Lociranje zemeljskostičnih okvar v srednjenapetostnem omrežju s kompenzirano nevtralno točko z uporabo sinhrofazorskih meritev in podatkov o topologiji omrežja

Vardić, Anja (avtor), Ažbe, Valentin (mentor), Curk, Jurij (komentor)

Zemeljski stiki predstavljajo približno 70 % okvar v elektroenergetskih omrežjih, zato sta njihova zanesljiva zaznava in pravočasna odprava pomembni za varno in zanesljivo obratovanje sistema ter preprečevanje poškodb opreme. Pri enofaznem zemeljskem stiku v resonančno ozemljenem omrežju Petersenova dušilka, ki je priključena na nevtralno točko transformatorja, s svojim induktivnim tokom kompenzira kapacitivni zemeljskostični tok omrežja. Posledično se tok na mestu okvare zmanjša, s čimer se omejijo njene posledice, hkrati pa postane zaznava s klasičnimi zaščitnimi metodami zahtevnejša. V okviru magistrskega dela sta zato detekcija in lokalizacija obravnavani v času, ko omrežje obratuje z vključeno Petersenovo dušilko. V obravnavanem omrežju je zato Petersenova dušilka uporabljena v kombinaciji z nizkoohmskim uporom. Dušilka med zemeljskim stikom omejuje okvarni tok, ob trajni okvari pa se vzporedno vključi upor, ki poveča njegovo delovno komponento in s tem omogoči zanesljivejšo zaznavo okvarjenega izvoda ter njegovo selektivno izključitev. Takšna izvedba hkrati omogoča uporabo obstoječega sistema zaščite, saj bi prehod na izključno resonančno ozemljitev zahteval obsežnejšo prilagoditev zaščitnih funkcij. Obstoječa zaščita tako omogoča zaznavo in izklop okvarjenega izvoda, ne zagotavlja pa natančne določitve mesta okvare vzdolž voda. Dodatno omejitev predstavlja spreminjanje konfiguracije omrežja. S spremembo stikalnih stanj se spremeni skupna dozemna kapacitivnost in s tem stopnja kompenzacije. Temu je treba prilagoditi nastavitev Petersenove dušilke, spremenjene razmere pa lahko vplivajo tudi na delovanje zaščite. To predstavlja izhodišče za razvoj metode, ki bi poleg merjenih električnih veličin upoštevala tudi trenutno topologijo omrežja. Uporaba časovno sinhroniziranih fazorskih meritev WAMS bi lahko v tem okviru predstavljala dodaten analitični sloj obstoječemu zaščitnemu sistemu ter omogočila natančnejšo detekcijo, določanje smeri in lokalizacijo zemeljskega stika. Cilj takšnega pristopa ni nadomestiti obstoječe zaščite, temveč izboljšati informacijo o mestu okvare in s tem podpreti njeno hitrejšo ter selektivnejšo izolacijo. Zaradi omejenih možnosti poseganja v obratovanje realnega omrežja in potrebe po preverjanju algoritma pri večjem številu scenarijev je bilo treba najprej vzpostaviti ustrezno testno okolje. To mora omogočati simulacijo različnih konfiguracij omrežja in poljubnih parametrov okvare, hkrati pa zagotavljati ponovljivost posameznih scenarijev ter ustrezno predstavitev fizikalnega obnašanja omrežja. V ta namen je bila razvita avtorska platforma, ki je zasnovana okoli simulacijskega jedra OpenDSS, ki omogoča pripravo, spreminjanje in izvajanje različnih obratovalnih scenarijev. Vzpostavitev takšnega okolja je predstavljala prvi del praktičnega dela in osnovo za nadaljnjo zasnovo metod za detekcijo in lokalizacijo okvare. Osnovna ideja detekcije temelji na tem, da posamezna PMU-naprava prevzame vlogo smernega merilnega elementa in na podlagi faznih odnosov določi prisotnost ter smer okvare. Časovna sinhronizacija PMU-naprav pri tem omogoča sočasno spremljanje istega dogodka z različnih lokacij v omrežju. Rezultati detekcije predstavljajo izhodišče za lokalizacijo okvare. Za vsako aktivirano PMU-napravo se upošteva zaznana smer okvare in na podlagi topologije določijo možne poti v tej smeri. S tem se iskanje lokacije omeji na del omrežja, ki je skladen z odzivom vseh aktiviranih naprav. Znotraj tako določenih poti se nato lokacija okvare ocenjuje na podlagi topoloških podatkov in sintetično generiranih fazorskih meritev. Rezultati so pokazali, da je uspešnost lokalizacije odvisna predvsem od števila in razporeditve PMU-naprav, oddaljenosti in upornosti okvare ter konfiguracije omrežja. Metoda je bila uspešnejša pri radialnem obratovanju in okvarah bližje RTP, z večanjem oddaljenosti pa se je njena natančnost zmanjšala. Pri zazankanem obratovanju je zaradi več možnih poti širjenja vpliva okvare za zanesljivo lokalizacijo potrebna večja merilna pokritost. Na uspešnost detekcije vplivata tudi okvarna upornost in izbira karakterističnega kota. V uspešnih primerih je metoda omogočila določitev mesta okvare znotraj zastavljenega kriterija natančnosti 300–500 m. Izračunane lokacije predstavljajo oceno mesta okvare, njihovo natančnost pa bi bilo mogoče dodatno izboljšati z upoštevanjem kapacitivnih prispevkov zdravih izvodov.

• Jezik: Slovenski jezik
• Ključne besede: PMU, enofazni zemeljski stik, Petersenova dušilka, lokalizacija, detekcija, OpenDSS
• Vrsta gradiva: Magistrsko delo/naloga
• Organizacija: FE - Fakulteta za elektrotehniko
• Leto izida: 2026
• PID: 20.500.12556/RUL-188947
• Datum objave v RUL: 30.09.2026

Sekundarni jezik

• Jezik: Angleški jezik
• Naslov: Localization of Earth Faults in a Medium-Voltage Network with a Compensated Neutral Using Synchrophasor Measurements and Network Topology Data

Izvleček

Single-phase-to-earth faults account for approximately 70% of faults in electrical power systems. Their reliable detection and timely clearance are therefore essential for safe and reliable system operation and for preventing equipment damage. In resonant-earthed networks, the inductive current of the Petersen coil connected to the transformer neutral compensates for the capacitive earth-fault current. Consequently, the current at the fault location is reduced, limiting the impact of the fault while making its detection by conventional protection methods more challenging. In this thesis, fault detection and localization are therefore considered while the network is operating with the Petersen coil connected. In the network under consideration, the Petersen coil is used in combination with a low-resistance resistor. During an earth fault, the coil limits the fault current, while in the event of a sustained fault, the resistor is connected in parallel to increase the active component of the fault current. This enables more reliable identification of the faulty feeder and its selective disconnection. Such an arrangement also allows the existing protection system to remain in use, as a transition to purely resonant earthing would require more extensive modifications to the protection functions. The existing protection system can therefore detect and disconnect the faulty feeder, but it does not provide an accurate estimate of the fault location along the feeder. An additional limitation is the changing network configuration. Changes in switching states alter the total phase-to-earth capacitance and, consequently, the degree of compensation. The Petersen coil setting must therefore be adjusted accordingly, while the changed operating conditions may also affect protection performance. This provides the motivation for developing a method that considers not only measured electrical quantities but also the current network topology. In this context, time-synchronized phasor measurements provided by a WAMS could serve as an additional analytical layer alongside the existing protection system, enabling more accurate fault detection, directional determination, and localization. The proposed approach is not intended to replace the existing protection system, but rather to provide more accurate information on the fault location and thereby support faster and more selective fault isolation. Due to the limited possibility of intervening in the operation of a real distribution network and the need to evaluate the algorithm under a wide range of scenarios, an appropriate testing environment first had to be established. The environment must support the simulation of different network configurations and arbitrary fault parameters while ensuring the repeatability of individual scenarios and an adequate representation of the network's physical behaviour. For this purpose, a custom platform was developed around the OpenDSS simulation engine, enabling the preparation, modification, and execution of different operating scenarios. Establishing such an environment constituted the first part of the practical work and provided the basis for the subsequent development of the fault detection and localization methods. The basic concept of the detection method is based on using each PMU as a directional measuring element that determines the presence and direction of a fault from the measured phasor relationships. Time synchronization between the PMUs enables the same fault event to be observed simultaneously from multiple locations across the network. The detection results then serve as the input for fault localization. For each activated PMU, the detected fault direction is combined with the network topology to identify all possible paths in that direction. The search area is thereby restricted to the part of the network consistent with the responses of the activated devices. Within the resulting candidate paths, the fault location is subsequently estimated using the network topology and synthetically generated phasor measurements. The results showed that localization performance depends primarily on the number and placement of PMUs, the fault distance and resistance, and the network configuration. The method performed better under radial operation and for faults located closer to the primary substation, while its accuracy decreased with increasing fault distance. Under meshed operation, greater measurement coverage was required for reliable localization due to the multiple paths through which the effects of the fault can propagate. Detection performance was also influenced by fault resistance and the selection of the characteristic angle. In successful cases, the method was able to locate the fault within the predefined accuracy criterion of 300–500 m. The calculated locations represent estimates of the actual fault location, and their accuracy could be further improved by accounting for the capacitive current contributions of healthy feeders.

• Ključne besede: PMU, single-phase-to-ground fault, Petersen coil, fault localization, fault detection, OpenDSS