Adaptive PI control and active tower damping compensation of a wind turbine

Abstract

This  paper  deals  with  the  control  problem  ofa  wind turbine model working in the nominal zone. This process is a    nonlinear    system    whose    dynamics    vary    strongly depending  on  the operation  point.  In  the  nominal  region, the wind turbine speed is controlled by means of the pitch angle    to    generate    the    nominal    power.    The    wind fluctuations  and  its  non-uniform  special  profile  act  as disturbances   on   the   power   generation   and   the   tower deflections. These oscillations must be reduced to improve the wind turbine lifetime. In  this  work,  an adaptive  control  structure  operating on the pitch variable is proposed. It is composed of a gain-scheduling     PI     control,     an adaptivefeedforward compensation  of  the  wind  speed  and  an adaptive  gain compensation  for  the  tower  damping.  The  tuning  of  the controller  parameters  is  formulated  as  an  optimization problem  that  minimizes  the  tower  fore-aft  displacements and  the  deviation  of  the  wind  turbine  speed  from  its nominal  value.  It  is  resolved  using  genetic  algorithms  for different   linear   models   that   are   obtained   from   the nonlinear model. The  proposed  controller  is    compared with  a  classical baseline   PI   (Proportional-integral)   controller   and the simulation  results  show  a  significant  improvement  of  the system performance when the proposed strategy is applied.