Nonparametric Probabilistic Total Transfer Capability Estimation with Deep Learning

A nonparametric deep-learning-based model is designed in this paper to estimate the probability distribution of the total transfer capability (TTC) in power systems. Differing from existing probabilistic TTC assessment approaches, the proposed method requires no a priori hypotheses on the estimated...

Full description

Saved in:
Bibliographic Details
Main Authors: Meng, Zichao, Guo, Ye, Tang, Wenjun, Sun, Hongbin, Li, Shiming, Dai, Yue
Format: Conference Proceeding
Language:English
Subjects:
adversarial training
Data models
Deep learning
graph convolutional network (GCN)
Maximum likelihood estimation
Network topology
Probabilistic logic
probabilistic total transfer capability (PTTC)
Probability distribution
Training
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:A nonparametric deep-learning-based model is designed in this paper to estimate the probability distribution of the total transfer capability (TTC) in power systems. Differing from existing probabilistic TTC assessment approaches, the proposed method requires no a priori hypotheses on the estimated probability distribution of TTC and the input data. In addition, the graph convolutional network (GCN) is used in the proposed method, which makes the algorithm effectively adaptive for the topology change in power systems. Moreover, the adversarial training is presented in the loss function formulation to improve the model's performance on "unseen" data. Numerical tests verify the proposed method on a modified 118-bus system.
ISSN:1944-9933
DOI:10.1109/PESGM48719.2022.9916886