Application of ANNs in Chemical Engineering

The ANN techniques can be successfully applied in the field of chemical engineering. The examples presented here give a brief overview of the capacity of ANNs to solve some chemical engineering problems. Readers interested in investigating this topic further can refer to Bulsari s book [5.41] or to other authors referenced in the bibliography [5.39, 5.54]. In addition, to complete the information presented here, an important number of Internet sites can be used as well as some of the current chemical engineering scientific publications, which have given important attention to this subject. 

Three major ways can be identified for the use of an ANN in chemical engineering  

as a substitute for the complicated models of transport phenomena or stochastic based models  

as data support, especially for the equilibrium and kinetic data needed by models based on transport phenomena. This kind of model is recognized as a hybrid neural -regression model  

In all the above-mentioned cases, once the learning processes have been completed, ANNs have an assistant function which gives one or many answers to an argument of the complex modeled process (parameters, factors or independent variables). When we use an ANN as a substitute for models for stochastic or complicated transport phenomena, the learning process must be as shown in Fig. 5.14 which shows the coupling of an ANN and a complicated mathematical process. 

---