| dc.description.abstract | The unplanned power interruptions and network faults are being main concerns of several network operators and customers as they cause economical loss, fires and life damages. Thus, the primary aim of this thesis is to investigate a potential use of power quality measurements to forecast some of network faults and power interruptions. Moreover, its secondary purpose is to develop an automatic algorithm that predicts some of the power interruptions and faults by monitoring and analysing the power quality measurements. So that the network operator would be able to take counteractions to avoid some of the power interruptions and faults before they occur.
The cycle by cycle RMS voltage trends of 300 power interruptions occurred in 41 distribution and transmission operators in Norway have been investigated. The measurements are registered in SINTEF Energy AS and Statnett databases using Elspec's investigator (power quality analysers). In addition, the fault statistics reports from the distribution and transmission network operators in which the faults occurred at have been analysed to verify the connection between the network faults (power interruptions) and their preceding pre-fault events by identifying the their source locations and causes.
Consequently, the results showed that 25 percent of the power interruptions whose fault statistics were analysed could have potentially been predicted before they occurred using their preceding pre-fault events as fault indicator signals (warnings) by monitoring and analysing the power quality measurements. The network faults that could have been forecasted were caused by tree falling on overhead lines, overhead line conductors' clashes and defected underground cables. Therefore, the findings of this study discovered a promising potential use of power quality measurements to forecast some of the network faults (power interruptions). In addition, based on the findings, an automatic algorithm that makes network fault predictions by analysing the power quality (RMS voltage) measurements automatically is proposed.
In conclusion, the findings of this study may be used as initiative for future works and studies on power quality and network faults as the previous studies on network fault predictions are very limited. This study can also contribute considerably to the development of an automatic fault prediction algorithm as it proposes a possible way of predicting some network faults
Key Words: Failure causes of power system components, classification of network faults, power quality measurements, predictable power interruptions, network fault prediction, pre-fault events, and network fault prediction algorithm. | en |
