Abstract

An artificial neural network for prediction of thermodynamic properties; Case study: saturated and superheated water

Author(s): Abdolreza Moghadassi, Fahime Parvizian, Sayed Mohsen Hosseini, Seyyed Jelaladdin Hashemi

Water is an important natural fluid that plays significant roles in many processes. Consequently, knowledge of the thermodynamic properties of water is necessary for the interpretation of physical and chemical processes. In this work a new method based on artificial neural network (ANN) for prediction of water thermodynamic properties such as specific volume, entropy and enthalpy for both superheated and saturated regions has been proposed. The needed data is taken from steam tables[PerryÂ’s Chemical Engineering Handbook]. The accuracy and trend stability of the trained networks, were tested against unseen data their. Different training schemes for the back-propagation learning algorithm, such as; scaled conjugate gradient (SCG), Levenberg-Marquardt (LM), gradient descentwithmomentum(GDM), variable learning rate back propagation (GDA) and resilient back propagation (RP) methods were used. The SCG algorithm with seven neurons in the hidden layer shows to be the best suitable algorithmwith theminimummean square error (MSE) of 0.0001517. TheANNÂ’s capability to predict thewater thermodynamic properties is one of the best estimating method with high performance.