محمدجواد ترابي

In this thesis, first, the nonlinear behavior of the reactive distillation tower is described and a neural network model with input, hidden and output layers is developed to predict certain variables for four-component feeds in the case study of normal propyl propionate, where the necessary training data set is obtained from multiple simulations using Aspen Plus software. The eva‎luation and selection of variables are carried out in such a way that the maximum possible process data is obtained in the post-processing stage. After examining the performance of the developed neural network, it will be used to examine the accuracy of predicting variables in each reactive distillation tower for the production of this material. Then, the use of the final model for the aforementioned simulation is examined and the results obtained with the help of the neural network will be compared with the results obtained from the exact simulation. Based on the comparison of the obtained results, it can be acknowledged that the neural network as a simulator shows very good performance in terms of the accuracy of predicting output variables. The MSE value obtained for each output variable varies from 3-10 to 4-10, which indicates the high accuracy of the developed networks. The study and interpretation of the results obtained from the application of the final neural network in simulation problems indicate the superiority of the neural network over the link mode of the two software Aspen Plus and MATLAB in terms of computational time. According to the results obtained, the neural network is able to reduce the number of times required to correctly solve the continuity, mass transfer and heat transfer equations in a complex simulation from 6912 to 1 time. Also, the accuracy of the neural network in performing calculations is satisfactory due to its nature, but the development of a more accurate model in these problems is still needed.

فرآيند تقطير واكنشي , شبكه‌هاي عصبي مصنوعي , مدل‌سازي و شبيه‌سازي

Reactive Distillation Process , Artificial neural networks , modeling and simulation

Mohammad Javad Torabi

Dr. Norollah Kasiri