Bootstrap Participation Screening for Photovoltaic Day-Ahead Bidding Under Alternative Price Scenario Models

A photovoltaic producer must select a day-ahead supply bid and decide whether to submit it. Positive revenue estimates obtained during bid selection may not persist under new price scenarios, motivating an independent assessment of the selected bid. This study separates bid selection by sample average approximation from participation screening and compares a multi-scenario bootstrap lower-quantile (MSBQ) rule with mean-based and standard-error alternatives. A case study of a 190 MW photovoltaic station in Shandong, China, combines controlled scenario comparisons with a 117-day retrospective evaluation using public hourly prices. The results show that conservative screening can forgo substantial positive-expectation opportunities under the empirical price model. MSBQ achieves lower model-conditional opportunity loss than a specified variance-corrected mean-minus-standard-error rule, while a development-calibrated standard-error rule achieves lower loss than MSBQ. Pooled zero-threshold screening earns higher observed incremental settlement revenue under expanding and rolling historical windows, and statistical evidence for an MSBQ revenue advantage remains inconclusive. Revenue rankings also change with the price model and sampling design. These findings connect the choice and calibration of participation rules to the representation of market prices and the economic consequences of rejected bids.