Zainaddin, Samir Ahmad (1994) A neural network approach for load flow estimation and var allocation. Masters thesis, King Fahd University of Petroleum and Minerals.
A load flow solution is estimated by a neural network. A one hidden layer neural network is trained to develop a non-iterative solution of the power flow problem. The effect of line outages and transformer taps on the load flow estimation is taken into consideration, during the training process. The trained neural network represents the power system model relating injected powers to output voltage amplitudes and angles. The neural network model is used for var/voltage control through a novel method of using the backpropagation learning algorithm. Power loss minimization is included as a var/voltage control requirement. The power systems studied in this thesis are a sample 6-bus and the IEEE 14-bus systems. Testing results of the neural network model have shown an excellent agreement with results obtained from the various iterative techniques. Moreover the neural network model was successfully used in a var allocation scheme that weighs the various var/voltage sources differently. The neural network model provides a simple non-iterative technique for the solution of the power system equations. The model can be extremely useful in such cases where the sensitivity matrix is required to be calculated in power system control applications.
|Item Type:||Thesis (Masters)|
|Divisions:||College Of Engineering Sciences > Electrical Engineering Dept|
|Creators:||Zainaddin, Samir Ahmad|
|Committee Advisor:||Abdel-Magid, Youssef L.|
|Committee Members:||El-Amin, Ibrahim Mohammad and Al-Baiyat, Samir A. and Abdel-Salam, M. S.|
|Deposited By:||KFUPM ePrints Admin|
|Deposited On:||22 Jun 2008 17:05|
|Last Modified:||30 Apr 2011 15:38|
Repository Staff Only: item control page