الفهرس Only 14 pages are availabe for public view
 |  | from | 120 |  |  |
| | | from | 120 | | |
Abstract
This thesis proposes a technique to restore lost loads after a fault event in the electrical distribution networks using the Q-Learning technique considering the different operating conditions of the network and the load priority factor that determines the importance of loading areas. It provides a suggestion of an optimum automated sequence of actions to transfer the network from unstable state (after a fault) to a stable state. Simulations are applied to a selected part of the Egypt{u2019}s 66kV Electricity Network (the end of transmission level and the start of distribution level). The Q-Learning Algorithm is divided into three steps: First: the information describing each possible case of the network is collected and the initial reward matrix is obtained. Second, the program uses data previously collected to obtain the trained Q matrix. Third: The trained Q matrix is used to get an optimum sequence of actions to restore the system to a stable state