Fault detection method and system of non-Gaussian industrial process

A fault detection and industrial process technology, applied in the direction of electrical program control, comprehensive factory control, etc., can solve problems such as only considering a single data set, no better solution, limited kernel function parameter selection, etc.

Active Publication Date: 2017-11-17
CENT SOUTH UNIV
View PDF4 Cites 21 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] On the other hand, for non-Gaussian processes, the existing non-Gaussian process detection methods based on multivariate statistical analysis are mainly divided into two categories, one is based on existing multivariate statistical analysis methods combined with threshold determination methods, the main threshold determination methods are Based on Gaussian mixture model (Gaussian mixture model, GMM), based on kernel function estimation (Kernel-based) and based on sequential quantile method (Sequential quantile estimation, SQE), although these methods have been applied, they are still limited by

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
  • Fault detection method and system of non-Gaussian industrial process
  • Fault detection method and system of non-Gaussian industrial process
  • Fault detection method and system of non-Gaussian industrial process

Examples

Experimental program
Comparison scheme
Effect test

Example Embodiment

[0040] Example 1

[0041] A non-Gaussian industrial process fault detection method, refer to figure 1 ,include:

[0042] The first step is to select a certain number of normal historical data sets, perform de-averaging preprocessing on the data sets, and establish a residual generator using the canonical correlation analysis method. This step can be performed offline and specifically includes:

[0043] Step A1. Use the historical database to select the process data under normal operating conditions, and use them respectively Represents the two variable data sets of the process, where l is the number of measured variables in the vector u, m is the number of measured variables in the vector y, and N is the number of independent sampling points; then u(i),i=1,...,N represents the vector The measured values ​​of l variables in u at the i-th sampling time, y(i), i=1,..., N represents the measured values ​​of m variables in the vector y at the i-th sampling time.

[0044] Step A2. Perfor...

Example Embodiment

[0096] Example 2

[0097] As an implementation of deterioration in the above-mentioned embodiment 1, in the above-mentioned detection process, the first residual generator or the second residual generator can be used for detection. Under the premise of reducing the amount of related data processing, the accuracy is compared with The combined detection method of the first and second residual generators is slightly reduced.

[0098] The non-Gaussian industrial process fault detection method disclosed in this embodiment (refer to Embodiment 1 above) includes:

[0099] The first step is to select a certain number of normal historical data sets, perform de-averaging preprocessing on the data sets, and establish a residual generator using the canonical correlation analysis method; specifically including:

[0100] Step A1. Use the historical database to select the process data under normal operating conditions, and use them respectively Represents the two variable data sets of the process, ...

Example Embodiment

[0121] Example 3

[0122] This embodiment is similar to the foregoing embodiment 2, and discloses a non-Gaussian industrial process fault detection method (refer to the foregoing embodiment 1), including:

[0123] The first step is to select a certain number of normal historical data sets, perform de-averaging preprocessing on the data sets, and establish a residual generator using the canonical correlation analysis method; specifically including:

[0124] Step A1. Use the historical database to select the process data under normal operating conditions, and use them respectively Represents the two variable data sets of the process, where l is the number of measured variables in the vector u, m is the number of measured variables in the vector y, and N is the number of independent sampling points; then u(i),i=1,...,N represents the vector The measured values ​​of l variables in u at the i-th sampling time, y(i), i=1,..., N represents the measured values ​​of m variables in the vector...

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 the field of industrial process monitoring and fault diagnosis, and discloses a fault detection method and system of a non-Gaussian industrial process in order to online detect the non-Gaussian industrial process conveniently and practically. The method includes: step 1, selecting a certain quantity of normal historical data sets, performing mean value preprocessing on the data sets, and establishing a residual generator by employing a canonical correlation analysis method; step 2, determining a corresponding threshold of the corresponding residual generator by employing a Monte Carlo method; and step 3, performing real-time online detection on industrial process data according to the determined threshold of the corresponding residual generator.

Description

technical field [0001] The invention relates to the field of industrial process monitoring and fault diagnosis, in particular to a non-Gaussian industrial process fault detection method and system. Background technique [0002] With the rapid development of information technology and data acquisition technology, factories and enterprises have abundant production data resources, and the era of industrial big data is slowly coming. In the process of industrial production, due to the harsh production environment and long-term operation, equipment will inevitably wear out or even fail. These abnormal conditions not only affect the quality of the product, but also affect the safe and stable operation of the system. Sometimes major production Accidents, relying entirely on the traditional detection methods of operators is not enough to solve complex process detection problems. [0003] The data in the actual production process is linear, nonlinear, Gaussian, non-Gaussian, etc. Fo...

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): G05B19/418
Inventor 陈志文彭涛阳春华袁小锋杨超杨笑悦
Owner CENT SOUTH UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap