Intelligent Modeling of Permeate Flux during Membrane Clarification of Pomegranate Juice

One of the problems in juice membrane clarification is the accumulation and deposition of rejected compounds on membrane surfaces or inside its pores which results in a membrane fouling.
Several parameters can have influence on fouling in one hand and prediction of juice permeates flux during the membrane processing is of importance in industrial applications on the other hand. Therefore, providing a model able to predict the permeate flux having the value of effective input parameters seems to be useful. In this regard, several artificial methods can be used. In contrast, the Fuzzy Inference System (FIS) has been proposed as a reliable and appropriate machine learning method to predict the output parameter with acceptable performance. In this study, a FIS will be used to model the permeate flux based on five input variables: the transmitted membrane pressure, feed flow rate, processing time, membrane pore size, and membrane type. For this purpose, a fuzzy system is trained using the laboratory data and then, appropriate membership functions for the input and output variables and fuzzy rules are extracted for the proper prediction of permeate flux.
Results were shown that the normalized mean squared errors for the prediction of permeate flux in the membrane processing are 0.0055 and 0.0081 for the Mixed Cellulose Ester (MCE) and the Polyvinylidene Flouride (PVDF) membranes, respectively.
In total, the fuzzy inference model which is presented in this study has been able to predict the pomegranate juice permeate flux with an acceptable error compared with the laboratory data.