and advanced
  optimisation tools can result in significant performance
  improvements


                                 11
           Conclusions & Further Work

Further work

• Different wind angles and speeds for full wind rose

• Alternative staggering

• Assessment of specific geographical topologies

• Allowance for local geographical features

• More complex investment models




                               12
                                   Questions ?




                                                         Stand V16
                                                        Boyd Orr Hall


                                       ...provides a unique, end-to-end service integrating
                                       computational engineering and optimisation tools,
                                       on demand over the web
...technologies have been developed over more
than 10 years by world-renowned professors
from the University of Southampton

								
To top