Under the umbrella of digital manufacturing, Meggitt Applied Research & Technology Group’s graduate engineer - global programme, Theo Abbott investigates how knowledge gleaned from large-scale data capture in a flexible production environment can be applied to high value manufacturing.
Traditional automation works well for large quantities of similar products, however, most aerospace parts are not made in sufficient volumes to require dedicated production lines. Instead they require a level of flexibility and operator dexterity not often found in high volume manufacturing. This leaves few options for improving cost efficiency; one option is exploiting intelligent, data-based manufacturing that supports, upskills, and shares the expertise of operators.
Digital manufacturing seeks to improve flexibility, traceability, and visibility; these improvements provide the opportunities and tools needed to increase production efficiency. Automation is often included to boost output by eliminating menial steps, allowing operators to focus expertise on core tasks. Systems better capitalise on the skillsets and knowledge of the most capable operators and share this learning with others across the workforce. This approach has been given many names recently: digital manufacturing, Industrial Internet of Things (IIoT) and Industry 4.0 (I4.0).
Meggitt has been leading a collaborative project between the Advanced Manufacturing Research Centre (Sheffield), the Manufacturing Technology Centre (Coventry), and IBM. This project is supported by the Aerospace Technology Institute. M4 (Meggitt, Modular, Modifiable, Manufacturing) investigates how knowledge gleaned from large-scale data capture in a highly flexible production environment can be applied to British high value manufacturing. This knowledge is being leveraged to produce a step change in the costs associated with maintaining quality and traceability while enhancing transparency and productivity. M4 is now in the integration phase where technologies that have been demonstrated individually are being proven as part of a pilot production environment.
Process and approach
The ethos of M4 is digital data capture at every stage of the manufacturing process. The data is automatically captured rather than burdening the operator with the additional work of logging it manually. Additional supporting systems save time thereby boosting productivity, even if the data is not used. Learning and planning algorithms facilitate increased visibility of product flow and provide more detailed, relevant and accurate key performance indicators, such as lead-time and equipment utilisation in real-time. Transport routes, tools, operations, and conditions are logged throughout the process giving detailed traceability and transparency for every product. Analysis is performed to identify issues and areas for improvement.
The work schedule is analysed using algorithms to find the best production plan before the required components are reserved, picked and validated using computer vision. Components are loaded into smart boxes which monitor them during the assembly process; these boxes are then delivered to workstations by autonomous intelligent vehicles.
A key benefit of M4 is increased product flexibility; this improves assembly lines by increasing the variety of work that can be produced efficiently. This is achieved by designing lines to take advantage of smart workbenches and intelligent distribution of work to minimise the risk and cost associated with reconfiguration to produce many different products. These workbenches can adapt to different parts in seconds and provide digital work instructions while automatically capturing production data. The data captured from skilled operators provides an opportunity to learn lessons such as the most effective order to torque bolts or the fastest ways to apply adhesive. These best practices are automatically flagged and shared with production and design engineers helping to improve build processes and facilitate knowledge transfer to new operators.
Alongside data capture, workbenches close the loop between instruction and action by validating every step. This prevents mistakes and ensures that experienced operators are aware of any subtle changes to a design. Digital work instructions allow operators to be more flexible, up-skilling them to work on other jobs with less dedicated training. The combination of interactive instruction and validation of correctly completed steps provides the elements required for learning and offers a unique option for teaching apprentices. Smart tools automatically apply and confirm the correct forces and data is captured about every operation at the individual part level. The depth of this part by part traceability can be thought of as ‘Product DNA’.
Existing solutions, from basic business management systems, to the most advanced planning systems, use production estimates to compare schedules on their ability to ship orders; there is little accounting for unexpected and unavoidable disturbances such as supplier lateness, machine breakdown, or scrapped components. M4 improves the robustness of production plans by utilising a dynamic, probabilistic simulation model which acts as a digital twin of the factory. This allows uncertainty to be modelled and managed using statistical methods enabling potential schedules to be evaluated for characteristics such as robustness to disturbances. The use of real historical production data, which is continuously captured, allows the simulation to iteratively improve. This enables sophisticated algorithms to produce ever more robust production plans.
The challenges ahead
A key novelty of M4 is the application of digital manufacturing to established production environments that specialise in low volume complex parts. This capriciousness requires large scale data capture; enormous datasets of thousands of operations must be produced requiring drastic increases to sensing capability.
It is inherently difficult to appraise the actual monetary value offered by differing degrees of digital traceability and more flexible production process; some benefits can be calculated from the start, and clearly there is an advantage from understanding flows and identifying issues earlier, though this is mostly qualitative until such a system is fully implemented. Much of the improvement offered by M4 comes from the single integrated system providing data and learning in a novel way. As such, the true value will not be completely appreciated until several years after deployment. That said, the question has evolved from ‘should we incorporate systems to improve traceability, transparency and flexibility?’ to ‘what level of integration and granularity provides the ideal payback period compared to a more quantitative investment?’
Share the learning
The learning from M4 is being shared to accelerate the adoption of digital manufacturing; to this end the M4 demonstration environment has been created to show the developed technologies and how they integrate together.
As M4 moves forward, the focus is shifting from R&D to industrialisation of key technologies, such as intelligent adaptive workbenches, smart boxes, digital twins, and simulation analysis. New challenges commensurate with moving from technology readiness level six through to nine, such as reliability and workforce acceptance must be overcome before M4 is deployed into a real production environment. In light of the recent announcement of a Meggitt supersite in Ansty Park, Coventry, now is a golden opportunity to begin with a clean sheet and build a true factory of the future. Such a factory, designed to leverage the full potential of M4, will support the UK’s continued place as a global leader in aerospace manufacturing.