20.500.12556/RUL-170655 Časovno-prostorsko modeliranje dvosmernega polnjenja električnih vozil za načrtovanje prožnosti v distribucijskem omrežju Spatial-temporal modelling of bi-directional electric vehicle charging for planning flexibility in the distribution grid E-mobilnost predstavlja enega izmed ključnih korakov pri zagotavljanju zelenega prehoda na področju prometa, vendar se bo njen vpliv zaradi naraščajočega števila električnih vozil (EV) najprej občutil v elektroenergetskem omrežju (EEO). V sklopu magistrskega dela raziskujemo vpliv e-mobilnosti na izbrani del EEO ter njeno vlogo pri razvoju storitev prožnosti. Na podlagi razvitega modela, ki temelji na časovno prostorskem modeliranju polnjenja EV, ocenjujemo prednosti in slabosti koncepta dvosmernega polnjenja (angl. Vehicle-to-grid, V2G). Ker področje e-mobilnosti leži na presečišču elektroenergetike, prometa in uporabnikov, je razviti model nastal kot rezultat multidisciplinarnega pristopa. Model smo na podlagi podatkov o topologiji in telemetriji razvili na primeru distribucijskega omrežja v Velenju, ki vsebuje 292 transformatorskih postaj (TP). Zaradi obsežnosti področja smo raziskavo zastavili v treh delih. Pričeli smo s časovno-prostorskim modeliranjem prometa, da smo ustvarili realistične profile voženj za 2000 EV in obdobje enega delovnega dne. Na podlagi tega smo določili lokacije vozil v opazovanem obdobju. Ta korak je temeljil na simulacijah Monte Carlo z uporabo Markovskih verig. Pri tem smo izhajali iz statističnih podatkov o dnevni mobilnosti potnikov v Sloveniji, ki smo jih s posebno metodo vzorčenja nadgradili s podatki iz tujih raziskav. V drugem delu smo razvili orodje za napovedovanje porabe električne energije na nivoju TP, ki je zasnovano na orodju Prophet. Izdelali smo napovedi za obdobje enega dne. Poleg natančnosti smo se osredotočili tudi na možnost razširitve rešitve, kar je ključno pri napovedovanju obremenitev večjega števila TP. V tretjem in hkrati osredjem delu raziskave smo se posvetili modeliranju različnih strategij polnjenja. V ta namen smo razvili model, ki pri generiranju profilov polnjenja EV za tri različne strategije polnjenja upošteva dnevne profile voženj EV in obremenitve omrežja. Za primerjavo vplivov teh strategij na omrežje smo poleg koncepta V2G pozornost posvetili še navadnem in pametnem polnjenju. Primerjava glavnih značilnosti generiranih profilov voženj s statistiko je pokazala, da so dobljeni profili reprezentativni in primerni za nadaljnjo uporabo. S tem je tudi potrjena metoda, ki je temeljila na časovno-prostorskem modeliranju prometa. V okviru napovedovanja obremenitve smo ugotovili, da model z uporabo orodja Prophet omogoča generiranje verodostojnih napovedi. Rezultati modeliranja polnjenja so na koncu pokazali, da je dolgoročno koncept V2G najugodnejša oblika polnjenja z vidika omrežja, saj zagotavlja največji potencial pozitivne in negativne prožnosti. Dokler V2G ne doseže masovne uporabe, kot primerno rešitev vidimo pametno polnjenje, saj z vidika omrežja predstavlja bistveno boljšo rešitev v primerjavi z navadnim polnjenjem. E-mobility represents a key step towards ensuring green transition in the transport sector. However, owing to the increasing number of electric vehicles (EVs), their impact is first felt in power systems. Consequently, this thesis investigates the impact of e-mobility on the selected area of the distribution grid and its role in the development of flexibility services. Based on the developed model that relies on spatiotemporal modelling of EV charging, we evaluated the advantages and disadvantages of the vehicle-to-grid (V2G) concept. As the field of e-mobility lies at the intersection of power systems, transportation, and end users, the developed model is the result of a multidisciplinary approach. Using real data on network topology and telemetry, we developed a model for a distribution network in Velenje, which comprises 292 substations. The research was structured into three parts owing to the broadness of this field. We began with spatial-temporal traffic modeling, which represents the basis for generating realistic driving profiles for 2000 EVs during one working day. Based on this, we determined the locations of vehicles during the observed period. This step was based on Monte Carlo simulations using Markov chains. We relied on statistical data on daily passenger mobility in Slovenia, which were then expanded using a special sampling method with data from foreign studies. In the second part, we developed a tool for forecasting electricity consumption at the substation level based on the Prophet tool in Python. Using this tool, we generated forecasts for one-day simulation period. In addition to accuracy, we also focused on scalability, which is crucial for predicting the loads of a large number of substations. In the third and central part of the research, we focused on modeling different charging strategies. For this purpose, we developed a model that generates EV charging profiles for three different charging strategies, considering the daily EV driving profiles and network loads. To compare the impacts of these strategies, in addition to the V2G concept, we examined uncontrolled and smart charging. A comparison of the main characteristics of the generated driving profiles with the statistical data showed that the obtained profiles were representative and suitable for further use. This also confirmed the method itself, which was based on spatial-temporal modeling. In the context of load forecasting, we found that the Prophet-based model enabled the generation of credible forecasts. The final results of the charging modeling showed that the V2G concept is the most favorable charging strategy from the grid perspective, while also providing the greatest potential for both positive and negative flexibility. Until V2G reaches its full development, we consider smart charging as an interim solution which, despite not allowing discharging, represents a significantly better solution than uncontrolled charging from the network’s perspective. e-mobilnost vehicle-to-grid vehicle-to-everything elektroenergetsko omrežje časovno-prostorsko modeliranje potencial prožnosti Monte Carlo Markovske verige e-mobility vehicle-to-grid vehicle-to-everything power grid spatial-temporal modelling flexibility potential Monte Carlo Markov chain true false false Slovenski jezik Angleški jezik Magistrsko delo/naloga 2025-07-11 11:15:01 2025-07-11 11:15:16 2025-07-31 14:39:08 0000-00-00 00:00:00 2025 0 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 63061 244487939 Golubovic_Dusan_-_Casovno-prostorsko_modeliranje_dvosmernega_polnjenja_elektricnih_vozil_za_nacr.pdf Golubovic_Dusan_-_Casovno-prostorsko_modeliranje_dvosmernega_polnjenja_elektricnih_vozil_za_nacr.pdf 1 80869296ECEE491C4A2DFA9D7C2D7C3F 39ea514a0cf5e50e517667e95f9206067b6ae50abd8fb1d7c0ba5f25f0328e9a a46e8cfc-5e35-11f0-b232-0050569b8976 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=212695 Fakulteta za elektrotehniko 0 0 0