Probabilistic Transient Stability Assessment of Renewable-Dominated Power Systems
Authors
Mahin, Shahariar ; Hossain, Anik
Term
4. term
Education
Publication year
2026
Submitted on
2026-05-27
Abstract
The shift from power grids dominated by synchronous generators to systems with high shares of renewable energy introduces two key challenges: increased operational uncertainty and reduced rotational inertia. Traditional, deterministic stability assessments use fixed assumptions about fault location, clearing time, and operating point. As a result, they cannot capture the real variability in system conditions and do not reveal which specific lines or operating situations actually drive instability. This thesis develops and implements a probabilistic transient stability assessment framework for the IEEE 39-bus test system in the software DIgSILENT PowerFactory. In each simulation, a Transient Stability Index (TSI) is used as the outcome, and on the system level the main metric is the probability of instability. Four sources of uncertainty are modeled and introduced step by step: fault clearing time, total load (demand), the transmission line on which the fault occurs, and the replacement of synchronous generators by wind farms. The faulted line is selected randomly with a probability proportional to its length, reflecting that longer lines are physically more exposed to faults. When load variability is added, the overall probability of instability for the system changes only modestly. However, the number of lines with non-zero conditional instability – meaning lines that can cause problems if the fault hits them – roughly doubles. This effect is hidden in the system-wide metric and only becomes visible in a per-line analysis. Wind farm integration consistently makes the system more unstable as synchronous inertia is replaced. Moreover, two wind integration scenarios with almost the same loss of inertia are found to lead to noticeably different stability outcomes and to shift the set of vulnerable lines to different geographical areas in the grid. This is reported as an empirical observation, and understanding the underlying mechanism is identified as future work. Overall, the findings demonstrate that probabilistic transient stability assessment can serve as a practical tool for risk-informed contingency analysis, dynamic security management, and risk-oriented power system planning under uncertainty.
Overgangen fra elnet, der hovedsageligt er baseret på synkrongeneratorer, til systemer med en høj andel af vedvarende energi skaber to centrale udfordringer: større driftsusikkerhed og lavere rotationsinerti i systemet. Traditionelle, deterministiske stabilitetsanalyser arbejder med faste antagelser om fejlens placering, udkoblingstid og driftspunkt. De kan derfor ikke beskrive den faktiske variation i driften og kan heller ikke vise, hvilke ledninger eller driftsbetingelser der konkret gør systemet ustabilt. Denne afhandling udvikler og implementerer en probabilistisk metode til vurdering af transient stabilitet på IEEE 39-bus testsystemet i programmet DIgSILENT PowerFactory. For hver simulering anvendes et Transient Stability Index (TSI) som resultatmål, og på systemniveau bruges sandsynligheden for ustabilitet som overordnet risikomål. Fire kilder til usikkerhed modelleres og tilføjes trinvis: fejlens udkoblingstid, det samlede forbrug (belastning), hvilken ledning fejlen optræder på, og hvor meget af synkrongeneratorkapaciteten der erstattes af vindmølleparker. Den fejlramte ledning vælges tilfældigt med en sandsynlighed, der er proportional med ledningens længde, for at afspejle at længere ledninger fysisk set er mere udsatte for fejl. Når variation i belastningen tilføjes, ændres den samlede sandsynlighed for ustabilitet kun moderat. Til gengæld fordobles omtrent antallet af ledninger, der har en ikke-nul betinget ustabilitet, det vil sige ledninger, hvor der kan opstå problemer, hvis fejlen netop rammer dem. Denne effekt kan ikke ses i det samlede systemmål, men bliver tydelig i en analyse ledning for ledning. Integration af vindmølleparker forværrer generelt systemets stabilitet, efterhånden som synkron inerti erstattes. To scenarier med vindintegration, der næsten mister den samme mængde inerti, viser sig desuden at give markant forskellige stabilitetsresultater og flytter de mest sårbare ledninger til andre geografiske områder i nettet. Dette konstateres som en empirisk observation, mens den præcise fysiske forklaring peges på som et emne for fremtidig forskning. Samlet set viser resultaterne, at probabilistisk transient stabilitetsvurdering kan være et praktisk redskab til risikobaseret beredskabsanalyse, dynamisk sikkerhedsstyring og planlægning af elnettet under usikkerhed.
[This abstract has been rewritten with the help of AI based on the project's original abstract]
Keywords