A Transformer Fault Diagnosis Method
A transformer fault and diagnosis method technology, applied in neural learning methods, biological neural network models, etc., can solve problems such as algorithm gradient disappearance, detection state quantity simplification, intelligent algorithm cannot carry out deep mining, etc., to achieve accurate prediction and prevent network The effect of vanishing gradients
Active Publication Date: 2021-08-06
SHENYANG HONGJI ELECTRICAL
View PDF3 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0004] In view of this, the purpose of the present invention is to provide a transformer fault diagnosis method to solve the problems in the prior art that the detected state quantity is too simplistic, the generated intelligent algorithm cannot be deeply excavated, and the algorithm has gradient disappearance and strong subjectivity.
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
[0040] The present invention will be further explained below in conjunction with specific embodiments, but the present invention is not limited thereto.
[0041] The invention provides a transformer fault diagnosis method, comprising the following steps:
[0042] S1: Divide the transformer historical data into a training set and a test set, where each historical data includes CH 4 ,C 2 h 6 ,C 2 h 4 ,C 2 h 2 ,CO,CO 2 ,H 2 , breakdown voltage, dielectric loss coefficient, acid value, furfural and water, a total of 12-dimensional transformer fault diagnosis state quantities;
[0043]The above-mentioned 12-dimensional state quantity has a great influence on the electrical fault degree, thermal fault degree and aging fault degree of the transformer. In this embodiment, the selected training set includes 800 transformer fault samples, and the test set includes 200 transformer fault samples, wherein the 800 samples in the training set include 200 labeled samples and 600 unla...
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 transformer fault diagnosis method. By extracting 12 kinds of transformer fault diagnosis state quantities of the transformer, combining the thermal fault, electrical fault and aging degree of the transformer, the fault of the transformer can be predicted from a wider level, which can be more accurate Predict transformer faults; use dimensionality reduction method to remove the correlation between parameters, and use parameters after removing redundant information to judge transformer faults more accurately; use single-layer training neural network method to determine through training Weights, rather than randomly assigning weights, can prevent the phenomenon of network gradient disappearance; compared with other intelligent methods, this method is objective and the prediction is more accurate.
Description
technical field [0001] The invention relates to the field of transformer fault diagnosis, and in particular provides a transformer fault diagnosis method. Background technique [0002] At present, transformer fault diagnosis methods mainly include ultrasonic fault detection method, partial discharge detection method, dissolved gas detection method in oil, etc., because the detection method of dissolved gas in oil is not affected by the strong electromagnetic environment of the transformer, so this method came into being Some intelligent algorithms (such as BP neural network, Bayesian algorithm, SVM, etc.) have also been widely used. However, this method only detects the dissolved gas content in oil, which is too simplistic and cannot better reflect the degree of transformer fault. According to The intelligent algorithm produced by it also has many disadvantages. For example, the gradient disappears easily when the BP neural network is trained. The essence of SVM is a binary ...
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): G06N3/02G06N3/08
CPCG06N3/02G06N3/08
Inventor 蒋辉马胤刚王巍孙鲜明
Owner SHENYANG HONGJI ELECTRICAL