Use of the Particle Swarm algorithm to estimate the error introduced by TIs in calculating the fault location distance using fundamental frequency phasors
Fault location, IT errors, nonlinear complex equations, PSO.
An Electric Power System (EPS) has the basic function of providing consumers with a continuous and quality product. For this, performance standards specified by ANEEL (National Electric Energy Agency) must be complied with by the concessionaire companies, in such a way as to guarantee the reliability of the energy supply and the conformity of voltage levels when disturbances occur and during operation. normal on a permanent basis. Thus, it ensures that the energy system will be able to keep working for a longer time, inhibiting failures due to equipment wear and avoiding the occurrence of financial losses. In order to meet reliability and quality requirements, the electrical system operation centers have data servers with oscillographic records from relays, oscillographs and qualimeters installed in the system's substations. These files are normally used for various functions, such as protection or power quality analysis. With regard to power quality, oscillographic records are normally used to classify disturbances in the electrical network.
From a protection point of view, the files are often used for post-operation analysis of the proper functioning of protective relays or for fault location algorithms. Therefore, it is essential that the records have a reliable time base, so that they can be considered synchronized, when more than one record is needed in a specific algorithm. Even considering that efficient algorithms for synchronizing records and locating faults are available, it is not possible to find in the literature an algorithm to estimate the error produced in the process of calculating the fault distance, a calculation performed according to the methodology that adopts phasors from the fundamental frequency.
It is noteworthy here that the oscillographic records available for calculating the phasors come from sampling signals, collected by means of TPs and CTs (TIs), which inherently have relation and phase errors, which depend on the accuracy class of these instruments. In addition to those related to TIs, other errors can be introduced by the sampling process or composition of the representative phasors of the sampled signals. Thus, it is necessary to develop an algorithm to estimate the error produced in determining the fault distance, even if this distance has been determined from reliable algorithms for synchronizing the records and locating the fault. In this Dissertation, we propose an algorithm based on Particle Swarm (PSO) to meet this purpose.