Study on damaged expanded water-repellent perlite insulation integrating neural network and simulation on temperature field and Boil-off Rate in a Liquefied Natural Gas Cargo Containment

Abstract

The temperature field and Boil-off Rate (BOR) of the cargo containment system (CCs) in the LNG (Liquefied Natural Gas) carrier are directly impacted by the thermal performance of insulation, which has an adverse effect on the structural safety and commercial feasibility of LNG transportation. However, the potential damage to hydrophobic layer of expanded perlite during the processes of filling, storage, and transportation can lead to undesirable increases in water absorption, thermal conductivity, even the risk of insulation failure. To investigate the variations in the thermal conductivity of damaged expanded water-repellent perlite used in CCs of LNG carrier, The variations in the thermal conductivity of expanded water-repellent perlite are tested using a guarded hot plate apparatus extending down to liquid nitrogen temperatures under the combined influence of moisture content and damage rates. Furthermore, heat transfer models are established, and the accuracy of the mathematical models is validated using experimental data collected from LNG carrier CCs. A Back-propagation Neural Network is used to conduct sensitivity analysis and predictions based on the experimental results. The results show that both the breakage rate and water content have detrimental effects on insulation performance. Specifically, under equatorial sailing conditions, 450 mm-thick expanded water-repellent perlite show respective BOR increases of 22.8% and 116.4%. With 530 mm-thick expanded water-repellent perlite, BOR increased by 20.7% and 112.1%, respectively, even exceeding the daily BOR limit for LNG cargo. Analyzing temperature field and BOR provides essential insight for evaluating the thermal insulation performance of the structural system in the NO 96 LNG carrier.