A Method for Early Diagnosis of Mechanical Faults Based on Time Series
A technology for early diagnosis of mechanical faults, applied to computer components, instruments, calculations, etc., can solve problems such as phase offset and difficulty in achieving high accuracy
Active Publication Date: 2021-03-09
HEFEI UNIV OF TECH
View PDF13 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, the current extraction and selection of shapelets are based on Euclidean distance, and in practical applications, mechanical fault sequences often have phase shifts on the time axis, and it is difficult to achieve high accuracy based on shapelets
[0008] Therefore, there is currently no time series early classification method that is well suited for early diagnosis of mechanical faults, that is, to diagnose as early as possible while ensuring the accuracy of mechanical fault diagnosis
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
[0046] In this example, refer to figure 1 , a mechanical fault early diagnosis method based on time series early classification is carried out as follows:
[0047] Step 1: Obtain a set of mechanical fault sequence data as training samples, the training samples are composed of mechanical fault characteristic data TS={TS 1 ,TS 2 ,···,TS i ,···,TS N} and mechanical failure classification label data Y={y 1 ,y 2 ,···,y i ,···,y N}, where N represents the number of training samples, TS i Represents the i-th piece of mechanical fault feature data in the training sample, and has: L i Indicates the length of the i-th piece of mechanical fault feature data in the training sample, t j Indicates the jth time in the training sample, x i,j Indicates the feature value corresponding to the i-th piece of mechanical failure feature data in the training sample at the j-th time; y i Represents the i-th piece of mechanical fault feature data TS in the training sample i The correspond...
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 method for early diagnosis of mechanical faults based on time series, comprising: 1. truncating the characteristic data of mechanical fault monitoring sequence data into s mechanical fault sequence data according to the length r; 2. selecting a mechanical fault classifier , train on each truncated training sample and give the mechanical fault classification results of all mechanical fault feature data; 3, count the mechanical fault classification accuracy rate of each class on each truncated training sample, and calculate the mechanical fault classification The trust degree of the result; 4. Determine the trust degree threshold θ of the mechanical fault classification result; 5. Use the selected mechanical fault classifier and the trust degree threshold θ to perform early diagnosis on the mechanical fault test data. The invention is not only suitable for the fault diagnosis of mechanical fault sequence data of equal length, but also suitable for fault diagnosis of mechanical fault sequence data of unequal length, and can predict the type of mechanical fault under the condition of ensuring the accuracy of mechanical fault diagnosis.
Description
technical field [0001] The invention relates to the field of mechanical fault diagnosis, in particular to a time series-based early diagnosis method for mechanical faults. Background technique [0002] Mechanical fault diagnosis is a technology that understands and masters the state of the machine during operation, determines its overall or partial normal or abnormality, detects faults and their causes early, and can predict the development trend of faults. Oil monitoring, vibration monitoring, noise monitoring, performance trend analysis and non-destructive testing are the main diagnostic techniques. [0003] The signals that can be used for mechanical condition monitoring and fault diagnosis include vibration diagnosis, oil sample analysis, temperature monitoring and non-destructive testing and flaw detection, supplemented by other technologies or methods. Among them, the field of vibration diagnosis is the most extensive, the theoretical basis is the most solid, and the ...
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 Patents(China)
IPC IPC(8): G06K9/62
CPCG06F18/24G06F18/214
Inventor 吕俊伟胡学钢李培培廖建兴
Owner HEFEI UNIV OF TECH