An experimental study of applying machine vision technology in on-line solid debris analysis of rolling oil of a cold rolling stand

摘要

This article describes an experimental study, in which a machine vision technology based sensor prototype for on-line solid debris analysis on oils is applied to condition monitoring measurements in an industrial application. Real-time analysis of solid particles in oil is useful in many preventive condition monitoring operations. The study consists of a measurement campaign and experiments in cold rolling mill at Outokumpu Tornio Works (Steel manufacturer in Tornio, Finland). The goal was to study the functionality and performance of the sensor prototype for producing on-line measurement data aiming mainly at preventive maintenance of the oil filters of a cold rolling stand, and in addition to find out if the measurement technique is capable of detecting particles affecting quality of the product. Results of the study show that the sensor prototype is functional in industrial environment. It can measure size and shape parameters of particles within the size range of 25 μm - 15 mm on-line at 3.3 Hz frame rate from oil volume flow of 1 litre per minute. The measurement results can be utilized in the decision making of changing the oil filters before the filtration performance decreases below the allowed limits. Presence of the particles, which could affect quality of the product, can be determined.