Practical Multimodel Dynamic Matrix Control for Power Plant Superheated Steam Temperature
IEEE Sensors Journal · 15.07.2026 · 26 (14) · 21057-21067
DOI: 10.1109/jsen.2026.3701052 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Sensors Journal, erschienen am 15.07.2026. Verfasst von Yuhan Zhang und 3 weiteren Autor:innen. Die Arbeit nennt 35 Literaturangaben.
Abstract (Originalsprache)
Controlling superheated steam temperature in thermal power plants is challenged by large inertia, time delays, and load-dependent dynamics, which severely limit traditional cascade PID controllers. To address these industrial issues, this article proposes a practical multimodel dynamic matrix control (DMC) strategy embedded in a cascade architecture. Step-response models are identified from operating data across 30% to 100% load conditions. A linear interpolation weighting mechanism dynamically schedules the DMC to mitigate model mismatch during load variations. The principal novelty of this work lies in performing load-indexed scheduling directly on nonparametric step-response coefficients, which guarantees a physically realizable blended model and preserves the lower-triangular Toeplitz structure required by the DMC formulation. Consequently, the scheduler introduces no additional online matrix inversion or optimization steps, ensuring full compatibility with the stringent CPU and memory constraints of industrial distributed control systems (DCSs). The strategy is rigorously validated through extensive MATLAB/Simulink simulations encompassing set-point tracking, disturbance rejection, and model–plant mismatch scenarios. Compared with five independently tuned cascade PI controllers, the proposed scheme reduces overshoot from approximately 12%–13% to below 1.3% and the 2% settling time by 52%–58% across the full 0%–100% load range while requiring only a single set of DMC tuning parameters. Results demonstrate that the multimodel DMC eliminates overshoot, significantly reduces settling time, and exhibits superior robustness against random disturbances and valve nonlinearities compared to well-tuned PID control, proving its strong potential for real-world distributed control system deployment.
Zitieren
APA
Zhang, Y., Zhou, X., Xiong, H., & Zhou, H. (2026). Practical Multimodel Dynamic Matrix Control for Power Plant Superheated Steam Temperature. IEEE Sensors Journal, 26(14), 21057-21067. https://doi.org/10.1109/jsen.2026.3701052
Vancouver
Zhang Y, Zhou X, Xiong H, Zhou H. Practical Multimodel Dynamic Matrix Control for Power Plant Superheated Steam Temperature. IEEE Sensors Journal. 2026;26(14):21057-21067. doi:10.1109/jsen.2026.3701052
BibTeX
@article{Zhang2026,
author = {Zhang, Yuhan and Zhou, Xingwei and Xiong, Hongkun and Zhou, Hong},
title = {Practical Multimodel Dynamic Matrix Control for Power Plant Superheated Steam Temperature},
journal = {IEEE Sensors Journal},
year = {2026},
volume = {26},
number = {14},
pages = {21057--21067},
doi = {10.1109/jsen.2026.3701052}
}
EUVIMED