June 2019

61 www.drivesncontrols.com June 2019 GAMBICA SUPPLEMENT n maintenance, where the machine provides an early warning of the need for attention, allows the vendors to make much more effective use of their service teams by planning site visits in advance and by ensuring that such visits are only made when definitely needed. Effective management of machine vendor service teams is important today and will be even more important in the future. This is because constant pressure to reduce costs means that fewer and fewer manufacturing companies can afford the luxury of an in-house maintenance team with the wide range of skills needed to maintain all of the equipment on a site. Instead, companies will in future rely on the machine vendors to fulfil their service needs, so those vendors that use industrial analytics to optimise service delivery will have an enormous competitive advantage. It’s clear that a strong case can be made for implementing industrial analytics, but what does this implementation involve? For most machine vendors, the first step is to choose an automation partner with proven expertise in this new and fast-developing field. Weidmüller, one of Europe’s leading and most experienced proponents of industrial analytics, is just such a partner. The actual implementation process involves multiple stages, typically spread over a period of several months. The first stage is a discussion with the machine manufacturer to explore possibilities, define the targets for the project, and carry out a preliminary investigation of the problems to be solved. The next step is for the industrial analytics partner to evaluate the data available from the machine to verify that that it is sufficient and of high enough quality. While, on the face of it, this may sound straightforward, in practice, data evaluation requires considerable expertise from the analytics partner to judge reliably where to draw the line between insufficient data and too much. It is also an interactive process that may, for example, lead to recommendations for adding sensors or enabling the built-in data collection features of machine components such as variable-speed drives. After the quality of the data has been verified, the project proceeds to the proof-of-concept stage, where experts from the industrial analytics company learn in depth about the operation of the machine and its characteristics, and work off-line with sample data to develop custom algorithms that will make it possible to achieve the project’s targets. This is one of the most crucial stages, and can only be carried out successfully by an analytics company with the wide experience of machines and control systems needed to understand all facets of their normal operation and potential failure modes. Algorithm development is a task that’s far from trivial and, as yet, there are very few companies with a proven track record in this area. The penultimate stage is to carry out a pilot study on a functional prototype machine and, during this stage, the algorithms are likely to be refined further. Also during this phase, the benefits accruing from the use of industrial analytics become fully tangible for the first time, and the machine vendor is in a position to start demonstrating them to potential machine purchasers, thus starting to reap the rewards of their foresight and investment. When it has been fully tested and proven during the pilot study, it is time for the last and most exciting phase: rolling out the industrial analytics system to production machines. For machines in series production – which are the main target for industrial analytics – this final step is typically easy and inexpensive, as the additional hardware requirements are minimal in most cases. As we’ve seen, thanks to the use of industrial analytics, machine vendors can offer their customers what they really want – productivity, reliability and, above all, profit. And the machine vendors can also benefit themselves, not least by making their service operations more efficient and effective. n BELOW: Machine builders and users can both benefit from industrial analytics

RkJQdWJsaXNoZXIy MjQ0NzM=