A neural network approach and thermodynamic model of waste energy recovery in a heat transformer in a water purification process

A theoretical comparison of neural network (NnM) and thermodynamic (ThM) models is carried out to estimate on-line the coefficient of performance (COP) in an absorption heat transformer integrated with a water purification process (AHT–WP). The NnM has been computed for16 variables measured by sensors (input and output temperatures for each of the four components absorber, generator, evaporator and condenser; input pressure parameters and LiBr + H2O concentrations). The ThM estimates the COP values with average temperatures of each component of the AHT–WP. This ThM has been designed for steady-state conditions while the NnM has been developed for steady- and unsteady-state conditions. Both models can be used to calculate the COP values on-line; nevertheless, each model has its own advantages and disadvantages in the AHT–WP design and control. The waste heat simulated for the experimental water purification system is lower than 1 kW, while temperature and lithium bromide concentration are invariable in huge systems.