الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 100 |  |  |
| | | from | 100 | | |
Abstract
The most reliable electric power grid is considered the main object of all countries in the world. But, the random variation of renewable energy sources (RES) and load fluctuations can lead to unbalanced grid frequency and tie-line power flow between interconnected areas. from here, the load frequency control (LFC) is considered as a vital role to solve these problems. However, there is more obstacles facing the designers when dealing with the LFC problem. These obstacles are represented in the optimization of controller parameters, the power system nonlinearities such as generation rate constraints (GRCs) and governor deadbands (GDBs), communication time delays and its parameter uncertainties. In last decades, the researchers designed the parameters of the model predictive control (MPC), MPC with superconducting magnetic energy storage (SMES) device, and MPC with plug-in hybrid electric vehicle (PHEV) based on their expertise which led to unacceptable performance. This optimization problem of design is overcome here by proposing the imperialist competitive algorithm (ICA), gravitation search algorithm (GSA) and bat inspired algorithm (BIA) as a new artificial intelligence (AI) techniques to tune the parameters of the proposed controllers. Also, the thesis compares between the different design of the proposed controllers by ICA, GSA and BIA with using different objective functions in order to minimize the frequency deviations and tie line powers changes against load perturbations. Thus, the best one from AI techniques and controllers are selected. This thesis deals with the design of a model predictive control (MPC), a MPC with superconducting magnetic energy storage (SMES) device and a MPC with plug-in hybrid electric vehicle (PHEV) for LFC in a multi-area power system. The proposed controllers are carried out on a linear interconnected two-area power system, a nonlinear interconnected two-area power system, a nonlinear interconnected three-area power system and a smart grid with a penetration of RES and load fluctuations. The considered nonlinearities are represented in GRCs and GDBs and communication time delays. Furthermore, the parametric uncertainties and the boiler dynamics are taken into consideration. Also, the thesis considers a parallel processing method as a real time simulation method to implement the controllers and the multi-area interconnected power system. In the parallel processing system, each area is represented as a microprocessor. Each microprocessor is connected and exchanges its data only with a central processor. The central processor exchanges the necessary data to each microprocessor. In addition, a multi-agents strategy is proposed as a real time organizational method to perform the controllers and the multi-area interconnected power system. In the multi-agents system, each controller in each area is represented as a control agent. Each control agent is connected and exchanges its data only with a central agent. The central agent distributes the necessary data to each control agent. The main purpose from using of the parallel processing method and the multi-agents strategy is the reduction of computation times, especially in complicated multi-area power system. Furthermore, the thesis proposes a new smart supervisor controller between two different controllers. The smart supervisor controller makes a smart selection between two different controllers to combine the advantages of each one.