How to manage uncertainty and improve production in one of the UK’s largest sectors.
How to manage uncertainty and improve production.
Manufacturing contributes over £6.7 trillion to the global economy and the UK is a major contributor. UK manufacturing directly employs 2.5 million people, generates half of all exports and accounts for three quarters of national business R&D.
In global tables, the UK is in the world’s top 10 in terms of manufacturing Gross Value Added (GVA). According to Innovate UK’s High Value Manufacturing Strategy, 2012 – 2015, the UK ranks second only to the US in the aerospace industry, and two out of the top six pharmaceutical companies are headquartered in the UK, where they support significant manufacturing assets.
The challenge:
Developing new products and improving complex production processes usually involves many time-consuming stages and extensive prototyping. Commercial success depends on careful management and control of risk in the face of many interacting uncertainties.
Historically, risk analysis has been carried out by human judgement, using years of experience. This carries a high personnel cost and also, as contemporary processes speed up, the time to build experience is reduced. Today’s market is fiercely competitive and highly regulated, so there is very little room for margin of error. Failure to understand and manage risks can result in severe financial penalties and damage a brand’s reputation.
However:
“All models are wrong – some are useful.” George Box, 1987
The solution:
We’re entering a new era of virtual design and engineering, underpinned by modelling and simulation, which offers the potential to deal with uncertainty in ways that were not previously possible. A systematic, formal approach allows uncertainties to be quantified and understood, so that designs and processes can be optimised against them. For example, using modelling and simulation to test predictions could reduce the need for expensive physical prototypes, and is a particularly attractive solution for high value manufacturing. It speeds up the production process and reduces costly design iterations.
The approach:
The challenges required to achieve the above benefits work across mathematics, statistical modelling, multi-disciplinary analysis, engineering, simulation and others, which fortunately are all areas of world-class expertise in the UK. Innovate UK and KTN have been leading an initiative to bring key players working on these challenges together, to share aspirations, knowledge and expertise. Several sectors are involved, from aerospace to automotive, energy and defence.
The Special Interest Group:
The Uncertainty Quantification and Management in High Value Manufacturing Special Interest Group (UQ&M SIG) is an industry-led and focused community which brings together industrial practitioners in high value manufacturing with developers of capability in the UK. The aim is to identify industrial requirement, outline current practice, map the UK capability, develop case studies of the application, and grow and maintain a UK community.
The challenges to be met in progressing to full industrial maturity in this field are substantial and include:
Modelling of Data: the uncertainty cannot be mathematically defined from measurement as it is due to lack of knowledge and must be modelled by expert judgement
Identifying and modelling dependency and co-variance of data
Functional dependency: covariance of uncertainty in parameters shared and operated upon by coupled tools and models
Cross-discipline propagation of uncertainty across a large chain of coupled analyses
Handling a large number of analyses across a high dimensional parameter space
Integration of multi-fidelity and multi-source analyses and data to secure realistic and practically useful estimates of uncertainty
Early-stage decision making underpinned by confidence in downstream mitigation strategies for achieving compliance with performance requirements
Inverse mapping across coupled and feedback loop processes: asking which uncertain inputs contribute most to uncertain output
Visualisation of uncertainty to aid the decision making process
Risk assessment: communication of risk and management of that risk
Cultural and educational: engineers need to be trained to an appropriate level in statistics
Our approach has been to tackle each of the challenges through specific industry use cases, below.
Developing early stage designs for compliance with flight performance requirements
Visualisation of uncertainty to aid the decision-making process
Integration of multi-fidelity and multi-source analyses and data
Applying inverse mapping across coupled and feedback loop processes
More information about upcoming UQ&M SIG group activities can be found here.