• 한국자동차공학회논문집
• /
• 제13권1호
• /
• pp.126-131
• /
• 2005

This paper presents experimental results by using the machinery fault simulator which is monitoring its conditions with an acceleration signal. Components of the machinery, for example, motor, belt pulley, belt, bearing, and gear, with artificial defects were used for the experiment.

• Journal of Mechanical Science and Technology
• /
• 제19권3호
• /
• pp.846-859
• /
• 2005

Support vector machines (SVMs) have become one of the most popular approaches to learning from examples and have many potential applications in science and engineering. However, their applications in fault diagnosis of rotating machinery are rather limited. Most of the published papers focus on some special fault diagnoses. This study covers the overall diagnosis procedures on most of the faults experienced in rotating machinery and examines the performance of different SVMs strategies. The excellent characteristics of SVMs are demonstrated by comparing the results obtained by artificial neural networks (ANNs) using vibration signals of a fault simulator.

• 한국소음진동공학회논문집
• /
• 제13권9호
• /
• pp.703-712
• /
• 2003

This paper suggests a fuzzy expert system for fault diagnosis of rotating machinery, based on modulated databases. In the proposed system, alarm and trip levels are set based on ISO, considering operating condition, machinery type and maintenance history. Input signals for diagnosis, such as sub-and super-harmonic components of vibration and mean value, are normalized from 0 to 1 under the threshold level and otherwise equal to one so that chronic faults slightly below the threshold level can be monitored. The database for diagnosis consists of two modules: the well-known Sohre's chart module and if-then type rules. The Sohre's chart is utilized for the most common problems of high-speed turbomachinery, while the rule-based module, which was collected from many papers and reports, is for diagnosing peculiar faults according to the type of machinery. To infer the results from two modules, a fuzzy operation of Yager sum was adopted. Using a simulator constructed in laboratory, experimental verification was performed for the cases of unbalance and resonance which were intended. The experimental results show that the proposed fuzzy expert system has feasibility in practical diagnosis of rotating machinery.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2003년도 춘계학술대회논문집
• /
• pp.303-312
• /
• 2003

This paper suggests a fuzzy expert system for fault diagnosis of rotating machinery, based on modulated databases. In the proposed system, alarm and trip levels are set based on ISO, considering operating condition, machinery type and maintenance history. Input signals for diagnosis, such as sub- and super-harmonic components of vibration and mean value, are normalized from 0 to 1 under the threshold level and otherwise equal to one so that chronic faults slightly below the threshold level can be monitored. The database for diagnosis consists of two modules: the well-known Sohre's chart module and if-then type rules. The Sohre's chart is utilized for the most common problems of high-speed turbomachinery, while the rule-based module, which was collected from many papers and reports, is for diagnosing peculiar faults according to the type of machinery. To infer the results from two modules, a fuzzy operation of Yager sum was adopted. Using a simulator constructed in laboratory, experimental verification was performed for the cases of resonance and housing looseness which were intended. The experimental results show that the proposed fuzzy expert system has feasibility in practical diagnosis of rotating machinery.

• 한국소음진동공학회논문집
• /
• 제22권12호
• /
• pp.1227-1236
• /
• 2012

The low speed machinery faults are usually caused by the bearing failure of the rolling elements. As the life time of the bearing is limited, the condition monitoring of bearing is very important to maintain the continuous operation without failures. A few monitoring techniques using time domain, frequency domain and fuzzy neural network vibration analysis are introduced to detect and diagnose the faults of the low speed machinery. This paper presents a method of fault detection for the rolling element bearing in the low speed machinery using the Wiener filtering and shock pulse counting techniques. Wiener filter is used for noise cancellation and it clearly makes the shock pulse emerge from the time signal with the high level of noise. The shock pulse counting is used to determine the various faults obviously from the shock signal with transient pulses not related with the bearing fault. Machine fault simulator is used for the experimental measurement in order to verify this technique is the powerful tool for the low speed machine compared with the frequency analysis. The test results show that the method proposed is very effective parameter even for the signal with high contaminated noise, speed variation and very low energy. The presented method shows the optimal tool for the condition monitoring purpose to detect the various bearing fault with high accuracy.

• 한국자동차공학회논문집
• /
• 제3권6호
• /
• pp.262-273
• /
• 1995

For rotating machinery with high speed and high efficiency, large labor and high expenses are required to conduct machine health monitoring. Therefore, it becomes necessary to develop new diagnosis technique which can detect abnormalities of the rotating machinery effectively. In this paper, it is identified that bispectrum analysis technique can be successfully applied to dectect the abnormalities of the roating machinery through computer simulation, and results of the bispectrum analysis are patterned in griding form. Further, pattern recognition technique using back propagation algorithm, which is one of neural network algorithm, being consisted of patterned input layer and output layer for abnormal status, is applied to detect the abnormalities of simulator which is able to make up various kinds of abnorml conditions(misalignment, unbalance, rubbing etc.) of the rotating machinery.