Successfully digitally network maintenance and production
The core topic of digital maintenance is the optimization of the runtime and thus the utilization of existing plants based on a wide variety of data. Kai Uwe Kuhn, Partner and Account Manager Hexagon EAM at RODIAS at Rodias GmbH, reports on the challenges and opportunities of networking.
What challenges do companies typically face when implementing digital maintenance?
The first challenge is certainly – if it has not already been done – a comprehensive consideration of the status quo: What assets are in place, what is their overall performance, where are the focal points of errors, and what would be realistic improvements? After such an examination, it usually becomes apparent that the answers to these questions are anything but trivial. Failures, underperformance, the human factor and structural bottlenecks are not very transparent, often occur randomly and can hardly be captured ad hoc in their entirety. However, the creation of such an assessment model is an essential prerequisite for optimization: What cannot be expressed in figures can hardly be used as the basis for a budget.
How can binding valuation bases be found?
In order to gain a first impression, figures from cost accounting as well as a consideration of production output are useful here. Here, the current output of a plant (e.g. 600 units per shift) is compared with the theoretically possible output (e.g. 800 units/shift), which allows an initial assessment of the optimization potential. In conjunction with cost accounting, the possible optimization potential in euros can then also be determined immediately.
Example: In the above example, you calculate an “overall plant effectiveness” (a current output of 600 units versus a target of 800 units per shift) of 75%; according to cost accounting, one hour of production with 40 employees (wages, depreciation, ancillary costs, excluding materials) costs approximately 5,000 euros. A ten percent increase in production output would achieve a monthly cost advantage of 80,000 euros here, or reduce manufacturing costs by 12%.
What is also often forgotten in this context – it is often not only the machine downtime, but also the quality of the output that drives up manufacturing costs. Even if the machine is doing its job reliably, the operating resources clamped in the machine are also subject to wear and tear, which must be absorbed by early maintenance. Depending on how much added value has already gone into the semi-finished product, the scrap costs can often add up unnoticed to considerable amounts.
In practice, it is advisable to start with such an estimate and to involve controlling at an early stage; later, these models can be refined with scenarios, since increases in productivity are often implemented in steps (e.g. by changing the shifts).
Why is condition-based monitoring so important for optimal maintenance?
In maintenance – as in all operational areas – the aim is always to optimize operating or manufacturing costs. This results in a field of tension.
Simplified considered
maintenance carried out too early leads to increased operating costs: maintenance carried out at 85% of the service life generates 18% higher maintenance costs, which are made up of working time and increased spare parts requirements
late maintenance leads to unplanned downtime, which is even more expensive – in addition to regular maintenance, there is often waiting time and overtime for production staff, additional freight costs and ultimately damage to reputation if delivery is too late.
So it goes – the earlier the maintenance, the lower my downtime risk, but the higher my maintenance costs. To stay with the example above, with maintenance already occurring at 85% of downtime (the ideal interval between two maintenance visits) – maintenance at 95% of downtime would generate only 5% higher maintenance costs compared to the ideal maintenance timing, instead of 18% – a saving of 13%. However, one would already come dangerously close to the failure of a machine in production.
So how can I achieve this additional 10% downtime (and thus 13% lower costs) without risking a loss of production?
Certainly, some of this potential can be captured through optimized inspections – as long as they don’t require as much intervention as the maintenance itself. Employees (or external companies) still have to be available for these inspections, and even then mistakes still happen because the evaluation criteria are often too subjective.
Here it is now much cheaper to let a suitable sensor system with downstream data interpretation do the job. The costs of such an installation (see also above) are in the long run a small fraction of what manual (and therefore often subjective) inspections can achieve. With suitable systems, you not only save on manual inspections, but also increase the information quality of the measurements, acquire knowledge and can thus gradually reduce time-consuming and costly maintenance.
The final last step in this chain is then the connection to a maintenance/maintenance system, which automatically triggers inspection and maintenance orders based on the interpreted and prioritized information received from the sensor system. This saves you having to monitor the data; the system then controls the maintenance on its own, without any further control effort.
What about networking right down to the machine control level?
In the first step, an evaluation of machine control data is particularly useful wherever machines are critical for production performance – the famous bottlenecks. This is especially true when these machines generate unplanned downtime, erratic performance or poor output quality due to their type, age, environment or other circumstances.
Machines are controlled and networked primarily based on production requirements – they “know” what feeds, forces, or other parameters are needed for the next batch of production, adjust them, and provide feedback on whether the adjustment was successful. With this – rather production-related – data, one can already make initial observations.
For example: if a release signal suddenly comes later than usual, this often has a mechanical cause – there is already a need for maintenance here. In everyday life, however, these subtle signs of an emerging problem are rarely recognized, and thus represent a first missed opportunity to service the machine at the optimal time.
In addition to this production-related data, there are other parameters on the machine (or on the operating equipment) for which sensors are not normally provided. The recording of maintenance-relevant data such as temperatures, pressures, distances or vibrations can generate significant added value here, since maintenance can be planned more precisely on this data basis, and maintenance costs as well as the risk of failure can be reduced.
How can this gap be closed?
As a first step, it is worth comparing the existing machine data with documented failures. If these are foreshadowed in the regular machine data, you have a first indication. The second step goes deeper: a failure in the regular machine data often occurs only after damage has already occurred, which could perhaps have been avoided with targeted sensor technology. This then has to be established.
Now, it is often the case that the first step already causes problems: the machine data is not collected and thus also not systematically evaluated. Based on IoT technologies as well as due to the price drop of cloud storage, this is no longer a technical problem today. One can implement data analysis and visualization solutions with specialized companies promptly and easily. Services such as automatic interpretation of this data via artificial intelligence can subsequently be built on top of it. All in all, this leads to impressive increases in performance, as it is now possible to warn and maintain quite accurately.
Why are many companies still finding it so difficult to drive forward the digital transformation in maintenance?
In practice, we see that many companies are still unaware of how little effort it takes to install these technologies. Of course, day-to-day business always comes first, and it’s also not easy to find a partner who understands the concerns of production, maintenance and in-house IT at the same time. Moreover, in the maintenance area (unlike in areas such as finance, marketing or production), it is still rather unusual to bring in external companies for consulting. Such optimization is also – contrary to frequent assumptions – not a matter that consumes a lot of money and time or blocks internal employees. Both in the recording and consideration of the problem, for a profitability calculation, as well as for the IT-technical realization, external companies can quickly make reliable statements through their experience and achieve first visible progress in a short time.
How can companies improve their starting position in maintenance and digital services through IoT and sensor technology? What previously often untapped opportunities lie in the data?
In addition to improving plant uptimes from a pure “machine view”, data can also help optimize production runtimes. – Not all downtime is caused by faults on the machines. With data analytics, correlations that are not clear at first glance also become visible.
For example, on Monday mornings, production does not start until an hour after the start of the shift because the employee first has to make production preparations. Such waiting times are usually invisible to the production or plant management, as they often only see the quantities booked into the ERP during the day. Everything seems to be going well, apparently there is no explanation as to why only 80 percent of the otherwise usual daily output was achieved.
In this case, the data from the control technology can quickly provide information about when the machines were stopped, whether there was an error, and whether it is the same every Monday morning. By capturing as many of the accompanying variables as possible, these – previously invisible – facts are made visible in corresponding dashboards.
With the help of the continued digitization of maintenance and production, as well as with the closing of the digital gaps between these two worlds, a significantly higher overall transparency and, as a consequence, significant production increases and cost savings are achievable.