Passive-scalar transfer-kernel modelling for thermal transit-time flow measurement in pipe flow
Measurement Science and Technology · 30.10.2026 · 37 (43) · 435303
DOI: 10.1088/1361-6501/aeb4d7 ↗Kurz gefasst
Fachartikel in der Zeitschrift Measurement Science and Technology, online erschienen am 30.10.2026. Verfasst von Taleb Moazzeni. Die Arbeit nennt 37 Literaturangaben.
Abstract (Originalsprache)
Thermal transit-time flowmeters are attractive in harsh environments, but shear, cross-stream mixing and sensor dynamics can bias the usual single-delay estimate. This study develops a passive-scalar transfer-kernel measurement model to separate transport-induced bias from instrumentation uncertainty and to guide sensor and heater design. Analytical kernels and a diffusion-corrected inverse operator are assessed using independent scalar discretisations, a newly implemented solved-momentum/scalar benchmark, published DNS integral quantities and Monte Carlo propagation. The new benchmark solves pressure-driven laminar or algebraic Reynolds-averaged mean momentum before transporting the scalar; it does not prescribe the velocity profile. Thirty-two operating cases span Reynolds numbers 500–96 000 and molecular Prandtl numbers 0.005–7. Three sensor layouts and three frequency bands give 288 inverse configurations. Noise-free full-band, three-radius inversion has a maximum absolute mean-speed error of 2.754% within the baseline matrix, whereas centreline-only and restricted-band designs can fail. At five representative corners, medium-to-fine mesh changes remain below 0.045% in the transfer vector. Closure perturbations increase the largest tested speed error to 4.78%; unmodelled wall heat loss causes a clear failure at low Peclet number. A separate, fixed-profile uncertainty scenario uses 4000 nominal trials and 1000 additional input-width sensitivity trials. Nominal three-radius mean-speed RMSE is 0.877%, conditional on the declared uncertainty model; its percentile endpoints have Monte Carlo standard errors of 0.024 and 0.039 percentage points. Likelihood-only Cramer–Rao bounds and model-matched efficiency checks distinguish ideal noise limits from model discrepancy. The framework supplies computational design evidence for thermal flow measurement, not a hardware accuracy certificate. Fully developed mean-flow closure, assumed uncertainty widths, adiabatic-wall operation and the absence of fluctuation-resolving coupled CFD/DNS remain explicit limits.
Zitieren
APA
Moazzeni, T. (2026). Passive-scalar transfer-kernel modelling for thermal transit-time flow measurement in pipe flow. Measurement Science and Technology, 37(43), 435303. https://doi.org/10.1088/1361-6501/aeb4d7
Vancouver
Moazzeni T. Passive-scalar transfer-kernel modelling for thermal transit-time flow measurement in pipe flow. Measurement Science and Technology. 2026;37(43):435303. doi:10.1088/1361-6501/aeb4d7
BibTeX
@article{Moazzeni2026,
author = {Moazzeni, Taleb},
title = {Passive-scalar transfer-kernel modelling for thermal transit-time flow measurement in pipe flow},
journal = {Measurement Science and Technology},
year = {2026},
volume = {37},
number = {43},
pages = {435303},
doi = {10.1088/1361-6501/aeb4d7}
}
EUVIMED