Predicting Performance of Expansion Turbines Using Different Working Fluids Based on the Artificial Neural Network

In order to obtain the efficiency curve of a helium Expansion Turbine (ET) in the factory, a method which can be used to transform the performance of an ET with one kind of Working Fluid (WF) to another must be investigated because of the lack of the helium in the factory. A performance prediction program based on an one-dimensional analysis of ETs has been developed and has proved valid. On the basis of the program, the Artificial Neural Network(ANN) (Back-Propagation Algorithm) has been used to deal with the transformation problem. The method has proved effective by the efficiency experiments of an ET using the CO2 and the air as the WF respectively.