According to the email confirmed by JIM, our paper entitled, "EBPC: a deep learning cloud computing framework for hybrid stack drilling monitoring", gets accepted.
Single-shot drilling of stacks composed of composite and aluminium is a common operation in aircraft assembly, where adaptive drilling that allows real-time adjustment of cutting parameters is crucial to improve assembly strength. Although deep learning approaches improve prediction accuracy, they also require significant investment in computational resources. This paper introduces a novel cloud computing framework to enable online and responsive process incident monitoring for composite/Al drilling. By measuring Signal-to-noise ratio of the harmonic components in thrust and torque, a bit depth limit for the signals is established, forming a basis for data minimisation in line with the signal sampling boundary theory. To reduce congestion and delay in the cloud computing system for online tool condition monitoring, a bit depth optimised EBPC cloud computing framework composed of exponential backoff adaptive client traffic control algorithm and priority queue based producer-consumer server request scheduling is proposed in this paper. Local network stress tests confirms the efficiency and resilience of proposed framework, while remote computing experiments demonstrate its capability to operate effectively across all Europe through different connectivities. This framework advances deep learning applications for cloud computing in tool condition monitoring, especially where low-latency response is essential.
We will make an special page to introduce the principles of our paper. Please keep updated for the project page
Link
Please view our paper in the link: