WIND FARM MEASUREMENTS AND MODELLING

1
WIND FARM MEASUREMENTS AND MODELLING
EWEC 2006 BW1 DYNAMIC MODELS OF WIND FARMS FOR
POWER SYSTEM STUDIES - IEA WIND RD ANNEX 21

• Kjetil Uhlen
• SINTEF Energy research

2
Objective

• To highlight main difficulties in modelling and
  model validation, related to
• measurement data
• choice of models
• choice and use of simulation tools
• Motivation Experiences and lessons learned from
  development, implementation and tuning of a
  doubly-fed induction generator (DFIG) model for
  use in PSS/E.

3
Measurements processing and analysis

• Processing of measured data is not a trivial
  task!
• Two approaches are shown
• Power and RMS voltage computed based on
  instantaneous three-phase data
• P u1i1 u2i2 u3i3
• Phasor approach First compute positive sequence
  voltage and current phasors and then compute
  power
• P real(UI)

Computed active power
4
- Measurements from instantaneous data-
Originally tuned model
- Measurements based on pos seq. phasors-
Realistically tuned DFIG model
5
Conclusions

• Same measurements and same model ? very different
  results!
• Why
• There are different ways of processing measured
  data!
• Models can be tuned to produce different
  characteristics!
• Third order DFIG model may not be a good
  choice
• Too simple for a good representation of voltage
  dips (in the millisecond range)
• Too detailed for a phasor model because of very
  fast closed loop dynamics
• To determine what is correct requires competence
  and knowledge about measurements and
  modelling/control!