Date of Award

Spring 1994

Document Type

Thesis - Restricted

Degree Name

Master of Science (MS)

Department

Mechanical Engineering

First Advisor

Feng

Second Advisor

Seem, John

Third Advisor

Heinen, James

Abstract

A new approach for detecting mechanical unbalance and shaft misalignment in rotating machinery is proposed. This approach uses information obtained from dual channel measurements to estimate the phase lag between two signals. Phase and spectral information for the frequencies that correspond to the first four multiples of the shaft turning speed are used as diagnostic information. This information is used as input to a neural pattern classifier which is capable of determining the type of mechanical fault that is present in the machine. The utilization of phase as diagnostic information has previously been considered only for the frequency which corresponds to the shaft turning speed. The techniques presented in this thesis utilize phase information from the frequencies which correspond to the first four multiples of the shaft turning speed. Phase values are not obtained from the use of a strobe light or other types of triggered devices but rather from dual channel data. This makes the diagnostic approach presented in this thesis applicable to a wider variety of machinery. The new approach for diagnosing misalignment and unbalance is demonstrated using acceleration data which has been collected from a specially built test machine. A structured experimental design approach is used to determine the optimal combination of techniques for diagnosing the aforementioned vibration faults using neural pattern classifiers. The results shown in this thesis demonstrate the successful utilization of phase, power spectral density values, and neural pattern classifiers in the detection of misalignment and unbalance in rotating machinery.

Download

Share

COinS

Restricted Access Item

Having trouble?