Thermal-capacity-aware admissible command generation and constrained tracking control for flexible pulverized-coal boiler–turbine units
High renewable penetration is shifting coal-fired units toward deep peak-shaving and automatic generation control (AGC)-like load following, where the key limitation is not only tracking speed but command thermal admissibility. This study proposes a thermal-capacity-aware admissible command generation and constraint-aware tracking framework for subcritical pulverized-coal boiler–turbine units. Rather than imposing raw load requests directly on the tracking controller, the framework screens them against thermal-capacity states, heat-storage margins, pressure/temperature envelopes and actuator constraints, and generates a thermally admissible command for downstream multirate model predictive control (MPC). A nonlinear 24-state boiler–turbine simulation plant, a load-scheduled linear parameter-varying (LPV) prediction bank, a migration observer, soft-sensed migration indices, a dynamic safety corridor and an admissible-command governor are integrated to convert thermal inertia from a post hoc explanation of slow response into an online command-admissibility variable. In the default and bidirectional heterogeneous cases, the electrical-output root-mean-square errors (RMSEs) relative to the original grid requests are 49.30 and 66.55 MW, while the corresponding tracking RMSEs relative to the admissible references are 0.809 and 1.263 MW. The governor reshapes the difficult requests into thermally admissible trajectories, with integrated raw-to-admissible command deviations of 32.69 and 53.12 MWh and route-active fractions of 74.2% and 65.7%, respectively. The tested AGC-like trajectory passes through unchanged, yielding zero route gap and an RMSE of 1.353 MW. Steam-temperature violations remain zero, while pressure and air-excess excursions are explicitly quantified. The framework therefore provides a constraint-aware supervisory interface that converts grid-side requests into thermally admissible boiler load trajectories for accurate downstream tracking.
Authors
- 陈德珍
- JINGKUAN HUANG
- Lijie Yin
- Zeyu Liu
Institutions
- Tongji University (CN)
Publication Details
- Journal
- Applied Thermal Engineering
- Published
- 2026-10-05
- DOI
- https://doi.org/10.1016/j.applthermaleng.2026.133478
- Primary Topic
- Advanced Control Systems Optimization
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Natural Science Foundation of Shanghai