Overview

Key figures in production: the data is there, the technology is not (always)

It’s a simple observation: with digitalization, the amount of data is increasing at a breathtaking pace. And those who make the best use of this data gain a huge competitive advantage. For example, because they know their customers down to the last detail – and can address them in a targeted manner. Or because they know in advance when a machine will break down – and can therefore act in good time. Predictive maintenance is just one data-based Industry 4.0 vision that most production managers would probably rather realize today than tomorrow – because it pays to use key figures in production.

Production figures are the exception

In many companies, however, this future is still a long way off. This is because often not even the data that is already sufficiently available today is consistently recorded, merged and summarized. In contrast to most other specialist areas, key figures in production are still the exception rather than the rule. The situation on the store floor is not very transparent. This makes it almost impossible to set up a continuous improvement process. As a result, performance in production falls well short of its potential. This is despite the fact that it would be extremely promising to use key performance indicators in production.

One of the main reasons for this is probably that many companies find it too costly to systematically generate key figures. And it’s no wonder, if you look at the typical situation on the shop floor: In order to obtain key performance indicators (KPIs), employees would have to document all machine and operating data manually – in many cases this is only possible with pen and paper. The information recorded in this way would have to be transferred to Excel spreadsheets, supplemented with commercial data and then individually condensed into the desired values.

MES is the prerequisite – the right integration is the success factor

Studies have shown that the use of manufacturing execution systems (MES ) has increased in recent years. However, the use of these systems is by no means a matter of course. In addition, the degree of implementation depends considerably on the size and sector of the respective company. What’s more, even if an MES has been introduced in principle, it is not suitable per se for generating key figures for production. If this is to succeed, an MES must be properly integrated into the automation pyramid: downwards to the machines and systems, upwards to the ERP system.

From the store floor to the system: machine – OPC UA – MES

The connection to all assets on the store floor is crucial because this is where the relevant data is generated. One challenge for companies is that machine parks usually grow over years and decades and are therefore very heterogeneous. New machines and systems collect a lot of data via built-in sensors and can often pass this on to downstream systems via interfaces. However, this also requires detailed knowledge of this data and its addresses on the machine control systems. This is the only way they can be accessed and processed. And this knowledge usually has to be researched, compiled or coordinated with the machine manufacturers at great expense. Old machines and systems may record no or only very rudimentary information – the start and end of a production process as well as faults – and have no data transmission options of their own. In these cases, however, analog-digital converters or external sensors – such as light barriers – can be retrofitted.

Ideally, the data from the individual machines and systems should be formatted in a standardized way so that it can be processed further. OPC UA (OPC Unified Architecture) has now established itself as an M2M machine protocol that describes machine data semantically and unambiguously. On an OPC UA server, the information supplied by the machines in different formats can be converted into a uniform standard, thus reaching the OPC UA client and from there to the MES, where it is logically processed. Modern control systems and controllers even provide their data directly in OPC UA format – the intermediate layer via the OPC UA server can therefore be omitted.

Use key figures in production: Commercial information completes the picture

However, it is not possible to generate meaningful key production figures from machine and operating data alone. This requires additional commercial information relating to the production order being processed. For example, the absolute throughput time says relatively little about the performance on the shop floor. However, it is much more interesting to know how the throughput time calculated for a specific order – and therefore also the calculated effort – relates to the actual throughput time and what consequences this has in terms of costs. Such target values are fully stored in the ERP system. From there, they should reach the MES unhindered. If the manufacturing execution system is an isolated solution, individually programmed interfaces are often necessary. If, on the other hand, the MES is an integral part of the ERP system, data is exchanged directly – this is generally the faster and more reliable way.

The KPIs themselves are ultimately created in the MES – by aggregating the available data according to the individual interest. In principle, a distinction can be made between real-time values, which allow an ad-hoc view of the store floor, and KPIs, which evaluate the past and therefore allow conclusions to be drawn about the future. Read on to find out which KPIs are useful and how they can be used.