Title page for ETD etd-03302007-232205

Type of Document

Dissertation

Author

Zhu, Dan

URN

etd-03302007-232205

Title

Electric Distribution Reliability Analysis Considering Time-varying Load, Weather Conditions and Reconfiguration with Distributed Generation

Degree

PhD

Department

Electrical and Computer Engineering

Advisory Committee

Advisor Name	Title
Broadwater, Robert P.	Committee Chair
Arthur, James D.	Committee Member
Cyre, Walling R.	Committee Member
Liu, Yilu	Committee Member
Tam, Kwa-Sur	Committee Member

Keywords

Time-varying Load.
Reconfiguration for Restoration
DG
Power Distribution System
Reliability Improvement
Storm Outage Prediction
Reliability Analysis

Date of Defense

2007-03-27

Availability

unrestricted

Abstract

This dissertation is a systematic study of electric power distribution system
reliability evaluation and improvement. Reliability evaluation of electric power systems
has traditionally been an integral part of planning and operation. Changes in the electric
utility coupled with aging electric apparatus create a need for more realistic techniques
for power system reliability modeling.
This work presents a reliability evaluation technique that combines set theory and
Graph Trace Analysis (GTA). Unlike the traditional Markov approach, this technique
provides a fast solution for large system reliability evaluation by managing computer
memory efficiently with iterators, assuming a single failure at a time. A reconfiguration
for restoration algorithm is also created to enhance the accuracy of the reliability
evaluation, considering multiple concurrent failures. As opposed to most restoration
simulation methods used in reliability analysis, which convert restoration problems into
mathematical models and only can solve radial systems, this new algorithm seeks the

reconfiguration solution from topology characteristics of the network itself. As a result
the new reconfiguration algorithm can handle systems with loops.
In analyzing system reliability, this research takes into account time-varying load
patterns, and seeks approaches that are financially justified. An exhaustive search scheme
is used to calculate optimal locations for Distributed Generators (DG) from the reliability
point of view. A Discrete Ascent Optimal Programming (DAOP) load shifting approach
is proposed to provide low cost, reliability improvement solutions.
As weather conditions have an important effect on distribution component failure
rates, the influence of different types of storms has been incorporated into this study.
Storm outage models are created based on ten years’ worth of weather and power outage
data. An observer is designed to predict the number of outages for an approaching or on
going storm. A circuit corridor model is applied to investigate the relationship between
power outages and lightning activity.

Files

Filename		Size	Approximate Download Time (Hours:Minutes:Seconds)
Filename		Size	28.8 Modem	56K Modem	ISDN (64 Kb)	ISDN (128 Kb)	Higher-speed Access
	dissertation_V6.pdf	932.67 Kb	00:04:19	00:02:13	00:01:56	00:00:58	00:00:04