Non-Contact Real-time Nondestructive Evaluation Technology for NDE 4.0

无损评估（NDE）对各个制造部门都至关重要，但目前的做法需要人工操作人员参与数据收集、传输和分析的各个方面，这导致了许多安全、速度和准确性方面的问题。随着工业4.0的到来，NDE技术需要升级到“NDE 4.0”，包括自动和自主的NDE、数据通信的互联性、利用人工智能进行实时数据分析等基本方面，这些都是目前技术无法实现的。在这个项目中，我们将提出一种新的混合NDE方法，可以解决这些方面，以实现NDE 4.0。目前使用的无损检测方法多种多样，包括x射线、电磁、超声、声发射、热成像和激光超声；然而，它们都有严重的局限性。最大的限制是人的参与；所有方法都依靠检查员进行检查和解释数据；因此，它们速度慢，不能应用于危险环境，结果是主观的，容易受到人为错误的影响。在这个项目中，我们将开发一种新的基于人工智能的混合无损检测方法，该方法可以执行非接触、在线、远程和自动检测，并具有使用嵌入式人工智能的实时数据解释能力。该计划包括三个相互关联的目标：1)开发一种相位激光超声技术，该技术可以聚焦并引导超声波以高信号强度到达感兴趣的区域。这可以实现非接触和无耦合的无损检测，使远程和自动无损检测适用于危险环境和几何复杂的部件。2)研制基于石墨烯的超声波接收器，用于接收反射的超声波。利用石墨烯的超低质量和高机械强度，我们将开发一种比现有超声波接收器更小、更灵敏的宽带接收器。3)开发基于fpga的NDE控制器，控制NDE进程，使用AI算法对数据进行解释，并与云服务器进行实时通信。基于fpga的控制器由于其快速的计算速度和强大的并行能力，可以利用人工智能实现实时数据分析。每个目标都是前沿和创新的，而所有目标的整体实现将开发新的NDE技术，可以立即应用于在线，自动和安全的NDE，这将引领技术发展，朝着工业4.0的方向发展。拟议的研究将使加拿大在NDE研究、结构健康监测以及NDE与实时人工智能的整合方面具有竞争优势。这种整合将实现更高水平的无损检测技术，使加拿大工业在新兴的无损检测服务和设备市场中获得动力，同时确保产品、结构和基础设施的安全。

Nondestructive evaluation (NDE) is crucial for various manufacturing sectors, but its current practices require human operators to be involved in all aspects of data collection, transfer and analysis, causing many issues on safety, speed and accuracy. With the advent of Industry 4.0, NDE technology needs to be upgraded to "NDE 4.0" comprising essential aspects such as automatic and autonomous NDE, interconnectivity for data communication, and real-time data analytics using AI, which cannot be achieved with current technology. In this program, we will advance a new hybrid NDE approach that can address these aspects to realize NDE 4.0. Currently various NDE methods are used, including X-ray, electromagnetic, ultrasound, acoustic emission, thermography and laser ultrasonics; however, all have critical limitations. The greatest limitation is human involvement; all the methods rely on inspectors to conduct the inspection and interpret the data; thus they are slow, cannot be applied to hazardous environment and the results are subjective and vulnerable to human errors. In this program, we will develop a new hybrid AI-based NDE method that can perform non-contact, in-line, remote and automatic inspections, with real-time data interpretation capability using embedded AI. This program consists of three inter-related objectives: 1) to develop a phased-laser ultrasonic technology that can focus and steer the ultrasound waves to the region of interest with a high signal strength. This can achieve a non-contact and couplant-free NDE, enabling remote and automatic NDE suitable for hazardous environments and geometrically complex components. 2) to develop a noble graphene-based ultrasonic receiver to receive the reflected ultrasonic waves. Utilizing the ultra-low mass and high mechanical strength of graphene, we will develop a wideband receiver that is much smaller and more sensitive than existing ultrasonic receivers. 3) to develop a FPGA-based NDE controller that controls the NDE process, interprets the data using AI algorithm, and communicates with the cloud server in real-time. Owing to its fast computation speed and powerful parallelism capability, a FPGA-based controller is able to achieve real-time data analytics using AI. Each objective is cutting-edge and innovative, while the achievement of all objectives as a whole will develop the novel NDE technology, that can be immediately applicable to the in-line, automatic and safe NDE, which would lead technological development for the progression toward and through the Industry 4.0. The proposed research will give Canada a competitive edge in the study of NDE, structural health monitoring and the integration of NDE with real-time AI. This integration will enable a higher level of NDE technology, allowing Canadian industry to gain momentum in emerging NDT service and equipment markets, while ensuring the safety of the products, structures, and infrastructure.