The degree of automation in the industry has been increasing in the last decades. It is extremely important to avoid faults or detect them on-time. Today, machines and production systems operate at full capacity and, thus, downtimes due to a failed component could represent high costs not only on maintenance but also on costs derived from the production stop. Under this scenario, condition-based maintenance procedures help to avoid faults and to reduce costs. The importance of diagnostics and condition monitoring (CM) techniques for machines has been increasing in order to support maintenance procedures.
Pneumatics is a drive technology that is widespread in automation environments for its well-known advantages: robustness and low-cost. CM and diagnostics are also gaining more importance in this area. On the one hand, worn components and the cause of wear must be detected; on the other hand, the pneumatic drive technology is often close to the process. Therefore, faults in the process are noticed in the behavior of the pneumatic actuators. This means that a pneumatic actuator can be used as a sensor to detect faults in the process.
In most cases, there are only few sensors available in pneumatic equipment, regularly only end-position sensors that are used to control the actuators. These available signals together with the valve control signals of pneumatic drives are enough to monitor symptoms. With them it is possible to detect changes, though not to diagnose the cause of the fault. However, for the cases in which a pneumatic drive is crucial for the process and hence requires a fault-diagnosis, additional sensors are needed. This raises the question as to which sensors, combined with mathematical models, are required to diagnose faults.
The few signals in pneumatics and the several diagnostic scenarios led to the development of a diagnostic concept that covers the different requirements and the level of detail of the diagnostic information. The diagnostic concept classifies the diagnostic possibilities both at system and at component level. In this way, several diagnostic methods are organized and can be used according to the application.
In this work, some examples of robust diagnostic methods will be presented. The diagnostic methods offer a good compromise between the level of detail of the diagnostic information and the additional required sensors.
The pneumatic drive technology is only a part of a machine or equipment so, based on the diagnostic concept for pneumatics, an overall CM-concept will be proposed. The CM-concept is open and neutral, while it also integrates the several views of CM: CM-Function View, Machine/Facility Hierarchical View and Automation System Hierarchical View. This could be the foundation to integrate the diagnostic information of CM-systems from different drive technologies.