摘要: |
本文对含有PAG润滑油的CO2气体冷却器建立了理论模型,通过人工神经网络对含油的CO2对流换热系数进行预测,通过与实验结果进行比较,验证了模型的准确性。进一步分析和计算结果表明气冷器的热效率随管道的长度的增加而增大,但增幅逐渐放缓。对于实际使用的气冷器,考虑设备运行的经济性,应合理设计气冷器的长度。对于小管径的气冷器 (Dh ≤ 2 mm),少量的润滑油(ω ≈ 1%)即会明显削弱气冷器的传热系数与热效率;但对于大管径的气冷器 (Dh ≥ 4 mm),少量的润滑油对气冷器的影响不明显。在实际小管径气冷器运行过程中,要尽量降低进入气冷器中的润滑油含量,以保证气冷器较高的换热性能。气冷器的热效率随CO2侧压力的增加而增大,但增幅逐渐变缓。CO2侧压力越接近临界压力,由于润滑油的存在导致气冷器热效率降低的幅度越大,并且气冷器的管径越小,降低幅度越明显。 |
关键词: CO2气体冷却器 传热系数 润滑油 神经网络 |
DOI: |
|
基金项目: |
|
Effect of Lubricant Oil on Thermal Performance of CO2 Gas Cooler |
Dai Baomin,Li Minxia,Ma Yitai,Wang Pai,Li Sha |
(Key Laboratory of Efficient Utilization of Low and Medium Grade Energy, MOE, School of Mechanical Engineering, Tianjin University;School of Mechanical Engineering, Tianjin Polytechnic University) |
Abstract: |
A theoretical model of CO2 gas cooler with PAG lubricant was developed. The heat transfer coefficient was predicted based on artificial neural network, and then the model was validated with experimental results. Further calculation shows that the thermal effectiveness increases with the length of the gas cooler tube, but the increase rate decreases gradually. Therefore, the length of the gas cooler tube should be properly designed considering the economical running in the practical applications. For the gas cooler with a smaller diameter (Dh ≤ 2 mm), a small amount of lubricant significantly deteriorates the thermal effectiveness while for that with a larger diameter (Dh ≥ 4 mm), the influence is not so apparent. Consequently, the oil concentration should be decreased as far as possible to keep high heat transfer performance of the small diameter gas cooler. The gas cooler performances better at higher CO2 side pressure. However, the increase rate of the thermal effectiveness decrease gradually as the pressure increases. At pressure approaching to the critical pressure, the degradation of the thermal effectiveness is more significant. Additionally, the deterioration is more apparent as the diameter decreases. |
Key words: CO2 gas cooler heat transfer coefficient lubricant oil artificial neural network |