Skip to content
Open access

Calibrated Soft Network-Security-Aware Predictive Dispatch for Renewable-Rich Distribution Networks With LinDistFlow and AC Validation

2026 · IEEE Access · Vol 14, pp. 156559-156570 · 0 citations · 25 references

Abstract

Renewable-rich distribution dispatch must coordinate grid import, storage, curtailment, and controllable demand while accounting for feeder-security risk. This study evaluates a calibrated soft network-security-aware model predictive control (MPC) framework for radial feeders. A branch-flow and voltage proxy is fitted using a 14-day calibration window that is separate from all evaluation periods. Branch ratings are derived from calibrated AC apparent-flow envelopes, and security-penalty weights minimize a prespecified equal-weight normalized root-mean-square score. AC loss statistics provide an operating reserve for network losses. Comparisons include aggregate deterministic MPC, the calibrated soft proxy, standard LinDistFlow, and variants with inverter reactive-power support. All realized schedules undergo validation using nonlinear radial AC power flow. The evaluation covers three 7-day operating cases, ten independent forecast-error realizations, voltage-factor and voltage-threshold sensitivity, and horizon and initial-state-of-charge tests. It also includes four seasonal measured-load windows and IEEE 33- and 69-bus feeders. In held-out IEEE-33 cases, the calibrated proxy and LinDistFlow variants reduce AC-validated branch exposure relative to aggregate MPC. They also eliminate voltage-shortfall exposure at the study’s 0.90-p.u. screening threshold. A post-hoc 0.92/0.95-p.u. analysis shows that this voltage result depends on the threshold, while the benefit of reactive-power support varies with operating conditions. On IEEE-69, the proxy increases branch exposure despite feeder-specific calibration, whereas LinDistFlow-Q reduces it. These improvements entail higher operating cost or energy not served (ENS) in some settings. The contribution is a calibrated benchmark and empirical evidence on security–cost trade-offs.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.