Optimization and neural network model for induction motors
作者:
Kamel Idir
所属专业方向:
电力系统
摘要:
This thesis addresses two major issues in electrical machine design process, namely
design optimization and machine modelling. nie h tpa rt of this thesis attempts to overcome
the main drawback of using local optimization techniques, Le., the local solution. A newly
proposed optimization approach for electric machine design problems is presented and
discussed. The advantage of such an approach is its ability to obtain the global solutions for
an induction motor. Its algorithm uses an error function which is the difference behveen the
actual objective function value, at a given computational step, and its bounded limit value
to search for the global optimum. The error fbnction provides a good indication of how far
or close the objective hction is approaching its ultimate solution. The design variables are
updated such that the error is progressively reduced to its minimum value. This approach is
suitable for bounded functions and its algorithm does no? use function derivatives. Both
feanires make it suitable for motor design problems. The proposed method has been tested
on various typical benchmark multimodal functions and has been particularly applied to a
100 horsepower squinel cage induction rnotor design optimization problem. This algorithm
is characterized by its simplicity and easy implementation, yet it is effective in achieving the
best motor designs, as indicated by the test results. A comparative study has also indicated
that the proposed method outperformed the conventional methods such as Hooke and Jeeves
and Powell methods.
赣公网安备36040302000210号 )|网站地图
狗仔卡
发表于 2008-5-20 20:43:48
提升卡
置顶卡
沉默卡
喧嚣卡
变色卡
显身卡