Predictive maintenance

What is the challenge?

Machines and systems are productive assets that ideally manufacture products permanently, consistently and with stable quality. In reality, however, defects, quality deviations or expensive downtimes interrupt production. Maintenance and servicing are therefore essential in production. However, fixed maintenance intervals can generate unnecessarily high costs. Short-term maintenance measures are based on historical data and serve to eliminate failures as quickly as possible, but cannot prevent them. 

What data can help?

• Machine data (temperature, humidity, noise)
• Data about the manufacturing process, the products and the production environment
• Historical machine and production data

How can companies use their data?

Machines that need maintenance or even break down show certain patterns before that. These can be, for example, trends in the development of temperatures or the strength of vibrations. The data collected is regularly compared with these patterns in order to estimate the need for maintenance. A large amount of data is required to make reliable predictions. If a maintenance requirement is identified, a message is generated. In this way, the corresponding component can be serviced or replaced before damage occurs. In addition to predictive maintenance, there is prescriptive maintenance. Prescriptive maintenance not only anticipates the problem, but takes steps to solve it. For example, it automatically adjusts parameters (such as temperature, speed) to keep production conditions optimal. Predictive maintenance, however, is currently more widespread.

Where is this use of data already being applied?

The aerospace engineering group Boeing has developed the “Boeing AnalytX” software. By processing data from flight operations, it supports the maintenance of machines and thus ensures efficient operations.

The technology group Siemens uses the “Railigent” software to avoid train failures and optimize maintenance. Sensor data, historical data and rail-specific product data are utilized for this purpose.

The technology group General Electric collects, stores and evaluates device data from power supply companies. With this information, necessary repairs can be identified at an early stage and carried out in a timely manner.

Elevator manufacturer Schindler uses predictive maintenance in its elevators. For this purpose, the elevators are connected to the internet and continuously supply the base with data. This data is compared with historical data on failures. This enables maintenance requirements to be identified at an early stage.

The technology group Georg Fischer Machining Solutions (GFMS) offers its customers services on the “rConnect” platform. These include the digital networking of machines, which enables permanent machine monitoring, predictive maintenance and remote access to the machines.

How does this use of data contribute to value creation?

With data-based predictive maintenance, production is optimized compared to interval- or time-based maintenance measures. The cost of repairs is reduced. At the same time, the efficiency and service life of the machine increases. The early anticipation of failures prevents loss of profit or damage to the image customers have.

Sources: Boeing (2018), Siemens (2019), General Electrics (2016), Schindler (2020), GF Machining Solutions GmbH (2020)