This is the easiest maintenance to implement from the company's point of view, but it carries the greatest risk: the wait-and-see maintenance. After a malfunction has occurred, an error analysis is carried out and finally processed. Depending on the malfunction, this can result in considerable downtimes because the problem may not be detected immediately or spare parts may have to be ordered at length.
In order to avoid such downtimes, more and more machines were being maintained and parts replaced as a precaution on the basis of fixed intervals. If this procedure is to be optimized, experience is required to ensure that maintenance is not carried out too late, but also not too early. A maintenance carried out too early is not as harmful as a failure, but there is no question of maximizing resources in this case.
What are the challenges of predictive maintenance?
Predictive maintenance is therefore not only a desirable, but also a worthwhile form of maintenance for machines and systems. In order for this to be effective, however, a few challenges have to be mastered.
Challenge 1: Networking
A precise prediction of maintenance due dates is only possible if the machine manufacturer can regularly receive data from the installed base and process this into information using logic, algorithms, etc. Prerequisite No. 1 for this is a comprehensive networking of machines and systems.
Our experience has shown that it quickly pays off for machine manufacturers to simply deliver every machine "Ready-to-connect". Thus, the first hurdles, such as retrofitting or installation, are overcome by the manufacturing company right from the start.
Challenge 2: Data transparency
Predictive maintenance approaches are always about data. These data are, at least in part, very sensitive for production machines.
Which data types play a role in predictive maintenance?
Machine condition data
Since the maintenance of machines is involved, data on the condition of the machine is of course essential. This can include various aspects such as operating performance, produced quantities, but also information on temperature or oil pressure. The points mentioned are to be seen as examples, since the data vary depending on the machine type and use.
Not only the machine as such can influence the maintenance cycles, but also the environment of a machine. Different environmental characteristics such as temperature, humidity, air pressure, etc. have a direct effect on the maintenance cycle.
The process data covers all data relating to the production process. These can be machine settings, but also data on the recipe of a product or on a special production process. When looking at time periods, trends can be read off and, if necessary, countermeasures can be taken at an early stage.
As already mentioned, all or some of the data is often highly sensitive and manufacturing companies are reluctant to release process data in particular - at least if it is not shown transparently which data is used for which purpose.
The first thing to do here is to define, as a service provider, exactly which data is needed, i.e. which factors are displayed when maintenance requirements are imminent. The next step is to provide the manufacturing company with plausible and transparent insight into the processes that generate a preventive maintenance message from the data.
It certainly also makes sense to involve the operators in the data collection process of a predictive maintenance approach - because nothing is worse than a predictive maintenance that is incorrectly displayed too early or, even worse, too late.
With symmedia's predictive maintenance approach, both sides benefit. The local maintenance manager helps the operator to perform certain maintenance tasks himself and to manage maintenance information. In return, the manufacturer also receives this maintenance information and can evaluate it.
During implementation, teething troubles cannot usually be avoided in an initial phase. However, the understanding of the operator will certainly increase due to a high degree of transparency and the corresponding benefit prospects.
Challenge 3: Using existing data effectively
Certain data is relatively easy to obtain and may already exist with machine manufacturers. For example, we would like to mention the service data in the form of service histories or the service archive from digital service reports. In many cases and with certain logical processes, this can already be used to identify the initial maintenance behavior of machines and systems. The data is not always easy to interpret, which is why a close exchange with the development engineers of the manufacturers is very important.
From the existing data, conclusions can be drawn as to which criteria led to which failures in the past. The larger this data base is, the higher the probability that these conclusions can be generalized in order to derive a more general predictive maintenance approach.
What are the advantages of the predictive maintenance approach?
We have already mentioned a few points. However, since this question is of central importance, we would like to list some possible advantages here:
- better planning of maintenance and service intervals
- fewer field service deployments of service employees
- less unscheduled machine failures
- optimized spare parts management
- higher machine performance
- better machine knowledge through evaluation of the collected data
As you can see, there are different requirements and therefore not one universal algorithm. This algorithm is as individual as your company. We are happy to support you in mastering the challenges of preventive maintenance. What exactly does symmedia's predictive maintenance approach look like? Let's take a look.
What does predictive maintenance look like at symmedia?
Central maintenance database
The basis for predictive maintenance approaches in symmedia SP/1 is formed by corresponding data. Information provided by the machine manufacturer is stored in a central database and delivered to the operator. In this way, the operator has direct access to current information in his Customer Cockpit at any time and can also transmit this back to the manufacturer. This allows manufacturers to create maintenance plans and conditions that can be called up worldwide. Be it the definition of triggers as direct initiators of a maintenance event, information on maintenance performance including the tools required or notifications of upcoming manufacturer overhauls.
Local maintenance manager
On site, the machine operator plans his maintenance tasks for the respective machines on the basis of the existing information. symmedia SP/1 Maintenance supports him with a simple traffic light system that clearly displays the maintenance status for the individual machines. When maintenance is complete, the traffic light system is automatically reset to green so that it is clear that there is currently no maintenance on the machine.
Of course, the operator can enter new data or modify existing information. Using data from remote service and monitoring, the operator can adjust trigger specifications and thus optimize maintenance cycles. You can find more information on monitoring here.
Covering spare parts requirements
Machine maintenance is not always possible without the right spare parts. In order to avoid delays and minimize machine downtimes, the integrated Parts Agent suggests spare parts procurement according to demand and enables automatic ordering processes. All required parts are provided by the manufacturer exactly at the planned maintenance time.
Conclusion and summary
Predictive maintenance has proven to offer significant benefits to machine manufacturers and operators by optimizing maintenance cycles and minimizing downtime. The resulting cost savings lead to a quick payback of the setup effort. The sensitivity of the evaluated data is safeguarded by responsible and transparent data protection. The different departments of the machine operators are directly involved in all processes in order to enable an effective use of data.