![]() | يوجد فقط 14 صفحة متاحة للعرض العام |
المستخلص Recently, artificial intelligence (AI) that mImICS the adaptive distributed architecture in the human brain provides potential alternatives to tackle different control problems in the industry. When the structure of the plant is unknown or the parameter variation is excessive; the effectiveness of modem control theory diminishes. Even, if it is possible to develop a asonably accurate model, the resulting control algorithm is so computationally intensive that it becomes very difficult, and some times infeasible to implement in a real-time control vironment. So, it is believed that a controller should be designed to have abilities to learn from experience and to use the knowledge gained. The applications of AI techniques such as artificial neural networks (ANN) promise high computation rate provided by the massive parallelism, the distributed representation, and the adaptation, learning and generalization to improve performance. A separately excited de motor drive has been used in the industrial field when a wide range and high accurate speed control is necessary. The aim of the thesis is to design a controller that has abilities to learn from experience and to use the knowledge gained to improve the total |