6. ULUSLARARASI MÜHENDİSLİK, MİMARLIK VE TASARIM KONGRESİ, İstanbul, Türkiye, 17 - 18 Aralık 2020, ss.693-700
Coil pitch angles have an important place in the design of rotary electrical machines. When
the stator windings of induction motors are wound with shortened coil pitch, some low level harmonics
can be suppressed. Therefore, determining the pitch angle of the coil can be considered an important
step in induction motor design. In the design process, this stage can be handled by mathematical calculation, but the parameters of each motor can be different and therefore, current and voltage harmonics
may also be different, so it is necessary to make separate calculations for each motor. In this study, in
order to find a solution to this problem, the coil pitch angles of three-phase cage induction motors were
estimated with support vector machines, a popular artificial intelligence method and classifier, and tested with a ten fold cross-validation technique. As a result, an SVM estimation model has been developed
that can estimate the coil pitch angle with minimum error without the need for a mathematical model of
an induction motor fed by a PWM inverter. The data used in the system have been obtained experimentally, all five three-phase cage asynchronous motors used in the experiment are identical, have the same
coil lengths, have windings wound at different coil pitch, all have the same power, and finally all are
four poles. This study is also a unique source of information and application that can be used in active
filter design, motor design, and winding design.