MétaCan
Menu
Back to cohort
Record W6996040694

Reinforcement learning approach for optimizing load shedding and controller parameter tuning in active islanded power systems

2024· dissertation· en· W6996040694 on OpenAlex

Why this work is in the frame

A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.

fundA Canadian funder is recorded on the work.
aboutThe title or abstract carries a Canadian signal from the geographic lexicon.
no affNo Canadian affiliation: this work is invisible to an affiliation-only frame.
No Canadian affiliation. An affiliation-only frame, the usual design, would never have seen this work. It is one of the works that make the case for inverting the frame.

Bibliographic record

VenueMspace (University of Manitoba) · 2024
Typedissertation
Languageen
FieldEngineering
TopicFrequency Control in Power Systems
Canadian institutionsnot available
FundersManitoba Hydro
KeywordsIslandingElectric power systemLoad SheddingDecentralised systemContext (archaeology)Reinforcement learningControl theory (sociology)Function (biology)Controller (irrigation)
DOInot available

Abstract

fetched live from OpenAlex

The rising frequency of extreme weather events associated with climate change has made the resilient design an important feature for power systems. Incorporating the capability of parts of the power system to operate as active decentralized systems is one way of increasing the resiliency and enhancing the system reliability. This thesis addresses two significant issues related to decentralized active networks. The first issue concerns the challenge of maintaining the balance between generation and load within an islanded system through under-frequency load shedding (UFLS) during islanding events. The optimal design of UFLS schemes in active networks presents a complex problem, especially with the high penetration of inverter-based resources (IBR) that exhibit complex dynamics and low inertia. It is also important to ensure that the system operates effectively at any given operating point, necessitating the consideration of multiple operating points in the design process. Conventional schemes and control methods may not be sufficient and identifying a suitable scheme can be both time-consuming and labor-intensive, often requires conducting many simulations at various operating points. This thesis presents an optimal load shedding method in active decentralized power systems using reinforcement learning. The effectiveness of these proposed methodologies is demonstrated through the implementation of a UFLS system on a segment of the Manitoba Hydro power network. The second significant issue addressed in this context is the tuning of controllers for active, decentralized systems. These controllers need to be adjusted to function optimally across a wide range of operating conditions, both in grid-connected and islanded modes. This poses numerous challenges due to the systems' complexity and dynamic nature and traditional heuristic controller tuning methods often prove inadequate in managing such complexities. In response, this thesis proposes a novel approach that extends the ideas of deep reinforcement learning to tune controllers across multiple operating scenarios. The validity of the proposed method is illustrated through tuning of Proportional-Integral (PI) controllers for battery energy storage systems within the modeled systems. While these developed methods are showcased on a section of the network, it can be expanded to larger power networks, providing scalability to the operation of power systems.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.000
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesMeta-epidemiology (narrow)
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Simulation or modeling · Consensus signal: Simulation or modeling
GenreCandidate signal: Empirical · Consensus signal: Empirical
Teacher disagreement score0.481
Threshold uncertainty score1.000

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0000.000
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0010.000
Bibliometrics0.0000.000
Science and technology studies0.0000.000
Scholarly communication0.0000.000
Open science0.0000.000
Research integrity0.0000.001
Insufficient payload (model declined to judge)0.0000.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.010
GPT teacher head0.197
Teacher spread0.186 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it