| Abstract |
This paper proposes a novel fuzzyneural
network for control and diagnosis of
AC motor drives and interruptible power
supplies (UPS) . Firstly,soft computing which
is the fusion or combination of fuzzy systems,
neural networks and genetic algorithms are
studied.
Then, chaos (fractal) computing, wavelet
transformation and immune network theory
are added to propose extended soft computing.
Intelligent control is defined using this.
Next, by taking advantages of fuzzy systems
and neural networks a novel fuzzy-neural
network with a general parameter learning
algorithm is developed.The general parameter
calculus is based on GMDH (Group Method
of Data Handling) suitable to system structure
determination. Local basis function networks
are introduced resulting in real time operated
autonomous controllers. The developed fuzzy
neural network is implemented using
HITACHI RISC + DSP processor, SH 7410.
Then, examples of this developed fuzzy
neural network applied to control and
diagnosis for AC drives and UPS systems are
given.
Lastly, examples of intelligent control with
other soft computing methods are described. |
