Folko Flehmig, Rüdiger von Watzdorf, Wolfgang Marquardt:
Dynamic MSF plant modeling and simulation: non-equilibrium effects and heat transfer
In: Encyclopedia of Desalination and Water Resources (DESWARE), EOLSS Publishers Co. Ltd., Oxford, UK, EOLSS Publishers Co. Ltd., Oxford, UK, 2000
This article addresses two topics related to the thermal efficiency of MSF desalination plants which are non-equilibrium effects and discrepancies between the nominal and the observed heat transfer situations. Non-equilibrium is addressed in the first part of this article focusing on simulation models where the tradeoff between model complexity and prediction accuracy is of major relevance. Based on measured data from a 20 stage industrial size MSF desalination plant it is shown that accounting for non-equilibrium in a simulation model is insignificant. This is due to the fact that uncertainties prevailing in the measurements as obtained from common industrial plant instrumentation are in the same order of magnitude as one expects the non-equilibrium temperature losses. The insignificance of non-equilibrium is also illustrated by presenting dynamic simulation results obtained from an equilibrium model in comparison to those obtained from a model including a state-of-the art correlation to predict non-equilibrium temperature losses. These findings need to be clearly distinguished from the design problem where the relation between non-equilibrium losses and stage design parameters is of key interest. Modeling of the heat transfer is addressed in the second part of this article. The heat transfer situation as observed during plant operation deviates from the nominal situation due to scaling on the tube side and due to non-condensible gases on the shell side. Both effects are not accounted for in the model since validated formulas to calculate scale formation or the release of non-condensible gases are lacking. It is shown that due to the non-condensible gases a compensation for non-idealities in the heat transfer must depend on the plant operating conditions if one expects predictions of reasonable accuracy from the simulation model. Therefore, the typical approach for compensation by means of constant fouling factors cannot be applied.