EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A novel support vector clustering-based data-driven robust optimization method for minimum energy consumption of the chiller-pump system

Jianyang Cai, Jianjun Liu

Thermal Science · 2026 · (00) · 190-190

DOI: 10.2298/tsci260509190c ↗

Kurz gefasst

Fachartikel in der Zeitschrift Thermal Science, 2026. Verfasst von Jianyang Cai und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Due to the deviation between the actual and predicted values of the cooling load(CL), it may lead to a mismatch between the commissioning strategy for optimizing the chiller-pump (CP) system energy consumption and the CL. Thus, variations in the CL of buildings should be considered when optimizing the energy consumption of the CP system. Uncertainty modeling of the CL is necessary. However, existing uncertainty modeling approaches do not fully consider the statistical information of historical CL data. Therefore, a novel support vector clustering(SVC)-based data-driven robust optimization method for the CP system is proposed in the work, which comprehensively takes into account the CL uncertainty of buildings and optimizes the operating status of chillers and pumps to minimize energy consumption. This method constructs the uncertainty set using historical CL data, ensuring that the worst-case scenario corresponds to one of the historical data points, thereby making the simulation results more realistic. Using the actual CP system as case studies, the outcomes showed that the created optimization method can effectively eliminate uncertainty perturbation for this system and provide the optimal operating status of the chillers and pumps under CL uncertainty. Additionally, the impacts of changes in the control parameter of the robust on the optimal operating status of chillers and pumps are examined. Finally, the proposed method was compared with the box and polyhedral uncertainty set model. The outcomes of comparisons demonstrate the proposed approach’s suitability in an uncertain setting.

Zitieren

APA

Cai, J., & Liu, J. (2026). A novel support vector clustering-based data-driven robust optimization method for minimum energy consumption of the chiller-pump system. Thermal Science, 190-190. https://doi.org/10.2298/tsci260509190c

Vancouver

Cai J, Liu J. A novel support vector clustering-based data-driven robust optimization method for minimum energy consumption of the chiller-pump system. Thermal Science. 2026:190-190. doi:10.2298/tsci260509190c

BibTeX
@article{Cai2026,
  author = {Cai, Jianyang and Liu, Jianjun},
  title = {A novel support vector clustering-based data-driven robust optimization method for minimum energy consumption of the chiller-pump system},
  journal = {Thermal Science},
  year = {2026},
  number = {00},
  pages = {190--190},
  doi = {10.2298/tsci260509190c}
}
Kostenlos starten →