To infinity and beyond

Originally published by UK5G Innovation Briefing magazine here. 

Manufacturing has identified 5G as a crucial enabling technology, so it’s to be expected that one of the world’s most ambitious manufacturing projects has embraced it.

Factory of the future sees BAE Systems harnessing the best of UK innovation to develop a digitally connected, intelligent factory for future military aircraft production. The company envisages that air forces of the future will operate in a complex and rapidly evolving environment. As that environment changes, manufacturing must change with it. The company therefore needs to develop the capability to adapt manufacturing facilities quickly based on demand. The project supports BAE Systems’ ambitions to deliver a future combat aircraft to the UK in timescales that are more compressed than ever before.

More than an aircraft, the Future Combat Air System is a technological tour de force; it will develop cutting-edge technologies such as radar-defeating stealth, remote piloting and swarming drones. BAE Systems is looking to use augmented reality and artificial intelligence to operate future aircraft, as well as in the manufacturing processes used to produce them. 

The company has a goal to reduce waste and halve the cost and time of building the aircraft when compared with previous projects. The plane will be manufactured at factories in Warton and Samlesbury,  Lancashire, where BAE Systems currently produces the Typhoon and Hawk as well as F-35 sub-assemblies. BAE Systems is working with the University of Sheffield Advanced Manufacturing Research Centre (AMRC) North West, IBM, Machine Tool Technologies and Miralis Data. It’s a typical 5G Testbeds and Trials mix of big and small companies, academia and suppliers to industry.

Machine Tool Technologies is a leading provider of technical support for machine-tool users, while Miralis Data is a logistics software company with an emphasis on low carbon-footprint processes.

The project will use a secure, private 5G network. Aparajithan Sivanathan from the University of Sheffield AMRC North West explains: “The main thing is that UK manufacturing needs to change in a number of ways so we can be ready for the next generation of manufacturing that involves connectivity challenges. We need to have the data and the physical activity happening in a very coupled manner. 5G is not an incremental technology, we see it as a step-change. And these programmes are not just at BAE Systems; it’s a national programme of testbeds.”

The project will prove the value of 5G through a number of use cases; five to begin with. There is an aspiration to add more.

An initial use case is closed-loop control of machining processes. The machine will have sensors linked to the tools, which will monitor the process and generate vast amounts of data. As a part is machined, the data will stream to an artificial intelligence system, which will adapt the machine parameters in real-time to ensure optimum performance and prevent potential defects or quality issues. 

Sivanathan says: “This use case has been quite an iconic challenge for 5G testbeds across the world. Latency is very critical from the point the vibration happens. If you’re making a turbine blade and you pick up vibration from the tip of the spindle, it goes all the way to a sensor, through the 5G network, to a centralised artificial intelligence platform that looks at the vibrations and tries to understand it. Then it needs to make decisions on the fly and send another signal back to the machines to slow down. This happens within milliseconds, otherwise it will be too late to make a correction.”

Linked to this will be a programme to harvest the data from the system and build a digital chain. That’s the second use case: the whole factory is a digital chain. It involves collecting data on different parts of production: machines, how people move within the building, temperature controls, everything. BAE Systems is looking into the use of cobotics, which involves people and robots working together to build an aircraft. The use case employs 5G to link up lots of sensors and the results of the processes to aggregate the data on a central platform.

The third use case supports a reconfigurable factory and monitoring of factory ecosystems. Military aircraft manufacture is a low-volume business. Even a spectacularly successful plane such as the Hawk has been sold fewer than 1,000 times, and that’s with more than 20 derivative models. 

The concept of reusing the same manufacturing line and rapidly reconfiguring the different setups for different products reduces cost lead time. But that involves the rewiring of lots of things. This has to be done within the exacting tolerances of military aerospace. It’s far more demanding than automotive. Laying new data cables for just a few robots can take days. Current military aerospace manufacturing facilities use millions of pounds-worth of bespoke fixed tooling that could not be used for new designs of aircraft.

The fourth use case is asset tracking. As with other 5G projects, this is well understood, but in the exacting requirements of the product that is being made there is an extra level of rigour. The workflow is headed up by Miralis Data. Aircraft components are fragile and valuable. As they pass between several suppliers, each takes ownership or custody of the asset. 

The tracking software records how the part or the asset is handled and stored. If somebody drops a radar component and you only find out when it comes to the final assembly stage and then it fails, the part needs to be re-ordered, which introduces a significant lead time. It’s like having a digital passport; the software will maintain continuous connectivity with the parts, which is called a “chain of custody”. This extends to external suppliers, which will run a virtual version of the BAE Systems network, perhaps using a 5G slice on a public network.

The final use case is virtual reality and augmented reality with distributed setups. Many people will work on the same area of assembly, albeit in different locations, and will be able to see what each is doing. Augmented reality will also be used for inspection tasks.

The 5G factory of the future will be built by the 5G consortium partners. The central core and the AI installation will be in AMRC North West. That’s in Preston at the moment, but will move to Samlesbury Aerospace Enterprise Zone in the next 12 months. The SAEZ sits alongside BAE Systems’ Samlesbury site and it is hoped that it will become home to some of their suppliers. 

The testbed will be a huge boost for the region and will enable Lancashire’s manufacturing industries to increase productivity through the early adoption of 5G. Companies will be able to de-risk their investment by utilising the testbed for trials, which will further benefit BAE Systems’ supply chain. The network will link to Warton and to another one at the AMRC in Sheffield. IBM is working on the AI platform. Of course, security is a prime concern. The project will have its own 5G core, and there will be a low-latency private core within AMRC North West.

The DCMS element of the project runs until March 2022, but this factory of the future is looking a lot further ahead than that. The initial timeline runs to 2025; Tempest will go into production a decade later.