Unsurpassed Innovation Award
Company: National Institute for Aviation Research (NIAR/WSU)
Description: The CAMX Award winners will be announced during Tuesday's General Session and will present in the theater at 12 pm. All entries will be on display throughout the week in the CAMX Awards Pavilion in the exhibit hall. In-process AFP Manufacturing Inspection System (IAMIS) attached to automated fiber placement (AFP) head for detecting manufacturing defects that are above the certification basis (or unacceptable) using machine-learning (ML) algorithms for reducing time-consuming and operator-dependent manual inspection processes that require significantly interrupting the manufacturing process. Use artificial intelligence (AI) for analyzing vast amount of digital processing data and geometry (stored in the digital backbone) at each defects location for identifying manufacturing anomalies for optimizing parameters (ex., lay down speed, heat input, compaction force, steering radii, etc.) in order to reduce manufacturing defects on subsequent parts.
Collaboration and Partnerships: Numerous discussions were carried out with multiple AFP equipment manufacturers such as Electroimpact, Coriolis, Mikrosam, M-Torres, and Fives regarding the implementation of in-process inspections systems, challenges, limitations, etc. In addition, inspection data gathered on prototype was included in several technical reports and conferences publications such NASA HiCAM (High-rate Composite Aircraft manufacturing) program final report, ACMA/TCC Composites Industrial Revolution Conference 2021, and SAMPE paper May 2021. Several demos were conducted to defense and advanced air mobility partners that are willing to integrate this system at their AFP machine to collect additional data and compare the automated inspection data against their current labor-intensive manual processes. In addition, currently negotiating with partners regarding the commercialization of various aspects of the system, ex., inspection system, software, and augmented-realty visualization module. Industry/production-scale investigation will be carried out by installing the system in an industry setting, but still carrying out traditional manual inspections for comparison. Automated and manual inspections are then compared for each part to establish the probability of detection (and records of false positive) to demonstrate the repeatability and reproducibility of automated system to quality control and certification authorities for detecting manufacturing. This will establish the framework for implementing this system in a factory setting as the primary inspection method for quality assessment of AFP parts.
Concept and Design: Quality assurance through inspections and process controls are essential to ensure that material is laid up and process according to specification with no process-induced defects. Although AFP significantly improve the production rates and quality, due to the lack of reliable in-process inspection techniques, AFP processes are currently interrupted intermittently (20-70% of the production time) for manual inspections, diminishing benefits of automation. In addition, manual inspection processes have deficiencies such as operator/training/environment-dependencies and inconsistencies. Main goal of this research is to develop a framework for implementing machine-learning algorithm and artificial intelligence for improving AFP manufacturing rate by reducing the labor-intensive manual inspection and improve the quality through big data analytics (history of inspection data). Most hardware associated with the inspection system was put together by off-the-shelf laser and camera systems. However, the integration of this system for various AFP head configuration is accommodated by additively manufactured compact light-weight bracket and wire management system. Information stored in the digital backbone can be used for airworthiness certification as a digital manufacturing twin and can be accessed for product life cycle management such as conducting a comprehensive analysis of an in-service damage for repair design accounting the allowable manufacturing defects associate with the digital twin of the physical part. - Since the inspection system provide AFP operators and quality inspectors the data in a reliable manner, secondary (manual) inspection time will be eliminated, except when a repair is needed. With the use of accompanying augmented reality visualization module (when the AFP operation is halted for repairs) provide the location and description of each defect to the quality inspectors and repair technicians in order to locate them quickly and accurately. Augmented reality visualization module also enables remote review by an inspection/quality specialist to significantly reduce the downtime waiting for repair/analysis decisions.
Additional Information: IAMIS can be implemented in other manufacturing processes with the selection of a suitable inspection techniques. The augmented reality system developed for IAMIS can be used for supporting various manufacturing activities (hand placement of fabric and adhesives in AFP parts) as well as for quality assurance.