Neural network based correlations for estimating temperature elevation for seawater in MSF desalination process

Abstract

Modelling played an important role in simulation, optimisation, and control of multi-stage flash (MSF) desalination processes. Top brine temperature (TBT) is one of the many important parameters that affect optimal design and operation of MSF processes. Within the MSF process model, calculation of TBT is therefore important. For a given pressure, TBT is a function of boiling point temperature (BPT) at zero salinity and temperature elevation (TE) due to salinity. In this work, we develop several neural network (NN) based correlations for predicting TE. It is found that the NN based correlations can predict the experimental TE very closely. Also predictions by the NN based correlations were good when TE values, obtained using existing correlations from the literature are compared. Due to advancement of the microcomputer, plant automation becomes reliable means of plant maintenance. NN based correlations (models) can be updated in terms of new sets of weights and biases for the same architecture or for a new architecture reliably with new plant data.