Skip to main content

UQ&M Study Group 4

The fourth Uncertainty Quantification and Management Study Group with Industry will take place at Warwick Centre for Predictive Modelling from 20-22 March 2019.

Organisers: James Kermode and Peter Brommer (WCPM); Alex Diaz De la O (Liverpool); Matt Butchers (KTN)

Registration is open via EventBrite - £15 for Warwick participants and £35 for non-Warwick participants.

The Warwick Centre for Predictive Modelling, the University of Liverpool Institute for Risk and Uncertainty and the Knowledge Transfer Network are delighted to host the Fourth Uncertainty Quantification & Management Study Group with Industry, in association with the HetSys Centre for Doctoral Training. This Study Group will be number four in a highly successful series of events tackling some of the hardest problems in UK manufacturing.

Industry Use Cases

  • GKN – Understanding Metal Spray Deposition Variation. This challenge aims to explore the relationship between particles coming from a wire-fed flame spray deposition gun and the resultant variation and repeatability of the deposited material. The current processes are heavily reliant on either skilled operators or automation using a measure - spray - measure feedback loop. Efforts have been made to reduce variation by better control of process inputs. This challenge wishes to use machine learning to better understand the effect of input variables such as temperature, humidity, nozzle condition, particle characteristics.
  • Rolls Royce – Probabilistic Frameworks for Nuclear Integrity Assessments. This challenge is to explore probabilisitic integrity assessment approaches for a nuclear use case, an industry that typically has not used structural reliability methods as they are perceived to be less safe than deterministic design methods. This challenge specifically considers a typical nuclear piping arrangement subject to thermal and pressure cycling. The system has an initial design life based on anticipated operational transient history and a deterministic view of cyclic loading (pressure and temperature). This deterministic view is also used to set inspection intervals and the ensuing inspections are used to justify continued operation - this could be a highly conservative approach.

  • AWE – A Computational Framework to Address Materials Ageing Behaviour. This challenge is to investigate determining the likely breakdown products for a given generic polymeric material using a combination of AI, machine learning and modern computational chemistry. Incorporating the long (years) timescale ageing behaviour of any new material is a considerable aspect of the qualification process. This challenge focusses on the use case of a self-contained decomposition mechanism for nitrocellulose ageing. A framework is desired which can efficiently and accurately reproduce this reaction mechanism and then incorporate the effects of uncertainty on the process. The second stage would aim to identify key ageing mechanisms, and to predict probabilities of individual reaction products appearing.
  • BAE Systems Maritime – Developing a Framework for Uncertainty Management in Cyber-Physical Power Distribution Test Environments Details TBC

Context: Manufacturing contributes over £6.7 trillion to the global economy and the UK is a major contributor. Virtual design through modelling and simulation is extremely attractive for manufacturing and design; it can significantly speed up the development cycle, and use simulations to substitute some real-life tests necessary in many cases for approval of new materials.

For virtual design to be useful, estimates of uncertainty should be present. A particular challenge is in developing decision support tools which are able to realistically incorporate uncertainty estimates from multiple sources. How uncertainties propagate between scales (atomic, macroscopic, component / system level) and ultimately how these feed into a decision-making process in an industrial and commercial context is a difficult, open challenge; what might be termed as models to decisions. What is needed is input from not only the mathematical sciences but also psychologists, data visualisation experts, and engineers.

The Study Group: The Study Group to be held at the University of Warwick between the 20th - 22nd March 2019 will bring together a multidisciplinary group of researchers and industrialists from across the UK to work on 3-4 problems over three intensive days to come up with solutions. These industry problems will focus around a central theme of decision making in virtual design and will be sourced from sectors as diverse as semiconductor fabrication, welding applications, gas turbine blade manufacture, drug discovery etc. Previous workshops of this scale have attracted 70 researchers and 4 problems per workshop.

The buzz of conversations that took place to refine the problem statements, propose solutions, and test ideas was exciting – equally so for the industrialist, both as observers and as participants, because we were treated as an integral part of the study group. Through these conversations we jointly developed a better understanding of the problem whereby the researchers could seek appropriate solutions. A stimulating aspect of the self-organised researchers was that they were from different universities and different disciplines; yet by working together perhaps they have created opportunities for new collaboration beyond the study group.
Sanjiv Sharma, Expert for Modelling and Simulation, Airbus

This workshop will follow the highly successful study group format which was highlighted in Professor Philip Bond’s Review into Knowledge Exchange in the Mathematical Sciences and an effective mechanism which should be ‘expanded in scope’.

The general format for these workshops is that representatives from industry will present their problems on the morning of the first day. The multidisciplinary research audience question the industrial representatives and decide which challenge they could offer input into. The group breaks into teams (with a nominated academic leader to facilitate the work) and over the next three days code, model, sketch solutions to the problem in an open collaboration. The group meets again at the end of the three days where the teams feedback the progress made.

Get involved: There are two ways to be involved in this group

  • Researchers: please complete the registration form. There is a nominal charge to participate. This will cover your accommodation, lunches and dinners for the three days.
  • Companies: if you are a UK business and wish to pose a problem, please contact matt.butchers@ktn-uk.org with a short description of the challenge.

Participant Information

Information Booklet

Study Group Programme

The Study Group will run from the morning of Wednesday 20th March to early afternoon on Friday 22 March. An outline programme is given below.

Tuesday 19 March
18:30 Pre-study Group Drinks (Varsity Bar, on Campus)

Wednesday 20 March
09:30 Registration WCPM, University of Warwick
10:00 Welcome, Introductions and Problem Presentations
11.30 Work on Problems
17.30 Dinner at WCPM

Thursday 21 March
09.00 Work on Problems
18.30 Workshop Dinner

Friday 22 March
09:00 Work on Problems
14.00 Final Summary Presentations

General Information

All participants will be provided (at no further cost other than the initial registration fee) with lunch, dinner and coffee from Wednesday morning until the workshop finishes on Friday afternoon.

Accommodation

Accommodation will be provided in the Arden Conference Centre at the University of Warwick. If you have requested accommodation you will have been sent details of how to reach the conference centre and when you can check in.

Financial Support

The event is supported by EPSRC under Warwick's Impact Acceleration Account, the Uncertainty Quantification and Management in High Value Manufacturing special interest group of the Knowledge Transfer Network and the School of Engineering at Warwick.