A New Approach to Machinery Monitoring and Diagnostics Using Self-Organizing Maps

1995 ◽  
Author(s):  
Siyu Zhang ◽  
R. Ganesan ◽  
T. S. Sankar
Keyword(s):  
Neural Network ◽  
Service Life ◽  
Trend Analysis ◽  
New Approach ◽  
Machine System ◽  
Network Algorithm ◽  
On Line ◽  
Neural Network Algorithm ◽  
Monitoring And Diagnostics ◽  
Self Organizing

Abstract The problem of estimating an unknown multivariate function from on-line vibration measurements, for determining the conditions of a machine system and for estimating its service life is considered. This problem is formulated into a multiple-index based trend analysis problem and the corresponding indices for trend analysis are extracted from the on-line vibration data. Selection of these indices is based on the simultaneous consideration of commonly-observed faults or malfunctions in the machine system being monitored. A neural network algorithm that has been developed by the present authors for multiple-index based regression is adapted to perform the trend analysis of a machine system. Applications of this neural network algorithm to the condition monitoring and life estimation of both a bearing system as well as a gearbox are fully demonstrated. The efficiency and computational supremacy of the new algorithm are established through comparing with the performance of Self-Organizing Mapping (SOM) and Constrained Topological Mapping (CTM) algorithms. Further, the usefulness of multiple-index based trend analysis in precisely predicting the condition and service life of a machine system is clearly demonstrated. Using on-line vibration signal to constitute the set of variables for trend analysis, and employing the newly-developed self-organizing neural algorithm for performing the trend analysis, a new approach is developed for machinery monitoring and diagnostics.

Multivariable Trend Analysis Using Neural Networks for Intelligent Diagnostics of Rotating Machinery

1997 ◽  
Vol 119 (2) ◽  
pp. 378-384 ◽  
Author(s):  
S. Zhang ◽  
R. Ganesan
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Trend Analysis ◽  
Condition Monitoring ◽  
Diagnostic Procedure ◽  
On Line ◽  
Neural Network Algorithm ◽  
Intelligent Diagnostics ◽  
Bearing System ◽  
Self Organizing

The objective of this paper is the development of an efficient intelligent diagnostic procedure that considers several diagnostic indices for the quantification of developing faults and for monitoring machine condition. In this procedure, the condition monitoring is performed based on the on-line vibration measurements, and further, the fault quantification is formulated into a multivariate trend analysis. Self-organizing neural networks are then deployed to perform the multivariable trending of the fault development. The attributes for the disordering of “knots” in the trend analysis are determined. The disordering of neural network units is then eliminated by suitably altering the self-organizing neural network algorithm. Applications of this diagnostic procedure to the condition monitoring and life estimation of a bearing system are fully developed and demonstrated. The efficiency and advantages of the intelligent diagnostic procedure in precisely monitoring and quantifying the fault development are systematically brought out considering this bearing system.

Self-Organizing Neural Networks for Automated Machinery Monitoring Systems

1995 ◽  
Author(s):  
Siyu Zhang ◽  
R. Ganesan ◽  
T. S. Sankar
Keyword(s):  
Neural Network ◽  
Trend Analysis ◽  
Real Life ◽  
Vibration Signal ◽  
Statistical Parameters ◽  
Data Set ◽  
Machine System ◽  
On Line ◽  
Machinery System ◽  
Self Organizing

Abstract Two fundamental problems that are frequently encountered in automated machinery monitoring and diagnostics are formulated into their corresponding mathematical problems of clustering and trend analysis. The need for and the efficiency of multiple-index based trend analysis, in both precisely evaluating the current conditions of a machine system using on-line vibration measurements and obtaining a reliable prediction about its future behaviour, are systematically brought out. Neural network solutions to these problems, particularly the solutions using Self-Organizing Maps (SOM) are sought. Statistical parameters of the on-line vibration signal such as peak-to-peak value, absolute mean value, crest factor etc., are used to form the data set depending on the machinery system being monitored and diagnosed. Self-organizing mapping algorithm is then employed to perform the clustering and feature extraction which takes as the input the multi-dimensional data set and provides as the output the condition of the machinery system. Associated one-layer neural network is developed during the process of SOM and the training of this network is performed in an unsupervised learning mode. A new efficient neural network algorithm that has been previously developed by the present authors for multiple-index based regression is adapted to perform the trend analysis of a machine system. Applications of the above neural network algorithms to the condition monitoring and life estimation of both a bearing system as well as a rotor system are fully demonstrated using real-life data.

Predicting Pavement Roughness Using Deep Learning Algorithms

2021 ◽  
pp. 036119812110237
Author(s):  
Qingwen Zhou ◽  
Egemen Okte ◽  
Imad L. Al-Qadi
Keyword(s):  
Neural Network ◽  
Service Life ◽  
Pavement Performance ◽  
Series Data ◽  
Statistical Regression ◽  
Network Algorithm ◽  
Pavement Roughness ◽  
Impacting Factors ◽  
Neural Network Algorithm ◽  
Maintenance Activities

Transportation agencies should measure pavement performance to appropriately strategize road preservation, maintenance, and rehabilitation activities. The international roughness index (IRI), which is a means to quantify pavement roughness, is a primary performance indicator. Many attempts have been made to correlate pavement roughness to other pavement performance parameters. Most existing correlations, however, are based on traditional statistical regression, which requires a hypothesis for the data. In this study, a novel approach was developed to predict asphalt concrete (AC) pavement IRI, utilizing datasets extracted from the Long-Term Pavement Performance (LTPP) database. IRI prediction is categorized by two models: (i) IRI progression over the pavement’s service life without maintenance/rehabilitation and (ii) the drop in IRI after maintenance. The first model utilizes the recurrent neural network algorithm, which deals with time-series data. Therefore, historical traffic data, environmental information, and distress (rutting, fatigue cracking, and transverse cracking) measurements were extracted from the LTPP database. A long short-term memory network was used to solve the vanishing gradient problem. Finally, an optimal model was achieved by setting the sequence length to 2 years. The second model utilizes an artificial neural network algorithm to correlate the impacting factors to the IRI value after maintenance. The impacting factors include maintenance activities; initial (new construction), milled, and overlaid AC thicknesses; as well as IRI value before maintenance activities. Combining the two models allows for the prediction of IRI values over AC pavement’s service life.

A New Approach to Analyzing Gas Mixtures Based on Improved BP Neural Network Algorithm

2012 ◽  
Vol 260-261 ◽  
pp. 548-553
Author(s):  
Teng Li ◽  
Xiao Mei Yuan ◽  
Shi Liang Yang ◽  
Xin Hui Zhang
Keyword(s):  
Neural Network ◽  
Numerical Simulation ◽  
Bp Neural Network ◽  
Array Signal Processing ◽  
Sensor Arrays ◽  
Gas Mixtures ◽  
New Approach ◽  
Gas Concentration ◽  
Network Algorithm ◽  
Neural Network Algorithm

A new approach is presented for analyzing gas mixtures by transforming the problem into a pattern classification one to reduce the effect of the poor repeatability of sensor response on the prediction of gas concentration. The aim of numerical simulation is to determine how successfully the approach using the combination of artificial neural networks with multi-sensor arrays can analyze multi-component gas mixtures. The results indicate that the new approach is realistic for gas mixture analysis, and numerical simulation is a powerful tool to determine the architecture of a network. By constructing improved BP neural network algorithm and basic BP neural network into sensor array signal processing and extracting 6 component as the input of neural network, Our investigation results indicated that recognition results obtained from improved BP neural network algorithm more accuracy than the results obtained from basic BP neural network.

Multivariable Trend Analysis for System Monitoring Through Self-Organizing Neural Networks

1997 ◽  
Vol 119 (2) ◽  
pp. 223-228
Author(s):  
Siyu Zhang ◽  
R. Ganesan
Keyword(s):  
Trend Analysis ◽  
Diagnostic Procedure ◽  
Vibration Signal ◽  
Machine Condition Monitoring ◽  
Diagnostic Indices ◽  
Condition Monitoring Systems ◽  
On Line ◽  
Neural Network Algorithm ◽  
Diagnostic Index ◽  
Self Organizing

For precise and reliable fault detection it is essential to consider simultaneously the changes in several diagnostic indices that are extracted from the on-line vibration signal. Existing machine condition monitoring systems consider each diagnostic index separately. Development of an automated diagnostic procedure that considers simultaneously several diagnostic indices is the objective of the present paper. The multivariable trend analysis of on-line vibration signals is deployed as the basis for this procedure. An efficient self-organizing neural network algorithm that is highly suitable to this diagnostic procedure is developed and deployed. Applications to both a bearing system as well as a gearbox system are fully developed and demonstrated.

Sign in / Sign up

Export Citation Format

Share Document