Neural-network-based modeling of structural parametric evaluation with dynamic responses in time domain

2003 ◽  
Author(s):  
Bin Xu ◽  
Zhishen Wu ◽  
Koichi Yokoyama
Keyword(s):  
Neural Network ◽  
Time Domain ◽  
Dynamic Responses

scholarly journals Diamond Grinding Wheel Condition Monitoring Based on Acoustic Emission Signals

Sensors ◽  
2021 ◽  
Vol 21 (4) ◽  
pp. 1054
Author(s):  
Guo Bi ◽  
Shan Liu ◽  
Shibo Su ◽  
Zhongxue Wang
Keyword(s):  
Neural Network ◽  
Acoustic Emission ◽  
Condition Monitoring ◽  
Time Domain ◽  
Diamond Wheel ◽  
Feature Representation ◽  
Grinding Wheel ◽  
Time Waveform ◽  
Original Time ◽  
Whole Life Cycle

Acoustic emission (AE) phenomenon has a direct relationship with the interaction of tool and material which makes AE the most sensitive one among various process variables. However, its prominent sensitivity also means the characteristics of random and board band. Feature representation is a difficult problem for AE-based monitoring and determines the accuracy of monitoring system. It is knottier for the situation of using diamond wheel grinding optical components, not only because of the complexity of grinding process but also the high requirement on surface and subsurface quality. This paper is dedicated to AE-based condition monitoring of diamond wheel during grinding brittle materials and feature representation is paid more attention. AE signal of brittle-regime grinding is modeled as a superposition of a series of burst-type AE events. Theory analysis manifested that original time waveform and frequency spectrum are all suitable for feature representation. Considering the convolution form of b-AE in time domain, a convolutional neural network with original time waveform of AE signals as the input is built for multi-class classification of wheel state. Detailed state division in a wheel’s whole life cycle is realized and the accuracy is over 90%. Different from the overlapping in time domain, AE components of different crack mechanisms are probably separated in frequency domain. From this point of view, AE spectrums are more suitable for feature extraction than the original time waveform. In addition, the time sequence of AE samples is essential for the evaluation of wheel’s life elapse and making use of sequential information is just the idea behind recurrent neural network (RNN). Therefore, long short-term memory (LSTM), a special kind of RNN, is used to build a regression prediction model of wheel state with AE spectrums as the model input and satisfactory prediction accuracy is acquired on the test set.

Neural network broken bar detection using time domain and current spectrum data

2012 ◽  
Author(s):  
Pedro Gardel ◽  
Daniel Morinigo-Sotelo ◽  
Oscar Duque-Perez ◽  
Marcelo Perez-Alonso ◽  
Luis A. Garcia-Escudero
Keyword(s):  
Neural Network ◽  
Time Domain ◽  
Current Spectrum ◽  
Broken Bar Detection ◽  
Spectrum Data

A Deep Neural Network Model for Learning Runtime Frequency Response Function Using Sensor Measurements

2021 ◽  
Author(s):  
Yongzhi Qu ◽  
Gregory W. Vogl ◽  
Zechao Wang
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Fourier Transform ◽  
Time Domain ◽  
Frequency Response Function ◽  
Frequency Component ◽  
Operating Conditions ◽  
Input Output ◽  
The Fourier Transform ◽  
The Time Domain

Abstract The frequency response function (FRF), defined as the ratio between the Fourier transform of the time-domain output and the Fourier transform of the time-domain input, is a common tool to analyze the relationships between inputs and outputs of a mechanical system. Learning the FRF for mechanical systems can facilitate system identification, condition-based health monitoring, and improve performance metrics, by providing an input-output model that describes the system dynamics. Existing FRF identification assumes there is a one-to-one mapping between each input frequency component and output frequency component. However, during dynamic operations, the FRF can present complex dependencies with frequency cross-correlations due to modulation effects, nonlinearities, and mechanical noise. Furthermore, existing FRFs assume linearity between input-output spectrums with varying mechanical loads, while in practice FRFs can depend on the operating conditions and show high nonlinearities. Outputs of existing neural networks are typically low-dimensional labels rather than real-time high-dimensional measurements. This paper proposes a vector regression method based on deep neural networks for the learning of runtime FRFs from measurement data under different operating conditions. More specifically, a neural network based on an encoder-decoder with a symmetric compression structure is proposed. The deep encoder-decoder network features simultaneous learning of the regression relationship between input and output embeddings, as well as a discriminative model for output spectrum classification under different operating conditions. The learning model is validated using experimental data from a high-pressure hydraulic test rig. The results show that the proposed model can learn the FRF between sensor measurements under different operating conditions with high accuracy and denoising capability. The learned FRF model provides an estimation for sensor measurements when a physical sensor is not feasible and can be used for operating condition recognition.

Sign in / Sign up

Export Citation Format

Share Document