The Present and Future Value of Quantitative As-built Data
Year: 2025
Type: conference-paper
Abstract: As composite production floors transition to automatic monitoring and in-process inspection methods (such as profilometry and machine vision), they generate large amounts of quantitative as-built data. Currently, composites manufacturers may use these data lakes for as-built documentation – but for little else. By applying data analytics and artificial intelligence/deep learning (AI/DL) to these vast stores of data, the composites industry has an opportunity to leverage manufacturing insights to develop improved designs and processes, boost product quality and production rate, lower manufacturing costs and extend product service life. This paper reviews the present value of in-process inspection data, including immediate feedback and error correction, documentation and the as-built digital twin. Automatic data collection, analysis and feedback to the shop floor help reduce inspection time up to 95% compared to manual inspection, and demonstrated six sigma reliability of the inspection data enables fabricators to avoid poor quality processes that raise the risk of project cancellation. The paper then outlines opportunities to use the data beyond the current production run. It covers the kinds of data analytics and AI/DL used to generate manufacturing insights and describes applications of these insights to closed-loop continuous improvement initiatives: problem-solving and root-cause analysis, design for quality, reduced uncertainty and opportunities to reduce allowables/overdesign, design for manufacture, more accurate digital simulations and accelerated new product/process introductions (NPIs).
Authors: Aligned Vision, Scott Blake
Language: English
Subfield: Industrial and Manufacturing Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4411598047