论文部分内容阅读
室内空气的温度和湿度都是影响居住者热舒适性和空调运行效率的重要因素,研究同时控制室内空气的温湿度意义重大。对于直膨式空调这样一个多变量、强耦合和非线性的系统而言,引入湿度将会显著增加系统的建模难度。单一的物理建模或者经验建模都无法同时满足精度和灵敏度的要求。本文提出了混合建模的方法,对与室内空气直接传热传质的蒸发器进行物理建模,对除蒸发器之外的系统其他部件利用神经网络(ANN)进行数学建模,模拟在变工况下空调系统产生的显热冷量和潜热冷量。结果表明,相比于单一的神经网络模型,混合模型在变工况下的模拟结果具有很好的稳定性。
Indoor air temperature and humidity are affecting the thermal comfort of occupants and air-conditioning operating efficiency of an important factor in the study while controlling indoor air temperature and humidity of great significance. For a multivariable, strongly coupled, and nonlinear system such as direct expansion air conditioning, the introduction of humidity will significantly increase the modeling difficulty of the system. A single physical modeling or empirical modeling can not both meet the requirements of accuracy and sensitivity. In this paper, a hybrid modeling method is proposed to physically model the evaporator with direct heat and mass transfer of indoor air. The other parts of the system except the evaporator are mathematically modeled by neural network (ANN) Working conditions under the air-conditioning system, the sensible heat capacity and latent heat capacity. The results show that compared with a single neural network model, the hybrid model has good stability under varying conditions.