Adaptive Angular-Velocity Vold-Kalman Order Tracking Used in Characterizing Dynamic Signals Measured From Rotary Machines

Author(s):  
Min-Chun Pan ◽  
Wei-Chen Chu

In this paper we propose and implement an order tracking (OT) technique evaluating in an adaptive manner for signal interpretation and diagnosis of rotary machinery during operation. The paper comprises theoretical derivation, numerical implementation and experimental justification. Comparisons of the adaptive Vold-Kalman filtering OT (VKFOT) scheme to its conventional form are made through synthetic signal processing. Parameters such as the weighting factor, Q1 parameter, sampling frequency and process noise, which influence tracking performance, are investigated. Moreover, this adaptive OT scheme can be computed on-line and implemented as a real-time processing application.

Author(s):  
Min-Chun Pan ◽  
Cheng-Xue Wu

This paper proposes and implements an adaptive Vold-Kalman filtering order tracking (VKF_OT) approach to overcome the deficiencies of the original VKF_OT scheme for condition monitoring and diagnosis of rotary machinery. This article comprises theoretical derivation, numerical implementation and experimental validation. Comparisons of the adaptive scheme to the original are accomplished through processing a synthetic signal composed of close order components. Parameters such as the weighting factor and the correlation matrix of process noise, which influences tracking performance, are investigated in the study. The adaptive OT scheme based on the Kalman filter can be computed on-line and implemented as a real-time processing application. This paper also illustrates experimental validation through the separation of two close orders arising from a transmission test bench.


Author(s):  
Ting-Chi Yeh ◽  
Min-Chun Pan

When rotary machines are running, acousto-mechanical signals acquired from the machines are able to reveal their operation status and machine conditions. Mechanical systems under periodic loading due to rotary operation usually respond in measurements with a superposition of sinusoids whose frequencies are integer (or fractional integer) multiples of the reference shaft speed. In this study we built an online real-time machine condition monitoring system based on the adaptive angular-velocity Vold-Kalman filtering order tracking (AV2KF_OT) algorithm, which was implemented through a DSP chip module and a user interface coded by the LabVIEW®. This paper briefly introduces the theoretical derivation and numerical implementation of computation scheme. Experimental works justify the effectiveness of applying the developed online real-time condition monitoring system. They are the detection of startup on the fluid-induced instability, whirl, performed by using a journal-bearing rotor test rig.


Author(s):  
Min-Chun Pan ◽  
Cheng-Xue Wu

Dynamic signals acquired from rotary machines can be characterized by the order tracking (OT) techniques. The extracted order components correspond to the operation of specific machine elements and reflect their current healthy or faulty states. The study extends the angular-velocity Vold–Kalman OT scheme to simultaneously extract multiple order components. Theoretical derivation is illustrated with simulation of processing three synthetic signals to show its merit. Additionally, as an example to validate its effectiveness, the improved OT scheme is used to process pass-by noise emitted from an electric scooter with a planetary-gear-set transmission system. The gear-meshing orders are effectively decoupled from structure-borne resonances.


2002 ◽  
Vol os-11 (3) ◽  
pp. 1558925002OS-01 ◽  
Author(s):  
Randall R. Bresee ◽  
Uzair A. Qureshi

On-line and off-line measurements were obtained to gain an understanding of fly production during multi-hole melt blowing at commercial speed. These measurements allowed us to describe the effects of common processing parameters on fly production and develop a model for fly formation that begins to account for experimental measurements.


Author(s):  
Norikazu Ikoma ◽  
◽  
Gefan Zhang

Decorations of face such as enlarging eyes, whitening skin, rendering face slim, and so on are commercially successful in amusement arcades especially in Japan for still image and off-line processing. This paper proposes to decorate human face in video on-line and in real-time processing. Face posture estimation using particle filter plays a key role to decorate the face by precisely determining position of the eyes as well as determining regional position of face. Our proposed method conducts two decorations, enlarging eyes and whitening skin, based on the estimation result of face posture. Real-time implementation of the proposed method has been demonstrated for real scenes of indoor situation.


2011 ◽  
Vol 32 (2) ◽  
pp. 299-331 ◽  
Author(s):  
LEAH ROBERTS ◽  
CLAUDIA FELSER

ABSTRACTIn this study, the influence of plausibility information on the real-time processing of locally ambiguous (“garden path”) sentences in a nonnative language is investigated. Using self-paced reading, we examined how advanced Greek-speaking learners of English and native speaker controls read sentences containing temporary subject–object ambiguities, with the ambiguous noun phrase being either semantically plausible or implausible as the direct object of the immediately preceding verb. Besides providing evidence for incremental interpretation in second language processing, our results indicate that the learners were more strongly influenced by plausibility information than the native speaker controls in their on-line processing of the experimental items. For the second language learners an initially plausible direct object interpretation lead to increased reanalysis difficulty in “weak” garden-path sentences where the required reanalysis did not interrupt the current thematic processing domain. No such evidence of on-line recovery was observed, in contrast, for “strong” garden-path sentences that required more substantial revisions of the representation built thus far, suggesting that comprehension breakdown was more likely here.


Sign in / Sign up

Export Citation Format

Share Document