化工学报 ›› 2025, Vol. 76 ›› Issue (11): 5828-5841.DOI: 10.11949/0438-1157.20250592

• 热力学 • 上一篇    

低GWP制冷剂与润滑油混合物的相平衡关联与预测

王慧荣1(), 孙玲玲1, 戚文端2, 胡熠暹2, 邵艳坡2, 杨智3, 赵延兴4()   

  1. 1.河南科技大学车辆与交通工程学院,河南 洛阳 471003
    2.广东美的制冷设备有限公司,广东 佛山 528311
    3.广东工业大学材料与能源学院,广东 广州 510006
    4.中国科学院理化技术研究所,北京 100190
  • 收稿日期:2025-06-03 修回日期:2025-08-04 出版日期:2025-11-25 发布日期:2025-12-19
  • 通讯作者: 赵延兴
  • 作者简介:王慧荣(1991—),女,博士,讲师,hrwang@haust.edu.cn
  • 基金资助:
    国家自然科学基金项目(52322602);国家自然科学基金项目(52036010)

Correlation and prediction of phase equilibria for mixtures of low-GWP refrigerants with lubricants

Huirong WANG1(), Lingling SUN1, Wenduan QI2, Yixian HU2, Yanpo SHAO2, Zhi YANG3, Yanxing ZHAO4()   

  1. 1.College of Vehicle and Traffic Engineering, Henan University of Science and Technology, Luoyang 471003, Henan, China
    2.Guangdong Midea Refrigeration Equipment Co. , Ltd. , Foshan 528311, Guangdong, China
    3.College of Materials and Energy, Guangdong University of Technology, Guangzhou 510006, Guangdong, China
    4.Technical Institute of Physics and Chemistry, Chinese Academy of Sciences, Beijing 100190, China
  • Received:2025-06-03 Revised:2025-08-04 Online:2025-11-25 Published:2025-12-19
  • Contact: Yanxing ZHAO

摘要:

制冷剂与润滑油所形成的混合物存在于制冷系统各部件中,其相平衡特性对热力循环及部件的设计优化具有重要意义。采用基于统计缔合流体理论的PC-SAFT状态方程,对R1234yf、R600a、R744等7种低GWP制冷剂与润滑油PEC4~PEC9的二元混合物气液相平衡进行计算,结果显示PC-SAFT模型与实验结果一致性较好,计算的平均压力偏差绝大部分在6%以内。同时,为进一步拓展模型对复杂多元混合物相平衡的预测性能,采用BP神经网络模型对PC-SAFT状态方程的二元交互参数进行训练,并将预测得到的二元交互参数用于缺乏实验数据的二元体系R1233zd(E)/PEC润滑油、多组分的复杂体系R744/POE润滑油和R290/POE润滑油的相平衡计算。采用神经网络模型预测得到的kij 对多组分复杂体系进行预测对比,平均压力偏差为8.99%,可用于实际工程计算。

关键词: 制冷剂, 润滑油, 相平衡, 混合物, PC-SAFT, BP神经网络

Abstract:

The mixture formed by refrigerants and lubricating oils is present in various components of the refrigeration system, and its phase equilibrium characteristics are of significant importance for the design optimization of thermodynamic cycles and components. Using the PC-SAFT state equation based on statistical associating fluid theory, the gas-liquid phase equilibrium of binary mixtures of seven low-GWP refrigerants such as R1234yf, R600a, R744, etc., with lubricating oils PEC4—PEC9 was calculated. The results show that the PC-SAFT model has good consistency with experimental results, and the average pressure deviation is mostly within 6%. Meanwhile, to further expand the model's predictive performance for complex multicomponent mixture phase equilibrium, a BP neural network model was used to train the binary interaction parameters of the PC-SAFT state equation, and the predicted binary interaction parameters were used for the phase equilibrium calculation of binary systems lacking experimental data, such as R1233zd(E)/PEC lubricating oil, multicomponent complex systems R744/POE lubricating oil, and R290/POE lubricating oil. The kij predicted by the neural network model was used to predict these complex multicomponent systems, with an average pressure deviation of 8.99%, indicating its applicability to practical engineering calculations.

Key words: refrigerant, lubricant oil, phase equilibria, mixtures, PC-SAFT, BP neural network

中图分类号: