Machine Health Monitoring and Life Management Using Finite-Element-Based Neural Networks

1996 ◽  
Vol 118 (4) ◽  
pp. 830-835 ◽  
Author(s):  
M. J. Roemer ◽  
C. Hong ◽  
S. H. Hesler
Keyword(s):  
Neural Network ◽  
Finite Element ◽  
Real Time ◽  
Health Monitoring ◽  
Element Model ◽  
Network Technology ◽  
Element Analysis ◽  
Recent Advances ◽  
Machine Health Monitoring ◽  
Neural Network Technology

This paper demonstrates a novel approach to condition-based health monitoring for rotating machinery using recent advances in neural network technology and rotordynamic, finite-element modeling. A desktop rotor demonstration rig was used as a proof of concept tool. The approach integrates machinery sensor measurements with detailed, rotordynamic, finite-element models through a neural network that is specifically trained to respond to the machine being monitored. The advantage of this approach over current methods lies in the use of an advanced neural network. The neural network is trained to contain the knowledge of a detailed finite-element model whose results are integrated with system measurements to produce accurate machine fault diagnostics and component stress predictions. This technique takes advantage of recent advances in neural network technology that enable real-time machinery diagnostics and component stress prediction to be performed on a PC with the accuracy of finite-element analysis. The availability of the real-time, finite-element-based knowledge on rotating elements allows for real-time component life prediction as well as accurate and fast fault diagnosis.

Machine Health Monitoring and Life Management Using Finite Element Based Neural Networks

1995 ◽  
Author(s):  
Michael J. Roemer ◽  
Chi-an Hong ◽  
Stephen H. Hesler
Keyword(s):  
Neural Network ◽  
Finite Element ◽  
Real Time ◽  
Health Monitoring ◽  
Element Model ◽  
Network Technology ◽  
Element Analysis ◽  
Recent Advances ◽  
Machine Health Monitoring ◽  
Neural Network Technology

This paper demonstrates a novel approach to condition-based, health monitoring for rotating machinery using recent advances in neural network technology and rotordynamic, finite-element modeling. A desktop rotor demonstration rig was used as a proof of concept tool. The approach integrates machinery sensor measurements with detailed, rotordynamic, finite-element models through a neural network which is specifically trained to respond to the machine being monitored. The advantage of this approach over current methods lies in the use of an advanced neural network. The neural network is trained to contain the knowledge of a detailed finite-element model whose results are integrated with system measurements to produce accurate machine fault diagnostics and component stress predictions. This technique takes advantage of recent advances in neural network technology that enable real-time machinery diagnostics and component stress prediction to be performed on a PC with the accuracy of finite-element analysis. The availability of the real-time, finite-element based knowledge on rotating elements allows for real-time component life prediction as well as accurate and fast fault diagnosis.

Synthesis Error Analysis of NC Manufacturing and Error Prediction Modelling Based on BP Neural Network Technology

2011 ◽  
Vol 314-316 ◽  
pp. 1254-1257
Author(s):  
Hao Fan ◽  
Hang Li ◽  
Dong Hong Si ◽  
Yu Jun Xue ◽  
Guo Feng Wang ◽  
...  
Keyword(s):  
Neural Network ◽  
Finite Element ◽  
Prediction Model ◽  
Cutting Force ◽  
Bp Neural Network ◽  
Error Prediction ◽  
Network Technology ◽  
Element Analysis ◽  
Machining Center ◽  
Neural Network Technology

The method was proposed by use of the finite element analysis software ABAQUS and the BP neural network technology to build a synthesis error prediction model of a machining-center. Firstly the finite element model of a vertical machining center CINCINNATI ARROW750 was created by use of ABAQUS ,and the cutting force induced error was analyzed which resulted from the deformation of the machining-center’s components that was caused by the cutting force ;Secondly the geometric error of the machining-center was measured by use of the laser interometer,and the sample of synthesis error was obtained. Finally the synthesis error prediction model was obtained by use BP neural network,and through the comparison of predicted value and actual value of 25 groups of samples, the feasibility of error prediction model was verified.

On real-time creep damage prediction for steam turbine

2022 ◽  
pp. 014233122110689
Author(s):  
Yongjian Sun ◽  
Bo Xu
Keyword(s):  
Finite Element ◽  
Steam Turbine ◽  
Real Time ◽  
Creep Damage ◽  
Element Model ◽  
Turbine Rotor ◽  
Theoretical Research ◽  
Element Analysis ◽  
Damage Prediction ◽  
Time Calculation

In this paper, in order to solve the calculation problem of creep damage of steam turbine rotor, a real-time calculation method based on finite element model is proposed. The temperature field and stress field of the turbine rotor are calculated using finite element analysis software. The temperature data and stress data of the crucial positions are extracted. The data of temperature, pressure, rotational speed, and stress relating to creep damage calculation are normalized. A real-time creep stress calculation model is established by multiple regression method. After that, the relation between stress and damage function is analyzed and fitted, and creep damage is calculated in real-time. A creep damage real-time calculation system is constructed for practical turbine engineering. Finally, a numerical simulation experiment is designed and carried out to verify the effectiveness of this novel approach. Contributions of present work are that a practical solution for real-time creep damage prediction of steam turbine is supplied. It relates the real-time creep damage prediction to process parameters of steam turbine, and it bridges the gap between the theoretical research works and practical engineering.

Structural damage identification via physics-guided machine learning: a methodology integrating pattern recognition with finite element model updating

2020 ◽  
pp. 147592172092748 ◽  
Author(s):  
Zhiming Zhang ◽  
Chao Sun
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Finite Element ◽  
Finite Element Model ◽  
Health Monitoring ◽  
Damage Identification ◽  
Model Updating ◽  
Element Model ◽  
Data Driven ◽  
Finite Element Model Updating

Structural health monitoring methods are broadly classified into two categories: data-driven methods via statistical pattern recognition and physics-based methods through finite elementmodel updating. Data-driven structural health monitoring faces the challenge of data insufficiency that renders the learned model limited in identifying damage scenarios that are not contained in the training data. Model-based methods are susceptible to modeling error due to model idealizations and simplifications that make the finite element model updating results deviate from the truth. This study attempts to combine the merits of data-driven and physics-based structural health monitoring methods via physics-guided machine learning, expecting that the damage identification performance can be improved. Physics-guided machine learning uses observed feature data with correct labels as well as the physical model output of unlabeled instances. In this study, physics-guided machine learning is realized with a physics-guided neural network. The original modal-property based features are extended with the damage identification result of finite element model updating. A physics-based loss function is designed to evaluate the discrepancy between the neural network model output and that of finite element model updating. With the guidance from the scientific knowledge contained in finite element model updating, the learned neural network model has the potential to improve the generality and scientific consistency of the damage detection results. The proposed methodology is validated by a numerical case study on a steel pedestrian bridge model and an experimental study on a three-story building model.

Real-time multi-channel monitoring of burst-suppression using neural network technology during pediatric status epilepticus treatment

2016 ◽  
Vol 127 (8) ◽  
pp. 2820-2831 ◽  
Author(s):  
Christos Papadelis ◽  
Seyedeh Fatemeh Salimi Ashkezari ◽  
Chiran Doshi ◽  
Sigride Thome-Souza ◽  
Phillip L. Pearl ◽  
...  
Keyword(s):  
Neural Network ◽  
Status Epilepticus ◽  
Real Time ◽  
Burst Suppression ◽  
Network Technology ◽  
Neural Network Technology

A Real-Time Neural Network Estimator for Workpiece Thermal Expansion Errors

2000 ◽  
Author(s):  
A. D. Yoder ◽  
R. N. Smith
Keyword(s):  
Neural Network ◽  
Finite Element ◽  
Thermal Expansion ◽  
Finite Element Model ◽  
Real Time ◽  
Element Model ◽  
Precision Machining ◽  
Machining Processes ◽  
The Neural Network ◽  
The Finite Element Model

Abstract The importance of predicting and reducing thermal expansion errors in workpieces is becoming greater as better precision machining processes are developed. An artificial neural network model to estimate the workpiece thermal expansion errors in real-time during precision machining operations is developed and compared with experimental results. A finite element model of workpiece thermal expansion has been created to predict expansions in a thin cylinder undergoing a turning process. The neural network has been trained using finite element model solutions over a range of conditions to allow for changing machining parameters. To realize “on-line” capability, the measurable values of heat flux into the workpiece, surface heat transfer coefficient, and tool location are used as inputs and the expansion as the output for the neural network. The estimations of the network are compared with experimental results from a turning process on a large diameter aluminum cylinder. There is reasonable agreement between measured and estimated expansions with an average error of 18%. The neural network has not been trained at the cutting conditions used during the experiment. The speed of the neural network estimation is much greater than the solution to the finite element model. The finite element model required over 15 minutes to solve on a Pentium 133Mhz computer. The neural network calculated the expansions easily at 1 Hz during the experiment on the same computer. With real-time estimation using measurable data, compensation can be made in the tool path to correct for these errors. The application of this method to precision machining processes has the capability of greatly reducing the error caused by workpiece thermal expansions.

scholarly journals Method for determining load magnitude and location from the plastic deformation of fixed beams using a neural network

2021 ◽  
Vol 104 (1) ◽  
pp. 003685042110033
Author(s):  
Junqing Yin ◽  
Jinyu Gu ◽  
Yongdang Chen ◽  
Wenbin Tang ◽  
Feng Zhang
Keyword(s):  
Neural Network ◽  
Plastic Deformation ◽  
Finite Element ◽  
Element Model ◽  
Training Data ◽  
Ann Model ◽  
Element Analysis ◽  
Data Set ◽  
Sample Set ◽  
Fixed Beam

Fixed beam structures are widely used in engineering, and a common problem is determining the load conditions of these structures resulting from impact loads. In this study, a method for accurately identifying the location and magnitude of the load causing plastic deformation of a fixed beam using a backpropagation artificial neural network (BP-ANN). First, a load of known location and magnitude is applied to the finite element model of a fixed beam to create plastic deformation, and a polynomial expression is used to fit the resulting deformed shape. A basic data set was established through this method for a series of calculations, and it consists of the location and magnitude of the applied load and polynomial coefficients. Then, a BP-ANN model for expanding the sample data is established and the sample set is expanded to solve the common problem of insufficient samples. Finally, using the extended sample set as training data, the coefficients of the polynomial function describing the plastic deformation of the fixed beam are used as input data, the position and magnitude of the load are used as output data, a BP-ANN prediction model is established. The prediction results are compared with the results of finite element analysis to verify the effectiveness of the method.

A Neural Network-Based System for Bridge Health Monitoring

2012 ◽  
Vol 452-453 ◽  
pp. 557-563 ◽  
Author(s):  
Tzu Kang Lin ◽  
Ming Chih Huang ◽  
Jer Fu Wang
Keyword(s):  
Neural Network ◽  
Transfer Function ◽  
Monitoring System ◽  
Health Monitoring ◽  
Network Technology ◽  
Health Monitoring System ◽  
Bridge Health Monitoring ◽  
Analytical Results ◽  
Neural Network Technology

A bridge health monitoring system based on neural network technology is proposed in this paper. Two major ground excitations recorded in Taiwan were used to establish the NARX-based system. Analytical results from different methods including transfer function, ARX-based model, and the proposed neural network-based system were used to evaluate the efficiency in health monitoring. The result shows that the proposed system can be used successfully with superior advantages after major earthquakes for bridge health monitoring.

Using a Multibody Dynamic Simulation Code With Neural Network Technology to Predict Railroad Vehicle-Track Interaction Performance in Real Time

2007 ◽  
Author(s):  
Timothy P. Martin ◽  
Khaled E. Zaazaa ◽  
Brian Whitten ◽  
Ali Tajaddini
Keyword(s):  
Neural Network ◽  
Real Time ◽  
Nonlinear Equations ◽  
Equations Of Motion ◽  
Measured Data ◽  
Network Technology ◽  
Simulation Code ◽  
Nonlinear Equations Of Motion ◽  
Railroad Vehicle ◽  
Neural Network Technology

Recently, there has been a large demand for predicting, in real time, the performance of multiple railroad vehicle types traversing existing track as the geometry of the track is being measured. To accurately predict a railroad vehicle’s response over a specified track requires the solution of nonlinear equations of motion and extensive calculations based on the suspension characteristics of the vehicle. To realize the real time goal, codes are being implemented that use linear approximations to the fully nonlinear equations of motion to reduce computation time at the expense of accuracy. Alternatively, neural network technology has the ability to learn relationships between a mechanical system input and output, and, once learned, give quick outputs for given input. The training process can be done using measured data or using simulation data. In general, measured data is very expensive to gather due to the instrumentation requirements and is most often not available. In this paper, the use of multibody simulation code as a training tool for a neural network is presented. The example results estimate the vertical and lateral forces at the wheel-to-rail interface as a function of the geometry of the track and the suspension characteristics of the vehicle type by using a multibody code with neural network technique.

Sign in / Sign up

Export Citation Format

Share Document