A new estimation algorithm of physical properties based on a group contribution and support vector machine

There are two ways to evaluate the properties of unknown chemical compounds. One is by traditional approaches, which measure the desired data from the experiments and the other is by predicting them in the theoretical approaches using a kind of prediction model. The latter are considered to be more effective because they are less time consuming and cost efficient, and there is less risk in conducting the experiments. Besides, it is inconvenient to conduct experiments to obtain experimental data, especially for new materials or high molecular substances. Several methods using regression model and neural network for predicting the physical properties have been suggested so far. However, the existing methods have many problems in terms of accuracy and applicability. Therefore, an improved method for predicting the properties is needed. A new method for predicting the physical property was proposed to predict 15 physical properties for the chemicals which consist of C, H, N, O, S and Halogens. This method was based on the group contribution method that was oriented from the assumption that each fragment of a molecule contributes a certain amount to the value of its physical property. In order to improve the accuracy of the prediction of the physical properties and the applicability, we extended the database, significantly modifying the existing group contribution methods, and then established a new method for predicting the physical properties using support vector machine (SVM) which is a statistical theory that has never been used for predicting the physical properties. The SVM-based approach can develop nonlinear structure property correlations more accurately and easily in comparison with other conventional approaches. The results from the new estimation method are found to be more reliable, accurate and applicable. The newly proposed method can play a crucial role in the estimation of new compounds in terms of the expense and time.     

估算有机物正常沸点的元素和化学键贡献法

首次提出了一种基于元素和化学键的估算有机物正常沸点的新方法,在4060种有机物实验数据的基础上,回归出了7个正常沸点估算式和估算式的参数值。对4068种有机物的正常沸点进行了估算,其平均相对误差为3.63%,总体估算精度明显优于相比较的基团贡献法,对炔烃、氯衍生物、碘衍生物,尤其是对烷烃类、溴衍生物、芳烃、脂环烃和含硫化合物的估算误差明显减小。     

Effect of pH on the rejection of inorganic salts and organic compound using nanofiltration membrane

Nanofiltration (NF) has recently received increased attention as a possible tertiary treatment process providing high rejection of solutes and high water flux rate. In this research, solute separation experiments using NF membranes were made with inorganic salts including heavy metal and organic compounds in different pH levels. The rejection of inorganics from feed solution was found to be dependent on the electric charge of membrane as well as the ionic radius and valence of ion. The divalent cation appeared to reduce the potential of negatively charged membrane to lower the rejection of ion. The results of organic compounds showed that the rejection could be estimated from the pKa value and molecular weight of organics, and the pH of the feed solution.     

An extension of the group contribution method for estimating thermodynamic and transport properties. Part III. Noble gases

Earlier work on the group contribution method applied to the Kihara potential is extended to noble gases for the estimation of second virial coefficients, dilute gas viscosities and diffusivities with a single set of gas group parameters. Group parameters are determined when second virial coefficient and viscosity data of pure gases are satisfactorily fitted within the experimental uncertainties. Parameters for gas groups (He, Ne, Ar, Kr and Xe) are found to provide good predictions of mixture properties: second virial cross coefficients, mixture viscosities, and binary diffusion coefficients. Application of the model shows that second virial coefficient data are represented with good results comparable to the values by means of the corresponding states correlation. The reliability of the present model in viscosity predictions is proved by comparison with the Lucas method. Prediction results of diffusivity are in excellent agreement with literature data and compare well with values obtained by means of the Fuller method.     

An extension of the group contribution method for estimating thermodynamic and transport properties. Part IV. Noble gas mixtures with polyatomic gases

Earlier work on the group contribution method applied to Kihara potentials is extended to noble-polyatomic gas mixtures for the calculation of second virial cross coefficients, mixture viscosities and binary diffusion coefficients of dilute gas state using a single set of gas group parameters. Previously estimated parameter values for pure gas groups by our work [Oh, 2005; Oh and Sim, 2002; Oh and Park, 2005] were used. Assuming that noble-polyatomic gas mixtures examined are chemically dissimilar, a group binary interaction coefficient, k _{ij, gc} , was assigned to each interaction between noble-polyatomic gas groups, and 25 group binary interaction parameter values (k _{He-H2, gc} , k _{He-N2, gc} , k _{He-CO, gc} , k _{He-CO2, gc} , k _{He-O2, gc} , k _{He-NO, gc} , k _{He-N2O, gc} ; k _{Ne-H2, gc} , k _{Ne-N2, gc} , k _{Ne-CO, gc} , k _{Ne-CO2, gc} , k _{Ne-O2, gc} ; k _{Ar-H2, gc} , k _{Ar-N2, gc} ; k _{Ar-CO, gc} , k _{Ar-CO2, gc} , k _{Kr-O2, gc} ; k _{Kr-H2, gc} , k _{Kr-N2, gc} , k _{Kr-CO, gc} , k _{Ar-CO2, gc} ; k _{Xe-H2, gc} , k _{Xe-N2, gc} , k _{Xe-CO, gc} , k _{Xe-CO2, gc} ) were determined by fitting second virial cross coefficients data. Application of the model shows that second virial cross coefficient data are represented with good results comparable to values predicted by means of the corresponding states correlation. Reliability of the model for mixture viscosity predictions is proved by comparison with the Lucas method. And prediction results of binary diffusion coefficients are in excellent agreement with literature data and compared well with values obtained by means of the Fuller method. Improvements of the group contribution model are observed when group binary interaction coefficients are adopted for mixture property predictions.     

An extension of the group contribution method for estimating thermodynamic and transport properties part II. Polyatomic gases (F2, Cl2, CS2, H2S, NO and N2O)

Earlier work on the group contribution method applied to the Kihara potential is extended to polyatomic gases for the calculation of second virial coefficients, viscosities and diffusivities of dilute gases with a single set of gas group parameters. Functional group parameters are evaluated from the simultaneous regression of second virial coefficient and viscosity data of pure gases. Parameters for gas groups (F₂, Cl₂, CS₂, H₂S, NO, nd N₂O) are found to provide good predictions of second virial cross coefficients, mixture viscosities and binary diffusion coefficients of gas-gas mixtures. Application of the model shows that second virial coefficient data can be represented with good results comparable to the values by means of the corresponding states model. The reliability of the present model in viscosity predictions is proved by comparison with the Lucas method. Predictions of binary diffusion coefficients are in excellent agreement with experimental data and compare well with values obtained by means of the Fuller method.