Neural Network-Based Gear Failure Prediction in a Brushless DC Actuation System

Yan Chen, Christina Lucas, Jay Lee, and Michael Buehner
Submission Type: 
Full Paper
AttachmentSizeTimestamp
phmc_15_066.pdf815.72 KBAugust 30, 2015 - 6:06pm

Due to its inherent efficiency and reliability, brushless DC (BLDC) driven actuation systems are widely used in a varietyof industries such as aerospace, electric transportation and industrial positioning. However, it is inevitable that various types of faults can develop in the actuator either from the BLDC motor or geared positioning systems. This paper, focusing on actuator load positioning system failures, proposes a data-driven based failure prediction method. Run-to-failure data is first collected from test-beds of specific BLDC actuation systems and then critical features representing system performance are extracted. There are also dynamic behavioral tests used, which are designed to provide discrete measurements reflecting system health conditions. Ultimately, based on optimized mapping between the two groups of information, a general neural network model is developed to establish a nonlinear trajectory model for failure progression. The model also allows for prediction of gear failure without the interruption of performing dynamic behavioral tests during continuous working condition. This approach provides for real time monitoring of system behavior as well as possibility of the predicting the Remaining Useful Life (RUL) of the actuation system. Although many efforts have been done to predict gear wear based on vibration signal, the proposed method is formulated within a "sensor-less" environment and makes full use of existing on-board sensing information, which provides the possibility of a closed-loop control system for life management. are required for all papers. Ensure that all symbols used in the abstract are defined, and all abbreviations are expanded. Do not cite references in the abstract.

Publication Year: 
2015
Publication Volume: 
6
Publication Control Number: 
066
Page Count: 
7
Submission Keywords: 
Remaining useful life; prediction; BLDC actuator; gear wear; General Regression Neural Network
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