Transformer Tap-changer Control using Artificial Neural Network for Cross Network Connection

F. Islam, J. Kamruzzaman, and G. Lu (Australia)


Power, Transmission, Distribution, Tapchanger, Neural Network.


In recent years Artificial Neural Network (ANN) has found increased applications in power system because of its computational simplicity and robust generalization ability. Tap-changer operation of parallel transformers for controlling secondary voltage in power system is a complex problem. Existing auto controlling methods have many restrictions in operating configuration, including paralleling transformers connected across power network [1]. These methods are implemented on complex circuits and have many limitations [1]-[4]. An ANN controlled tap-changer can be a better replacement of the currently used methods to overcome the problems. Previous study demonstrated the applicability of neural network for tap control operation in closed bus system [5]. This paper briefly describes the issues of the transformer parallel operational complexities when primary sides are connected across the power network. It presents a technique to prepare the data set necessary to train neural network based on the analysis of an equivalent circuit for cross connected power network. Experiments were conducted with the data prepared using the system data collected from a power utility in Victoria, Australia. The effects of incoming source voltage variations in magnitude and phase angle are considered. Results show that ANN based tap changer can accurately determine the appropriate tap positions for in-phase incoming voltage. For phase displaced incoming voltages the overall performance is satisfactory, however, further improvements are necessary and currently under investigation.

Important Links:

Go Back