Showing papers on "Condition monitoring published in 2004"

Vibration Based Condition Monitoring: A Review:

Abstract: Vibration based condition monitoring refers to the use of in situ non-destructive sensing and analysis of system characteristics –in the time, frequency or modal domains –for the purpose of detecting changes, which may indicate damage or degradation. In the field of civil engineering, monitoring systems have the potential to facilitate the more economical management and maintenance of modern infrastructure. This paper reviews the state of the art in vibration based condition monitoring with particular emphasis on structural engineering applications.

Residual life predictions from vibration-based degradation signals: a neural network approach

Abstract: Maintenance of mechanical and rotational equipment often includes bearing inspection and/or replacement. Thus, it is important to identify current as well as future conditions of bearings to avoid unexpected failure. Most published research in this area is focused on diagnosing bearing faults. In contrast, this paper develops neural-network-based models for predicting bearing failures. An experimental setup is developed to perform accelerated bearing tests where vibration information is collected from a number of bearings that are run until failure. This information is then used to train neural network models on predicting bearing operating times. Vibration data from a set of validation bearings are then applied to these network models. Resulting predictions are then used to estimate the bearing failure time. These predictions are then compared with the actual lives of the validation bearings and errors are computed to evaluate the effectiveness of each model. For the best model, we find that 64% of predictions are within 10% of actual bearing life, while 92% of predictions are within 20% of the actual life.

A taxonomy and catalog of runtime software-fault monitoring tools

Abstract: A goal of runtime software-fault monitoring is to observe software behavior to determine whether it complies with its intended behavior. Monitoring allows one to analyze and recover from detected faults, providing additional defense against catastrophic failure. Although runtime monitoring has been in use for over 30 years, there is renewed interest in its application to fault detection and recovery, largely because of the increasing complexity and ubiquitous nature of software systems. We present taxonomy that developers and researchers can use to analyze and differentiate recent developments in runtime software fault-monitoring approaches. The taxonomy categorizes the various runtime monitoring research by classifying the elements that are considered essential for building a monitoring system, i.e., the specification language used to define properties; the monitoring mechanism that oversees the program's execution; and the event handler that captures and communicates monitoring results. After describing the taxonomy, the paper presents the classification of the software-fault monitoring systems described in the literature.

Fault classification and fault signature production for rolling element bearings in electric machines

Abstract: Most condition monitoring techniques for rolling element bearings are designed to detect the four characteristic fault frequencies This has lead to the common practice of categorizing bearing faults according to fault location (ie, inner race, outer race, ball, or cage fault) While the ability to detect the four characteristic fault frequencies is necessary, this approach neglects another important class of faults that arise in many industrial settings This research introduces the notion of categorizing bearing faults as either single-point defects or generalized roughness These classes separate bearing faults according to the fault signatures that are produced rather than by the physical location of the fault Specifically, single-point defects produce the four predictable characteristic fault frequencies while faults categorized as generalized roughness produce unpredictable broadband changes in the machine vibration and stator current Experimental results are provided from bearings failed in situ via a shaft current These results illustrate the unpredictable and broadband nature of the effects produced by generalized roughness bearing faults This issue is significant because a successful bearing condition monitoring scheme must be able to reliably detect both classes of faults

An amplitude Modulation detector for fault diagnosis in rolling element bearings

Abstract: The purpose of this research is to identify single-point defects in rolling element bearings. These defects produce characteristic fault frequencies that appear in the machine vibration and tend to modulate the machine's frequencies of mechanical resonance. An amplitude modulation (AM) detector is developed to identify these interactions and detect the bearing fault while it is still in an incipient stage of development (i.e., to detect the instances of AM when the magnitude of the characteristic fault frequency itself is not significant). Use of this detector only requires machine vibration from one sensor and knowledge of the bearing characteristic fault frequencies. Computer simulations as well as machine vibration data from bearings containing outer race faults are used to confirm the proficiency of this proposed technique.

Power efficient monitoring management in sensor networks

Abstract: Optimizing the energy consumption in wireless sensor networks has recently become the most important performance objective. We assume the sensor network model in which sensors can interchange idle and active modes. Given monitoring regions, battery life and energy consumption rate for each sensor, we formulate the problem of maximizing sensor network lifetime, i.e., time during which the monitored area is (partially or fully) covered. Our contributions include (1) an efficient data structure to represent the monitored area with at most n/sup 2/ points guaranteeing the full coverage which is superior to the previously used approach based on grid points, (2) efficient provably good centralized algorithms for sensor monitoring schedule maximizing the total lifetime including (1+ln(1-q)/sup -1/)-approximation algorithm for the case when a q-portion of the monitored area is required to cover, e.g., for the 90% area coverage our schedule guarantees to be at most 3.3 times shorter than the optimum, (4) a family of efficient distributed protocols with trade-off between communication and monitoring power consumption, (5) extensive experimental study of the proposed algorithms showing significant advantage in quality, scalability and flexibility.

Real-time fault diagnosis [robot fault diagnosis]

Abstract: This article presents a number of complementary algorithms for detecting faults on-board operating robots, where a fault is defined as a deviation from expected behavior. The algorithms focus on faults that cannot directly be detected from current sensor values but require inference from a sequence of time-varying sensor values. Each algorithm provides an independent improvement over the basic approach. These improvements are not mutually exclusive, and the algorithms may be combined to suit the application domain. All the approaches presented require dynamic models representing the behavior of each of the fault and operational states. These models can be built from analytical models of the robot dynamics, data from simulation, or from the real robot. All the approaches presented detect faults from a finite number of known fault conditions, although there may potentially be a very large number of these faults.

The design of a multi-agent transformer condition monitoring system

Abstract: Online diagnostics and online condition monitoring are important functions within the operation and maintenance of power transformers. This paper describes how a multi-agent system (MAS) for transformer condition monitoring has been designed to employ the data generated by the ultra high frequency (UHF) monitoring of partial discharge activity. It describes the rationale behind the use of multi-agent techniques, and the problems overcome through this technology. Every aspect of the MAS design is discussed. In addition, the design and performance of the intelligent interpretation techniques are detailed.

An investigation of self-powered systems for condition monitoring applications☆

Abstract: Over recent years there has been a growing interest in the field of micro-systems and their applications across a wide range of areas, including sensor-based systems able to operate with full galvanic isolation. This paper details the development of a self-powered system, specifically for sensor applications that can be energised on a test rig by an electromagnetic vibration-powered generator. This enables wireless operation without the use of a battery with a finite service life. The results of two systems designed for remote sensing in condition monitoring applications are discussed. The first system uses a liquid crystal display to provide the system output; the second uses an infra-red link to transmit the data output.

An evidential reasoning approach to transformer condition assessments

Abstract: This paper presents an evidential reasoning (ER) approach to the transformer condition assessment. An ER algorithm is briefly introduced which is used to combine evidences and deal with uncertainties. The methodology of transferring the transformer condition assessment problem into a multiple-attribute decision-making (MADM) solution under an ER framework is then presented. Several solutions to the transformer condition assessment problem, using the ER approach, are then illustrated, highlighting the potential of the ER algorithm. Based on the outputs of the ER approach, system operators can obtain an overall evaluation of the observed unit's condition; also, several units may be ranked in order of severity for system maintenance purposes. It can be seen from the results developed that the ER approach is a suitable solution to tackle the MADM problem of transformer condition assessment.

Complexity as a measure for machine health evaluation

Abstract: This paper presents a machine health evaluation technique using the Lempel-Ziv complexity as a numerical measure. Comparing to conventional techniques such as spectral and time-frequency analysis, the presented approach does not require a linear transfer function of the physical system to be evaluated, and is thus suited for the condition monitoring of machine systems under varying operation and loading conditions. Theoretical foundation of the technique is introduced, and its performance is investigated through experimental study of realistic vibration signals measured from a rolling bearing system. The results demonstrated that complexity provides an effective measure for machine health condition evaluation.

ANN approach for condition monitoring of power transformers using DGA

Abstract: Power transformer being a major apparatus in a power system, monitoring of its in-service behavior is necessary to avoid catastrophic failures, costly outages. Dissolved gas analysis (DGA) is an important tool for transformer fault diagnosis. The ratio methods used in the DGA have an advantage that they are independent of volume of gases involved. But the main draw back of the ratio methods is that they fail to cover all ranges of data. ANN approach is adopted as a remedy for the drawback of ratio methods in this paper.

Similarity based modeling of time synchronous averaged vibration signals for machinery health monitoring

Abstract: Monitoring rotating machinery is often accomplished with the aid of vibration sensors. The vibration sensor signals contain a wealth of complex information that characterizes the dynamic behavior of the machinery. Transforming this information into useful knowledge about the health of the machine can be challenging due to the presence of extraneous noise sources and variations in the vibration signal itself. This is particularly true in situations in which the rotating machinery is monitored under varying loads and/or speeds. In order for any gained knowledge or insight into the health of machinery to be useful, it must be actionable. This is achieved by detecting incipient faults as early as possible. A novel approach to vibration monitoring that employs a multivariate similarity-based modeling (SBM) technique to characterize the expected behavior of time synchronous averaged spectral features is shown to enable the detection in rotating machinery. This in turn facilitates the assessment of machine health and enables fault diagnostics and ultimately prognostics. SBM has been applied successfully to a variety of non-vibration related multi-sensor, health monitoring applications. Our new approach builds off of these experiences and a combination of signal processing algorithms to expand the overall applicability of SBM into single sensor vibration monitoring. We discuss an approach to gearbox fault monitoring using vibration data and SBM. This new approach is described in detail and is applied to actual H-60 gearbox vibration data acquired from seeded fault tests conducted by U.S. Naval Air Systems Command (NAVAIR) at the Helicopter Transmission Test Facility (HTTF) in Patuxent River, MD in 2001 and 2002.

A neuro-fuzzy approach for estimating mean residual life in condition-based maintenance systems

Abstract: This paper presents a framework for online reliability estimation of physical systems utilising degradation signals. Most prognostics methods promoted in the literature for estimation of mean-residual-life of individual components utilise trending or forecasting models in combination with mechanistic or empirical failure definition models. In the absence of sound knowledge for the mechanics of degradation and/or adequate failure data, it is not possible to establish practical failure definition models. However, if there exist domain experts with strong experiential knowledge, one can establish fuzzy inference models for failure definition. This paper presents a neuro-fuzzy approach for performing prognostics under such circumstances. The proposed approach is evaluated on a cutting tool monitoring problem. In particular, the method is used to monitor high-speed-steel drill-bits used for drilling holes in stainless steel metal plates.

A neuro-fuzzy approach to gear system monitoring

Abstract: The detection of the onset of damage in gear systems is of great importance to industry. In this paper, a new neuro-fuzzy diagnostic system is developed, whereby the strengths of three robust signal processing techniques are integrated. The adopted techniques are: the continuous wavelet transform (amplitude) and beta kurtosis based on the overall residual signal, and the phase modulation by employing the signal average. Three reference functions are proposed as post-processing techniques to enhance the feature characteristics in a way that increases the accuracy of fault detection. Monitoring indexes are derived to facilitate the automatic diagnoses. A constrained-gradient-reliability algorithm is developed to train the fuzzy membership function parameters and rule weights, while the required fuzzy completeness is retained. The system output is set to different monitoring levels by using an optimization procedure to facilitate the decision-making process. The test results demonstrate that the novel neuro-fuzzy system, because of its adaptability and robustness, significantly improves the diagnostic accuracy. It outperforms other related classifiers, such as those based on fuzzy logic and neuro-fuzzy schemes, which adopt different types of rule weights and employ different training algorithms.

Monitoring of induction machines by maximum covariance method for frequency tracking

Abstract: Motor current signature analysis (MCSA) has been widely investigated in order to monitor fault conditions of induction machines. On the other hand several solutions were proposed for the detection of rotor speed of induction motor for sensorless control. Another deeply investigated field of research is the detection of supply frequency of power lines, for the diagnosis of the distribution network. A common root of these three key topics is the need of accurately stating specific spectrum frequencies. Several techniques were presented in the literature in order to perform accurate tracking of frequencies for different purposes. They are modified versions of the traditional discrete Fourier transformation (DFT), or novel spectrum estimation techniques. This paper presents a novel procedure based on the statistical analysis of the current signal in the time domain, referred to as maximum covariance method for frequency tracking (MCMFT), that allows to obtain high frequency resolution accuracy independently of the sampling frequency and of the time acquisition period. Therefore those spectrum lines related to supply frequency or to slip can be detected with extreme accuracy within a wide range of sampled data conditions. Then either an accurate diagnosis of the machine electric faults or sensorless control, or distribution network diagnosis can be performed. Comparison between the proposed method and the literature are reported, in order to critically analyze its performances. An induction machine with two artificially broken bars was used for the experiments.

Fuzzy estimation of feed-cutting force from current measurement-a case study on intelligent tool wear condition monitoring

Abstract: It is very important to use a reliable and inexpensive sensor to obtain useful information about manufacturing processing, such as cutting force for monitoring automated machining. In this paper, the feed-cutting force is estimated using inexpensive current sensors installed on the ac servomotor of a computerized numerical control (CNC) turning center, with the results applied to the intelligent tool wear monitoring system. The mathematical model is used to disclose the implicit dependency of feed-cutting force on feed-motor current and feed speed. Afterwards, a neuro-fuzzy network is used to identify the cutting force with current measurement only. This hybrid math-fuzzy approach will reduce the modeling uncertainty and measurement cost. Finally, the estimated cutting force is applied in the tool-wear monitoring process. Successful experiments demonstrate robustness and effectiveness of the suggested method in the wide range of tool-wear monitoring applications.

Intelligent System for Condition Monitoring of Underground Pipelines

Abstract: Pipeline infrastructure is decaying at an accelerating rate due to reduced funding, insufficient quality control resulting in poor installation, little or no inspection and maintenance, and a general lack of uniformity and improvement in design, construction and operation practices, among other things. Developing an intelligent system can provide fast and reliable decision-making tools that are needed to handle the large volume of deteriorating buried pipeline infrastructure systems, particularly water and wastewater pipelines, that pose serious threats to environment if they fail. The focus of this article is to develop state-of-the-art concepts and technology for buried pipeline system data acquisition, data interpretation, and utilization of the data for an intelligent renewal of buried infrastructure. Such a system could automate in real time much of the pipeline data acquisition, interpretation, and evaluation process, and capture the experience and judgment of expert utility engineers in performing condition assessment and identification of appropriate rehabilitation and maintenance strategies.

Vision modules for a multi-sensory bridge monitoring approach

Abstract: This work presents the use of multiple sensor modalities in order to perform traffic analysis for health monitoring of transportation infrastructure. In particular, testbeds containing video and seismic sensors giving complementary information about vehicles are described. Computer vision algorithms are used to detect and track the vehicles and extract their properties. This information is combined with the data from seismic sensors for robust classification of vehicles. Experimental results obtained with our testbeds are described.

Signal processing of vibrations for condition monitoring of an induction motor

Abstract: Vibration monitoring is studied for fault diagnostics of an induction motor. Several features of vibration signals are compared as indicators of broken rotor bar of a 35 kW induction motor. Regular fast Fourier transform (FFT) based power spectrum density (PSD) estimation is compared to signal processing with higher order spectra (HOS), cepstrum analysis and signal description with autoregressive (AR) modelling. The fault detection routine and feature comparison is carried out with support vector machine (SVM) based classification. The best method for feature extraction seems to be the application of AR coefficients. The result is found out with real measurement data from several motor conditions and load situations.

A real-time predictive maintenance system for machine systems

Abstract: This paper describes a novel real-time predictive maintenance system for machine systems based upon a neural network approach. The ability of a neural network to learn non-linear mapping functions has been used for the prediction of machine system parameters using the motion current signature. This approach avoids the need for costly measurement of system parameters. Unlike many neural network based condition monitoring systems, this approach is validated in an off-line proof of concept procedure, using data from an experimental test rig providing conditions typical of those used on production machines. The experiment aims to classify five distinct motor loads using the motion current signature, irrespective of changing tuning parameters. Comparison of the predicted and actual loads shows good agreement. Generation of data covering all anticipated machine states for neural network training, using a production machine, is impractical, and the use of simulated data, generated by an experimentally validated simulation model, is effective. This paper demonstrates the underlying structure of the developed simulation model.

Product Quality Estimation and Operating Condition Monitoring for Industrial Ethylene Fractionator

Abstract: In this industry–university collaboration, a soft sensor for measuring a key product quality and an on-line monitoring system for testing the validity of the soft sensor were developed to realize highly efficient operation of the ethylene production plant. To estimate impurity concentration in the ethylene product from on-line measured process variables, a dynamic partial least squares (PLS) model was developed. The developed soft sensor can estimate the product quality very well, but it does not function well when the process is operated under conditions that have never been observed before. Therefore, an on-line monitoring system was developed to judge whether the soft sensor is reliable or not. The monitoring system is based on the dynamic PLS model designed for estimating the product quality. The present research provides a PLS-based framework for developing a soft sensor and monitoring its validity on-line. In addition, simple rules were established for checking the performance of a process gas chromatograph by combining the soft sensor and the monitoring system. The soft sensor and the monitoring system have functioned successfully in the industrial ethylene plant.

Dynamic gas sensor network for air pollution monitoring and its auto-calibration

Abstract: The use of a dynamic gas sensor network is proposed for air pollution monitoring, and its auto-calibration is discussed to achieve the maintenance-free operation. Although the gas sensor outputs generally show drift over time, frequent recalibration of a number of sensors in the network is a laborious task. To solve this problem, instead of the static network proposed in the related works, we propose to realize a dynamic gas sensor network by, e.g., placing sensors on vehicles running on the streets or placing some of them at fixed points and the others on vehicles. Since each sensor in the dynamic network often meets other sensors, calibration of that specific sensor can be performed by comparing the sensor outputs in such occasions. The sensors in the whole network can thus be calibrated eventually. The simulation results are presented to show that adjusting the sensor outputs to the average values of the sensors sharing the same site improves the measurement accuracy of the sensor network.