Diagnosis of Pneumatic Systems on Basis of Time Series and Generalized Feature for Comparison with Standards for Normal Working Condition

Faults are unwanted events in any industrial production system. Early detection and diagnosis of faults automated systems is important order to prevent equipment damage, loss performance profits. For this purpose, more sophisticated complex for observation monitoring basic characteristics processes being built. Preconditions increasing their efficiency processing analysis process information obtained through a significant number sensors. pneumatic addition the identification certain that may affect normal process, it consider possibilities improve energy efficiency. In regards, work focuses on leaks. The fault based measurement compressed air consumption at inlet module synchronization with signal from PLC valve, controlled cylinder. experimental study aims develop methods automatic classification leaks be used machine learning algorithms.