Do you already know today which machine or component of a system may cause problems tomorrow? Condition monitoring makes it possible. Sensors record important measured values such as temperature, vibration or humidity and thus the technical condition of a device.

Early fault detection prevents unplanned downtime and saves time and money. The associated predictive maintenance is therefore rightly regarded as a key innovation of the Industrial Internet of Things (IIoT).

Complete solution

With the Condition Monitoring Toolkit (CMTK), sensor and automation specialist Balluff launched an independent complete solution for condition monitoring on the market last year. "From the sensors to the base unit to the pre-installed software, which enables monitoring and setting of limit values, our system bundles all the necessary components for condition monitoring," explains Balluff Program Manager Dr. Thomas Meißner.

Whether motor, fan or conveyor system: the areas of application of the system are diverse. Balluff itself uses the condition monitoring toolkit in its own company building for predictive maintenance. Since the middle of last year, the vibration and temperature of two pumps in the company's own air conditioning system have been recorded and monitored. If these fail, the cooling stops - it would then no longer be possible to work in the overheated building in summer.

But that doesn't have to be the case: "If the limit values are exceeded, the system informs you with an automatically set warning message," says Meißner. This is particularly advantageous for remote monitoring, which is now possible from any location: "Our pump application is somewhat remote in the building, technicians almost never get past there. Thanks to the CMTK, however, we still have the system under constant observation. We no longer have to worry about an unplanned outage."

AI recognizes the ideal state

So far, so good - but not enough for Balluff. "Our customers' expectations in the area of condition monitoring and predictive maintenance have risen sharply in recent years," reports Dominik Rotter, Data Scientist at Balluff. In order to save costs due to downtime or damage, devices need to run ever more reliably. The demand for comprehensive sensor-based machine monitoring is increasing. "This is where artificial intelligence, or AI for short, comes into play."

According to Rotter, Balluff's decision to integrate additional AI functionalities into the CMTK was a quick one: "The software contained in the CMTK collects sensor data and visualizes and evaluates it on the dashboard. So what could be more obvious than automatically analyzing and evaluating the collected information?"

With the help of AI, this would open up completely new possibilities for deriving concrete benefits from the data. Where the user previously had to manually define and set individual threshold values, the AI now records the crucial vibration data of the pump and takes over the monitoring. As the AI model was previously trained with data from the machine's normal operation, it knows its "healthy" condition. If this changes, the model notices the deviation - the so-called anomaly score increases. The advantages: less effort for the user. In addition, emerging problems can be detected more quickly compared to conventional condition monitoring, consequential damage can be prevented more reliably and production figures can be increased as a result.

The two pumps in the Balluff building have been working with integrated AI in the CMTK for several months now. Balluff is planning to launch its first anomaly detection product on the market this year. Based on the knowledge gained from this, a CMTK with integrated AI functionality will then be available to customers at the beginning of next year. Initial talks are already underway with interested companies who want to try out and use the new features.

Valuable information through artificial intelligence

For data scientist Rotter and his colleagues, it goes without saying that AI plays an important role in numerous activities at the company from Neuhausen auf den Fildern: "As a manufacturer of sensors and network technology, Balluff reliably supplies data for industrial automation." With a view to the Industrial Internet of Things, the team is naturally also concerned with the question of what information and benefits can be gained from the data with the help of AI - and how AI can make the sensors even better. "Our goal is always one thing: to generate concrete benefits for our customers."