Page 59 - ASME DSCC 2015 Program

Technical Program

Smartphone-Based Wheel Imbalance Detection
ConTRIBuTED SESSIon
1-19-1 TP2 Diagnostics and Detection Contributed regular paper. DSCC2015-9716
George Bellows B 4:00pm–6:00pm Joshua Siegel, Rahul Bhattacharyya, Sanjay Sarma, Massachusetts
Institute of Technology, Cambridge, MA, United States, Ajay Deshpande,
Session Chair: Jason Kolodziej, Rochester Institute of Technology IBM Research, Yorktown Heights, NY, United States
Session Co-Chair: Ioannis Raptis, University of Massachusetts Lowell Onboard sensors in smartphones present new opportunities for vehicular
sensing. In this paper, we explore a novel application of fault detection in
Design and Analysis of a Scale-Sized Electromechanical Actuator for
wheels, tires and related suspension components in vehicles. We present
unsteady Condition Monitoring Applications
a technique for in-situ wheel imbalance detection using accelerometer
Contributed regular paper. DSCC2015-9688
data obtained from a smartphone mounted on the dashboard of a vehicle
Jason Kolodziej, Rochester Institute of Technology, Rochester, NY, United having balanced and imbalanced wheel conditions. The lack of observable
States, William Craig, University of Maryland, College Park, MD, United distinguishing features in a Fourier Transform (FT) of the accelerometer data
States
necessitates the use of supervised machine learning techniques for imbal-
Growing interest in using Electromechanical Actuators to replace current hy- ance detection. We demonstrate that a classification tree model built using
draulic actuation methods on aircraft control surfaces has driven significant Fourier feature data achieves 79% classification accuracy on test data. We
research in the area of prognostics and health management. Non-stationary further demonstrate that a Principal Component Analysis (PCA) transforma-
speeds and loads in the course of controlling an aircraft surface make fault tion of the Fourier features helps uncover a unique observable excitation
identification in EMAs difficult. This work presents a time-frequency analysis frequency for imbalance detection. We show that a classification tree model
of EMA thrust bearing vibration signals using wavelet transforms. A lab sized trained on randomized PCA features achieves greater than 90% accuracy
EMA system is designed and fabricated to allow for quick and repeatable on test data. Results demonstrate that the presence or absence of wheel
component replacement. Indentation faults from moderate and heavy loads imbalance can be accurately detected on at least two vehicles of different
are seeded in the thrust bearings and are then tested to generate data. An make and model. Sensitivity of the technique to different road and traffic
artificial neural network achieves 95% classification accuracy in a two class conditions is examined. Future research directions are also discussed.
scenario using healthy and moderately spalled thrust bearings.
observer Based Diagnostic Scheme for lithium-Ion Batteries
full-order Distributed fault Diagnosis for large-Scale nonlinear Contributed regular paper. DSCC2015-9913
Stochastic Systems Zoleikha Abdollahi Biron, Pierluigi Pisu, Beshah Ayalew,
Contributed regular paper. DSCC2015-9927 Clemson University, Greenville, SC, United States
Elaheh noursadeghi, Ioannis Raptis, University of Massachusetts Lowell, This paper presents an observer based fault diagnosis approach for Lith-
Lowell, MA, United States
ium-ion battery. This method detects and isolates five single faults in the
This paper deals with the problem of designing a distributed fault detection system which includes sensor faults in current, voltage and temperature sen-
and isolation algorithm for nonlinear large-scale systems that are subjected sors, failure in fan actuator and disturbance in battery State-of-Charge (SOC)
to multiple fault modes. To solve this problem, a network of detection nodes dynamics. Current, voltage and temperature of the battery are used as the
is deployed to monitor the monolithic system. Each node consists of an only possible measurements to design two different observers; Kalman filter
estimator with partial observation of the system’s state. The local estima- and sliding mode observer. Three residuals derived from the observers gen-
tor executes a distributed variation of the particle filtering algorithm; that erate fault signature to detect and isolate the faults and SOC disturbance in
process the local sensor measurements and the fault progression model the system. Simulation results show the effectiveness of the approach.
of the system. In addition, each node communicates with its neighbors by
sharing pre-processed information. The communication topology is defined
using graph theoretic tools. The information fusion between the neighboring
nodes is performed by a distributed average consensus algorithm to ensure
the agreement on the value of the local estimates. The simulation results
demonstrate the efficiency of the proposed approach.
Vibration Detection of Clutch Dragging Process for Multi Clutch
Transmission
Contributed regular paper. DSCC2015-9695
Man Chen, Biao Ma, Changsong Zheng, Beijing Institute of Technology,
Beijing, China

Vibration Detection of Clutch Dragging Process for Multi Clutch
Transmission

59