Font Size: a A A

High Voltage Current Measurement And Power Quality Analysis In Distribution Networks

Posted on:2015-10-19Degree:MasterType:Thesis
Country:ChinaCandidate:Z GongFull Text:PDF
GTID:2272330422471092Subject:Detection Technology and Automation
Abstract/Summary:PDF Full Text Request
Power system, which is one of the most important infrastructure and public utilities,plays a critical role in economic and social development, and its stability is an importantguarantee of the productivity and people’s living. With the development of industry andtechnology, more and more electric devices with high power, high precision and highsensitivity come into use, which demands higher power quality and makes the problemsmore complex. To address the problem, power quality analysis has received thewidespread attention. Real time fault monitor and diagnose play a critical role in powerquality maintain. This paper makes researches on high voltage measurement system andpower quality analysis. Given the insufficiency oft the traditional system, a new type ofhigh voltage current measurement system based on optical current transducer wasdesigned, which can provide necessary information for electrical measurement, relayprotection, and power system health assessment. Modeling analysis was conducted toresearch the power disturbance and an effective method on feature extraction andrecognition was proposed.First, based on the analysis of the overseas and domestic research status of opticalcurrent transducer, a structure of the high voltage measurement system was designed. Theprinciple of optical current transducer with Rogowski coil was introduced. In view of theproblem on the power supply for the high voltage side, an improved laser power systemwas built, which could effectively guarantee the stability of the power supply.To address the problem of feature extraction and classification of power qualitydisturbance, S transform was used to extract the time frequency characteristic ofdisturbance signals, and extreme learning machine (ELM) was introduced to patternrecognition. Given the large amount of calculation in S transform, A pre-fault-discoverymethod based on similarity measurement was designed and an adaptive threshold settingsmethod was proposed to improve the algorithm performance, which effectively reducedthe unnecessary calculation and improved the running speed of the system.The practicability of the presented approach is verified through the simulation and real power quality disturbance data from a certain power network.
Keywords/Search Tags:Optical current transducer, The laser feedback regulation, Power quality, Time series similarity, Extreme learning machine
PDF Full Text Request
Related items