Industry 4.0 promises improved performance and enhanced maintenance activities, but many companies are often unsure of where to start. Mike Lomax, electrification manager at Bosch Rexroth, spoke at the Maintec show at the NEC in November to offer some practical advice on how to add sensors to existing equipment

OE January 2019 -Mike Lomax.jpg

Maintenance engineers are presented with new ways of working through Industry 4.0 – monitoring assets, picking up on trends, and carrying out maintenance activities before those assets experience problems and ‘go down’.

But, “how can we get started with industry 4.0 and monitoring?” asked Mike Lomax, electrification manager at Bosch Rexroth, during the Maintec show. “I’m sure when you go to your finance people they say ‘yes, have an unlimited amount of money. There are no limitations to how much money you can have’. No: they actually want to see something working.

“Let’s just consider one aspect of maintenance. You have to sense what is going on in the process, you have to analyse it, and you have to act. That is basically what we want to do.”

Delegates were shown a video of a ‘traditional sensor’ whereby an engineer walked around a baggage handling line and put his hand on a motor to check the temperature and vibration. “Did you catch that?” asked Lomax. “He put his hand on the motor, checked its temperature, checked its vibration, and that is what we all do instinctively.

“To sum all that up, when you’ve got a vibration sensor, you have got two ways to look at it. You can look at the RMS and that is what generally the British standards are applying. In other words: how loud and fierce is the vibration. Or you could do something a little more sophisticated, a fast Fourier transform on the vibration and convert it to a series of data points and amplitudes. It’s a way of converting sound into a series of numbers that is compact.”

Lomax gave an example of a device from Bosch Sensor Technology – it comes from the sensors that are in smartphones and allows the user to “write an analysis in the unit”. “From that, you can take the information, move it around and it transmits straight to the cloud – an IBM Watson System – but it can be any cloud or a server in your factory. In fact, you don’t even need an actual server, you can just use your phone to transmit it to the server where you can analyse it.”

Another option is through the use of free mobile applications such as Bosch’s Virtual Connected Industrial Sensor Solution (CISS) app (https://is.gd/mabifo). Lomax asked a delegate to cup a sensor and breathe on it, while Lomax held his smartphone on stage. The active mobile app, which Lomax didn’t name, showed an increase in humidity from around 60% to 71%. “You see, really minimum costs to get you started, and that handles eight different functions, including temperature, humidity, magnetism, orientation, light level and vibration,” said Lomax.

Despite the wide availability of sensing technology, Lomax does not advise starting there. He continued: “What I am in favour of is using information that is already on your machines. You already have a variety of hydraulics, perhaps pneumatics and electric motors. You can use those as virtual sensors to take data into your information system.”

Next step: analysis
Once there, the numbers can be crunched by a system such as Bosch Rexroth’s IoT Gateway software (https://is.gd/lifola). “What that does is it takes sensor information and passes it to either visualisation and some data analytics,” Lomax said.

Finding meaning in the endless rows of figures is the real goal of Industry 4.0. “If you’ve got five machines with 10 monitoring points each, that soon becomes a huge amount of data,” Lomax added. “You could employ data miners, [or] you could employ teams of PhD students to analyse that data.”

Another option is to create a machine ‘health index’, either by tracking key parameters on one’s own, or by using software such as Bosch’s Nexeed Production Performance Manager (https://is.gd/yomufa). The software gives a complete overview of the current state of individual machines and production system.

For engineers operating some 50 machines, Performance Manager will take data inputs, collect them together in a cloud server in a local server and automatically notify if a parameter exceeds upper or lower limits, as well as performing trends, Lomax explained.

Once this has been done, the intelligence can be sent on to maintenance engineers and teams for action. There are different ways of doing this, but some systems automatically create instructions and task descriptions (https://is.gd/egapim).

So, in this simple implementation, sources of data from production machines are recorded and analysed to determine what needs to be done. Now the responsibility for physical intervention transfers to on-site staff, Lomax concluded.