ANN-based classification system for different windows of voltage dips in a power network

Voltage dip has been identified as one of the common types of power quality disturbance in electrical power networks and drew a lot of attention in voltage quality research presently. It is regarded as the most costly power quality problem. It can be caused by starting electrical motor, switching of generators or bulk loads, transformer energizing and short circuits in the power networks. In recent time power utilities and customers have made an effort to improve the reliability of power network, but it has been so difficult to control the external factors that cause voltage dip. As a result voltage dip must be classified and diagnosed accurately so that proper mitigation measures can be implemented. Classification of voltage dips plays an important role in voltage dip mitigation, investigations, assessment and equipment immunity specifications. In this context, this paper develops a technique using artificial neural network (ANN) for voltage dip classification, which is based on the voltage dip windows defined by the South African utility ESKOM. The test is carried out on IEEE 9-bus system through simulation in DigSILENT Power Factory 14.0 software and the ANN model is trained, tested and validated in Matlab environment using neural network Toolbox.