Modeling and Optimization of Crude Oil Fouling Based on Artificial Neural Networks and Genetic Algorithms
Systematic methods and tools for managing the complexity
Advances in Computational & Numerical Methods (T4-4P)
Keywords: Crude oil Fouling, Modeling, Optimization, Neural networks, Genetic algorithms
Javad Aminian1 , Shahrokh Shahhosseini1, , Mehdi Azarmi2, Mazaher Molaei1
1-Department of Chemical Engineering, Iran University of Science and Technology, Narmak, Tehran, Iran P.O. Box 16765-163
2-Department of Chemical Engineering, School of Engineering, Shiraz University, Shiraz, Iran
Abstract
For industrial preheat train, the feedstock variability and presence of different species and operating conditions increase the fouling influence on heat exchanger performance. Prediction of crude oil fouling behavior in industrial preheat exchangers have been extensively studied by various researchers. However, a general, effective and robust method has yet been demanded. This paper illustrates a method based on combination of artificial neural networks (ANN) and genetic algorithms (GA). In this approach, the optimal structure of ANN is constructed using a GA code. The ANN model was then applied to a set of experimental data for predicting crude oil fouling behavior in an industrial shell and tube heat exchanger. The overall mean relative error of ANN was 94.1% lower than the model presented by Panchal. Another GA optimization algorithm was developed, to minimize fouling formation. The results indicate significant improvement in reduction of fouling formation at 5 constant tube diameters.
Presented Tuesday 18, 13:30 to 15:00, in session Advances in Computational & Numerical Methods (T4-4P).
