Application of artificial neural networks (ANN) for vapor‐liquid‐solid equilibrium prediction for CH₄‐CO₂ binary mixture

Abstract

The study of the frosting behavior of CO2 in the binary CH4‐CO2 is very important for energy minimization and for the smooth operation of the cryogenic purification process for natural gas due to its extensive cooling requirements. The present study focuses on the solid region of the phase envelope and the development of a predictive model using the artificial neural network (ANN) technique. It validates the model using available experimental data. The model points out the outlying data points. The ANN prediction method developed in this work can be successfully used for the vapor‐solid (V‐S) and vapor‐liquid‐solid (V‐L‐S) equilibrium of a CH4‐CO2 binary mixture for CO2 concentration of 1 to 54.2% and a temperature range of −50°C to −200°C. The use of the model for the liquid‐solid (L‐S) region in its current form is not recommended because the model was not validated due to lack of experimental data in this region. © 2018 Society of Chemical Industry and John Wiley & Sons, Ltd.

Comparative study of artificial neural network versus parametric method in COVID-19 data analysis

“…Another hyper parameter that affects the training and learning capabilities of neural networks is the training algorithms [58] . Levenberg-Marquardt training algorithm, which is one of the algorithms with deep learning and training performance, is employed in designed neural network model [59] .…”

Investigation of flow parameters of Reiner–Philippoff nanofluid flow with higher‐order slip properties, activation energy, and bioconvection by artificial neural networks

“…37 In the developed ANN models, the Levenberg-Marquardt training algorithm, which has high performance and is frequently used in MLP networks, is preferred. 38,39 The Tan-Sig transfer functions used in the hidden layer and the Purelin transfer functions used in the output layer are given below. 40,41 After completing the training stages of ANN models, the next step is performance analysis.…”

Artificial Neural Networking (ANN) Model for Drag Coefficient Optimization for Various Obstacles

“…Training algorithm with named Levenberg-Marquardt holds high training and learning performance, is used in the ANN model [52,53]. Pureline and Tan-Sig transfer functions are considered in output and hidden layers.…”

“…Performance parameters were chosen in order to evaluate the training and learning stages [56]. The mathematical relationships [57,58] used in calculating the coefficient of determination (R) parameters and mean squared error (MSE) are reported as Training algorithm with named Levenberg-Marquardt holds high training and learning performance, is used in the ANN model [52,53]. Pureline and Tan-Sig transfer functions are considered in output and hidden layers.…”