Empowering Digital Manufacturing Teams With Knowledge

Empowering Digital Manufacturing Teams

The fourth industrial revolution has provided the opportunity to obtain huge amounts of real-time data from manufacturing processes, and the potential for almost every surface to be transformed into a sensor for data collection. But does this wealth of data provide the knowledge digital teams need to optimize their activities?

In this film, Senseye's Alexander Hill and Rob Russell are joined by Dr. Hannah Edmonds from the Manufacturing Technology Centre, Make UK's Jim Davison, and Peter Gagg from MCP Consulting Group, to discuss what data manufacturers should collect and how, how existing data sets can be repurposed to drive new insights and greater value, and how processes like maintenance can be redesigned to meet the needs of engineers on the shop floor.



Alexander Hill, Senseye: I would love to say it’s as simple as ‘you have the data and away you go’. Unfortunately, what we see is that a lot of organisations have too much data. So finding out what the correct data is and what you need to understand from that data is the challenge.

Dr. Hannah Edmonds, Manufacturing Technology Centre: There is a challenge that comes with having a wealth of data, we can get to a stage of data saturation - too much of a good thing.

Rob Russell, Senseye: From the perspective of predictive maintenance, the key thing is gathering condition monitoring data that we can assess the state of machines from. But as important as that sensor information is, it’s really important that data is collected with context and at an appropriate point in time.

Dr. Hannah Edmonds: Condition monitoring and analytics can enable a more proactive approach to maintenance. It’s not just waiting until an error occurs, but trying to identify when the condition of pieces of equipment is deteriorating over time.

We can instrument equipment, collect the data, then monitor their operations to get production insights in order to prevent this failure - intervene in advance of the issue.

Peter Gagg, MCP Consulting Group: Everybody’s first thought is that data from the sensors is all you need to start a predictive maintenance program. However, there are many other types of data already existing in a business.

Data from the building management systems, the equipment history, data from PLCs, manufacturing execution systems, OEE systems, etc. Along with the usual vibration, temperature, pressure, flow, current, and voltage.

It’s really about pulling it all together, identifying the parameters that you really need to use within the predictive regime, and looking at how they’re all interconnected.

Jim Davison, Make UK: It is all about really understanding the critical parameters within your process, machinery or product. Understand what are the inputs that can impact whether a part is good or bad, or a machine can run at nameplate capabilities reliably and repeatedly.

Once you can understand that, and know your process, you can capture the data that is important and can influence the parameters you’re looking for. Then display that in a way that is useful.

Alexander Hill: In modern industrial equipment, there is a lot of data already available and already captured. Things like current, temperature, pressure, cycle times. These are things that can actually be used for condition monitoring purposes but often aren't - they’re instead only used to control the process.

We can take these measurements and use them to understand machine health. We build models of the machines automatically, using the available data. Then we help our customers understand what they need to pay attention to from the machine perspective. Without saying ‘look at all of this raw data,' we actually take it away, abstract it, and make sure they can understand which assets - out of hundreds and thousands - they need to pay attention to.

View All Videos Here