A method for predicting failure trend of gas turbine compressor equipment
A technology for equipment failure and trend prediction, which is applied to computer parts, instruments, calculations, etc., can solve problems such as measurement concentration, and achieve the effect of saving time and cost
Active Publication Date: 2020-11-06
CHINA DATANG CORP SCI & TECH RES INST CO LTD EAST CHINA BRANCH
View PDF3 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
At the same time, in high-dimensional space, there is a phenomenon of metric concentration, and the traditional intra-class variance based on geometric distance is not suitable for Gaussian kernel function space.
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
[0023] First give the expression of the Gaussian kernel function:
[0024]
[0025] Use Maclaurin series to expand the formula, in order not to lose generality, shilling σ=1.
[0026]
[0027] where k is the dimension of vectors x and y, means satisfying n 1 +...+nk = all n of j 1 ,...,n k The number of combinations of sequences, all numbers in the sequence are non-negative integers,
[0028] It can be seen from the derivation of the above formula that the radial basis kernel function The definition formula is:
[0029]
[0030] From It can be seen that it is an infinite-dimensional vector.
[0031] Suppose x is a four-dimensional fault feature vector, and the feature value of each dimension is generally a decimal smaller than 1.
[0032] after mapping in the first five dimensions They are 1, 4, 10, 20 and 35 respectively.
[0033] As the number of dimensions increases, the number of combinations of sequences increases dramatically.
[0034] When k>...
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 predicting the failure trend of gas turbine compressor equipment, which selects specific operating parameters in a certain operating state of the compressor to form a feature vector, characterizes the operating state, and maps the feature vector to the Gaussian kernel function space through a Gaussian kernel function The classification is completed in , and the fractional norm is selected as the criterion of the distance measure in the high-dimensional space, and the separability index of the Gaussian kernel function space sample points is solved based on the fractional norm. The invention has the advantages that: the formula for solving the separability index of Gaussian kernel function space sample points is constructed based on the fractional norm, and the form of the mapping point of the feature vector in Gaussian space is modified to meet the requirements of engineering calculation accuracy And save time cost.
Description
technical field [0001] The invention relates to the technical field of high-dimensional data processing methods, in particular to a method for predicting failure trends of gas turbine compressor equipment. Background technique [0002] At present, there are few gas turbine units in service, and there are few equipment operating state parameters, so the fault trend prediction of gas turbine compressor equipment belongs to the category of small sample identification. The failure process of gas turbine compressor equipment is a complex process. A single operating parameter is not enough to reflect the performance status of the equipment. It needs to combine multiple parameters for combined prediction. The prediction of multi-dimensional data belongs to the category of high-dimensional space classification and identification. Therefore, there are two typical characteristics in the fault trend prediction of gas turbine compressor equipment—small sample size and high-dimensional s...
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
CPCG06F18/213G06F18/24147G06F18/2431
Inventor 徐搏超韩宏洲阮圣奇吴仲王松浩许昊煜李强胡中强任磊蒋怀锋陈开峰邵飞徐钟宇
Owner CHINA DATANG CORP SCI & TECH RES INST CO LTD EAST CHINA BRANCH