Technology

What is Predictive Maintenance?

Predictive Maintenance is a sensor-based form of maintenance. The process uses big data analyses to predict failures and malfunctions in a plant and thus plan maintenance operations more efficiently. As an Industry 4.0 application, predictive maintenance can thus significantly reduce costs and downtimes compared to conventional maintenance strategies.

1. how does predictive maintenance work?

1.1 Collect data

The term Predictive Maintenance can be translated into German with predictive maintenance or predictive maintenance translated. To enable this "look-ahead", relevant data must first be determined in the machine by sensors. For the implementation of a Predictive-Maintenance-application, a wide variety of data and thus a wide variety of sensors can be used. Examples of this are:

– electrical sensors

– Vibration sensors

– Temperature sensors

– Pressure sensors

– Humidity sensors

– ...and many more

This multitude reflects the complexity of the topic. Unfortunately, there is no one-size-fits-all solution for implementing projects. It takes a capable teamwhich creates customized solutions for the application area. Wind turbines require completely different concepts than machine tools.

That's why companies are reluctant to do so, Predictive Maintenance to implement comprehensively. Most market participants are aware of the importance of the field, but lack experience in implementation.

1.2 Analyze data

The counterpart to data collection is data analysis. By evaluating the measurements from the running process, problems are to be detected at an early stage. Service technicians are then dispatched before a failure occurs.

This requires large amounts of data, for the evaluation of which approaches from the Big Data-environment. Even before the sensors are installed, it should of course be clarified which data indicate that the machine will soon fail, so that everything goes smoothly with your predictive maintenance implementation.

Implementation of Predictive Maintenance

Since many companies don't know how to get started in predictive maintenance, we've created an 8-week plan for implementing an initial pilot in this area. More info?

AI and Predictive Maintenance

Artificial intelligence can also be used to analyze data and maintain plants more effectively. Machine learning offers great future opportunities here. More info?

2. advantages of predictive maintenance

In a Roland Berger Study 79% of the respondents said that they saw the benefits of predictive maintenance mainly in a Performance improvement of production technology. This results above all from the higher Availability and the longer Lifetime of the plants. Also the improvement of the Process quality is a decisive factor.

21% percent, on the other hand, expected primarily a Reducing their costs, by lower expenditures for spare parts and repairs. Even if some work and investment is required initially, the implementation is definitely worthwhile in the long term. Predictive Maintenance has advantages compared to conventional maintenance strategies:

2.1 Predictive maintenance vs. preventive maintenance

The concept of preventive maintenance is actually familiar to anyone who has ever ridden an elevator. There is usually a TÜV seal with the date of the last inspection next to the button. Specialized personnel check the functionality and safety of the elevator at specified intervals. By Predictive Maintenance technicians could save themselves some missions in this example.

With the help of comprehensive data analyses, it is possible to determine which equipment requires maintenance and which does not yet need to be inspected at the specified interval. In comparison with the preventive maintenance can Predictively maintained plants therefore increase safety and at the same time avoid costs for unnecessary maintenance operations. Likewise, the running time and service life are significantly improved.

2.2 Predictive maintenance vs. reactive strategy

The reactive strategy usually takes place in less safety-relevant areas. Actually, this is not really a maintenance strategy either. This is because only repairs and analyses are carried out in the event of a malfunction. The disadvantages are obvious.

Running the system to failure without maintenance can result in far greater damage in the event of a failure. In addition, failures occur more frequently. Adherents of this strategy also often forget to include the lower resale value of the equipment in their profit calculations. A poorly maintained checkbook has no less of a negative impact in the industry than in the private car market. Those who save on maintenance often end up paying more.

Predictive maintenance has not yet really arrived in many companies. The ServiceLobby has already outlined the extent to which this technology will Future of the service will influence.

3. examples for applications of predictive maintenance

Possible application areas of Predictive Maintenance are diverse. In some industries, implementation is also easier than in others. The complexity of the plants varies. However, the individual steps of the conversion process are always the same. In some markets, however, players are somewhat more intensely resistant to new types of maintenance technologies.

Wind Energy IndustryVibration analyses provide suitable measurement data to sound the alarm in good time when a wind turbine is no longer running smoothly. This is comparatively easy to do with today's analysis and sensor technology. The industry is therefore already well positioned and can be considered a benchmark industry against which others must measure themselves.

Rail industry: A system developed by Siemens System detects and documents cavities in the subgrade based on the acceleration behavior of a train. Even slight subsidence of the track bed causes unevenness in the course of the rails, which means that the power of the drive can no longer be transmitted evenly. Severe subsidence, on the other hand, can become a safety risk. Predictive maintenance helps to better coordinate track maintenance work.

Automotive industry: In the automotive environment, engine and drive component failures are becoming increasingly predictable.

Mechanical engineering: In mechanical engineering, we find a wide variety of different plant concepts. Predictive maintenance solutions can therefore take a wide variety of forms. There are almost unlimited possibilities to become active here. However, a little more work is required in the development and planning of projects, since machines typically combine very many different technology concepts. In addition, the quantities are significantly lower than in the automotive environment, for example. This makes the development of dedicated monitoring concepts considerably more expensive, as it has to be allocated to a much smaller number of units.