June 2021

33 www.drivesncontrols.com June 2021 COMMUNICATIONS, SECURITY AND NETWORKING n production visualisation lend themselves well to the use of on-premises servers because less data needs to be stored, and less data processing is needed. Predictive maintenance requires huge amounts of data and processing power, so it lends itself to the cloud to churn the data to get meaningful outcomes. Manufacturing traceability probably also lends itself best to the cloud as it requires vast amounts of information to be stored. Likewise, product quality improvement projects are also suited to the cloud because real-time access is not required. Moving to the edge Today, more operators are also seeing the benefit of processing data at the edge – close to the devices themselves – making it possible to gain real-time operational benefits. Using the example of Omron's Sysmac controller again, its local artificial intelligence (AI) engine can analyse data in real-time, allowing for fast reactions to the information at the point of its creation, enabling real-time predictive analysis. However, the limitation of edge technology is that it cannot handle large amounts of data. The best data management approach – which allows OT and IT environments to maximise the use of data created – is often a hybrid approach that sees data processed at the edge and then escalated for storage and further analysis in on-premises or cloud-based servers to gain the most benefit from data. It is useful to know that you can work with the same data in more than one place – for example, using it at the edge first for real-time processing, and then escalating the same data for post-processing purposes. This is possible because data is time-stamped, so it can be referenced and synchronised in different applications. For added data security, it is also helpful to use controllers that can spool data so, in the event of a connection or data corruption issue, there will always be a backup – a particularly important consideration in FMCG sectors where compliance and product traceability is a high priority. For most manufacturers, on-premises servers will offer the best first step on the digitalisation ladder, as it is usually the easiest to implement. Indeed, many manufacturers will already have some degree of IT infrastructure on the factory floor, giving easy access to data. It is also possible to achieve such an approach more cost-effectively if you already have a controller that has the capability to escalate data. Once there, it can be analysed using a choice of low-cost software. Data is opening many manufacturers’eyes as they begin to understand what it can achieve and which traditional processing issues can now be solved. When it comes to storage and processing options, there is a choice of platforms – edge, on-premises or cloud-based servers – to suit all applications. However, bear in mind that storing data for the sake of it is a costly exercise, so make sure that you use the data and don’t let it use you! For engineers, it is reassuring to know that that digitalisation does not require fundamental changes in manufacturing – its purpose is to gather and handle data. You should consider digitalisation just another tool that can give you greater insight into your processes to help improve efficiencies. n Data is opening many manufacturers’eyes as they start to understand what it can achieve and which traditional processing issues it can solve. Machine-builder finds 16-port switches are better than eight The German automation specialist HandlingTech develops and implements modular handling systems for applications from entry-level robotic cell to complex multi- cell systems. One example is a large system for machining shafts, where a machine tool is positioned between two robot cells. The first cell removes the shafts from a grid box in which they are placed randomly, and loads them into the machine tool. The robot uses path planning based on a point cloud to grip the components.The cloud is generated using data froma vision system located above the grid box, and used to calculate the best approach point for amagnetic gripper.The robot, equippedwith an interchangeable gripper system, picks up the component using themagnetic gripper and deposits it on a clipboard, where it is aligned. It then loads the tool that machine the parts. Another robot in the second cell unloads the machined shafts from the tool and packs them into cartons with blister layers. The camera data is processed autonomously by a separate computer. HandlingTech is using 16-port switches to communicate between the robot systems after fing that conventional eight-port switches can reach their limits when handling the large amounts of data involved. For this task, it has chose Murrelektronik Tree16 TX unmanaged switches which have two Gigabit Ethernet uplink ports and 14 Fast Ethernet downlink ports. The Gbit ports are used to communicate with the higher-level controller, while the 100 Mbit Ethernet ports collect data from the various participants in the cell. The two switches used for each robot cell support separate Ethernet and Profinet networks, even if only four or five of the ports are occupied. The industrial PC, robot, and safety and remote maintenance systems are linked to the switch via an Ethernet TCP/IP interface on the outside of the control cabinet which avoids the need to open the cabinet for service work. Most of the ports are occupied, leaving one or two in reserve for measuring tasks. Murrelektronik claims that the 16-port switches have advantages over using two interconnected eight-port switches. For example, the connections between eight-port switches are typically not designed to handle large amounts of data. Although the 16-port switch may be more expensive, it has dedicated controller chip with the necessary computing power. HandlingTech has found that 16-port switches are ideal for handling large amounts of data