Experimental verification of the novel transcritical CO2 heat pump system and model evaluation method

This work proposes a novel method for verifying the accuracy of the numerical simulation model of the compression/ejector transcritical CO2 heat pump system using convolutional neural networks. The method focuses on converting the original unequal input conditions into equal input conditions and com...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Renewable energy 2024-02, Vol.222, p.119936, Article 119936
Hauptverfasser: Qin, Xiang, Shen, Aoqi, Duan, Hongxin, Wang, Guanghui, Chen, Jiaheng, Tang, Songzhen, Wang, Dingbiao
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:This work proposes a novel method for verifying the accuracy of the numerical simulation model of the compression/ejector transcritical CO2 heat pump system using convolutional neural networks. The method focuses on converting the original unequal input conditions into equal input conditions and comparing the simulation results with the experimental prediction results of each component. The validation results reveal the following findings: 1) The root mean square errors of the gas cooler and the water source evaporator are 9.16 °C and 11.47 °C respectively, indicating that future work should focus on correcting of the CO2 heat transfer coefficient calculation method; 2) The verification results of the evaporator indicate that suggesting the need to incorporate an air dynamic change module in the air source input; 3) The root mean square error of the CO2 outlet pressure in the ejector is 462.45 kPa. It can be inferred from the pressure variation trend that the limit pressure ratio of the compressor is the main factor affecting the accuracy of the ejector model. Overall, this article presents a novel and effective method for verifying the accuracy of numerical models.
ISSN:0960-1481
1879-0682
DOI:10.1016/j.renene.2024.119936