創新及科技解決方案

解決方案編號

S-0568

解決方案名稱

基於人工智能的跨材料超聲波結構健康監測方法

解決方案描述

The development of modern industries has put forward new requirements for structural health monitoring, including high efficiency, high accuracy, and high reliability. Existing physical-model-based evaluation methods, such as using ultrasonic waves, developed on one specific material cannot be directly applied to other materials due to the different mechanical properties of the materials. Therefore, the application of the physical-model-based damage evaluation in different materials requires a large amount of modelling work, which is time-consuming and inefficient. Our proposed AI-based method can automatically determine the health state of the monitored structures (intact or damaged) by making use of the wave energy distribution features in the frequency domain. In conjunction with a proper imaging algorithm, the location of the damage can be identified. Independent of the mechanical properties of the material and wave propagation characteristics in the structures, this method can be applied to different materials, including metallic and composite materials. The AI-based cross-material structural health monitoring using ultrasonic waves focuses on solving the key requirements of potential partnerships including automotive, aviation, and ship manufacturers, civil infrastructure systems, energy industries, etc., developing effective communication and technical solution.

應用領域

基礎設施
Automotive Production, Aircraft Maintenance, Material Testing

使用的技術

機器學習
人工智能
深度學習
數據分析
Ultrasonic Testing

使用例子

Different from the conventional ultrasonic wave-based structural monitoring and inspection method which requires a vast amount of modelling work for multiple materials, this newly proposed method can be directly applied to different materials, which is much more efficient by circumventing complicated physical modelling of wave propagation in structures. According to previous test results, the damage detection accuracy for aluminum plate and glass fibre-reinforced plastic laminate is 92.86% and 95.45%, respectively. The method can be applied in various industry fields, such as automotive, aviation, and ship manufacturers, civil infrastructure systems, energy industries, etc.

公司／機構名稱

香港生產力促進局

電郵地址

mingge@hkpc.org

電話號碼

27886007

網址

www.hkpc.org

地址

HKPC Building, 78 Tat Chee Avenue, Kowloon, Hong Kong.

如政府部門對創新及科技解決方案進行概念驗證或驗證技術請聯繫智慧政府創新實驗室。