Bearing remaining useful life prediction under starved lubricating condition using time domain acoustic emission signal processing

It is known that adequate knowledge of the sensitivity of acoustic emission signal parameters to various experimental parameters is indispensable. According to the review of the literature, a lack of knowledge was noticeable concerning the behavior of acoustic emission parameters under a broad range of machining parameters. This becomes more visible in milling operations that include sophisticated chip formation morphology and significant interaction effects and directional pressures and forces. To remedy the aforementioned lack of knowledge, the effect of the variation of cutting parameters on the time and frequency features of acoustic emission signals, extracted and computed from the milling operation, needs to be investigated in a wide aspect. The objective of this study is to investigate the effects of cutting parameters including the feed rate, cutting speed, depth of cut, material properties, as well as cutting tool coating/insert nose radius on computed acoustic emission signals featured in the frequency domain. Similar studies on time-domain signal features were already conducted. To conduct appropriate signal processing and feature extraction, a signal segmentation and processing approach is proposed based on dividing the recorded acoustic emission signals into three sections with specific signal durations associated with cutting tool movement within the work part. To define the sensitive acoustic emission parameters to the variation of cutting parameters, advanced signal processing and statistical approaches were used. Despite the time features of acoustic emission signals, frequency domain acoustic emission parameters seem to be insensitive to the variation of cutting parameters. Moreover, cutting factors governing the effectiveness of acoustic emission signal parameters are hinted. Among these, the cutting speed and feed rate seem to have the most noticeable effects on the variation of time–frequency domain acoustic emission signal information, respectively. The outcomes of this work, along with recently completed works in the time domain, can be integrated into advanced classification and artificial intelligence approaches for numerous applications, including real-time machining process monitoring.

Download Full-text

The Research of Plants' Water Stress Acoustic Emission Signal Processing Methods

2011 Third International Conference on Measuring Technology and Mechatronics Automation ◽

10.1109/icmtma.2011.802 ◽

2011 ◽

Cited By ~ 3

Author(s):

Yang Liu ◽

Zhang Junmei ◽

Luo Xiaoli ◽

Kan Jiangming ◽

Yang Kai

Keyword(s):

Signal Processing ◽

Acoustic Emission ◽

Water Stress ◽

Acoustic Emission Signal ◽

Processing Methods ◽

Emission Signal ◽

Signal Processing Methods

Download Full-text

A Scheme with Acoustic Emission Hit Removal for the Remaining Useful Life Prediction of Concrete Structures

Sensors ◽

10.3390/s21227761 ◽

2021 ◽

Vol 21 (22) ◽

pp. 7761

Author(s):

Tuan-Khai Nguyen ◽

Zahoor Ahmad ◽

Jong-Myon Kim

Keyword(s):

Neural Network ◽

Acoustic Emission ◽

Recurrent Neural Network ◽

Life Prediction ◽

Deep Neural Network ◽

Remaining Useful Life ◽

Support Vector ◽

Health Indicator ◽

Removal Process ◽

Useful Life

In this study, a scheme of remaining useful lifetime (RUL) prognosis from raw acoustic emission (AE) data is presented to predict the concrete structure’s failure before its occurrence, thus possibly prolong its service life and minimizing the risk of accidental damage. The deterioration process is portrayed by the health indicator (HI), which is automatically constructed from raw AE data with a deep neural network pretrained and fine-tuned by a stacked autoencoder deep neural network (SAE-DNN). For the deep neural network structure to perform a more accurate construction of health indicator lines, a hit removal process with a one-class support vector machine (OC-SVM), which has not been investigated in previous studies, is proposed to extract only the hits which matter the most to the portrait of deterioration. The new set of hits is then harnessed as the training labels for the deep neural network. After the completion of the health indicator line construction, health indicators are forwarded to a long short-term memory recurrent neural network (LSTM-RNN) for the training and validation of the remaining useful life prediction, as this structure is capable of capturing the long-term dependencies, even with a limited set of data. Our prediction result shows a significant improvement in comparison with a similar scheme but without the hit removal process and other methods, such as the gated recurrent unit recurrent neural network (GRU-RNN) and the simple recurrent neural network.

Download Full-text