Online Monitoring System for Macro-Fatigue Characteristics of Glass Fiber Composite Materials Based on Machine Vision

Glass fiber-based composites are widely used in various fields due to the excellent properties of glass fiber. Although glass fiber has the advantages of high strength and corrosion resistance, it also has the weakness of brittleness and low toughness. When glass fiber-based materials are loaded under alternating stresses for a long time, there is a high risk of fatigue-induced deformation damage. The accidents caused by fatigue damage (FD) are all over the field and have caused serious accidents. Most of the existing methods for fatigue deformation monitoring rely on high-precision sensors, which are expensive and poorly robust, and do not have the ability to be extended to industrial production. In this article, we first proposed an in situ online monitoring system purely based on machine vision, which identifies the key frames at the same position in different periods in the experimental video stream and extracts the changing regions through a series of image processing operations to judge whether fatigue deformation occurs. The proposed monitoring method can be applied to the task of monitoring fatigue deformation tests of rigid materials before service. After on-site testing and evaluation by professionals, the proposed monitoring system can completely capture the deformation characteristics of the material during the entire test cycle with a deformation accuracy of 0.07 mm and an average false alarm rate of 2.14 & x0025;. The complete deformation video data are provided by the monitoring system that can be used for material deformation characteristics modeling.

Automated summary

Glass fiber-based composites are widely used due to their excellent properties, but they are prone to fatigue-induced deformation damage. Existing monitoring methods are not suitable for industrial production. This article proposes a machine vision-based online monitoring system that can identify fatigue deformation in real-time. The system has a high deformation accuracy and a low false alarm rate, as confirmed by on-site testing and evaluation.