Scholarlink

<p>This investigation examined the density and viscosity of five binary mixtures of propyl butyrate with 1-alkanols (C<sub>5</sub>–C<sub>10</sub>). The properties were measured across the complete composition range at temperatures from 293.15 to 323.15&#xa0;K. Experimental analysis showed positive excess molar volumes, and the magnitude of these values increased with both temperature and the chain length of the 1-alkanol. These results are consistent with volume expansion upon mixing for the systems studied. A Physics-Informed Neural Network (PINN) was also developed to predict the excess molar volumes. This model, which incorporates thermodynamic constraints, was applied to the mixtures. For the datasets in this study, the model yielded an Average Absolute Deviation (AAD) of 0.0053 cm<sup>3</sup>/mol and a mean relative error of 3.23&#xa0;%. Furthermore, <i>R</i><sup>2</sup> values for the model’s predictions exceeded 0.99 for all tested systems. This indicates a strong correlation between the predicted and experimental data under these specific conditions, particularly for the longer-chain alkanols and at higher temperatures.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Physics-Informed Neural Network Prediction of Thermophysical Properties for Propyl Butyrate + 1-Alkanol (C6–C10)

  • 期刊论文

  • 发表日期：

    2026年1月12日

查看全文
  • Adel Noubigh,
  • Mohammad Almasi,
  • Nasim Rahmani-Ivriq

摘要

This investigation examined the density and viscosity of five binary mixtures of propyl butyrate with 1-alkanols (C5–C10). The properties were measured across the complete composition range at temperatures from 293.15 to 323.15 K. Experimental analysis showed positive excess molar volumes, and the magnitude of these values increased with both temperature and the chain length of the 1-alkanol. These results are consistent with volume expansion upon mixing for the systems studied. A Physics-Informed Neural Network (PINN) was also developed to predict the excess molar volumes. This model, which incorporates thermodynamic constraints, was applied to the mixtures. For the datasets in this study, the model yielded an Average Absolute Deviation (AAD) of 0.0053 cm3/mol and a mean relative error of 3.23 %. Furthermore, R2 values for the model’s predictions exceeded 0.99 for all tested systems. This indicates a strong correlation between the predicted and experimental data under these specific conditions, particularly for the longer-chain alkanols and at higher temperatures.