Monte Carlo-optimised circumferential gamma-ray scanning for measuring angular film thickness profiles in...

Accurate quantification of condensate film thickness in large-diameter wet gas transmission pipelines represents one of the more challenging problems in nuclear measurement technology. In misty flow, liquid is distributed as fine aerosol droplets across the full cross-section of the pipe, making conventional gamma densitometry unreliable. A preconditioner system - consisting of a static helical impeller and a circular nozzle - converts this misty regime into a quasi-annular configuration in which the liquid coalesces as a continuous film on the inner pipe wall, making the geometry tractable for circumferential gamma scanning. This paper presents a Geant4/GATE Monte Carlo radiation transport study of a 360-detector circumferential scanning array designed for a full-scale 32-inch (DN-800, Schedule 40) steel pipe with an outer diameter of 813 mm, inner diameter of 778.04 mm, and wall thickness of 17.48 mm. A ¹³⁷Cs source of 40 mCi activity, positioned at the geometric centre of the pipe, was modelled with an axial lead collimator reproducing the laboratory pencil beam. Three condensate-equivalent calibration phantom sectors - 15 mm, 20 mm, and 27 mm thick, wood, density 750 kg/m³, angular opening 67° - were placed against the inner wall at angular separations of 20°, 51°, and 88°. The model was implemented in two campaigns: the original thesis configuration, and a fully-documented reconstruction (Geant4 11.0.0/GATE 9.2) incorporating explicit lead collimation, a 7% detector energy-resolution model, and 1.92×10⁹ primary histories. The reconstructed run gives sector relative count rates of 0.9255 ± 0.0011, 0.9021 ± 0.0011 and 0.8696 ± 0.0011 for the 15, 20 and 27 mm phantoms, referenced to the condensate-free gas region at a no-film baseline of 13144 cps per detector element. Weighted regression yields an effective linear attenuation coefficient of μ=0.00517±0.00003mm^-1 for the wood phantom at 0.750 g/cm³, and inverting this calibration recovers the three film thicknesses to within 0.06 mm of nominal. Angular detection errors across six measurement zones range from 0.42% to 2.06%, with a total circuit error of only 0.04% for the complete 360° rotation. Agreement between the legacy and reconstructed campaigns is 2.1% in normalised RMSE across the sampled regions. The central-source configuration yields a baseline count rate approximately seven times higher than an equivalent external-source arrangement, owing to the halving of steel wall traversal. To the authors' knowledge, this is the first reported Monte Carlo optimisation study of a circumferential gamma scanning array at DN-800 scale, providing a validated computational framework for annular film profiling in large-bore wet gas infrastructure.