A Multiscale Digital Twin-Driven Smart Manufacturing System for High Value-Added Products

Driven by the ever-increasing demand for performance enhancement, light weight and function integration, more and more next-generation products/components are designed to possess 3D freeform shapes (i.e. non-rotational symmetric), to integrate different shapes/structures and/or to be made of multi-materials. Examples are seen in freeform lens array photovoltaic concentrators, integrated car head-up displays for improving road safety; Lidar (light detection and range) devices for autonomous vehicle; minimal invasive surgery tools for curing aging related diseases such as cataract blindness, osteoarthritis, and saving lives, to name a few. The ratio of required product tolerance to its dimension is less than 1 part in 10e-6, i.e. in the ultra-precision manufacturing domain. The design, manufacture assembly and characterisation challenges for these products are considerable, requiring a step change in the current manufacturing system to achieve the ambitious target of securing industrial efficiency gains of up to 25% (Industrial Digitalisation Interim Report, 2017) as Britain's productivity has long lagged behind that of its competitors. The project will start from an established baseline in a unique flexible and reconfigurable hybrid micromanufacturing system developed from a recently completed EPSRC project (EP/K018345/1) and advance beyond state-of-the-art of system modelling, digital, control and automation technologies. It will research and develop the underlying science and technology for the creation of a new generation smart digital twin-driven manufacturing system that can sense consumer needs and actively self-optimise for customised next-generation high performance 3D products with enhanced productivity in a sustainable way. It will break new ground in understanding intrinsic links among product design, manufacturing and metrology with a novel product/process fingerprint approach. For the first time, a digital twin-driven automation approach which combines feedback and feed forward control algorithms with inputs from high-frequency digital twins of manufacturing process at machine level will be developed to bridge the real and virtual systems and eliminate dynamic errors and thermal errors which cannot be measured by machine encoders even the machine is running at an extremely high operational frequency to meet the required product performance through predictive control. As such, this project will make a step change in manufacturing automation which is based on linear control theory using semi-closed-looped feedback from encoders. As building blocks of the smart manufacturing system, smart multi-sense in-line surface metrology and smart assembly system will be developed to measure complex and high dynamic surface and to precision assemble large variety of parts that are difficulty to achieve before. A novel multiscale business modelling and system analysis approach will also be developed to allow integration of these smart systems and take the live data, model, predict product quality, delivery time, cost, emission, waste, and optimise the performance into the future in different scenarios. The effectiveness of the SMART will be demonstrated through manufacturing the selected demonstrators including minimal invasive surgery tools, Head-up displays, Lidar and solar cell concentrators. The consortium will transform the research outcome to industry and our society through knowledge exchange, training, industrial demonstration and deployment. A unified expertise pool in smart manufacturing established in this project will be a "one-stop-shop" for the UK industry, particularly SMEs, who are keen to exploit the benefit of the project.