Industrial system working condition monitoring method and system based on static and dynamic conjoint analysis
A technology of joint analysis and working conditions, applied in neural learning methods, character and pattern recognition, instruments, etc., can solve problems such as inapplicability, and achieve the effect of improving accuracy and enhancing robustness
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0043] As the industrial system is getting more complex, single from static or dynamic feature does not accurately distinguish the correct working condition, only the inclusion of the two will be more realistic to change condition changes. On the other hand, with the continuous operation of the industrial process, there will be many new unknown working information. Traditional division methods may not be updated because the model cannot be updated, and the unknown works may get an error division result, which in turn affects subsequent monitoring and control.
[0044] Inspired by the above practical industrial process, this patent proposes an incremental working condition division method based on static and dynamic combined analysis. First, the method of the present invention consider static features and dynamic features in the industrial process, and can effectively grasp the characteristics of the industrial process. And according to the changes in each type of characteristics, ...
Embodiment 2
[0101] This embodiment provides an industrial system working condition monitoring system based on static and dynamic combined analysis, including feature extraction modules, SOM network training modules, control limit computing modules, and working conditions change time point judgment modules;
[0102] The feature extraction module is used to obtain an industrial process data sequence of a stable working condition i = a prior to the current industrial process, and the industrial process data of each sampling point is used as one training data sample, using a slow feature analysis method to extract training data. Sample static feature vector S and dynamic feature vector
[0103] The SOM network training module is used to use a static feature vector S using all training data samples, training 1 SOM network to obtain a static SOM network corresponding to the stable working case i = a, add it to the static SOM network group SOM sta Remember the static SOM network group SOM sta Stati...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


