Artificial neural networks as a predictive tool for vapor-liquid equilibrium

doi:10.1016/0098-1354(94)80011-1

Computers & Chemical Engineering

Volume 18, Supplement 1, 1994, Pages S63-S67

https://doi.org/10.1016/0098-1354(94)80011-1 Get rights and content

Abstract

A new type of group-contribution method for calculation of liquid phase activity coefficients is presented. The method is implemented by using an artificial neural network. Calculated results are compared with the UNIFAC method and experimental data.

References (12)

K.V. Kurmanadharao et al.
Binary vapour-liquid equilibria of the systems methyl ethyl ketone — cyclohexane and acetone - cyclohexane
Chem.Eng. Sci.
(1957)
T. Boublik et al.
Vapor-liquid equilibria of the cyclohexane — cyclohexanone system at 323.15 K and 348.15 K
J. Chem. Eng. Data.
(1977)
Aa. Fredenslund et al.
G. Geiseler et al.
Thermodynamisches verhalten der mischsysteme methyläthylketoxim /n-heptan, diäthylketon / n-heptan und methyläthylketoxim / diäthylketon
Berg. Bunsenges. Phys. Chem.
(1968)
D.O. Hanson et al.
Alteration of the relative volatiliy of n-hexane - 1-hexene by oxygenated and chlorinated solvents
J. Chem. Eng. Data.
(1967)
R. Hecht-Nielsen

There are more references available in the full text version of this article.

Cited by (50)

Thermodynamically consistent vapor-liquid equilibrium modelling with artificial neural networks
2023, Fluid Phase Equilibria
Citation Excerpt :
The black-box denomination is because the model is not transparent and implies that it is only based on the data and not on a priori knowledge [24] (on the contrary, mechanistic models are known as white-boxes). The first formal attempt in the literature at modelling VLE with ANNs was done by Petersen et al. in 1993 [25]. Back then, they concluded that the UNIFAC model estimations were better than the ANN because the mechanistic model had more knowledge embedded into it.
An integration of Artificial Neural Networks (ANNs) and thermodynamics through the application of Neural Network Programming (NNP) is proposed. Thermodynamic consistency is achieved because the thermodynamic relationships and constraints are transcribed into a specially crafted ANN. Moreover, the developed models allow predicting and extrapolating the model outside the experimental data boundaries.
The Wilson and NRTL models are used as case studies. Modifications to these models based on sigmoid functions are rigorously assessed in order to perform the simultaneous modelling of VLE and excess enthalpy. The automatic differentiation together with the ANN optimization algorithms can find sets of parameters that are better than the ones obtained with traditional gradient-based optimizers.
The frequently disregarded concepts of thermodynamic modelling with ANNs are discussed in-depth. A mathematical analysis of the impossibility of typical fully connected ANNs to formulate thermodynamically consistent equilibrium models is discussed and their use is discouraged (e.g., VLE, LLE, or adsorption.).
A generalized machine learning-assisted phase-equilibrium calculation model for shale reservoirs
2022, Fluid Phase Equilibria
Citation Excerpt :
In recent years, machine learning (ML) techniques have been utilized in petroleum engineering to facilitate production forecasting [40–43], parameter optimization [44–47], compositional simulation [39,48–50], and image recognition [51–53]. Specifically, many researchers have applied ML techniques in the stabilization and acceleration of vapor-liquid equilibria (VLE) calculations [54–57]. Gaganis and Varotsis [39,58] used ML to speed up the compositional reservoir simulation.
In compositional reservoir simulation, a significant portion of the CPU time is consumed in phase equilibrium calculations. Previous studies have incorporated the machine learning (ML) technique to accelerate and stabilize the phase equilibrium calculations. However, there are two main limitations: 1) previous work mainly focuses on conventional reservoirs, which cannot be extended to unconventional reservoirs; 2) previous studies are limited to fluid compositions with specific hydrocarbon components that narrows their application. In this paper, we propose a novel ML-assisted framework for phase equilibrium calculations in shale reservoirs. A general set of pseudo-components is considered to allow users to customize the composition of hydrocarbon mixtures. A pore size-dependent EOS is applied to simulate the fluid phase behavior in nano-scale conditions. In the stability test, the multilayer perceptron (MLP) is trained to predict the fluid phase state: single-phase or two-phase. For the fluid labeled as two-phase condition, the phase-split computation is performed to obtain the equilibrium ratio. Instead of using the initial estimate from the stability test, the MLP and the physics-informed neural network (PINN) are applied to obtain the initial estimates for the minimization program. The results show that, with the assistance of ML technique, we are able to reduce the computation time needed for the nano-scale phase equilibrium calculations by more than two orders of magnitude while maintaining 97 $%$ accuracy. Compared with MLP, PINN can accurately predict the equilibrium ratios with a limited range of input variables but require more training time. The progress of this study present a ML-assisted framework for phase equilibrium calculations and the generalized proxy phase-equilibrium calculator can be compiled into reservoir simulator to accelerate flash calculation.
Vapor–liquid equilibrium estimation of n-alkane/nitrogen mixtures using neural networks
2022, Journal of Computational and Applied Mathematics
Understanding fluid phase behavior, like VLE, in high P&T conditions is crucial for developing high-fidelity simulations of chemically reacting flows in liquid-fueled combustion systems and also forms an integral part of the design-modeling of the control processes in chemical industries. Two data-driven models have been proposed in this study, each of which was competent in estimating VLE for the Type III binary systems of $C_{10} / N_{2}$ and $C_{12} / N_{2}$ , at pressures ranging up to 50–60 MPa. Both models showed better performance in predicting equilibrium pressure as compared to VLE modeled using PR-EOS. A modified model has also been proposed, capable of estimating the full phase envelope for the binary systems of $C_{10} / N_{2}$ and $C_{12} / N_{2}$ across a wide range of temperatures, and thus exhibit the mixture critical pressure at the concerned temperature. The diverse applicability of the proposed network architecture was further exhibited while estimating the VLE of a ternary system of $C_{1} / C_{10} / N_{2}$ .
A note on artificial neural network modeling of vapor-liquid equilibrium in multicomponent mixtures
2019, Fluid Phase Equilibria
Application of artificial neural networks (ANNs) for modeling of vapor-liquid equilibrium in multicomponent mixtures is considered. Two novel ANN-based models are introduced, which can be seen as generalizations of the Wilson model and the NRTL model. A unique feature of the proposed approach is that an ANN approximation for the molar excess Gibbs energy generates approximations for the activity coefficients. A special case of the ternary acetic acid–n-propyl alcohol–water system (at 313.15 K) is used to illustrate the efficiency of the different models, including Wilson's model, Focke's model, and the introduced generalized degree-1 homogeneous neural network model. Also, the latter one-level NN model is compared to the Wilson model on 10 binary systems. The efficiency of the two-level NN model is assessed by a comparison with the NRTL model.
Computing thermodynamic properties of ammonia–water mixtures using artificial neural networks
2019, International Journal of Refrigeration
Artificial neural networks (ANN) provide a computationally efficient pathway for solving complex non-linear problems. Cascaded ANN are used to compute thermodynamic properties of the ammonia–water mixture working-pair commonly used in vapor absorption heat pumps. Thermodynamic property routines can affect the accuracy and pose a significant computational bottleneck for steady-state and transient cycle simulations of ammonia–water. The properties computed using the proposed method agree within 0.5% of equation of state data over a wide range of operating parameters. It is observed that the ANN based property routines, developed as explicit functions, offer ∼60% decrease in computational time over currently used property routines. As a case study, the property calculation modules developed using ANN are employed in simulating the dynamic response of a representative ammonia-water condenser for a vapor absorption cycle. The models predict the transient behavior accurately with an ∼80% computational speedup compared to conventional property routines.
Prediction of gas hydrate formation using empirical equations and data-driven models
2019, Materials Today: Proceedings
Gas hydrates are formed in various natural gas transmission facilities and gas processing equipment in oil and gas fields, refineries, petrochemical plants and chemical industries when water and natural gas combine at sufficiently low temperatures and high pressures condition. Estimation of the temperature and pressure at which hydrates are formed is the first measure that is usually taken to prevent hydrate formation. In this paper, two data-driven models namely artificial neural network (ANN) and Adaptive Neuro Fuzzy Interference System (ANFIS model), were developed as an alternative instruments that draw on empirical data to estimate hydrate formation pressure for various gas systems. To this end, statistical parameters were used to determine the optimum structure of each data-driven model applied on these systems. The results obtained from neural network ANN and ANFIS models were compared to the experimental equations According to the performance criteria, the ANFIS model outperforms the ANN model in all cases. The results also showed that the ANFIS model was in more agreement with experimental data than empirical equations.

View all citing articles on Scopus

View full text

Artificial neural networks as a predictive tool for vapor-liquid equilibrium

Abstract

Chem.Eng. Sci.

Vapor-liquid equilibria of the cyclohexane — cyclohexanone system at 323.15 K and 348.15 K

J. Chem. Eng. Data.

Thermodynamisches verhalten der mischsysteme methyläthylketoxim /n-heptan, diäthylketon / n-heptan und methyläthylketoxim / diäthylketon

Berg. Bunsenges. Phys. Chem.

Alteration of the relative volatiliy of n-hexane - 1-hexene by oxygenated and chlorinated solvents

J. Chem. Eng. Data.