Design and Validation of a Machine Vision and Ultrasound System for Diagnosing Internal Wear in Warman Pump Circuit Pipes

Galarza A.M.M.; Alvino J.M.M.; Pino L.E.V.; Moreno A.H.G.; Daga R.F.J. · Proceedings of the International Conference on Automation, Robotics and Applications, ICARA (2026) · DOI: 10.1109/icara69401.2026.11480409

Resumen

This research presents a mechatronic internal inspection system designed to detect and map erosion in pipes used to transport pulp using Warman pumps. The device integrates a wheel-centering mechanism, ultrasonic sensors, and a machine vision module, allowing for continuous, non-intrusive wear measurements. To characterize flow conditions and their relationship to erosion, the Reynolds, Darcy-Weisbach, and critical velocity models were applied, while material loss was evaluated using the Finnie and Oka models. Digital image processing, implemented in Python/OpenCV, generated pixelated maps, adaptive segmentation, and 20 × 20 mesh analysis to quantify deteriorated areas and classify the level of severity. Ultrasonic measurements allowed the remaining thickness to be estimated and the correlation between surface texture and actual wear to be validated. The results show agreement between computer vision, UT measurement, and theoretical predictions, demonstrating that the proposed system is an accurate and low-cost alternative for predictive maintenance in pipes subject to abrasion.

Instituciones: Universidad Continental