摘要
本文研究了数据中心空调系统的制冷剂泄漏故障。在虚拟制冷剂充注量传感器模型的基础上,结合数据中心空调系统液体管路长的特点,提出了一种增加液管压降特征指标的改进型充注量估计灰箱模型。结合构建的神经网络模型,建立了一种基于混合模型的故障诊断方法。结果表明,混合模型在60%制冷剂充注量情况下仍能将预测误差控制在5%以内,极大地改善了模型在大故障情况下的计算精度。混合模型整体计算偏差大幅度降低,对不同充注量预测的平均误差为2.73%。
The refrigerant leakage problem existed in air conditioning system of data center is studied in this paper.Considering the long liquid line in the targeted system,an improved virtual refrigerant charge grey box model is proposed by adding the index of liquid line pressure drop to the original virtual refrigerant charge sensor based on the model of virtual refrigerant charge sensor.Furthermore,a hybrid model is developed by combining neural network and gray box model.The results indicate that,the hybrid model has an error of 5%in the 60%refrigerant charge level,and significantly improves the performance of the original model under large fault intensity.The proposed hybrid model is much more accurate,and its mean error under various charge levels is 2.73%.
作者
陈志杰
朱旭
黄小清
杜志敏
CHEN Zhijie;ZHU Xu;HUANG Xiaoqing;DU Zhimin(Institute of Refrigeration and Cryogenics,Shanghai Jiao Tong University,Shanghai 200240,China)
出处
《制冷技术》
2019年第6期9-14,共6页
Chinese Journal of Refrigeration Technology
基金
国家自然科学基金（No.51376125）
关键词
虚拟制冷剂充注量传感器
神经网络
数据中心
故障诊断
Virtual refrigerant charge sensor
Neural network
Data center
Fault detection and diagnosis
