Abstract
It is well known that Wavelet Networks (WN) are powerful tools
for handling problems of large dimensions. The integration of Wavelet
Network and Fuzzy Logic (FL) enable a tool condition monitoring system
to have a high monitoring success rate and fast training feed over a wide
range of cutting conditions in drilling applications. To overcome offline
learning and to perform efficient tracking behavior, an Auto Tuning
Adaptive Fuzzy Wavelet Network (ATAFWN) controller is proposed. It
was shown that such structure don’t need offline learning to govern the
system in stable regions. It can be handle also a wide range of parameter
changes in comparison with the conventional controller as well as such
controller is simple to configure since it doesn’t need a process model and can be easily adapted to the existing controller and plants.
for handling problems of large dimensions. The integration of Wavelet
Network and Fuzzy Logic (FL) enable a tool condition monitoring system
to have a high monitoring success rate and fast training feed over a wide
range of cutting conditions in drilling applications. To overcome offline
learning and to perform efficient tracking behavior, an Auto Tuning
Adaptive Fuzzy Wavelet Network (ATAFWN) controller is proposed. It
was shown that such structure don’t need offline learning to govern the
system in stable regions. It can be handle also a wide range of parameter
changes in comparison with the conventional controller as well as such
controller is simple to configure since it doesn’t need a process model and can be easily adapted to the existing controller and plants.
Keywords
Fuzzy logic
fuzzy wavelet network.
online controller
wavelet network
Abstract
مما لا شك فيه أن تقنية شبكة المويجه استطاعت التغلب على مشكلة الأبعاد الكبيرة للاشاره. وكان لدمج تقنية شبكة المويجه مع تقنية المنطق المضبب الأثر الواضح في تحسين مراقبة الأنظمة وسرعة في التدريب للانظمه المعقدة منقوصة المعلومات. للتغلب على مشكلة التدريب المسبق ولزيادة كفاءة سلوك النظام، تم اقتراح مسيطر توليف ذاتي لشبكة المويجة المضببة المتكيفة. وأظهرت النتائج أن هذا المسيطر لا يحتاج إلى تدريب مسبق لجعل النظام مستقرا. كما وأظهرت النتائج أن المسيطر المقترح استطاع الصمود بوجه التغيرات الطارئه على متغيرات النظام، بالاضافه إلى كل ما سبق فان المسيطر المقترح يعتبر سهل التنصيب خاصة وانه لا يحتاج لمعرفه مسبقة بالنظام المراد السيطرة عليه.