In this work, 2099 experimental data of binary systems composed of CO₂ and ionic liquids are studied to predict solubility using a multilayer perceptron. The dataset includes 33 different types of ionic liquids over a wide range of temperatures, pressures, and solubilities. The main objective of this work is to propose a procedure for the prediction of CO₂ solubility in ionic liquids by establishing four stages to determine the model parameters: (1) selection of the learning algorithm, (2) optimization of the first hidden layer, (3) optimization of the second hidden layer, and (4) selection of the input combination. In this study, a bound is set on the number of model parameters: the number of model parameters must be less than the amount of predicted data. Eight different learning algorithms with (4,m,n,1)-type hidden two-layer architectures (m = 2, 4, …, 10 and n = 2, 3, …, 10) are studied, and the artificial neural network is trained with three input combinations with three combinations of thermodynamic variables such as temperature (T), pressure (P), critical temperature (Tc), critical pressure, the critical compressibility factor (Zc), and the acentric factor (ω). The results show that the 4-6-8-1 architecture with the input combination T-P-Tc-Pc and the Levenberg–Marquard learning algorithm is a very acceptable and simple model (95 parameters) with the best prediction and a maximum absolute deviation close to 10%.

In this work, four hundred and forty experimental solubility data points of 14 systems composed of methane and ionic liquids are considered to train a multilayer perceptron model. The main objective is to propose a simple procedure for the prediction of methane solubility in ionic liquids. Eight machine learning algorithms are tested to determine the appropriate model, and architectures composed of one input layer, two hidden layers, and one output layer are analyzed. The input variables of an artificial neural network are the experimental temperature (T) and pressure (P), the critical properties of temperature (Tc) and pressure (Pc), and the acentric (ω) and compressibility (Zc) factors. The findings show that a (4,4,4,1) architecture with the combination of T-P-Tc-Pc variables results in a simple 45-parameter model with an absolute prediction deviation of less than 12%.

<![CDATA[Se presenta un Vétodo de Contribución de Grupos (VCG) para la estimación de propiedades de sustancias y se plantea como un método básico que podría ser incluido en cursos avanzados de termodinámica, fisicoquímica y física de fluidos. La idea es mostrar a los alumnos que las propiedades físicas y fisicoquímicas de las sustancias están relacionadas con su estructura química y características físicas y químicas de los átomos y grupos químicos que forman una molécula. Por lo tanto el método considera que una molécula está formada por grupos definidos a los que se les asigna un determinado valor como contribución al valor de una propiedad. Luego se supone que dicha contribución es la misma en todo compuesto donde dicho grupo esté presente. Este trabajo presenta algunos ejemplos que pueden ser mostrados en un curso formal y propone temas de tareas cortas y semestrales. Las tareas involucran conceptos generales asociados a la termodinámica, conocimiento sobre planillas de cálculo y manejo de bases de datos. Se discute sobre la utilidad del VCG como una herramienta en la metodología de aprendizaje basado en problemas. Se concluye que este tipo de conceptos y métodos se ajustan bien a cursos avanzados de pregrado o en tópicos y cursos de postgrado, en disciplinas como física, química, ingeniería química y otras afines.]]>

The solubilities of 498 datasets of N₂O and ionic liquid systems were predicted using a multilayer perceptron. The data used to train the artificial neural network was subjected to the Gibbs–Duhem test to analyze their thermodynamic consistency. The Peng–Robinson cubic equation of state, combined with the Kwak–Mansoori mixing rule, was used as the thermodynamic model to implement the test. The analysis indicated that 71.9% of the data were declared thermodynamically inconsistent. The ability of artificial neural networks (ANNs) to predict the solubility of these systems using experimental datasets that do not satisfy the thermodynamic consistency criteria based on the Gibbs–Duhem equation was studied. The multilayer perceptron model achieved an average absolute deviation of 1.81% and a maximum individual deviation of 7.56%. These results highlight the potential of ANNs as robust predictive tools even when the available data do not fully satisfy thermodynamic consistency criteria.