Anti-freezing time prediction of fire pipe covered with insulation material in low temperature environment

Abstract

The reliable operation of fire pipes in cryogenic environments is essential to ensure safety in residential and industrial environments. This study introduced a systematic approach that combined experimental data, computational fluid dynamics (CFD) simulations and machine learning techniques to investigate the thermal performance of insulated fire pipes. A key contribution of this study is the derivation of a calibrated analytical expression for the heat transfer coefficient in natural convection that accounts for the dynamic effects of insulation thickness and ambient temperature. The validated CFD model was used to simulate the freezing process, while the random forest (RF) algorithm was used to predict the optimal insulation thickness to achieve a specific antifreeze time. The results showed that the type, thickness and ambient temperature of the insulation material had considerably affect the antifreeze performance, with rubber sponges outperforming aluminium silicate shells in cryogenic environments. The optimized RF model exhibited high accuracy and robustness, providing a practical tool for insulation design. This study provides valuable insights and technical guidance for improving the reliability of fire pipes under extreme conditions.