Data Acquisition and Signal Analysis from Measured Motor Currents for Defect Detection in Electromechanical Drive Systems

Christian Lessmeier, Olaf Enge-Rosenblatt, Christian Bayer, and Detmar Zimmer
Submission Type: 
Full Paper
phmce_14_021.pdf687.1 KBJune 22, 2014 - 11:17pm

This paper presents the development of a diagnostic method which uses the measurement of motor currents in order to detect defects in electromechanical systems. It focusses on two main topics: the acquisition of experimental data, and the development of the diagnostic method. The data acquisition was crucial for the successful development of a dedicated signal analysis method. For this purpose, a test rig for generating experimental training data was created. The rig provides the ability to simulate a wide range of defects experimentally. Different types of artificial defects, such as bearing damage or misalignments, were used; these are described in detail in the second section of the paper. The experimental data was obtained under varying operational conditions. Using all possible settings of operational parameters for data generation would mean excessive experimental time and effort. Therefore, a special approach using the theory of “Design of Experiments” was applied. By using a fractional factorial design based on orthogonal arrays, the number of experiments could be reduced significantly. Details of this approach are given in the third section. The main ideas of the classification algorithm, including some of the results, are summarized in the fourth section. A special method using a combination of Principal Component Analysis and Linear Discriminant Analysis was designed for the correct detection of damage or misalignments. With this method, a successful classification of the systems’ health state could be obtained.

Publication Year: 
Publication Volume: 
Publication Control Number: 
Page Count: 
Submission Keywords: 
Data Based Diagnostics
bearing defect diagnosis
experimental test stand
feature selection
Submission Topic Areas: 
Data-driven methods for fault detection, diagnosis, and prognosis
Submitted by: 

follow us

PHM Society on Facebook Follow PHM Society on Twitter PHM Society on LinkedIn PHM Society RSS News Feed