Local-to-rack-scale temperature responses of sodium heat pipes inserted into industrial graphitization-furnace insulation

Abstract

Post-shutdown cooling of industrial graphitization furnaces is constrained by the large thermal inventory and low conductivity of packed petroleum-coke/graphite insulation. This study examines the local-to-rack-scale thermal response produced when sodium high-temperature heat pipes (HTHPs) are inserted into that insulation. The experimental program comprised depthwise and paired spatial baseline measurements, three simultaneously inserted and individually monitored HTHPs at 0.2–0.4 m depth, small-array tests at 1.0 and 0.5 m pitch, and a six-rack deployment of 96 HTHPs. The day-2/day-12 baseline consisted of paired spatial snapshots rather than a continuous cooling curve: 8 of 18 locations were cooler on day 12 and 10 were warmer, so no furnace-wide natural-cooling rate was inferred. At points 5 cm from the three individually inserted sections, regression rates were 1.690 ± 0.338, 1.757 ± 0.320 and 2.543 ± 0.341 °C/min, respectively (expanded uncertainty, k c = 2). Insertion depth, absolute position and initial local temperature varied together. At the centroid of an equilateral 1.0 m-pitch three-pipe array, the insulation-temperature trend was 0.1667 ± 0.1740 °C/min and is interpreted cautiously. The 0.5 m-pitch 4 × 4 array produced post-settling HTHP-wall rates of 0.04165–0.07914 °C/min across 12 retained channels. Reanalysis of 82 valid industrial channels separated advancement transients from holding intervals. Five 2000 min racks yielded segment-weighted HTHP-tip wall-temperature rates of 0.0713–0.0873 °C/min, with a between-rack mean ± standard deviation of 0.0791 ± 0.0078 °C/min. The results demonstrate repeatable local and rack-scale temperature responses under the tested boundary conditions; direct mechanism identification and whole-furnace process benefits require synchronized control and full-cycle measurements.