神经网络模型在制冷剂物性参数计算中的应用

介绍采用BP，RBF和Elman神经网络计算制冷剂物性参数的方法。以R11，R134a和近共沸混合制冷剂R410A为研究对象，分别建立三种制冷剂的BP，RBF和Elman网络饱和物性参数计算模型。根据该模型由已知温度求各制冷剂在饱和气和饱和液状态下的其他物性参数值，通过与REFPROP软件计算结果进行对比，证明BP，RBF和Elman神经网络物性计算模型具有很高的精度，可以用于物性参数的计算，是一种新的物性计算方法。

Application of neural network model to refrigerant thermodynamic and transport properties calculation

The neural network models including a back-propagation neural network model, a Radial Basis Function neural network model and an Elamn neural network model have been set up to calculate the refrigerant thermodynamic and transport properties. And based on the model, the thermodynamic and transport properties of refrigerant Rll, R134a and R410A are calculated as an example. Compared with results calculated by the software REFPROP, satisfied results have beet got. The results prove that the neural network can be used to calculate the refrigerant thermodynamic and transport properties, which is a new method in the refrigerant thermodynamic and transport properties calculation field.