A semantic infrastructure for advanced manufacturing

We are stepping into a new era of digitalisation. In this new era machines will communicate and exchange large amounts of data to ensure they can work harmoniously and collaboratively with little human intervention. Current machines use symbolic language to represent the data, but they cannot directly interpret its meaning. As a result, information loss and incorrect interpretation can often happen during communication. To improve manufacturing intelligence, we need the manufacturing system to "understand" the data, which we refer to as "semantics" of the data. If the manufacturing system can be represented at a semantic level, the data will become knowledge to the machine and enable it to be ready for exchange, interrogation and reuse. There is current work taking place to upgrade manufacturing systems to a semantic level but this is still at an early and enabling stage. This fellowship aims to effect a step change in manufacturing intelligence, to support rigorous semantic exchanges between different manufacturing phases, and to allow formalisation and reuse of new/existing knowledge from advanced manufacturing.

The proposed research will build a novel semantic infrastructure for advanced manufacturing by supporting knowledge representation, interrogation, reasoning and exchange for smart design, manufacturing and measurement of advanced products. The focus will be on the development of a toolbox to formalise knowledge in/between design, manufacturing and measurement, especially for additive manufacturing (AM). The resulting semantic infrastructure will allow the machine to "interpret" the meaning of the data/information. To be more specific: how the design parameters (geometries, tolerances and materials) are related to each other; how the design parameters relate with the AM process/post process parameters (layer thickness, build orientation); and how the design and process parameters contribute to the measurement details (methods, calibration, etc.). The work will provide a new universal language for any data/information involved in a manufacturing value chain, and will permit a comprehensive infrastructure to digitalise the fast growing AM industry.

It is expected that the proposed work will have economic impact as well as benefiting industrial and academic communities. The novel semantic infrastructure will provide a smart language to formalise AM knowledge at different stages, providing a well-suited knowledge management tool to enable right-first-time AM production. The proposed technology may open an untraveled path in providing a smart toolbox to enable designers, AM engineers and inspectors to truly communicate with each other in an intelligent, robust and productive way.

Working closely with industry partners (Reliance Precision, Renishaw and NPL) the proposed programme will develop customised knowledge blocks to support decision-making and semantic exchange. Using the industry partner's expertise in AM design, process/post process control and measurement solutions as an input, customised knowledge blocks can then be developed using the semantic toolbox. The developed semantic exchange protocol will ensure that the knowledge blocks can be exchanged between different production stages. Equipped with the developed reasoning mechanism, AM designers and engineers can interrogate the knowledge blocks to retrieve answers such as how each design parameter is related with specific process parameters, and how to choose the series of measurement parameters from the specification. It will also enable an accurate interpretation from the design semantics to the process control semantics, and from the design semantics to the measurement semantics. Right-first-time of designing and process parameters can then be ensured, therefore greatly reducing power usage and failure rate, lowering the cost of production and making a more environmentally-friendly manufacturing technology. It is also expected that the developed semantic infrastructure will facilitate a quicker adoption for SME's in manufacturing, as the development of customised knowledge blocks will reduce the cost of employing skilled and experienced AM designers, engineers and inspectors of which there is a national and shortage.

Advanced manufacturing and industrial digital technologies will directly benefit from the novel semantic infrastructure. The proposed work will also be beneficial to those in the interdisciplinary areas where advanced informatics is applied in different disciplines such as biological informatics, chemical informatics, healthcare informatics and smart cities. To demonstrate the great potential of the semantic infrastructure to a wider academic community, visits to academic partners (University of Sheffield, Huazhong University of Science and Technology, China) are planned to exchange expertise and share knowledge. Throughout the duration of the fellowship, academic engagement and dissemination activities such as publications in multidisciplinary journals, presentations at conferences, seminars and workshops are scheduled.

To promote the general public's understanding of our research and how the research will impact the economy, the PI and team will engage with the general public via hosting workshops, presenting our research outputs at public research activities at UoH, seeking media opportunities through UoH's Press Office, and sharing our research highlights and vision through the Women in Science and Engineering (WISE) network, LinkedIn and other engineering and science news bulletins.