Maintenance of critical production line equipment at Ford Motor Co. can slow down the manufacturing process, costing significant time and money. Too much maintenance time can be as bad as too little, so Ford and its equipment suppliers are seeking ways to optimize maintenance schedules through analysis of data from 5G-connected sensors attached to the machines. Combined with related technologies such as machine learning and augmented reality, 5G has the potential to transform the way manufacturing equipment is maintained.
Intelligent maintenance is a key use case for the 5G Enabled Manufacturing (5GEM) project, a consortium of eight partner organisations:
Are exploring use cases for 5G private networks in different manufacturing environments. The project, supported by the UK government’s 5G Testbeds and Trials Program, is developing 5G-enabled intelligent maintenance for vacuum furnaces in a test environment at TWI near Cambridge, UK.
Car manufacturers use vacuum furnaces to bond two metals using a filler metal that melts at a lower temperature than the metals being joined. Properly maintaining the furnace is vital to ensure a high-quality bond, but maintenance typically requires the machine to be taken out of service, interrupting production. Ford and its manufacturing partner Vacuum Furnace Engineering are using 5G connected sensors to remotely monitor the vacuum furnace’s performance, state of health and environmental factors in order to streamline the maintenance process.
As Ian Jenner, Director of Control Systems at Vacuum Furnace Engineering (VFE), describes in this Q&A, predictive maintenance based on data can save a manufacturer time and money.
What problems are you trying to solve with intelligent maintenance?
There are many expensive machines with complicated maintenance schedules involved in automobile manufacturing processes. We don’t want to carry out unnecessary maintenance which causes downtime on the production line, but equally we don’t want to wait until there is a fault, which can cause longer and more expensive delays.
It’s much more efficient to use predictive maintenance, whereby data from sensors attached to a machine tells you when it’s about to need servicing. With this project, we’re collecting large quantities of data from vacuum furnaces in different environments and processing that data centrally at VFE using machine learning and AI to develop predictive maintenance algorithms together with Lancaster University which can be deployed in the factory.
Project partner ATS Global uses the 5G network to deliver a hybrid cloud directly to the shopfloor, enabling the 5GEM consortium partners to deploy AI at the network edge. 5G-enabled sensors attached to machines on the production line then feed the predictive maintenance algorithms in real time to indicate the optimal maintenance schedule for each machine.
Why are you using 5G-connected sensors?
Firstly, 5G provides a highly secure connection, which is important because we are collecting commercially sensitive data from our customers’ sites and processing it centrally. In the 5GEM project we are using a 5G mobile private network (MPN) supplied by Vodafone, which gives Ford full control over who has access to their network and their data. They can work confidently with trusted suppliers and partners such as VFE and TWI.
Secondly, in order to feed the predictive maintenance algorithms deployed in the factory we need to collect large quantities of data from multiple sensors in real-time. 5G provides high bandwidth connectivity to support these requirements. With an MPN we are not limited by the network design choices relevant for public mobile network users, so in principle we can tune the network to support the higher uplink bandwidth needed for our use cases.
SENSOR MEASUREMENTS FROM A VACUUM FURNACE
What are the other applications of 5G for intelligent maintenance?
We’re also using 5G-connected AR/VR headsets to provide remote expert support to our customers. With this setup we can guide on-site staff to fix relatively simple faults which would previously have required a callout of a VFE engineer to the factory. 5G’s low latency is critical for a practical and comfortable VR experience.
What challenges and opportunities lie ahead?
We are still somewhat constrained by the lack of standardized, in-built connection points to install and power sensors. In the future, we envisage supplying machines with such connection points ready for 5G sensors to be added as and when needed. For example, if a machine is performing poorly, we can quickly add extra sensors to monitor its behaviour remotely in real-time to diagnose the fault.
We are also looking to build digital twins of processes to help remotely located experts work out solutions without stopping the production line.
Downtime can cost £100,000 [$138,000] per day in lost production for high-value vacuum furnace processes such as diffusion bonding. Ultimately, we are looking at 5G-enabled intelligent maintenance to save our customers significant money in downtime and scrapped loads.
This is the second in a series of interviews about the 5GEM project. In the first, Chris White, Manager of 5GEM at Ford, explains why enabling real-time process analysis and control in manufacturing is critical.