Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Deep-learning-based method for intermittent process fault detection

A technology of fault detection and deep learning, applied in electrical testing/monitoring, testing/monitoring control systems, instruments, etc., can solve problems such as slow modeling speed, slow fault detection speed, complex modeling data, etc., to improve detection accuracy efficiency, rapid modeling and detection

Inactive Publication Date: 2019-01-01
HUZHOU TEACHERS COLLEGE
View PDF7 Cites 14 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] In order to solve the problem of the existing technology, the present invention is to solve the problems of intermittent process faults, the existing technology has complex modeling data in the detection process, and the slow modeling speed leads to slow fault detection speed, and proposes a deep learning-based intermittent process fault detection method

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Deep-learning-based method for intermittent process fault detection
  • Deep-learning-based method for intermittent process fault detection
  • Deep-learning-based method for intermittent process fault detection

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0036] Assuming that the total number of batches of a batch of equal-length data X is I, the number of samples in each batch is J, and the number of variables is K, then the method of batch expansion is used. Such as figure 1 As shown, the three-dimensional data (J×K×I) ​​is expanded into a two-dimensional matrix (JK×I) in the batch direction. Among them, each column of the expanded matrix is ​​a batch of data, and finally the training data is obtained: X={x 1 ,x 2 ,...,x I}∈R JK×I .

[0037] Unlike Principal Component Analysis (PCA), the self-encoding network uses a nonlinear activation function to perform nonlinear transformations, such as nonlinear functions such as sigmoid function or tanh function, in order to enable the self-encoding network to extract features and reconstruct data, the original data needs to be scaled, otherwise the autoencoder network will not be able to reconstruct the data in a non-linear manner. Take the tanh activation function as an example: ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention relates to a deep-learning-based method for intermittent process fault detection. According to the method not requiring original data hypothesis, equal-length and scaling processing is carried out on original data, training is carried out on a deep neural network with convolution and a plurality of intermediate layers based on a minimum reconstruction error principle, and stage division and feature extraction are carried out in a non-linear manner automatically and precisely; a Gaussian mixture model is built on a coding layer of the network and clustering is carried out, so thatthe model building calculation load is reduced substantially while features are extracted; and then a global probability detection index is put forward by combining a mahalanobis distance, thereby realizing fault detection. On the basis of the simulation experiment of one kind of semiconductor etching process, the result demonstrates that the fault detection rate can be increased effectively by using the method.

Description

technical field [0001] The invention relates to the field of fault detection, in particular to a deep learning-based intermittent process fault detection method. Background technique [0002] Batch production process is a kind of complex industrial process, which means that the production process is carried out in batches at the same location and at different times. It has been widely used in industrial production fields such as biopharmaceuticals, food, and semiconductor processing. Changes with time. Compared with continuous production, the process is more complex and changeable. Even a small abnormal condition in any process will affect the quality of the final product. Therefore, it is of great significance to find an effective process monitoring method for fault detection in batch processes. [0003] Since different operating stages have different process characteristics, the monitoring variables will be affected by the time dimension. Therefore, for the batch productio...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G05B23/02
CPCG05B23/024
Inventor 王培良王硕蔡志端徐静云周哲钱懿
Owner HUZHOU TEACHERS COLLEGE
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com