Rotating Machine Fault Diagnostics Using Vibration and Acoustic Emission Sensors
Rotating machines are widely used in various industrial applications. Implementation of condition based maintenance for rotating machines is becoming necessary in order to prevent failure, increase availability, and decrease maintenance cost. Rotating machine fault detection and diagnostics is a critical component of condition based maintenance. In order for condition based maintenance to work for rotating machines, especially for new designs and materials, effective and advanced rotational machine fault detection and diagnostic methods and tools need to be developed. Currently, vibration signal based techniques are the most widely used techniques in rotating machinery fault detection and diagnosis. However, the current vibration signal processing methods for rotating machine fault detection and diagnostics have their own limitations. Investigating and developing new advanced signal processing methods based vibration signal fault feature extraction methods and fault detection methods and tools is necessary. To do so, two advanced signal processing methods, empirical mode decomposition and interference cancellation algorithm has been investigated and developed in this dissertation. Recently, acoustic emission signal based methods are attracting researchers’ interests because acoustic emission signals have some advantages over the vibration signals. Unlike vibration signals, advanced signal processing techniques for extracting fault features from acoustic emission signals have not been well developed for rotating machine fault detection and diagnostics. Development of acoustic emission signal quantification methods and diagnostic methods and tools are in a great need. In this dissertation, acoustic emission signal quantification methodologies using Laplace wavelet and empirical mode decomposition have been developed. To validate the effectiveness of the developed diagnostic methods and tools, seeded fault test data are needed. To obtain seeded fault test data, a split-torque type gearbox test rig has been designed and developed. The seeded fault experiments have been designed and conducted on both the split-torque gearbox test rig and the bearing run-to-failure test rig. The developed fault detection and diagnostic methods and tools have been validated using the seeded fault test data collected.
condition based maintenance