Automatic Validation of Converter Behaviour According to Grid Codes, and Optimisation of Their Behaviour
Authors
Kristensen, Tobias Hyldgaard ; Shen, Zhili
Term
4. term
Education
Publication year
2024
Abstract
This thesis develops a method to automatically tune the control parameters of an inverter-based source - a power device that feeds electricity to the grid through power electronics - so its responses comply with grid codes (the technical rules for grid-connected equipment). The method couples a surrogate optimization engine written in Python with PSCAD simulations of the electrical grid. System responses are converted into numerical scores through cost functions, and a weighted-sum utility combines multiple objectives into a single measure to direct the search. We examined several parts of the workflow to improve optimization quality. Parameters were optimized for different grid strengths using different procedures. For a grid with a short-circuit ratio (SCR) of 3, sets of optimized parameters were validated on a nine-bus grid model. The validated responses met the relevant grid code requirements. Overall, the work shows that carefully designed cost functions can drive an optimization engine to produce grid-code-compliant responses, enabling automated tuning of inverter-based sources.
Dette speciale udvikler en metode til automatisk at indstille (tune) kontrolparametre for en inverter-baseret kilde - et effektelektronisk anlæg, der leverer strøm til elnettet - så dens respons opfylder elnettets regler (grid codes). Metoden kobler en surrogat-optimeringsmotor i Python med PSCAD-simulationer af elnettet. Systemets responser omsættes til tal via omkostningsfunktioner, og en vægtet sum-nyttefunktion samler flere mål til ét styrende kriterium for søgningen. Vi undersøgte flere dele af arbejdsgangen for at forbedre optimeringsresultaterne. Parametrene blev optimeret for forskellige netstyrker med forskellige fremgangsmåder. For et net med kortslutningsforhold (SCR) på 3 blev sæt af optimerede parametre valideret på en ni-bus elnetmodel. Valideringen viste, at responserne opfyldte de relevante netregler. Samlet viser arbejdet, at veludformede omkostningsfunktioner kan styre en optimeringsmotor mod responser, der overholder netreglerne, og dermed muliggøre automatisk tuning af inverter-baserede kilder.
[This apstract has been rewritten with the help of AI based on the project's original abstract]
Keywords