# Automated Local Measurement of Wall Shear Stress with AI-Assisted Oil Film Interferometry

## TL;DR

This paper introduces an AI-powered method to measure wall shear stress accurately and automatically using oil film interferometry.

## Contribution

The novel AI-OFI system automates oil film interferometry with deep learning for real-time, high-resolution wall shear stress sensing.

## Key findings

- AI-OFI achieves local WSS detection with average deviation below 5% compared to reference measurements.
- The system uses YOLO and VGG16 models for real-time interference pattern analysis and WSS estimation.
- A smart interrogation-window algorithm ensures robust performance under varying conditions.

## Abstract

Accurate measurement of wall shear stress (WSS) is essential for both fundamental and applied fluid dynamics, where it governs boundary-layer behavior, drag generation, and the performance of flow-control systems. Yet, existing WSS sensing methods remain limited by low spatial resolution, complex instrumentation, or the need for user-dependent calibration. This work introduces a method based on artificial intelligence (AI) and Oil-Film Interferometry, referred to as AI-OFI, that transforms a classical optical technique into an automated and sensor-like platform for local WSS detection. The method combines the non-intrusive precision of Oil-Film Interferometry with modern deep-learning tools to achieve fast and fully autonomous data interpretation. Interference patterns generated by a thinning oil film are first segmented in real time using a YOLO-based object detection network and subsequently analyzed through a modified VGG16 regression model to estimate the local film thickness and the corresponding WSS. A smart interrogation-window selection algorithm, based on 2D Fourier analysis, ensures robust fringe detection under varying illumination and oil distribution conditions. The AI-OFI system was validated in the high-Reynolds-number Long Pipe Facility at the Centre for International Cooperation in Long Pipe Experiments (CICLoPE), showing excellent agreement with reference pressure-drop measurements and conventional OFI, with an average deviation below 5%. The proposed framework enables reliable, real-time, and operator-independent wall shear stress sensing, representing a significant step toward next-generation optical sensors for aerodynamic and industrial flow applications.