Estimation of vapor pressures, compressed liquid, and supercritical densities for sulfur dioxide using artificial neural networks

Abstract

Homo-and copolymers of 2-(N-phthalimido)ethyl methacrylate (NPEMA) and p-chlorophenyl methacrylate (PCPMA) were prepared in N,N-dimethyl formamide (DMF) solution at 70 °C using 2,2-azo-bisisobutyronitrile (AIBN) as initiator. The nano-CdS-doped polymer composite of NPEMA and PCPMA was prepared via in situ technique. The homo-and copolymers of NPEMA and PCPMA were characterized using FT-IR spectroscopy and gel permeation chromatography (GPC). The polymer nano composites were characterized using FT-IR spectroscopy, X-ray diffraction, and transmission electron microscopy. The reactivity ratios (r 1 and r 2 ) were obtained from the various linear graphical methods. The values of r 1 (NPEMA) = 0.55 and r 2 (PCPMA) = 1.30 were found from the same graphical methods. The copolymer microstructures were found from the mean sequence length, run number, and dyad fraction. Thermal behavior of polymers and polymer nano composites under nitrogen atmosphere was studied. The activation energies of neat polymers were varied in the range of 56-85 kJ/mol, while 28-56 kJ/mol energies were found for nano-CdS-doped polymer composites. The thermodynamic parameters of thermal degradation were also obtained. Kinetic and thermodynamic parameters were confirming the stability of the neat polymers than polymer nano composites. The polymers were assessed on different microorganisms for obtaining the antimicrobial properties. Overall, the polymers permit 10-52, 20-58, and 18-56% growth of bacteria, fungi, and yeast, respectively.

Predicting the Coefficients of Antoine Equation Using the Artificial Neural Network (TECHNICAL NOTE)

“…The work showed that this method could estimate the density of hydrocarbons with good accuracy. In addition, Moghadassi et al [8] presented a model by a neural network that could estimate the sulfur dioxide density. The best network arrangement for this case was the Levenberg-Marquardt training algorithm with 15-10-1 topology and had good performance.…”