According to the manufacturer, the new HST (High Speed Turbo) compressors represent a milestone in the further development of Class 0 oil-free compressed air generation. Applications for the compressors include the pharmaceutical and food industries, industrial paint stores and semiconductor production. In all areas of application, machine downtimes have fatal consequences. Preventive fault avoidance is a prerequisite for increased process reliability and optimum energy management for the customer. This is why Boge is continuing to drive forward its Compressed Air 4.0 strategy in the area of service and maintenance.

Boge relies on predictive maintenance software from Weidmüller to detect potential faults and operating anomalies before they occur. This software solution is part of Boge Analytics, the Industry 4.0 package for intelligent data evaluation and use. The software package also includes the Boge airstatus remote monitoring solution, the Continuous Improvement Program and 24-hour recovery.

The predictive maintenance software detects faults and critical deviations in technological parameters at an early stage. The operational experience of all Boge compressed air solutions is incorporated into the data evaluation. This allows predictions to be made about future maintenance requirements during operation and service calls to be optimally planned. Boge then replaces affected components before any predicted damage occurs. Cost-intensive aftercare and long downtimes are therefore a thing of the past. With predictive maintenance, the company is focusing on a key value driver for a competitive and continuous compressed air supply in accordance with Industry 4.0 standards.

Boge equips the HST compressors with an Industry 4.0 service package for the intelligent evaluation of operating data. This makes it possible to make predictions about upcoming maintenance requirements during operation and optimally plan service calls. (Image: Weidnüller)

Boge as a user was at the center of the analyses - his know-how is very important. Although the software can predict an error with a certain probability, the prerequisite for this is always that it has previously been classified as relevant by an application expert. Only the user can assess whether an anomaly should actually be classified as a critical error.

Data patterns are crucial

The recorded data all comes from measurement technology components that are already present in the compressor. No new sensors had to be retrofitted. The following data was included in the evaluation and analysis: Engine temperature, air pressure, engine speed. The data quality was carefully assessed and a decision was made together with Boge as to its relevance. Non-relevant data sources were then sorted out. It is important to note how the measured values interact with each other, as it is not the individual value that is decisive, but the existing data pattern. This results in a complex data model of normality, which can be used to precisely predict the case of damage if the values deviate from the learned model in a certain way.

One decisive factor is that the system is constantly learning. The model changes with every new error message and based on feedback from the operator. This produces the following analytics results: What is the current and past status of the compressor, when will the status change to a critical fault and how efficient has the compressor been, how much compressed air has it delivered in the last few days. The analytics solution is designed in such a way that it can also learn about faults that are not yet known - calculated predictions about compressor failure therefore become increasingly precise over the entire operating time of the compressor. as