Back to search

University of Denver

Deep Learning for Power Flow Estimation and High Impedance Fault Detection

Abstract

dc:description.abstract

My thesis is divided into two parts. The first part is: “Optimal Power Flow Estimation Using One-Dimensional Convolutional Neural Network [1]“. Optimal power flow (OPF) is an important research topic in power system operation and control decisions. Traditional OPF problems are solved through dynamic optimization with nonlinear programming techniques. For a large power system with large amounts of variables and constraints, the solving process would take a long time. This paper presents a new method to quickly estimate the OPF results using a one-dimensional convolutional neural network (1D-CNN). The OPF problem is treated as a high-dimensional mapping between the load inputs and the generator dispatch decisions. Therefore, through training the neural network to learn the mapping between loads and generator outputs, we can directly predict the OPF results with the load information of a system. In this paper, we built and trained a 1D-CNN to learn the mappings between system loads and generator outputs, and the 1D-CNN model was tested using IEEE 30, 57, 118, and 300 Bus systems. Extensive test and sensitivity study results have validated the effectiveness of using the 1D-CNN to estimate the OPF results. This part is from chapter 1 to chapter 6; The second part is: “Synthetic High Impedance Fault Data through Deep Convolutional Generated Adversarial Network [2]“. High impedance faults (HIFs) have always been significant challenges in the power grids. Researchers have developed some advanced protective methods to detect the HIFs. To test and validate these methods, large amounts of HIF data are required. This paper presents a synthetic HIF data generating method using the deep convolutional generated adversarial network (DCGAN). The DCGAN includes a generator module to create synthetic HIF waveform from random noises; and a discriminator module to identify the flaws of those synthetic data, which ultimately helps improve the quality of the synthetic data created by the generator. To test the fidelity of the generated synthetic HIF data, two different HIF-detection methods have been applied. Extensive simulation results have validated the effectiveness of using the DCGAN to create synthetic HIF data. This part is from chapter 7 to chapter 11.

Degree

thesis:*
Name thesis:degree_name
M.S.
Level thesis:degree_level
Masters Thesis

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Yang, Kun
Contributors dc:contributor
  • Rui Fan
  • Yunbo Yi
  • Mohammad Abdul Matin
  • Wenzhong Gao

Subjects

dc:subject × 7

Rights

dc:rights
Statement dc:rights
  • <p>Copyright is held by the author. User is responsible for all copyright compliance.</p>
  • Kun Yang
Language dc:language
en

Identifiers

dc:identifier.*
Repository record dc:identifier
https://digitalcommons.du.edu/etd/2180

Chain of custody

source
Harvested from
University of Denver
Base URL
digitalcommons.du.edu/do/oai/
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Yang, Kun. Deep Learning for Power Flow Estimation and High Impedance Fault Detection. Masters Thesis thesis, https://digitalcommons.du.edu/etd/2180