Enhanced Predictionof CO₂–BrineInterfacial Tension at Varying Temperature Using a Multibranch-Structure-BasedNeural Network Approach

Interfacial tension (<i>IFT</i>_<i>C–B</i>) between CO₂ and brine depends on chemical components in multiphase systems, intricately evolving with a change in temperature. In this study, we developed a convolutional neural network with a multibranch structure (MBCNN), which, in combination with a compiled data set containing measurement data of 1716 samples from 13 available literature sources at wide temperature and pressure ranges (273.15–473.15 K and 0–70 MPa), was used to quantitatively explore the correlation of various chemical components with <i>IFT</i>_<i>C–B</i> at varying temperature, aiming to achieve accurate predictions of <i>IFT</i>_<i>C–B</i> under complex conditions. Our multibranch neural network analysis yielded some important insights: (1) Leveraging the convolutional and multibranch structure, MBCNN effectively mitigates the adverse effects of sparse matrices resulting from the absence of certain basic components, exhibiting higher prediction accuracy particularly for low <i>IFT</i>_<i>C–B</i> scenarios (MAE = 0.47, and R² = 0.9921) than other AI models. (2) The multibranch structure allows MBCNN to additionally capture the interattribute relationship between temperature and each chemical component. Such interattribute relationships are quantitatively correlated with <i>IFT</i>_<i>C–B</i>, demonstrating that varying temperature significantly influences the dependence of <i>IFT</i>_<i>C–B</i> on chemical components in gas and brine by causing the variation in their solubility. Specifically, the ratio of <i>IFT</i>_<i>C–B</i> to the molality of monovalent cations (Na⁺ and K⁺) and bivalent cations (Ca²⁺ and Mg²⁺) in brine, as well as to the mole fraction of non-CO₂ components (CH₄ and N₂) in the gas phase, varies with increasing temperature, approximately following a quadratic function. (3) By formulating the effect of each attribute on <i>IFT</i>_<i>C–B</i> and quantifying their respective weight, we derived a new piecewise function for predicting <i>IFT</i>_<i>C–B</i> at three temperature intervals (<i>T</i> ≤ 293.15 K, 293.15 K < <i>T</i> ≤ 324.4 K, and <i>T</i> > 324.4 K), with high prediction performance (MAE = 2.3672, R² = 0.9263) across a wide temperature range in saline aquifers.