Optimized Artificial Neural Network for Evaluation: C4 Alkylation Process Catalyzed by Concentrated Sulfuric Acid

Abstract

In this work, an artificial neural network was first achieved and optimized for evaluating product distribution and studying the octane number of the sulfuric acid-catalyzed C4 alkylation process in the stirred tank and rotating packed bed. The feedstock compositions, operating conditions, and reactor types were considered as input parameters into the artificial neural network model. Algorithm, transfer function, and framework were investigated to select the optimal artificial neural network model. The optimal artificial neural network model was confirmed as a network topology of 10-20-30-5 with Bayesian Regularization backpropagation and tan-sigmoid transfer function. Research octane number and product distribution were specified as output parameters. The artificial neural network model was examined, and 5.8 × 10 –4 training mean square error, 8.66 × 10 –3 testing mean square error, and ±22% deviation were obtained. The correlation coefficient was 0.9997, and the standard deviation of error was 0.5592. Parameter analysis of the artificial neural network model was employed to investigate the influence of operating conditions on the research octane number and product distribution. It displays a bright prospect for evaluating complex systems with an artificial neural network model in different reactors.

A Pilot-Scale HiGee-Aided Fixed Bed Reactor: Size Characteristics of Microbubbles in Diesel

“…As a typical sort of deep learning, ANN was popularly and widely employed in chemical engineering and industry. − Unlike empirical correlation, ANN was a black box that did not require equations as priors, which meant that the essential mechanism of process or phenomenon did not need to be considered when ANN was applied . Especially for some complex problems with multi-input, multi-output, and strong nonlinear, the advantage of ANN was quite obvious . There were three main factors to be considered in the process of designing and training an ANN, namely, the architecture of ANN, the transfer function, and the training algorithm.…”

“…37 Especially for some complex problems with multi-input, multi-output, and strong nonlinear, the advantage of ANN was quite obvious. 38 There were three main factors to be considered in the process of designing and training an ANN, namely, the architecture of ANN, the transfer function, and the training algorithm. The architecture of ANN mainly focused on the number of layers of hidden layer and the number of neurons in each layer.…”

Predicting the Properties of High-Performance Epoxy Resin by Machine Learning Using Molecular Dynamics Simulations

“…It makes deductions by identifying patterns and relationships in data [ 23 , 24 ]. A significant advantage of ANNs is their fast speed and accurate processing provided in large parallel implementations through high-level algorithms because of which they have being gaining increasing attention in research [ 25 , 26 , 27 , 28 ]. However, hundreds or thousands of combinations are possible for epoxy resins with multiple components, resulting in an experimental overload when the formulation of the epoxy system becomes complicated.…”