Use of in-process monitoring and ultrasound to detect defects in thermoplastic AFP-produced parts

Automated manufacturing methods such as automated fibre placement (AFP) have led to a significantly increased production rate of fibre-reinforced composite components. However, this increased production is accompanied on an increased reliance on autonomous quality control or post-manufacturing inspection methods. The rising popularity of thermoplastic-matrix composite materials and out-of-autoclave manufacturing pushes this reliance even further, with the primary objective of many manufacturers being a single-step manufacturing process. To achieve this goal, defects or deviations during the manufacturing process must be identified as soon as possible, preferably without the need for additional destructive or non-destructive testing steps. In this work, a unidirectional carbon fibre-reinforced PPS laminate was manufactured containing a series of AFP-inherent defects; charring from excessive heating, poor consolidation from insufficient heating, foreign material inclusions, and missing tow material. The visibility of these defects was then assessed using in-process monitoring, via temperature readings from a thermal camera, and subsequent ultrasonic scanning of the laminate. The in-process monitoring displayed sound detection of the missing material and poorly consolidated tow, as well as a variation in the substrate surface caused by charring during deposition of the previous ply. However, the camera was unable detect temperature variations caused by the foreign aluminium inclusions. This shortcoming is attributable to the current (coarse) averaging used when outputting temperature values. Future work will aim to quantify the current minimum detectable length and reduce it through modifications to the collection software. The ultrasound scanning method was able to detect all defects within the laminate, with a combination of amplitude-based and timedelay-based read-outs producing the greatest visibility. The results from this work will be further developed in order to increase the detectability of defects during manufacturing and hence deliver more robust automated quality control capabilities.