Procjena stanja elektroenergetskog sustava u vremenskoj domeni pomoću mjerenja kutova napona

doctoral thesis

• Title (english): Power system state estimation with time tags using voltage angle measurements
• Author: Domagoj Peharda
• Mentor: Ivica Pavić (mentor)
• Committee member: Ivica Pavić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; (Department of Energy and Power Systems); Zagreb
• Defense date and country: 2014, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering; Power Engineering
• Universal decimal classification (UDC): 621.3 - Electrical engineering

Abstract

Algoritam postupaka za procjenu stanja dugo je poznat i implementiran u većini centara za vođenje elektroenergetskog sustava. Klasična statička procjena stanja koristi mjerenja snage i napona propuštene kroz niskopropusni filter, te najčešće računa procjenu stanja s periodičnošću od jedne minute ili na promjenu topologije s počekom. Sve veća povezanost regionalnih elektroenergetskih sustava znači i brže promjene u sustavu, što dovodi do zahtjeva za češćom procjenom stanja. Uređaji za mjerenje... More fazora mjere veličine svake periode, te nije moguće miješati klasična mjerenja s mjerenjima fazora u klasičnim tokovima snaga. Razvijen je model mjerenja koji modelira težinski faktor mjerenja u vremenskoj domeni, poštujući karakteristike mjerenja i pretvornika (točnost, mrtva zona, periodičnost) i međupodručne oscilacije. Tako razvijeni model težinskog faktora omogućava upotrebu klasičnih mjerenja s mjerenjima fazora u statičkoj procjeni stanja svake sekunde. Promjene topologije u klasičnoj procjeni stanja iniciraju poček u kojem se vremenu očekuje da stignu promijenjena mjerenja. Predložena procjena stanja kreira pseudomjerenja na bazi novih tokova snaga. Također, predložena procjena stanja koristi mjerenja fazora da utvrdi suvislost rješenja procjene stanja i da otkrije komforna pogrešna mjerenja koja su česta pojava prilikom topoloških promjena u nenadziranoj mreži. U svrhu uključenja mjerenja fazora napona i struje u algoritam za procjenu stanja kao ravnopravnih mjerenja, korištene su kompleksne veličine svih mjerenja, te kompleksne funkcije nad njima. Problem procjene stanja postavljen pomoću kompleksnih jednadžbi nije moguće riješiti klasičnom Newtonovom metodom jer funkcije ne zadovoljavaju Cauchy-Riemannove uvjete, zbog čega je modificirana Newtonova metoda tako da se koristi usmjerena derivacija funkcije s kompleksnim varijablama, gdje je vektor po kojem se derivira jediničan vektor, okomit ili uzdužan na vektor napona. Ukoliko je mreža potpuno observabilna pomoću kombinacije mjerenja fazora napona i struje, tada metoda konvergira u jednom koraku, tj. prelazi u direktnu metodu. Razvijena programska podrška je testirana na IEEE testnoj mreži s 24 čvorišta, izračunata su 604800 stanja i na stvarnoj mreži od 329 čvorišta. Less

Abstract (english)

Power system state estimation is implemented in most Energy Management System software. Classic static state estimation uses measurements of power and voltage from SCADA system and usually calculates state estimate, i.e. voltages of all nodes and tap ratios, with periodicity of one minute or on topology change. There are more measurements than state variables to be calculated so algorithm for state estimation solves overdetermined system according to some minimization function. The most... More commonly used algorithm is WLS – Weighted Least Squares. The classic static state estimation has a period between two calculations usually set at one minute which is too slow for dispatchers who must act according to the current state of the system. Measurements in SCADA are not synchronized in time (measurements reflect different states of the power system) and take some time to come to SCADA because they are traveling through various communication channels. Classic state estimation does not take the time tags in consideration. Classic state estimation does not have a benchmark for accuracy of its estimate. As the regional power systems are becoming more and more interconnected, changes in the system are becoming faster thus expressing the need for faster state estimation. It was believed that the Wide Area Measurement Systems would solve the problem of state estimation by measuring states of all nodes on each period. Phasor measurements are very sensitive to angle measurement errors (which are common because of errors of measurement transformers and calibration problems of phasor measurement units) and small error in angle measurement translates to large error in estimate of the power through the line which makes them unsuitable for static state estimation. On the other hand phasor measurement units have self-diagnostic algorithms which make sure that phasor measurements do not have gross errors. Proposed model of static state estimation uses time tags to determine weights of each measurement at each state estimation calculation and uses phasor measurement for calculating the accuracy of the state estimate. For this purpose a model of measurements with time tags is developed for determining of weights in each estimate according to time of the measurement and measurement transformer/converter characteristics (accuracy, dead zone, periodicity) and inter-area oscillations. Using this weight model it is possible to mix classic measurements with phasor measurements in a static state estimation with periodicity of up to one second. Topology change in classic state estimation initiate a waiting period so all measurements changed by topology change would be updated. Proposed method for state estimation creates pseudo-measurements using load flow made on changed topology but using injections and generator voltages of last correct state. State estimation is done in two steps. In first step phasor measurements are used as equality constraints. Average normalized residual is calculated and used as a proxy to the accuracy of the state estimation solution. If the solution is satisfactory, i.e. classic measurements are in accordance with phasor measurements, then the second step of the state estimation is performed. In the second step, phasor measurements are used as regular measurements with weights. Use of phasor measurements as a equality constraints allows for finding conforming bad measurements which are common after topology changes in unobservable part of the network. A new algorithm is proposed which includes phasor measurements of voltage and current as equal to other measurements. Proposed algorithm uses complex values for measurements and complex valued functions. This problem is not solvable using classic Newton method because functions do not conform to Cauchy-Riemann’s rules. Newton method is modified to use directional derivation of a function with complex variables where directional vector is unity vector with direction parallel or vertical to state vector of voltages. If the network is completely observable using only phasor measurements of voltage and current, method converges in one step, i.e. it becomes a direct method. For purpose of this thesis a software was developed and tested on IEEE 24 bus test system for 604800 states, and also on a real network with 329 busses. Less

• Keywords: elektroenergetski sustav; procjena stanja; mjerenja fazora; Newtonova metoda; usmjerena derivacija; težinski faktori; metoda najmanjih vaganih kvadrata; topološka promjena; observabilnost; digitalni filteri; filteri
• Keywords (english): power systems; state estimation; phasor measurements; Newton method; directional derivation; weighting factors; weighted least squares; WLS; topology change; observability; digital filters; filters
• Language: croatian
• URN:NBN: urn:nbn:hr:168:620499
• Study programme: Title: Electrical Engineering and Computing; Study programme type: university; Study level: postgraduate; Academic / professional title: Doktor znanosti elektrotehnike i računarstva (Doktor znanosti elektrotehnike i računarstva)
• Type of resource: Text
• Extent: 160 str. ; 30 cm.
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-04-19 13:32:08