Cognitive Sensor: Listen to Your Plant

Need support?
Get in touch

How Voice Recognition Revolutionizes Maintenance

Anyone who is well acquainted with industrial systems will hear immediately if something is running differently than normal. This simple principle was the starting point for the development of the Bilfinger Cognitive Sensor. The digital technology supports plant operators and maintenance experts in identifying changes in the operation mode as well as faults and maintenance requirements.

Challenge

Changes in the machine condition cannot be detected efficiently all the time with regular inspection rounds.

Solution

The Cognitive Sensor continuously records the sounds through deployed microphones and evaluates the status of the plant.

Benefit

Noise related machine mode detection reduces reactive maintenance efforts and reduces downtime.

Challenge

In the process industry, avoiding machine downtime is the highest priority. The method often used to monitor and ensure machine health and smooth running processes are inspection rounds. However, changes in the machine condition cannot be detected efficiently all the time with regular inspection rounds as there is a chance that the change does not appear when the round is carried out or already started when no one was around. If the change is not detected early enough, reactive maintenance and production changes is the only option left which costs the company money in spare parts for the machine and revenue lost in down time.

„Sound is the most apparent sign of mechanical failure. “

Solution: Cognitive Sensor

The developed method in the field of machine state detection and noise based anomaly detection is based on artificial intelligence. Based on the specific noise data, the Cognitive Sensor is trained according to the customer's requirements. The system continuously records the sounds through the deployed microphones, evaluates them using the trained algorithms and visualizes the results in the Bilfinger BCAP (Bilfinger Connected Asset Platform) platform. The customer has access to the results and, if required, to the raw sounds via a web-based dashboard. The service offers warnings and notifications via text messages and e-mail. The notification identifies the system and the individual sensor (location, component) and contains the respective event. With increasing experience and feedback from the customer, the potential of the AI-based algorithm requires continuous improvement of the trained models.

The system continuously records the sounds through the deployed microphones, evaluates them using the trained algorithms and visualizes the results in the Bilfinger BCAP (Bilfinger Connected Asset Platform) platform

From noise to effective machine state detection

  • Automated anomaly analysis of various machines and rotating equipment´s based on sound
  • Customized Dashboard with all relevant KPI´s
  • Customized classification of different machine states based on sound
  • Continuous improvement of trained detection models leading to early prediction models
  • Remote listening to machine

Benefit: Improved Maintenance

Noise related machine mode detection helps significantly in maintenance and production. Using the noise to predict failure modes will decrease the reaction time significantly and improves your planning and execution of maintenance activities. With a non-invasive data recording process the sensor ensures no interruption in the normal running of the process and the decision makers get the ability to make more informed decisions which will reduce the overall risk factor in the plant. With the integration with the BCAP platform it gives the end user functionality to import and analyze data from different sources and compile them to get new insights which they previously would not have access to.

Any questions? We answer.

Give us a call: +49 6221 427894-0