About this Research Topic
With the advent of Industrial 4.0 and now the birth of Industrial 5.0, machine learning (ML) is more integrated into manufacturing technology than ever before. One area of manufacturing that is using more and more machine learning is that of inspection and especially, Non-destructive Testing where safety critical parts must be evaluated but not to the point of destruction when such parts are needed to enter service at a specific time and, for a specific period. Due to the nature of NDT and especially multi-spectral NDT, machine learning is required to provide understanding through patterns where no traditional methods can be used due to the very non-linearity within such problem spaces. This is why this title has been selected for a greater focus within Frontiers.
This research topic aims to highlight and address the gaps within ML when applied to NDT and manufacturing systems. Such gaps that have been identified, huge uncertainty and variation encountered in data sets. There is a need therefore for greater transparency and to ensure ‘black boxes’ are more transparent so users gain greater confidence with such technologies. The learning of datasets should be displayed in terms of decision boundaries and where improvements can be made or where potential decisions are not trusted due to being in-between data boundaries for example. Such measures would tackle uncertainty where the transparency would break down how a prediction or decision has been made.
This Research Topic aims to address the above context and apply such technologies to as many manufacturing processes as possible. We welcome contributions on/exploring Topics of interest including but not limited to the following :
• ML techniques applied to NDT sensors to determine salient patterns for inspection or control.
• Use ML models to provide control for manufacturing processes (machining, wire arc additive manufacturing and polymer transformation, reinforced composite auto enclaves technologies to name a few).
• How to make black-box classifiers such as Neural Networks (NNs)/DNNs more transparent in terms of learning data boundaries and towards a greater acceptance in terms of best practices or standardization when applied to NDT and manufacturing systems.
• Using multiple ML models to give a higher confidence in pattern recognition or prediction of NDT systems applied to manufacturing systems (see above for examples), maintenance of manufacturing systems, and robotic processes used to automate the manufacturing process.
• How ML and sensors can provide a better understanding of the manufacturing process – in terms of material interactions, measured temperature, and measured stress/loads, and tracing the future of the manufacturing Industry.
This research topic aims to highlight and address the gaps within ML when applied to NDT and manufacturing systems. Such gaps that have been identified, huge uncertainty and variation encountered in data sets. There is a need therefore for greater transparency and to ensure ‘black boxes’ are more transparent so users gain greater confidence with such technologies. The learning of datasets should be displayed in terms of decision boundaries and where improvements can be made or where potential decisions are not trusted due to being in-between data boundaries for example. Such measures would tackle uncertainty where the transparency would break down how a prediction or decision has been made.
This Research Topic aims to address the above context and apply such technologies to as many manufacturing processes as possible. We welcome contributions on/exploring Topics of interest including but not limited to the following :
• ML techniques applied to NDT sensors to determine salient patterns for inspection or control.
• Use ML models to provide control for manufacturing processes (machining, wire arc additive manufacturing and polymer transformation, reinforced composite auto enclaves technologies to name a few).
• How to make black-box classifiers such as Neural Networks (NNs)/DNNs more transparent in terms of learning data boundaries and towards a greater acceptance in terms of best practices or standardization when applied to NDT and manufacturing systems.
• Using multiple ML models to give a higher confidence in pattern recognition or prediction of NDT systems applied to manufacturing systems (see above for examples), maintenance of manufacturing systems, and robotic processes used to automate the manufacturing process.
• How ML and sensors can provide a better understanding of the manufacturing process – in terms of material interactions, measured temperature, and measured stress/loads, and tracing the future of the manufacturing Industry.
Keywords: Industrial 4.0, Machine Learning, Non Destructive Testing, Neural Networks, Manufacturing Systems, Manufacturing Processes, measured temperature, manufacturing Industry
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.
