Elsevier

Energy

Volume 217, 15 February 2021, 119292
Energy

Molecular descriptors-based models for estimating net heat of combustion of chemical compounds

https://doi.org/10.1016/j.energy.2020.119292Get rights and content

Highlights

  • The capabilities of CSA-LSSVM, GP, PSO-ANFIS and GA-ANFIS models to estimate ΔHc are investigated.

  • The input parameters of the models are nC, Sv, ATS2m, and siege.

  • Two parameter models based on measureable variables of nC and Sv are developed.

  • 1714 data points were used for model development.

  • The models are found reliable and accurate (especially the forecast of PSO-ANFIS model).

Abstract

The heating values of fuels are determined by Heat of Combustion (ΔHC)in which the higher amount is more lucrative. Moreover, one of the best methods to compare the stabilities of chemical materials is using ΔHC. Therefore, improving precise and general models to estimate this property in different areas such as industries and academic perspective should be considered. In this study, three models namely Least Square Support Vector Machine optimized by Coupled Simulated Annealing optimization algorithm (CSA-LSSVM), Genetic Programming (GP) and Adaptive-Neuro Fuzzy Inference System optimized by PSO, and GA methods (PSO-ANFIS and GA-ANFIS) were applied to estimate ΔHC Also, ΔHC can be expressed by the GP model with an equation. The input parameters of the models are total carbon atoms in a molecule (nC), sum of atomic van der Waals volumes (scaled on carbon atom) (Sv), Broto–Moreau autocorrelation of a topological structure (ATS2m), and total Eigenvalue from electronegativity weighted distance matrix (siege). In addition, two parameter models based on measureable variables of nC and Sv are proposed. In a comprehensive set, 1714 data points were used to fulfill and develop the models. Results demonstrate that the models are trustworthy and accurate (especially the PSO-ANFIS model) in comparison with other recently developed literature models.

Introduction

The heat of combustion (ΔHC) of a substance is the amount of heat, which is released from complete combustion at the oxygen-rich environment and it is normally defined at 298.15 K in 1 atm. Throughout the process, which includes oxidation to final products, the common products are CO2 (g), F2 (g), Cl2 (g), Br2 (g), I2 (g), SO2 (g), N2 (g), H3PO4 (s), H2O (g), and SiO2 (cristobalite) [1]. To produce heat or power, havingΔHC from the fuel is vital to increase the efficiency of instruments. Measuring ΔHC is essential to investigate the potential of fire threat of chemical compounds, differentiate the fuel heating value, materials stability, and engine performance such as emission and efficiency [[2], [3], [4], [5], [6], [7], [8]].

These ΔHC values are gathered in some databases like API-TDB [9] and AIChE-DIPPR [1] for many chemicals. Because of time-consuming, being expensive experiments and even sometimes impossible procedure, determining ΔHC by modeling methods would be faster, accurate, and more accessible. Using reliable relations of the determined ΔHC could anticipate the heating value of other compounds and even un-synthesized materials. It is well known that the compound structure shows its property. Thus, well-established quantitative structure-property relationship (QSPR) technique has been extensively applied to estimate and correlate a variety of complicated and straightforward physicochemical properties of compounds in different areas. Until now, the ΔHC property modeling has been investigated in few reports [[10], [11], [12], [13]] and the models work with combination of linear molecular descriptors as, exploring non-linear effects might surpass their predictive potential.

In the current study, new models were developed to forecast ΔHC of many pure compounds, which belong to different chemical families. Moreover, the non-linear relationship between ΔHC and molecular descriptors have been surveyed. Accordingly, Coupled Simulated Annealing optimization algorithm (CSA-LSSVM), Genetic Programming (GP) and Adaptive-Neuro Fuzzy Inference System optimized by Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) (PSO-ANFIS and GA-ANFIS) methods were employed for modeling. Table 1 shows brief explanations about the models developed in this study.

Section snippets

ANFIS

For the first time, Zade [14] presented the fuzzy logic (FL). The FL and ANN approaches are common in the Adaptive Neuro-Fuzzy Inference System (ANFIS). These two algorithms are coupled by explaining the If-then rules and certain functions called membership functions (MFs) for a construction called Fuzzy Inference (FIS). The modifications of MFs are based on the ability of the ANNs [[15], [16], [17]].

To optimize the ANFIS in the current study, two optimization algorithms were utilized. These

Data acquisition

In this study, data were collected from Refs. [11]: The database of the physical properties of pure chemicals of AIChE (American Institute of Chemical Engineers) is DIPPR 801 that was used in this work. In this database, 1714 ΔHC data for pure compounds were found. Molecular descriptors are introduced as numerical features related to chemical structures. The molecular descriptor is the ultimate outcome of the rational and mathematical method that converts chemically encoded information in the

Model development

Simulated Annealing (SA) [[48], [49], [50]] is a precise algorithm to avoid local optima to expand local search methods. The basic approach leads to solve the lower quality than present solution to simplify the exit from local optimum. CSA, as a modification of SA, intends to far away from the local optima and advances the precision of solutions without reducing the rate of convergence, too much. Suykens et al. [51] displayed the method fundamentals that coupling in local optimization processes

Conclusion

In this work, CSA-LSSVM, GP, GA-ANFIS and PSO-ANFIS models with four and two inputs were developed for the estimation of ΔHc°. The parameters of LSSVM model were optimized to obtain an accurate model with the best performance using a simulated annealing algorithm. The ANFIS model was trained by PSO and GA algorithms to obtain high-accurate results and better performance. A comprehensive database (containing1714 points) of ΔHc° was used for model development. The data in the database was

Credit author statement

Amir Dashti: Conceptualization, Methodology, Software, Writing - original draft, Resources Omid Mazaheri.: Data curation, Formal analysis, Writing - review & editing. Farid Amirkhani: Visualization, Data curation, Investigation, Formal analysis. Amir H. Mohammadi: Supervision, Project administration, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.None.

References (60)

  • M.A. Halali et al.

    Application of a radial basis function neural network to estimate pressure gradient in water–oil pipelines

    J Taiwan Inst Chem Eng

    (2016)
  • A. Khajeh et al.

    Prediction of solubility of gases in polystyrene by adaptive neuro-fuzzy inference system and radial basis function neural network

    Expert Sys APP

    (2010)
  • A. Dashti et al.

    Accurate prediction of solubility of gases within H 2-selective nanocomposite membranes using committee machine intelligent system

    Int J Hydrogen Energy

    (2018)
  • M. Rezakazemi et al.

    Development of hybrid models for prediction of gas permeation through FS/POSS/PDMS nanocomposite membranes

    Int J Hydrogen Energy

    (2018)
  • A. Dashti et al.

    Estimation of biomass higher heating value (HHV) based on the proximate analysis: smart modeling and correlation

    Fuel

    (2019)
  • M. Rezakazemi et al.

    H2-selective mixed matrix membranes modeling using ANFIS, PSO-ANFIS, GA-ANFIS

    Int J Hydrogen Energy

    (2017)
  • A. Dashti et al.

    Rigorous prognostication and modeling of gas adsorption on activated carbon and Zeolite-5A

    J Environ Manag

    (2018)
  • A. Dashti et al.

    Estimation of CO2 adsorption in high capacity metal− organic frameworks: applications to greenhouse gas control

    J CO2 Utilization.

    (2020)
  • A. Eslamimanesh et al.

    Phase equilibrium modeling of clathrate hydrates of methane, carbon dioxide, nitrogen, and hydrogen+ water soluble organic promoters using Support Vector Machine algorithm

    Fluid Phase Equil

    (2012)
  • A. Dashti et al.

    Molecular dynamics, grand canonical Monte Carlo and expert simulations and modeling of water–acetic acid pervaporation using polyvinyl alcohol/tetraethyl orthosilicates membrane

    J Mol Liq

    (2018)
  • A. Dashti et al.

    Estimation of CO2 equilibrium absorption in aqueous solutions of commonly used amines using different computational schemes

    Fuel

    (2020)
  • A. Dashti et al.

    Efficient hybrid modeling of CO2 absorption in aqueous solution of piperazine: applications to energy and environment

    Chem Eng Res Des

    (2019)
  • A. Dashti et al.

    Estimating CH4 and CO2 solubilities in ionic liquids using computational intelligence approaches

    J Mol Liq

    (2018)
  • M. Raji et al.

    Efficient estimation of CO2 solubility in aqueous salt solutions

    J Mol Liq

    (2019)
  • T.M. Lee et al.

    Prediction of membrane fouling in the pilot-scale microfiltration system using genetic programming

    Desalination

    (2009)
  • A. Okhovat et al.

    Modeling of arsenic, chromium and cadmium removal by nanofiltration process using genetic programming

    Appl Soft Comput

    (2012)
  • C. Suh et al.

    Application of genetic programming to develop the model for estimating membrane damage in the membrane integrity test using fluorescent nanoparticle

    Desalination

    (2011)
  • M.M. Atiqullah et al.

    Reliability optimization of communication networks using simulated annealing

    Microelectron Reliab

    (1993)
  • V. Fabian

    Simulated annealing simulated

    Comput Math Appl

    (1997)
  • A. Vasan et al.

    Comparative analysis of simulated annealing, simulated quenching and genetic algorithms for optimal reservoir operation

    Appl Soft Comput

    (2009)
  • Cited by (0)

    View full text