Bearing characteristic data analysis method based on linear discriminant analysis
A linear discriminant analysis and characteristic data technology, applied in special data processing applications, electrical digital data processing, instruments, etc., can solve problems such as failure to distinguish, large error in fault classification, misjudgment, etc., to reduce coupling and improve The effect of discrimination
Inactive Publication Date: 2017-09-29
CHONGQING JIAOTONG UNIVERSITY +1
View PDF6 Cites 5 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Due to the harsh working environment of petrochemical equipment, the collected data has great uncertainty compared with the ideal environment. In addition, due to the existence of noise in the equipment itself and human operation errors and other factors, direct use of data for fault diagnosis often leads to misjudgment , can't even tell
In view of the large number of characteristic indicators in the original characteristic data of bearings and the large amount of sample data, the direct fault classification error of the samples is relatively large
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0050] Embodiment A bearing feature data analysis method based on linear discriminant analysis.
[0051] A method for analyzing bearing characteristic data based on linear discriminant analysis, comprising the following steps:
[0052] Step 1: Construct the bearing fault data matrix X from the number of bearing fault samples m and the number of bearing fault indicators n m×n , and define its optimal projection vector as w T ;
[0053] Step 2, define the mean μ of each type of fault samples i , and get the projected vector, and then introduce the hash matrix S i To measure the distribution between sample points in each class, and then get the sample hash value S in the class w , after the corresponding projection, it is obtained that each class sample point is relative to the center point of the class degree of hashing Then get the sample hash value projection of all classes
[0054] Step 3, take the number of sample points N in the category with more sample points ...
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
Login to View More Abstract
The invention discloses a bearing characteristic data analysis method based on linear discriminant analysis. Original data is projected toward the direction in which classification is most easily achieved based on the linear discriminant analysis to achieve dimensionality reduction, the projection direction is determined jointly through intra-class scattered matrices and inter-class scattered matrices of bearing sample data, and correctness of the projection direction is ensured. In addition, it is ensured that effective fault characteristic information can be retained while a large amount of error information is removed through linear discrimination and dimensionality reduction processing conducted on bearing fault data, meanwhile the dimensionality reduction direction is the projection direction in which fault classification can be performed more easily on data, and thus the accuracy rate of later BP neural network can be effectively improved.
Description
technical field [0001] The invention relates to the technical field of bearing feature data analysis, in particular to a bearing feature data analysis method based on linear discriminant analysis. Background technique [0002] In the bearing data, the following six indicators are often used to measure the operating state of the bearing, namely: vibration intensity, waveform, pulse, margin, peak value, and kurtosis indicators. Due to the harsh working environment of petrochemical equipment, the collected data has great uncertainty compared with the ideal environment. In addition, due to the existence of noise in the equipment itself and human operation errors and other factors, direct use of data for fault diagnosis often leads to misjudgment , can't even tell. In view of the large number of characteristic indicators in the original characteristic data of bearings and the large amount of sample data, the direct fault classification error of the samples is relatively large. ...
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
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06F19/00
CPCG16Z99/00
Inventor 黄大荣陈长沙孙国玺赵栋柴彦冲赵玲
Owner CHONGQING JIAOTONG UNIVERSITY