A novel intelligent protection system for power transformers considering possible electrical faults, inrush current, CT saturation and over-excitation

• Published in: International journal of electrical power & energy systems Vol. 64; pp. 1129 - 1140
• Main Authors: Yazdani-Asrami, Mohammad, Taghipour-Gorjikolaie, Mehran, Mohammad Razavi, S., Asghar Gholamian, S.
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.01.2015
• Subjects: Bayesian classifier; CT saturation; Improved gravitational search algorithm; Internal fault; Magnetizing inrush current; Over-excitation; Over-excitation; Magnetizing inrush current; Improved gravitational search algorithm; Internal fault; CT saturation; Bayesian classifier
• ISSN: 0142-0615; 1879-3517
• DOI: 10.1016/j.ijepes.2014.08.008

•Four condition identified in power transformers.•Bayesian classifier has high distinctive capability.•The sensitivity analysis show the effectiveness of approach.•Improved gravitational search algorithm has been introduced. Many electrical events can be easily damage electrical equipments in power systems. Such events or faults can be easily stopped at incipient steps but because of weakness of protecting systems they grow and extend, and consequently impose so many problems and cost to utilities. Power transformers are one of the vital equipments in electrical networks and industries, although many protecting systems have been implemented to prevent dangerous electrical faults, but most of them suffer many problems, such as; time wasting, computational burden, and low speed in response. In addition, whenever patterns of fault signals are similar, their discrimination from each other is so hard. Magnetizing inrush current, internal fault, CT saturation and over excitation are common electrical faults in power transformers so that most of the time are difficult to be separated. In this paper all considerations for designing perfect protecting system for power transformers are considered. Using intelligent approach, Artificial Neural Network (ANN) based method is designed. Indeed, proposed protecting system includes two major sections. In the first section, using Bayesian Classifier (BC) which works based on Bayesian rules and uses knowledge of training data directly; internal fault is detected and is discriminated from three other mentioned faults and normal condition. If the event is not internal fault, second condition of this intelligent system makes a decision. In this section, ANN trained by swarm based algorithms, namely; Improved Gravitational Search Algorithm (IGSA) or Particle Swarm Optimization (PSO) is used to discriminate magnetizing inrush current, current transformer (CT) saturation and over excitation. Obtained results show that proposed system can easily and precisely follow the electrical faults in power transformer and detect them at incipient steps. Such quick and accurate response helps to save so much energy, financial cost and time.