.png){kind=icon}
.png){kind=icon}
Return
Risk-Window Planning of HVDC-Connected Renewable Energy Bases with a Chronological-Replay-Protected Decision Gate
.png){kind=icon}
OA
AI
J
L
F
G
Y
D
DOI：10.3390/en19163801.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Planning HVDC-connected renewable energy bases is constrained by the cost of full 8760 h multi-scenario replay and the risk that temporal reduction obscures coupled renewable scarcity, ramping stress, storage depletion, and delivery recovery. This study develops an auditable and extensible workflow in which a seven-dimensional coupled state generates continuous preparation–shock–recovery risk windows that feed replay-based metric diagnosis, surrogate-guided query ordering, and replay-authoritative decision release. Decision-LCB prioritizes a finite 600-vector pool using exact investment costs, predicted variable costs, and heuristic tree dispersion; an information firewall, a separate 120-vector calibration pool, and a replay-all fallback prevent unreplayed predictions from certifying feasibility. Decision-LCB recovered the finite-pool oracle in 60/60 original cases and 59/60 pre-label-frozen follow-up cases; it recorded 0/60 empirical false stops in each layer and achieved 60/60 within-tolerance outcomes and 60/60 gate passes in the follow-up. Ridge cost-greedy recovered 60/60 follow-up oracles with the same pass and false-stop counts, supporting cost-aware acquisition but not a unique LCB advantage. A passing gate requires 360 total vector replays, 40% fewer than full enumeration, whereas failure requires 720, 20% more. Simultaneous lower-bound coverage of the 360 sealed candidates was 0/60, and the measured local implementation increased online and end-to-end wall-clock times by 346.75% and 408.53%; therefore, neither simultaneous regret protection nor runtime improvement is claimed. The principal contribution is therefore architectural rather than algorithm-specific: alternative acquisition policies and higher-fidelity network-constrained or project-specific replay engines can be integrated while the information firewall, chronological replay authority, and replay-all safeguard remain fixed. This modularity provides a route to larger finite design spaces, richer technology portfolios, and broader scenario sets when the simulator cost dominates workflow overheads, while measured-data and deployment-specific validation remain necessary.
Keywords:
renewable energy base planning
continuous risk window
Decision-LCB
chronological-replay-protected gate
aggregate chronological replay
energy storage
HVDC delivery
Journal
IF:
3.2
Papers:
1.4W
Citations:
14.2W
Organization
Cited Papers