Artificial Intelligence Techniques for Fault Location and Detection in Distributed Generation Power Systems

There are multiple reasons that distributed generation systems are exposed to faults and errors. Among them we can consider the most common ones being failure of the power system components or lighting strikes. Power system failure is usually caused by components failure, human error or equipment aging. The above-mentioned events affect the power system reliability and imply expensive repairs, lost of productivity and dissatisfied customers. Due to the fact that faults are unpredictable, fast fault location and isolation is necessary to minimize the impact of fault in distribution systems. Therefore, many methods have been previously developed by researchers to locate and detect these faults in distribution power systems that incorporate distributed generation. The main available methods can be split into two categories: conventional and artificial intelligence techniques. Conventional methods are composed of travelling wave method and impedance-based method while artificial intelligence techniques include Artificial Neural Network, Support Vector Machine, Fuzzy Logic, Genetic Algorithm and matching approach. However, fault detection using intelligent methods are difficult to implement due to the fact that it requires training data for processing and also are time consuming. This paper presents some of the newly introduced artificial intelligence techniques for islanding detection in distributed generation power networks. Research works in islanding detection, control algorithms, advantages and disadvantages are presented, hence the opportunities in islanding detection research area in power distribution system can be explored further.