Celik, Damla YilmazVaheddoost, BabakAras, EgemenSibil, Rahim2026-02-082026-02-0820260955-59861873-6998https://doi.org/10.1016/j.flowmeasinst.2025.103154https://hdl.handle.net/20.500.12885/5672The accuracy of field measurements obtained from aeration tanks is of critically important for the validation of Computational Fluid Dynamics (CFD) models. In many cases, the employed validation metrics serve as a fundamental keystone for evaluating both the credibility of experimental data and the accuracy of numerical simulations. In this study, a novel data refinement approach is developed to assess the physical plausibility of velocity measurements collected from a full-scale aeration tank. Unlike conventional validation approaches, the CFD model is utilized as a reference framework within a reverse-approach perspective to evaluate the reliability of field data. Measurement points affected by acoustic noise, surface sludge interference, and turbulence near static structures were identified and excluded through curve-fitting and statistical filtering techniques. Velocity data obtained with the help of an Acoustic Doppler Current Profiler (ADCP) across six lateral and 53 vertical layers were evaluated using the Coefficient of Determination (R2), Relative Error (RE), and Performance Index (PI) metrics. The maximum-elimination combined with polynomial fitting notably enhanced the model accuracy, reducing RE from -123 % to 20 %, increasing R2 from 0.054 to 0.96, and improving PI from 2.6 to 1.16. As a result, the refined dataset provided a more consistent and realistic representation of the flow structure and established a robust observational basis for the future calibration.eninfo:eu-repo/semantics/closedAccessCurve fittingData eliminationaeration tanksADCPArticle10.1016/j.flowmeasinst.2025.103154108WOS:0016388943000012-s2.0-105025197370Q2Q2