Weiyun Mao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Weiyun Mao has a total of 11 co-authors. Here are the top co-authors by number of shared publications:
- Bengang Wei
State Grid Shanghai Electric Power Research Institute, Shanghai 200437, China
1 shared publication - Chen Lu
Hangzhou International Innovation Institute, Beihang University, Hangzhou, China
1 shared publication - Chen Ren
State Grid Shanghai Electric Power Research Institute, Shanghai 200437, China
1 shared publication - Kai Gao
State Grid Shanghai Electric Power Research Institute, Shanghai, China
1 shared publication - Lu Chen
State Grid Shanghai Electric Power Research Institute, Shanghai 201199, China
1 shared publication - Tianyi Wu
Institute of Deep Learning Baidu Research Beijing China
1 shared publication
Latest Publications
- Open Access
Fault Diagnosis for Power Transformers through Semi-Supervised Transfer Learning
The fault diagnosis of power transformers is a challenging problem. The massive multisource fault is heterogeneous, the type of fault is undetermined sometimes, and one device has only met a few kinds of faults in the past. We propose a fault diagnosis method based on deep neural networks and a semi-supervised transfer learning framework called adaptive reinforcement (AR) to solve the above limitations. The innovation of this framework consists of its enhancement of the consistency regularization algorithm. The experiments were conducted on real-world 110 kV power transformers’ three-phase fault grounding currents of the iron cores from various devices with four types of faults: Phases A, B, C and ABC to ground. We trained the model on the source domain and then transferred the model to the target domain, which included the unbalanced and undefined fault datasets. The results show that our proposed model reaches over accuracy in classifying the type of fault and outperforms other popular networks. Our AR framework fits target devices’ fault data with fewer dozen epochs than other novel semi-supervised techniques. Combining the deep neural network and the AR framework helps diagnose the power transformers, which lack diagnosis knowledge, with much less training time and reliable accuracy.Open AccessFault Diagnosis for Power Transformers through Semi-Supervised Transfer Learning
The fault diagnosis of power transformers is a challenging problem. The massive multisource fault is heterogeneous, the type of fault is undetermined sometimes, and one device has only met a few kinds of faults in the past. We propose a fault diagnosis method based on deep neural networks and a semi-supervised transfer learning framework called adaptive reinforcement (AR) to solve the above limitations. The innovation of this framework consists of its enhancement of the consistency regularization algorithm. The experiments were conducted on real-world 110 kV power transformers’ three-phase fault grounding currents of the iron cores from various devices with four types of faults: Phases A, B, C and ABC to ground. We trained the model on the source domain and then transferred the model to the target domain, which included the unbalanced and undefined fault datasets. The results show that our proposed model reaches over accuracy in classifying the type of fault and outperforms other popular networks. Our AR framework fits target devices’ fault data with fewer dozen epochs than other novel semi-supervised techniques. Combining the deep neural network and the AR framework helps diagnose the power transformers, which lack diagnosis knowledge, with much less training time and reliable accuracy. - 6 September 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-4
DC Bias Phenomenon Recognition Based on Supervised Learning Model
dc bias phenomenon refers to the abnormal operation state of transformers whose vibration and noise signals increase due to the high voltage direct current (HVDC) transmission lines operating in ground return mode or the geomagnetically induced current (GIC). Shanghai Power Grid have built transformers' neutral point dc, vibration and noise monitoring platform since 2015. In this paper, Logistics Regression is firstly selected as a supervised learning model to judge whether the transformer has dc bias phenomenon based on the acquiring data. Two characteristic indexes representing dc bias phenomenon are found to improves the efficiency of model building. Simultaneously, appropriate regularization coefficients are chosen according to the evaluation effect and the physical meaning of model parameters. Considering most of data are unlabeled owing to the enormous effort of tagging data, a semi-supervised learning method is used to utilize the unlabeled data. Finally, an estimation model is established which calculates a probability value to indicate the risk of transformers' de bias phenomenon quantitatively. This model can remind operation and maintenance personnel strengthen devices inspection, and improves the efficiency of equipment management.
- 6 September 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-4
DC Bias Phenomenon Recognition Based on Supervised Learning Model
dc bias phenomenon refers to the abnormal operation state of transformers whose vibration and noise signals increase due to the high voltage direct current (HVDC) transmission lines operating in ground return mode or the geomagnetically induced current (GIC). Shanghai Power Grid have built transformers' neutral point dc, vibration and noise monitoring platform since 2015. In this paper, Logistics Regression is firstly selected as a supervised learning model to judge whether the transformer has dc bias phenomenon based on the acquiring data. Two characteristic indexes representing dc bias phenomenon are found to improves the efficiency of model building. Simultaneously, appropriate regularization coefficients are chosen according to the evaluation effect and the physical meaning of model parameters. Considering most of data are unlabeled owing to the enormous effort of tagging data, a semi-supervised learning method is used to utilize the unlabeled data. Finally, an estimation model is established which calculates a probability value to indicate the risk of transformers' de bias phenomenon quantitatively. This model can remind operation and maintenance personnel strengthen devices inspection, and improves the efficiency of equipment management.