Quality-Aware Feasibility-Preserving Unit Aggregation for Smart-Grid Production Simulation

High renewable penetration, distributed energy resources, and fast-varying electric loads are shifting smart-grid planning from energy-balance simulation toward quality-aware operational assessment. Full-unit benchmark models (FULL) preserve unit commitment, ramping memory, and reserve feasibility but are expensive for repeated annual studies, whereas conventional equivalent aggregation (EQ) can overstate the realizable flexibility of heterogeneous units. This paper proposes quality-aware flexibility-envelope aggregation (QFEA), which separates units by inherited boundary state, ranks them by renewable-following flexibility, constructs conservative cluster envelopes, and couples reduced optimization with feasible disaggregation and state write-back. The model coordinates renewable curtailment, reserve sufficiency, tie-line ramping, and a normalized quality-stress proxy without claiming to replace detailed power-flow, harmonic, or electromagnetic studies. In the nominal single-region case, QFEA reduces the number of commitment objects by 46.2% and computation time by 63.7%, while limiting total-cost deviation to 1.1% and renewable-curtailment deviation to 0.2 percentage points. In 20 matched 24-h stress scenarios, its mean quality-stress index is 2.56%, compared with 2.58% for FULL and 6.57% for EQ. A separate 13–104-unit simplified scaling test keeps inverse-mapping closure error below 5.2 × 10−9 MWh and disaggregation below 1.3% of measured end-to-end time. The results identify QFEA as a traceable intermediate model for renewable-integration screening when annual computational efficiency and implementable unit trajectories are both required.