Host Institution: The University of Bristol
Start Date: 1st April, 2019
Duration: 6 months
Lead Investigator: Robert Hughes
The rate of production is the most significant cost driver in composite manufacturing processes today. This process is dominated by two main activities; lay-up, and inspection & rework. Inspection & rework is integral for ensuring product quality and integrity but is performed 100% by eye, and as such consumes 63% of the manufacturing time for manual lay-up processes . This has been significantly improved through the introduction of automated fibre placement (AFP) systems, which have a higher reliability during layup and as such require less inspection, and have therefore reduced the inspection time by nearly half (32%) . However, these machine-laid components are still inspected 100% manually throughout the world today. Reducing the time spent inspecting the parts is the only way of increasing the rate of manufacture further, whilst strict quality and sensitivity conditions must still be met. It is into this context that in-process non-destructive evaluation techniques can be employed. The ultimate goal of this project is to explore the state-of-the-art advanced inductive sensing technology and automatic evaluation systems that can all but eliminate separate inspection times, thus bringing about a step-change in the rate of composite manufacturing. This will fill a much-needed hole in the Hubs inspection and in-process sensing priority research area.
The ideal time to evaluate the quality of these composite components is during the manufacturing process, where potential problems can be immediately reworked, or layups discarded quickly to reduce delays. At this stage the introduction of contaminants must be strictly guarded against to prevent potential sources of weakness or disbonding within the components. Unlike other non-destructive techniques, such as ultrasonic testing, eddy-current testing (ECT) inspections do not require physical contact or a coupling medium (i.e. water) to operate. In addition, the principles of electromagnetic induction mean that ECT sensors can probe multiple layers into the material in a single measurement [Hughes et. al. 2018]. Unlike rival optical methods which must be performed on each new layer, ECT inspections can therefore be applied after multi-layers have been applied, further reducing the time spent inspecting the components. As such ECT techniques present an attractive inspection method for composite manufacturing applications. However, to date, all attempts at utilising ECT inspection techniques on carbon fibre materials have been hampered by slow and unreliable data analysis techniques preventing significant confidence in such systems.
The research proposed will explore the potential for ECT inspection and analysis of carbon fibre composites, determining whether such technology is viable for a diverse range of in-process inspections, how it’s use can be exploited to advance the success of manufacturing plants of the future.
Recent preliminary investigations , , identified Radon transform analysis as a more sensitive and computationally less-demanding method of evaluating fibre orientation compared to current 2DFFT methods and showed depth dependent information encoded within the inspection data (Figure 2.3). I will use this grant to expand from this simple observation and develop the techniques, collaborations and knowledge required to design, build and operate advanced sensor and evaluation technology for real-time 3D structural mapping of composites for in-process applications.
The ultimate output of this project is a critical assessment of the inspection technology and data evaluation methods capable of automatically detecting and characterising target structural properties in pre-cured composites such as; material damage, fibre volume fraction (FVF) variations, fibre orientation deviations, stacking sequences, bridging and wrinkling, with the sensitivities equivalent, or superior, to current standards. The grant will be split into 2 key strategic objectives (see Figure 3), each representing key milestones, and containing specific work packages.