About BAE Systems
At BAE Systems, we help our customers to stay a step ahead when protecting people and national security, critical infrastructure and vital information. We provide some of the world’s most advanced, technology-led defence, aerospace and security solutions and employ a skilled workforce of 85,800 people in more than 40 countries.
From state of the art cyber threat detection to flight control systems that enable pilots to make better decisions, we never stop innovating to ensure that our customers maintain their advantage. This is a long-term commitment involving significant investments in skills. We also work closely with local partners to support economic development through the transfer of knowledge, skills and technology.
BAE Systems and Made Smarter Technology Accelerator
Andy Schofield – Technology Delivery Director, BAE Systems said: “BAE Systems is on a continuing mission to be a front line developer and instigator of future manufacturing technology. Our customers’ demands for ever more complex and flexible products along with ever shortening timelines and challenging cost targets makes alignment with the Made Smarter Technology Accelerator and Digital Catapult a clear match to achieving our long term goals, therefore sustaining us as a driving force in both UK and global manufacturing markets.”
Challenges brought to you by Industry Challenge Owner – BAE Systems
BAE Systems has two challenges for relevant startups. Applicants must choose only one challenge from the programme’s 14 challenges.
Read on to find out about our challenges:
- Challenge 1: Scalable artificial intelligence for visual inspection systems
- Challenge 2: Dynamic workflow management
Challenge 1: Scalable artificial intelligence for visual inspection systems
The Factory of the Future programme at BAE Systems is aimed at creating a blueprint of an intelligent manufacturing environment, production and demonstration facility for the design, manufacturing, assembly, testing and support of future products.
Low volume, extremely high complexity, tight tolerance levels and high facility flexibility are challenges faced during manual sub-assembly tasks. Currently, checks as to whether work has been done correctly are manual, leading to large amounts of non-value added time. This can lead to delays in assembling and potential problems down the line.
A prototype visual inspection artificial intelligence (AI) is in place on BAE Systems’ future collaborative workstations, that utilises camera inspection along with a machine learning (ML) model against which validates the current build state as either a pass or fail. BAE Systems plan to use this as a tool to aid and inform an existing highly skilled manual inspection process that is not being replaced. BAE Systems wants to vastly improve the efficiency of this process rather than fully automating it. However the main issue is that the ML model used, while accurate, is not flexible or scalable. It requires large amounts of data and compute-intensive retraining to extend to new components and build stages, with an increasing risk of catastrophic forgetting as the network is fine tuned. The manufacturing philosophy being pursued within the Factory of the Future revolves around agility and adaptability, the need to accommodate new inspection stages in a reliable and flexible manner is paramount, based on minimal data and fast turnaround.
For this challenge and from applications:
BAE Systems is interested in an AI vision inspection system which can, once proven, roll-out as a validation assistance tool across various assembly, manufacturing and build stages. This will encompass image categorisation, for build verification and quality checks, and inform go/no go decisions.
Most importantly of all, the solution will need to be extremely quick and easy to train up and deploy, requiring minimal datasets to run, as well as scalable to new assemblies and build stages. Overall success/fail accuracy is important but needs to be relevant in context that this will always be used to inform a manual inspector, and not fully automate the process. Therefore, it is the implementation and training aspect that needs to be made as lean as possible whilst maintaining the accuracy within a suitable boundary so it can be trusted as useful by the experienced inspector. Visualisation of the inspection output is also important and the solution must be user-friendly to fully support the remaining manual inspection process.
Challenge 2: Dynamic workflow management
At BAE Systems, supply chain disruptions and other variables including resource limitations and quality failures can have a large impact on the workflow management and scheduling of manufacturing production. Currently BAE Systems uses a logistic workflow management system that is not integrated with existing systems, so any disruptions and changes in orders need to be manually adjusted by local supervisors, and then shared across with suppliers. This results in delays, longer system downtime, and in turn operation slacks as well as lack of transparency to users across the full process.
For this challenge and from applications:
BAE Systems is interested in developing a manufacturing scheduling and workflow management capability that can dynamically react to disruptions and changes through: automatic updates to different variables including supply chain disruptions, resource availability (parts, consumables and processes) and production status optimised real-time scheduling and replanning to mitigate the impact on production informed advanced scheduling for future planning across the supply chain for example, pattern analysis, learning, applying this as continuous improvement
The solution can leverage or build on supply chain connectivity and IIoT, and the exploitation of data. In addition, the solution should consider data visualisation and interaction with operators, and how insights such as operational performance, utilisation rates and availability can be shared.
Involvement with the wider Made Smarter movement
BAE Systems is proud to have been a core player in the Made Smarter movement since the beginning on all key aspects. This includes: with our Technology Delivery Director being on the Made SmarterIndustrial Advisory Board and our Principle Technologist running a number of Made Smarter funded programmes.
In addition BAE Systems also chairs the Made Smarter North West pilot programme and supports the Royal Academy of Engineers on future skills development.
Furthermore, our Factory of the Future development facility is being utilised as a testbed facility for our role as an ongoing partner and Work Package Lead in the Made Smarter Smart Connected Shop Floor Programme.
BAE Systems has been an avid supporter of Industry 4.0 and industrial internet of things (IIoT) technology for some time, and as such we are part of multiple major initiatives to redevelop the company’s manufacturing capabilities in line with our core business drivers. This includes the ongoing BAE Systems’ direct investment and development of a cutting edge development facility known as the Factory of the Future in our Warton Facility.
Challenge focused FAQs
We’ve had a couple of questions related directly to our challenges and BAE Systems, please see below:
COVID-related questions: (a) What are your health and safety procedures for working with startups in the Made Smarter Technology Accelerator programme during COVID/lockdown with respect to physical engagement?* (b) Would the team have to work on-site? If so, where?
At BAE Systems’ core is a company-wide policy due to present restrictions that ‘unless the work in question is deemed business critical’ on-site work or physical engagement is avoided unless it becomes absolutely necessary. BAE Systems also has developed specific infrastructure improvements to enable remote working aspects which it continues to do.
At this moment in time, it is highly unlikely that at any point in the 12 week sprint a face-to-face or on-site scope of work will be requested or required. If any of the sprints are taken forwards to longer tasks however and COVID-19 restrictions are lifted, on-site work would be the Warton W25 Factory of the Future development facility (postcode PR4 1AX) where all the technology being discussed in our requirements is present as well as the key personnel. On site, extremely stringent distancing and safety policies are in effect and anyone attending will be required to undertake on boarding and training prior to entry.
What is your view on cloud computing?
BAE Systems believes cloud technology is core to aspects of our future manufacturing and integration technology development scopes, and we have been working, testing and developing confidence with a number of well-known cloud service providers for some time.
However, at present, our manufacturing specific developments are remaining firmly in the ‘fog’ or internal environments. This is because of the nature of data security requirements we need to handle which means we have less barriers to overcome by running our current technology development in this way. If the successful participant wishes to advise that their solution to our problem requires cloud utilisation, we will need to have a very careful conversation about firstly whether the data we supply can be sanitised for use in a cloud and secondly how the cloud is structured (for example; security in place, physical global location of servers, method of data deletion and proof, etc). We would advise caution in this on a further point, that such discussions can take time, and on a 12 week sprint may limit what can be achieved.
Do you have physical systems/hardware to connect or are you open to new ones?
Yes for both projects, although connections we may be able to offer the participants engaging will be limited for the sprint, based again around security.
Regarding new systems, yes we are open to discussing new approaches be it a package, protocol or overall solution. We have an IIoT structure that is being implemented in our Warton environment and have already down selected a number of core systems and infrastructure process, technologies, protocols, etc. However, we are developing new systems that connect to it all the time and each one brings some aspect that is new which we need to translate for integration.
What MES you have? Would it have API’s for real-time data access? To develop a dynamic scheduler, a good access to all data would be critical.
BAE Systems Air runs a Siemens MES which is in a very early stage of understanding and development with respect to integrating this into our Future Factory ISA95 ‘stack’. We are more focused at this stage in connecting levels 0 (cells) through to 2 (SCADA) to create a central data and processing layer for the manufacturing system, and we are in the process of looking as to how the MES and ERP layers could start being folded into this (hence the problem we have posed so we can learn the best way forward whilst developing this link).
We would argue that access to our full system based on the time allowed for the initial sprint and also current access restrictions as explained in regards to COVID-19, that we would instead of offering specific data offer the successful programme participant an understanding of what we have in place expect some research, testing and recommendations within the time period. The longer project, (if granted) would consider deeper access to our systems as time would not be so short.
Dynamic scheduling – Is the problem mainly with external suppliers or internal operations or both?
The construct as a whole- BAE’s supply chain is complex and full of unique problems. Firstly, unlike a typical mass manufacturing company with extreme levels of volume, BAE Systems has the challenge of high cost, complexity, tight tolerance and sign off requirements for everything supplied, with knock of any individual items supplied being problematic potentially having major impacts to something such as an aircraft build process. Add to that the problem that BAE Systems has specific suppliers, sometimes very small companies who are incredibly specific in what they manufacture or deliver for us, as well as the extreme complexity and long lead time of everything we assemble and manufacture. It is also important to remember the security constraints and that not only does data and information being shared between BAE Systems and its suppliers need to be trusted, it needs to be traceable and proven secure at all stages to the level of rigour that our typical customers and the UK Government demand. Depending on the supplier this can be anything from BAE Systems Proprietary, to beyond Secret classified designs and subsequent information.
Is all of the input data required to do the rescheduling available in your systems already
No, as per a previous question, we think we advised above, for the sprint tasks we would expect the successful programme participant to liaise directly with the industry sponsor stakeholder to identify the data required and its availability. They should come up with sensible proposals and ideas as to how to develop a dynamic scheduling approach (again, it would be unlikely to be able to get large sets of real data available within the sprint task)
If the base concept existed, would you be interested to create business intelligence and a digitalisation strategy that would encompass full end-to-end, immutable, collaborative virtual and physical environment connected digital twins and threads throughout product and application lifecycle management?
We are always interested to review and understand all proposals and approaches, but a more likely useful task here would be for BAE Systems to advise what components to our system already exist, what our roadmaps currently are, and then in turn get a view from the successful programme participant as to the approach they would recommend to joining and creating a system that would meet our requirements and offer a scope for further testing and development if the project is successful.
What are the “consequences” you are willing to see with this scheduling software?
We are interested in seeing and understanding other approaches if integrating them to what we already have is beneficial but what is more important is having someone review what we have and how to best leverage it (adding further aspects can be part of the proposal).