Tsukamoto’s defuzzification technique is chosen to
transform all the fuzzy outputs to a single mean value
ymcm. The Tsukamoto’s method is chosen because it is
very simple and makes the computation much easier and
faster. Fig. 5 shows how ymean is obtained from
Tsukamoto’s defuzzification process. Then the estimated
angular frequency and the adjustment in the torque
current will be given by:
o*(t) = kw* ymean(t)
Alq(t) = ki* ~mmn(t)
k,* and k,* are frequency conversion coefficient and
current conversion coefficient respectively. Their values
rae obtained by using simulations similar to that for
finding kw and ki in the neural network model. The
outputs of the fuzzy logic model will be fed into the
vector control model in the same manner as the outputs of
a neural network model.
IV THE VECTOR CONTROL MODEL
The vector control model has all the current values
adjusted in the two-phase domain. First in the
synchronously rotating stator frame, the flux current I,, is
taken as equal to 40% of the current limit of the motor. Id
is then maintained at this constant value and the value of
the torque current I, is adjusted by adding the value of
Al,(t), obtained from the fuzzy control model or the
neural network model, to its instantaneous value Itl(t>.
Adjustment of I, alone, as shown in fig. 6, will be
Tsukamoto’s defuzzification technique is chosen totransform all the fuzzy outputs to a single mean valueymcm. The Tsukamoto’s method is chosen because it isvery simple and makes the computation much easier andfaster. Fig. 5 shows how ymean is obtained fromTsukamoto’s defuzzification process. Then the estimatedangular frequency and the adjustment in the torquecurrent will be given by:o*(t) = kw* ymean(t)Alq(t) = ki* ~mmn(t)k,* and k,* are frequency conversion coefficient andcurrent conversion coefficient respectively. Their valuesrae obtained by using simulations similar to that forfinding kw and ki in the neural network model. Theoutputs of the fuzzy logic model will be fed into thevector control model in the same manner as the outputs ofa neural network model.IV THE VECTOR CONTROL MODELThe vector control model has all the current valuesadjusted in the two-phase domain. First in thesynchronously rotating stator frame, the flux current I,, istaken as equal to 40% of the current limit of the motor. Idis then maintained at this constant value and the value ofthe torque current I, is adjusted by adding the value ofAl,(t), obtained from the fuzzy control model or theneural network model, to its instantaneous value Itl(t>.Adjustment of I, alone, as shown in fig. 6, will be
การแปล กรุณารอสักครู่..