Optimal fault location

Abstract

Basic goal of power system is to continuously provide electrical energy to the users. Like with any other system, failures in power system can occur. In those situations it is critical that correct remedial actions are applied as soon as possible after the accurate fault condition and location are detected. This thesis has been focusing on automated fault location procedure. Different fault location algorithms, classified according to the spatial placement of physical measurements on single ended, multiple ended and sparse system-wide, are investigated. As outcome of this review, methods are listed as function of different parameters that influence their accuracy. This comparison is than used for generating procedure for optimal fault location algorithm selection. According to available data, and position of the fault with respect to the data, proposed procedure decides between different algorithms and selects an optimal one. A new approach is developed by utilizing different data structures such as binary tree and serialization in order to efficiently implement algorithm decision engine. After accuracy of algorithms is strongly influenced by available input data, different data sources are recommended in proposed architecture such as the digital fault recorders, circuit breaker monitoring, SCADA, power system model and etc. Algorithm for determining faulted section is proposed based on the data from circuit breaker monitoring devices. This algorithm works in real time by recognizing to which sequence of events newly obtained recording belongs. Software prototype of the proposed automated fault location analysis is developed using Java programming language. Fault location analysis is automatically triggered by appearance of new event files in a specific folder. The tests were carried out using the real life transmission system as an example.