Networked Manufacturing

9 May 2016

By Dieter Adam, Chief Executive of the NZMEA

Over the last year, we have talked a bit about ‘Industry 4.0’ – the next wave of developments which have the potential to substantially change the nature of manufacturing. Though its future is uncertain, it could represent huge opportunities for those who can leverage technology, and risks if we cannot keep up.

This month I want to talk more about networked manufacturing. At the core of this is the idea that the physical elements on the factory floor – machines, tools, transportation systems and products are all networked; they can send data to a central control point and to each other and receive instructions from a central control point and from each other.

Centralised monitoring and control of production equipment is not new – the interconnectivity of all elements and the potential to process and learn from large amounts of data, however, is new.

This technology could give manufacturers the ability to monitor and record the performance of pretty much any tool or machinery on a production line in terms of each individual action.

Beyond the immediate quality control, this will allow far-reaching retrospective analyses of potential mechanical failure when the cause of a vehicle accident is investigated, for example.

Other uses could include fully autonomous local logistics, selfmanaging the flow of materials through a process line and factory, or physically attaching further-processing instructions to semi-finished products on a line (via RFID) to allow automatic variations within the production process. It will provide more flexibility to use input from other parties to define customisation without having to manually change tools or adjust production lines.

At the macro level (systems level), Big Data systems will allow the analysis of ever more complex sets of data to identify and predict the occurrence of (rare) events. One of the most potent Big Data systems currently in action is the software that analyses the physical observations from experiments at CERN’s Large Hadron Collider research facility in Switzerland. You can also use that software to predict, for example, how much mincemeat is going to be sold at your local supermarket next Wednesday – or when complicated manufacturing equipment needs to be serviced to prevent failure. If our competitors are investing in this, it may be just a case of when New Zealand manufacturers follow.

There is no doubt that the technology in its current form would bring little benefit to most of our manufacturers, as was the case with very early adoption of steam engines and computers. They were expensive, big, and required special expertise to operate. In a recent survey among SME manufacturers in Germany, only 17% of respondents saw themselves as early adopters of the technology. But it’s not going to be decades away, either. The German company Trumpf, aleading manufacturer of sheet metal processing equipment, is one of the leaders globally in adopting the technology in a manufacturing environment that is quite typical for New Zealand.

However, the technology itself still faces a number of its own challenges, apart from price and the skills required to install and operate the technology. Things like standardisation of data formats and communication protocols still need more work, for example.

We are confident that our manufacturers will approach these new technologies based on practical considerations and impact on profitability, rather the getting seduced by ‘sexy high tech’.

The fundamental question for New Zealand manufacturers remains, however. The key sales pitch for networked manufacturing equipment and systems is the much enhanced flexibility it will introduce into manufacturing processes. Will that jeopardise the niche most of our manufacturers have carved out for themselves in international markets – being highly adaptive and responsive on the basis of high levels of product quality? And if so, do we need to become early adaptors of the technology to defend that niche, and how should policy respond? We don’t think anybody has got the answer to that question yet, but we better start looking for it soon.