Document Details
|
Document Type |
: |
Thesis |
|
Document Title |
: |
State Estimation of Power Systems using Adaptive Filtering Techniques تقدير الحالات لنظم القدرة باستخدام طرائق الترشيح المكيف |
|
Subject |
: |
Faculty of Engineering - Department of Electrical and Computer Engineering |
|
Document Language |
: |
Arabic |
|
Abstract |
: |
State estimation plays an important role in monitoring, control and stability analysis of electric power systems. For reliable operations of modern power systems it is essential to have accurate estimates of the states of the system. Traditionally, estimation was based on steady state model, and therefore, was unable to capture the system dynamics due to limitations of the supervisory control and data acquisition (SCADA) system which had slow update rates. However, due to the development of phasor measurement units (PMU), state estimation has revolutionized with high speed measurements and global positioning system (GPS) timed samples. This led to the development of dynamic state estimation (DSE) techniques. Various materials on DSE techniques can be found in the literature, including Kalman filtering (KF), Neural Networks, Particle filtering, etc. However, a very few deal with what is at the core of the power system i.e. the synchronous generator.
In this thesis, the DSE of a synchronous generator is discussed. For the purpose of estimation, 4th order nonlinear synchronous generator model was considered. The dynamic states of this model are the rotor speed, rotor angle, and the d/q – axis induced stator voltages. Existing literature deals with this problem using various Kalman filtering techniques, some techniques of Neural Networks, and Particle filters. All the aforementioned techniques are computationally extremely costly, and therefore, the aim of this thesis is to develop and implement computationally lighter as well as efficient algorithms for the mentioned task. In this regard, the first contribution in the thesis work is the development of the family of state space least mean (SSLM) algorithms which has very low computational complexity as compared to the existing model based KF techniques. Moreover, the convergence analysis is also provided to support the theoretical development of these algorithms. The second contribution is the implementation of extended Kalman filter (EKF) and the extended fractional Kalman filter (EFKF) in the DSE problem. The implementation of EFKF has not yet been investigated in the existing literature. In order to validate the performance of the aforementioned algorithms, extensive simulation experiments were conducted for various scenarios. The scenarios included the simulation and estimation of the dynamic states of the synchronous generator under normal condition, under varying input condition, and finally under short circuit fault conditions. For all these scenarios, different noise environments including Gaussian and non-Gaussian noises were investigated. Another contribution of this thesis is to provide an extensive comparison of computational complexity of all the investigated algorithms. |
|
Supervisor |
: |
Dr. Ubaid M. Al-Saggaf |
|
Thesis Type |
: |
Master Thesis |
|
Publishing Year |
: |
1436 AH
2015 AD |
|
Co-Supervisor |
: |
Dr. Muhammad Moinuddin |
|
Added Date |
: |
Monday, May 4, 2015 |
|
Researchers
| عارف .. أحمد | Ahmed, Arif .. | Researcher | Master | |
|
Back To Researches Page
|