Information Technology Reference
In-Depth Information
Consequently, output power has least variations. Based on Based on a perform
ability model, a control strategy is devised for maximizing energy conversion in
low to medium winds, and maintaining rated output in above rated winds while
keeping tensional torque
fluctuations.
Controlling and estimating of a parameter in variable speed wind turbines was
performed in this work using a new hybrid controller. MLPNN learns experimental
and optimal data and FRENGA can be easily used for extracting a set of fuzzy
if-then rules directly from NN black box. However, because of the good ability of
NN to extract noise its operating is advised. As plotted from simulation results, is
realized with change of turbine blades pitch angle, system can obtain most suitable
performance coef
fl
cient (Cp) and tip speed ratio (TSR). Proposed controller by
changing Pitch angle optimally Controls system. Experimental study has shown
that it works ef
ciently for both continuous and enumerates attributes.
Hence we use NN to control wind turbine and produce a training example for
the rule extraction method. Consequently output power has been regulated suc-
cessfully. Finally we have been compared between proposed approaches and two
different methods an MLPNN and FRENGA alone. We see our output power
almost is most proper than them.
10 Tuning of Membership Function
As we know, the knowledge Base (BS) is formulate by membership rule MR
Membership Function MF and fuzzy rule base RB. Thus ditty or intellect mem-
bership methods, or fuzzy mandate base or both of them are some resource to
originate genetic fuzzy system (Magdalena 2001 ).
When fuzzy statue base is violated in advance, for regulate membership function
an independent population denotes parameters of the membership province
appearance. Against for adapt fuzzy rule base, all of fuzzy laws possibility that
membership tasks is divert before have been personify the population Fig. 13 shows
these syllabify. In this section of work initial shapes for the membership function
have been shown in Fig. 8 and GFS will tune this shapes.
10.1 Proposed Tuned Fuzzy Genetic System (TFGS)
Thus this method is absorption in acquire the set of fuzzy If-Then rules that assuage
different criteria.
Search WWH ::




Custom Search