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

Monitoring data fault detection method and system

A technology for monitoring data and fault detection, applied in digital data information retrieval, electrical digital data processing, special data processing applications, etc., can solve problems such as ignoring local structure information of different data points

Pending Publication Date: 2020-11-24
肖姝君
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] Most of the existing fault detection methods only consider the global structural information, while ignoring the local structural information between different data points

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
  • Monitoring data fault detection method and system
  • Monitoring data fault detection method and system
  • Monitoring data fault detection method and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0066] see figure 1 , is a flow chart of the monitoring data fault detection method provided by the first embodiment of the present invention, including steps:

[0067] Step S10, performing standardized processing on the monitoring data to be tested;

[0068] Specifically, in this step, the step of standardizing the monitoring data to be tested includes:

[0069] Obtain each variable in the monitoring data to be tested;

[0070] Subtracting the variable from the mean value obtained in the corresponding training data to obtain the target difference;

[0071] Dividing the target difference by the variance obtained in the training data completes the standard process.

[0072] Step S20, sampling the monitoring data to be tested to obtain sampled data;

[0073] Step S30, construct past vector and future vector according to described sampling data, and calculate past kernel matrix and future kernel matrix about described sampling data;

[0074] Step S40, standardize the past ke...

Embodiment 2

[0078] see figure 2 , is a flowchart of the monitoring data fault detection method provided in the second embodiment of the present invention, including steps:

[0079] Step S11, obtain dimensionless data, respectively select the first lag and the second lag to construct past vectors and future vectors, and construct past Hankel matrix and future Hankel matrix;

[0080] Step S21, selecting a radial basis kernel function, and calculating a past kernel matrix and a future kernel matrix;

[0081] Carry out matrix calculation to described past kernel matrix and described future kernel matrix respectively, to obtain the first Laplacian matrix and the second Laplacian matrix;

[0082] Step S31, calculating the parameter values ​​of α and β according to the optimization formula;

[0083] Specifically, this embodiment proposes a fault detection method based on Local Kernel Canonical Variable Analysis (LPP-KCVA), which takes into account local structure information and non-local str...

Embodiment 3

[0133] see Figure 19 , is a schematic structural diagram of the monitoring data fault detection system 100 provided in the third embodiment of the present invention, wherein:

[0134] The data sampling module 11 carries out standardized processing to the monitoring data to be measured, and samples the monitoring data to be measured to obtain sampling data;

[0135] Matrix calculation module 12, for constructing past vector and future vector according to described sampling data, and calculate past kernel matrix and future kernel matrix about described sampling data;

[0136] Statistic calculation module 13, used to carry out standardization process to described kernel matrix in the past and described future kernel matrix, and calculate the corresponding statistic with state space and residual space according to preset off-line model;

[0137] Fault judgment module 14, is used for comparing described statistic with preset control limit, when judging that described statistic is...

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 provides a monitoring data fault detection method and system. The method comprises the following steps: standardizing to-be-monitored data; sampling the to-be-monitored monitoring data to obtain sampling data; constructing a past vector and a future vector according to the sampling data, and calculating a past kernel matrix and a future kernel matrix about the sampling data; standardizing the past kernel matrix and the future kernel matrix, and calculating statistics corresponding to a state space and a residual error space according to a preset offline model; and comparing the statistics with a preset control limit, and when it is judged that the statistics is greater than the control limit at any moment, judging that the to-be-monitored data has a fault. According to the embodiment, the KCVA method and the LPP method are combined, the defect that the KCVA method ignores local structure information is overcome, and the data mapped to the high-dimensional space is constrained to a certain extent through the LPP method, so that even if the data is mapped to the high-dimensional space, the manifold structure in the low-dimensional space can still be reserved.

Description

technical field [0001] The present invention relates to the technical field of monitoring data processing, in particular to a monitoring data fault detection method and system. Background technique [0002] KCVA can be regarded as a nonlinear data dimensionality reduction method that focuses on global structural information and ignores local structural information. Although it can solve the nonlinear and dynamic problems of data at the same time, in the KCVA method, the data is mapped to After the high-dimensional space, it is very likely to cause data divergence and lead to the loss of a large amount of detailed information, which will affect the effect of the KCVA method in fault detection. In recent years, the manifold learning algorithm that has emerged can effectively extract local feature information. Manifold is a space with local Euclidean space properties. Manifold learning refers to embedding a low-dimensional manifold structure in a high-dimensional space in orde...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06F16/215G06F17/16
CPCG06F16/215G06F17/16
Inventor 肖姝君
Owner 肖姝君
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