On-Policy Machine Learning Based-Disturbance Rejection Control for Grid-Tied PEC9 Inverter Under Parameters Mismatch and Distorted Grid Voltage
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.
Bibliographic record
Abstract
Thanks to higher power quality and performance efficiency, multilevel grid-tied inverters are the right choice for DC-to-AC conversion like the PV systems to the main power grid. However, the complexity of controlling the switching devices and capacitor voltages in these inverters presents significant stability challenges, particularly during grid-tied operation and when dealing with parameter mismatches. This paper proposes an optimized adaptive Active Disturbance Rejection Controller (ADRC) to stabilize the current of the grid-tied PEC9, serving as a multilevel inverter for PV applications. For this purpose, the PV system, connected to PEC9 as a main DC source to be integrated into the grid. The tunable coefficients of the ADRC controller are automatically adjusted using the on-policy reinforcement learning (RL) technique to effectively stabilize the grid-tied PEC9 with a PV inverter. In this approach, a reward function tailored to the inverter requirements guides the RL-agent in determining the optimal policy. Through maximizing the reward signal, the on-policy algorithm generates regulatory signals to adjust control gains accordingly. A laboratory prototype of PEC9 inverter is constructed by implementing OPAL-RT simulator to investigate the feasibility and applicability of suggested adaptive data-driven scheme. The experimental responses of grid-tied PEC9 equipped with the proposed adaptive ADRC demonstrate the effective performance under various operating conditions of grid-tied PV inverters, including change in the system’s references and parameter mismatches.
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 imitationNot 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.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it