Surface temperature before and after removable modular insulation on irregular industrial components: a CAD-reconstruction method validated by infrared thermography — case study on a 6-ton steam boiler

Abstract

Industrial steam boilers and thermal process equipment pose workplace hazards through high outer-surface temperatures. The incumbent approach to component-level surface-temperature assessment relies on equivalent-length approximations (ASTM C1129, ISO 12241) calibrated to cylindrical geometries, which introduces substantial uncertainty when applied to irregular access components (valves, flanges, doors) with non-standard surface morphologies. Radiometric infrared thermography (IRT) provides field-measured surface temperatures but offers no predictive capability for modified insulation configurations. This case-study paper presents a field-validated methodology for estimating outer-surface temperature reduction under modular removable insulation: (i) radiometric IRT survey of bare component surfaces (FLIR S62 Pro, ISO 18434-1 calibration, emissivity 0.90, ±2 °C uncertainty); (ii) CAD reconstruction extracting true outer surface areas with a +12% allowance for bolts, seams, and irregularities; (iii) ISO 12241 steady-state heat-transfer prediction for a standardized 100 mm mineral-wool panel (lambda = 0.045 W/m·K, external coefficient h = 10 W/m²·K, ambient 25 °C); and (iv) direct validation against FLIR-measured insulated-surface temperatures. The method is demonstrated on a single-flame-tube fire-tube steam boiler (about 6 t/h; surveyed in 2025, 38 pixel-level measurements across three component types). All insulated surfaces remained below the ISO 13732-1 touch-safety threshold of 45 °C: front door 30 °C (bare 96 °C, model 28.1 °C), burner flange 30 °C (bare 146 °C, model 30.2 °C), and steam valves 36 °C (bare 190 °C, model 32.1 °C). Measured insulated temperatures exceeded ideal model predictions by 0–4 °C (mean 1.9 °C), consistent with thermal bridging at seams and maintained access points. The dataset and reproducible CAD-to-prediction workflow provide a single-case, equipment-specific, field-validated reference for component-level surface-temperature estimation. This case study demonstrates feasibility but requires additional validation on other equipment types, geometries, and operational conditions before generalization to broader industrial classes.