A fault prediction method for oil-immersed transformer and related device
By combining multiple data features and sample equalization processing with an improved DCCNN model, the accuracy and adaptability issues of fault prediction for oil-immersed transformers are solved, achieving more efficient fault trend capture and prediction.
Patent Information
- Application Number
- CN202610260148.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2026-03-04
- Publication Date
- 2026-05-12
- Estimated Expiration
- 2046-03-04
AI Technical Summary
Existing methods for predicting faults in oil-immersed transformers are inadequate in terms of accuracy, adaptability, feature extraction, and output information, and cannot meet the real-time and stability requirements of power systems.
An improved DCCNN model is adopted, which combines dissolved gas analysis data, fault sensitivity ratio feature data and physical feature data of oil-immersed transformers. The receptive field is expanded by a dilated convolution module, and sample equalization and feature vector fusion are performed to improve the accuracy and adaptability of fault prediction.
It improves the accuracy and adaptability of fault prediction for oil-immersed transformers, reduces model bias and false negative rate, and meets the real-time and stability requirements of power systems.