University of Southern California Dissertations and Theses

Impacts of high solar penetration on power transmission systems

(USC Thesis Other)

Impacts of high solar penetration on power transmission systems

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Master's Thesis
IMPACTS OF HIGH SOLAR PENETRATION ON
POWER TRANSMISSION SYSTEMS
by
Bingtao GAO
A Thesis Submitted to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulllment of the
Requirements of the Degree of
MASTER OF SCIENCE
(ELECTRICAL ENGINEERING)
MAY 2017
© Copyright 2017 I Bingtao GAO
Master's Thesis
Table of Contents
1 INTRODUCTION 1
1.1 Impacts of High Solar Penetration on Power Systems . . . . . . . . . 2
1.2 Methodologies and Tools . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.1 Snapshot method . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.2 Analyzing software . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Thesis Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2 SYSTEM DESCRIPTION 11
2.1 IEEE RTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1.1 Bus Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1.2 Load Curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2 Base Line Load Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.3 Realization of PV penetration . . . . . . . . . . . . . . . . . . . . . . 19
3 POWER FLOW ANALYSIS 27
3.1 Voltage Prole . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.1.1 Impact of PV penetration on bus voltage . . . . . . . . . . . . 27
3.1.2 Simulation result . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.2 Real Power Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.3 N-1 Contingency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.3.1 Special action on L0610 open case . . . . . . . . . . . . . . . . 32
3.3.2 Contingency for base line . . . . . . . . . . . . . . . . . . . . . 34
3.3.3 Contingency with PV penetration . . . . . . . . . . . . . . . . 36
3.4 System Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . 38
II
3.4.1 Problematic system conguration . . . . . . . . . . . . . . . . 38
3.4.2 Improvement devices . . . . . . . . . . . . . . . . . . . . . . . 40
3.4.3 Real power curtailment due to voltage control inverters . . . . 42
3.5 Early evening load peak without PV . . . . . . . . . . . . . . . . . . 43
3.5.1 Supply shortage after sunset . . . . . . . . . . . . . . . . . . . 43
3.5.2 Capacity factor issue . . . . . . . . . . . . . . . . . . . . . . . 44
4 SHORT CIRCUIT ANALYSIS 47
4.1 Event Denition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.2 Base Line SC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.3 SC Current Variation due to PV Penetration . . . . . . . . . . . . . . 51
4.4 SC Analysis for Improved RTS . . . . . . . . . . . . . . . . . . . . . . 54
5 RELIABILITY ANALYSIS 56
5.1 Reliability Indices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.2 Reliability Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.3 Base Line Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.4 Change of Reliability Indices due to PV Penetration . . . . . . . . . . 60
5.5 Reliability Analysis for Improved RTS . . . . . . . . . . . . . . . . . 60
6 CONCLUSIONS 62
III
Master's Thesis
List of Tables
1.1 Summary of maximum PV penetration levels suggested in the litera-
ture [1] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.1 Data of Generators at Each Bus . . . . . . . . . . . . . . . . . . . . . 13
2.2 Data of Bus Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3 Simplied load data for reliability analysis . . . . . . . . . . . . . . . 16
2.4 Load Flow Result for Base Case . . . . . . . . . . . . . . . . . . . . . 17
2.5 Generation Mix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.6 Parameters of a Solar Panel under STC . . . . . . . . . . . . . . . . . 20
2.7 Number of PV plants needed on each bus . . . . . . . . . . . . . . . . 22
2.8 PV Penetration Level Variation with Replacement of Coal Plants . . 25
3.1 WECC Disturbance-Performance Table of Allowable Eects on Other
Systems [2] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.2 Impedance and Rating Data of RTS [3] . . . . . . . . . . . . . . . . . 32
3.3 Base Line RTS Contingency Violations . . . . . . . . . . . . . . . . . 35
3.4 38% Penetration RTS Contingency Violations . . . . . . . . . . . . . 37
3.5 Improved 38% Penetration RTS Contingency Violations . . . . . . . . 41
3.6 Approximation of Gas Turbine Operation Time . . . . . . . . . . . . 44
3.7 Hourly Peak Load in Percent of Daily Peak [3] . . . . . . . . . . . . . 46
4.1 Base Case SC Currents at Each Bus . . . . . . . . . . . . . . . . . . . 50
5.1 Part of the Reliability Indices for RTS Base Case . . . . . . . . . . . 59
5.2 EENS on Dierent Load States . . . . . . . . . . . . . . . . . . . . . 59
5.3 Comparison of RTS and Improved RTS in Aspect of Reliability Indices
(38% PV penetration) . . . . . . . . . . . . . . . . . . . . . . . . . . 60
IV
Master's Thesis
List of Figures
1.1 Renewable electricity generation from 2000 and projected to 2040 [4] 2
2.1 One-line diagrams of the two IEEE RTS models . . . . . . . . . . . . 12
2.2 Hourly peak load through 364 days . . . . . . . . . . . . . . . . . . . 15
2.3 One-line diagram of the RTS model . . . . . . . . . . . . . . . . . . . 18
2.4 Solar panel characteristics in ETAP . . . . . . . . . . . . . . . . . . . 21
2.5 Conguration and power generation of PV arrays . . . . . . . . . . . 23
2.6 Conguration and power output of inverters in solar plants . . . . . . 24
2.7 Normal real power output of solar plants under given conditions . . . 26
3.1 Voltage prole of each bus due to PV penetration . . . . . . . . . . . 28
3.2 Total real power loss of the system with PV penetration . . . . . . . 30
3.3 One line diagram of cable L0610 . . . . . . . . . . . . . . . . . . . . . 33
3.4 Voltage violations due to PV penetration . . . . . . . . . . . . . . . . 36
3.5 One line diagram of transmission line L0708 . . . . . . . . . . . . . . 39
3.6 One line diagram of Bus 3 . . . . . . . . . . . . . . . . . . . . . . . . 39
3.7 Bus 1 and Bus 2 power
ows with 38% PV penetration RTS . . . . . 42
4.1 Default X/R ratio settings for the 76 MW coal power steam turbine
in ETAP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.2 LG fault current variations due to PV penetration . . . . . . . . . . . 52
4.3 3 phase fault current variations in 138 kV area due to PV penetration 53
4.4 LL fault current variations in 138 kV area due to PV penetration . . 54
4.5 Comparison of RTS and improved RTS in aspect of LG fault current 55
5.1 The reliability conguration of a solar plant at Bus 1 . . . . . . . . . 58
5.2 The variations of SAIDI, SAIFI, and EENS in RTS due to PV pene-
tration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
V
Master's Thesis
NOMENCLATURE
RPS Renewable Portfolio Standards
CPP Clean Power Plan
AEO2017 Annual Energy Outlook 2017
EIA Energy Information Administration
PV Photovoltaic
DC Direct Current
AC Alternating Current
PF Power Flow
SC Short Circuit
RTS Reliability Test System
IEEE Institute of Electrical and Electronics Engineers
CB Circuit Breaker
UI User Interface
CTPG California Transmission Planning Group
NERC North American Electric Reliability Corporation
WECC Western Electricity Coordinating Council
PSO Public Service Company of Oklahoma
SCE Southern California Edison
LTC Load Tap Changing
STC Standard Test Conditions
AM1.5 Air Mass 1.5
MP Maximum Power
SVC Static VAR Compensator
ANSI American National Standards Institute
VI
LG Line-to-Ground
LL Line-to-Line
LLG Line-to-Line-to-Ground
FLA Full-Load Ampacity
SAIFI System Average Interruption Frequency Index
SAIDI System Average Interruption Duration Index
CAIDI Customer Average Interruption Duration Index
ASAI Average Service Availability Index
ASUI Average Service Unavailability Index
EENS Expected Energy Not Supplied
MTTF Mean Time To Failure
MTTR Mean Time To Repair
NREL National Renewable Energy Laboratory
VII
Master's Thesis
ABSTRACT
The studies of system with high solar penetration are highly demanded due to the
RPS requirement of rapid increase in renewable energy. In this thesis, a 24-bus
transmission RTS model is built using ETAP software. Totally four analyses are
conducted on the model including power
ow, contingency, short circuit, and reli-
ability analysis. The simulation result indicates that a high PV penetration level
of 38% is achievable in RTS if two capacitors are installed on Bus 3 and 6, several
voltage control inverters are implemented on Bus 1, 2, and 15, and the interruption
ratings of protection devices at buses with PV plants are redesigned to coincide with
the low fault current. This thesis is a preliminary evaluation of a power transmission
system with high solar penetrations.
X
CHAPTER 1: INTRODUCTION
The limited reserve of fossil-fuel energy resources and the rising environmen-
tal concerns about reduction of greenhouse gas emission and air pollution require a
widespread and rapid deployment of renewable energy systems to meet the increasing
electric power demand. Many countries require the power generation from renewable
sources to meet the specic Renewable Portfolio Standards (RPS) and Clean Power
Plan (CPP). In California, the Senate Bill X1-2 was signed in 2011, setting the RPS
goal at 33% by 2020 [5]. After that, in 2015, the Senate Bill 350 required that the
amount of electricity generated and sold to retail customers per year from eligible
renewable energy resources be increased to 50% by 2030 [6]. Such a rapid increase
in the renewable penetration level aects the control, protection and design of power
systems signicantly. Therefore, more accurate models of renewables and more ad-
vanced power system designs are needed to study the impacts of high renewable
penetration systems.
Renewable energy is generated from sunlight, wind, geothermal, biomass, tides
and other resources which can be naturally replenished in a rate that is comparable
or higher than human consumption. About 10 years ago, the major component of
renewable energy is from hydropower. With further study and observation of hydro
power plant, its sustainability and eect upon environment become quite contro-
versial. Therefore, some CPP and state-dened RPS strictly limit the denition of
\renewable" hydro facilities and many of them are demanding increments in solar
and wind electricity. According to a projection in Annual Energy Outlook 2017
(AEO2017) released by U.S. Energy Information Administration (EIA), encouraged
by decrease of cost, improvement of performance, and considerable tax credits, wind
and solar will become the predominant sources of renewable generation before 2040,
1
with each surpassing hydroelectric generation [4], as shown in Figure 1.1.
Figure 1.1: Renewable electricity generation from 2000 and projected to 2040 [4]
1.1 Impacts of High Solar Penetration on Power Systems
Despite of the benets of renewable energy, high solar penetration imposes strong
impacts and serious challenges on power systems in many aspects. Unlike conven-
tional fossil-fuels, solar power is largely depended on weather conditions which are
neither \controllable" nor fully predictable. Its inherent intermittency makes the
dispatch of power extremely dicult. Energy storage devices or backup combustion
turbines are able to deal with some problems concerning insolation variations, while
real-time data collection and well-designed controllers are also required, which would
greatly increase the cost.
The most common utilization of solar power is through photovoltaic (PV) panels,
which absorbs the quantum energy of photons in sunlight into electrons hence forms
a current through semiconductor materials. The generated direct current(DC) is
then inverted via inverters to alternating current (AC) which can be connected into
power grids. Generally, large scale solar power plants have two dierences from
2
conventional steam turbines that incur serious problems to power grids:
• Solar plants do not generate reactive power which provides voltage support for
the systems, which generates voltage regulation diculties.
• The lack of rotating inertia creates great problems on frequency regulation in
transient conditions.
The weakness in stability and resilience further aects the reliability of power sys-
tems. Besides, some aspects also studied in this thesis include real power losses and
SC currents. All of the above will be discussed thoroughly and detailedly in the fol-
lowing chapters. There are other respects concerning high solar penetration, such as
relay settings, harmonics, islanding and system protection. Also, some social factors
like cost and policy are very crucial to the construction and operation of the system.
They are beyond the reach of this thesis hence will not be analyzed in detail.
1.2 Methodologies and Tools
In this thesis, power
ow (PF) analysis, n 1 contingency analysis, short-circuit
(SC) analysis, and reliability analysis are conducted on a Reliability Test System
(RTS) proposed by Institute of Electrical and Electronics Engineers (IEEE) RTS
Task Force of the Application of Probability Methods (APM) Subcommittee [3]
using ETAP software. The RTS is discussed in detail in Chapter 2.
1.2.1 Snapshot method
Many of the analyses in power systems are based on snapshot method. A power
system is determined under a particular state like peak load or high PV penetration.
Then voltage prole and load
ow are obtained by iteratively solving PF equations
3
of the system. Snapshot method is important for power system analysis because
it re
ects the behavior of system under the specic state of interest such as steady
state, and it also indicates the inherent ability of power system in perspectives of
stability and reliability.
The most fundamental method to analyze a power system is through PF analysis.
PF analysis evaluates the thermal and electrical performance of the system through
iteratively solving the PF equations. The result of PF analysis tells the performance
of the system in aspects of voltage and current on each bus, active and reactive power
from the generator or to the load, and the power
owing through each branch. The
most commonly used iteration algorithms include Newton-Raphson, Gauss-Seidel,
and fast decoupled PF methods.They are expected to derive the same answer for the
system with only dierence in number of iterations before convergence.
PF analysis can be applied under dierent conditions of the system to conduct
the contingency analysis. Usually a base case will be determined to indicate the
referential performance of the system. Then the system will be switched to varieties
of contingencies to evaluate its resilience toward any line or transformer outage.
The n 1 contingency analysis is a standard approach of contingency analysis, in
which case an automatic opening of circuit breaker (CB) will be triggered due to
disturbances, SCs, or any kind of fault, and causes a single-line outage in the system
[7, 8]. It simulates the most common outage in power system and has a series of
well-regulated criteria to follow.
SC is often related to power system stability and security. If a SC occurs at a
certain bus, usually a voltage drop and a current increase will be observed. The CBs
then have to operate correctly and quickly in order to prevent any further damage to
the system, and the operators or some smart grid devices have to nd out the location
of the fault and be able to clear it. The SC analysis can be viewed as another form
4
of contingency analysis, with line outages being changed to bus grounding faults.
Another power system evaluation is through reliability analysis since reliable load
power supplies are essential in modern society. Instead of using electrical parameters
in PF analysis, reliability analysis is based on the element outage probability model.
The system elements are modeled as blocks in a diagram associated with dierent
states like operating (up) and failure (down), and transition time between (or prob-
abilities of) two states. The blocks are linked in the same way the corresponding
components are connected to form a path. Then a simulation is conducted for the
cascaded blocks to calculate the frequency of each path to failure as well as the aver-
age failure time. Usually a Monte Carlo method is used, which randomly determine
the states of all components on the basis of their probabilities. Sometimes, if the
component states are independent with other components, some simpler statistic
methods can be applied. A series of indices are computed to quantify the system
adequacy, such as the interruption frequency, the amount of load loss, and the inter-
ruption cost.
1.2.2 Analyzing software
ETAP is a versatile commercial program that integrates functions needed for
power system analysis from modeling to operation. Its graphical user interface (UI)
provides a suite of built-in models including PV panels and converters. Also it can
export a series of available outcomes in dierent le formats.
1.3 Literature Review
The evaluation of power system is a methodological process that follows a set
of recognized steps. Some graduate students have nished the study of high PV
5
penetration impacts on power systems [9, 10, 11]. In 2010 and 2011, the California
Transmission Planning Group (CTPG) has published four phased reports so as to ad-
dress the system requirements that allow electricity utilities to meet their obligations
in the 33% renewable penetration goal under California RPS[12, 13, 14, 15]. Based on
the reliability standards of North American Electric Reliability Corporation (NERC)
and Western Electricity Coordinating Council (WECC), the rst phase of CTPG re-
port developed a set of general guidelines and criteria for contingency analysis of
CTPG system [12]. After that, the second phase aims to rene the methodology
as well as produce a set of solutions to mitigate the violations [13]. The last two
phases then further study and determine the potential sites for state's RPS goals, and
claim that simple transmission line additions will not make the system to fully meet
California RPS requirements [14, 15]. The entire report of CTPG provided a good
demonstrative guideline of performing an evaluation of high renewable penetration
power system on transmission level.
Small penetration of PV is proved to be practical in a report from Southern Cal-
ifornia Edison (SCE) [16], which witnessed the operation of a central-station solar
plant with MW-size in SCE area and claimed good performance during the rst year.
As for higher penetrations of solar generation, despite of the mature analyzing meth-
ods that are used, the conclusions dier from system to system due to the distinctive
factors they took into consideration. A study in Arizona state [17] examined the in-
tegration of central-station solar plants into a utility system. It simulated the system
under dierent shading conditions and found out that the PV penetration could not
exceed 5% because of diculties in tracking PV changes. Meanwhile, a research with
dispersed PV generation in Public Service Company of Oklahoma (PSO) [18, 19] con-
cluded that even for the smallest installation of 15% solar generation, certain cloud
patterns could incur signicant impacts to the system. However, the so-called 15%
penetration is not exactly accurate since on the light load spring condition where the
6
main problem occurred, the actual penetration level was 45% according to [18]. An-
other paper [20] used a criteria of 1% maximum
uctuation to evaluate the in
uence
of PV output variation, and derived dierent penetration limits from 1.3% to 35.8%
depending on the dispersion level of PV generation (smallest for the most centralized
PV plant). In 2012, a group of researchers in Caltech suggest that the solar PV rate
in California could reach between 15-50% by 2020 [21]. Since some simulations and
experiments are conducted on transmission systems while others are on distribution
systems, the dierences of two power system levels are also crucial to the evaluation
of performance. Any result in performance of solar penetrated transmission system
will not necessarily be obtained in a small-scale distribution system [22].
Literature [1] from Sandia National Laboratories summarized some outcomes
of maximum penetration level from previous researchers together with the factors
causing the limitation, as listed in Table 1.1.
7
Table 1.1: Summary of maximum PV penetration levels suggested in the literature
[1]
Reference Maximum PV Cause of Upper Limit
Number Penetration Level
[17] 5%
Ramp rates of mainline generators.
PV in central-station mode.
[18] 15%
Reverse power swings during cloud transients.
PV in distributed mode.
[23] No limit found Harmonics.
No problems caused by clouds, harmonics, or
[24] >37% unacceptable responses to fast transients were found.
Experimental and theoretical study.
Varied from
1.3% to 35.8%
Unacceptable unscheduled tie-line
ows.
The variation is caused by the geographical extent of
[20] the PV (1.3% for central-station PV).
Results particular to the studied utility because of the
specic mix of thermal generation technologies in use.
Equal to minimum
load on feeder
Voltage rise.
[25] No load tap changes (LTCs) in the medium or
higher voltage transformer banks.
Primarily voltage regulation, especially unacceptably
[26, 27] <40% low voltages during false trips, and malfunctions of
step voltage regulators (SVRs).
Minimum distribution system losses occurred here.
[28] 5% Could be nearly doubled if inverters were equipped
with voltage regulation capability.
Voltage rise.
The limit of 33% is imposed by a very strict reading
[29] 33% or50% of the voltage limits in the applicable standard.
The excursion beyond that voltage limit at 50%
penetration was extremely small.
8
1.4 Thesis Organization
In this thesis, a modied IEEE RTS model is built in ETAP software with addition
of PV power plants and related elements. Load
ow analysis including contingencies,
SC analysis and reliability analysis are conducted on this model from 0 to a high PV
penetration level of 37%.
In Chapter 1, the impacts of hight PV penetration on transmission power systems
are discussed. Meanwhile, the dierent types of analytical methods for power system
with high PV penetration level are presented, together with a software platform to
realize them. A summarization of review of literature is also completed at the end
of this chapter.
In Chapter 2 a transmission power system model based on IEEE RTS is described.
A base line is set for the model with no PV penetration and peak load demand. A
base line load
ow analysis is conducted and the results are used as the reference
for other cases. Then the coal power plants are gradually replaced by PV power
plants with identical real power capacities unit by unit so as to reach a highest PV
penetration level of 37%. The parameters of the system, the operation rules, and
the replacement instructions are presented in this chapter.
In Chapter 3, RTS is test by PF analysis with increasing of PV penetrations. A
normal PF analysis is conducted to depict the voltage prole change of the system
as well as the real power loss variations due to inserted PV. Then a whole set of
n 1 contingency analysis is used on each penetration level. The reason for voltage
drop in PV penetrated system is discussed based on the results. Some fragile and
sensitive system settings are pointed out, and the corresponding improvements are
proposed. Then a special case of early evening load peak is mentioned in order to
further discuss the reliability of the system.
9
In Chapter 4, the SC analysis is conducted to determine the magnitude of fault
current of each bus following dierent fault types. A comparison between the im-
proved RTS and RTS with 38% penetration is performed in aspect of fault currents.
In Chapter 5, the system reliability is measured by reliability indices, and the varia-
tion of those indices due to PV penetration is discussed. The improved RTS is also
tested to evaluate its eect on reliability of the system. In Chapter ??, the results
from previous chapters are collected together, and a conclusion of impacts of high
PV penetration level is made. Some suggestions on the improvement of system in
voltage regulation are provided. The limitation and future works on this thesis is
then presented.
10
CHAPTER 2: SYSTEM DESCRIPTION
2.1 IEEE RTS
The system being studied in this thesis is built mainly on basis of IEEE RTS
model in 1979 [3], which is also called the RTS-79 model and is shown in Figure
2.1a. It is a transmission level power system with total generation capacity of 3405
MW and peak load of 2850 MW. It has 24 buses, 32 generation units, 17 loads, and
38 lines or transformers at 138 or 230 kV level. The data used in this RTS model
from [3] include hourly peak load through 364 days, generating unit capacity, bus
load and voltage correction device ratings, transmission line impedances and thermal
ratings, and reliability parameters of all the components in the system.
However, since the original data report [3] did not provide any restrictions on
bus data, it is usually used with combination of another IEEE RTS model built
in 1996 [30] which contains more information about bus settings and data. This
system is also called RTS-96, which is presented in Figure 2.1b. The reason they
have identical name is that, RTS-96 is an integration of three RTS-79 system with
5 extra transmission lines and 1 optional DC link. Sometimes, RTS-79 is also called
\One Area RTS-96" [30].
The model in this thesis has the same conguration as RTS-79. However, under
steady state, it is operating at the point where the rst area in RTS-96 is at [30],
and will gradually replace the fossil-fueled power plants in it by PV plants in order
to realize PV penetration. So basically, either the name of RTS-79 or RTS-96 cannot
perfectly indicate its characteristic. For convenience, it will be called the modied
RTS, or simply RTS.
11
(a) RTS-79 [3]
(b) RTS-96 [30]
Figure 2.1: One-line diagrams of the two IEEE RTS models
12
2.1.1 Bus Data
The generation capacity is very important for the system, which is determined
in Table 2.1 [30]. The types of data are dened in Table 2.2, along with the bus
levels and load ratings [3]. Under condition of peak load, all the generation units are
operating, and the swing bus is dened as the one with largest generation capacity
(Bus 13).
Table 2.1: Data of Generators at Each Bus
Bus ID Type V
op
P
G
Q
max
Q
min
# # p.u. (MW) (Mvar) (Mvar)
1 1 & 2 OCT 1.035 20 10 0
1 3 & 4 CST 1.035 76 30 -25
2 1 & 2 OCT 1.035 20 10 0
2 3 & 4 CST 1.035 76 30 -25
7 1 - 3 OST 1.025 100 60 0
13 1 - 3 OST 1.020 197 80 0
14 1 Sync.Cond. 0.980 0 50 -200
15 1 - 5 OST 1.014 12 6 0
15 6 CST 1.014 155 80 -50
16 1 CST 1.017 155 80 -50
18 1 NCL 1.050 400 200 -50
21 1 NCL 1.050 400 200 -50
22 1 - 6 HYD 1.050 50 16 -10
23 1 & 2 CST 1.050 155 80 -50
23 3 CST 1.050 350 150 -25
(OCT: Oil Combustion Turbine, OST: Oil Steam Turbine, CST: Coal Steam
Turbine, Sync.Cond.: Synchronous Condenser, NCL: Nuclear, HYD: Hydropower)
13
Table 2.2: Data of Bus Loads
Bus Type V
base
P
L
Q
L
Bus Type V
base
P
L
Q
L
# (kV) (MW) (Mvar) # (kV) (MW) (Mvar)
1 G 138 108 22 13 S 230 265 54
2 G 138 97 20 14 G 230 194 39
3 L 138 180 37 15 G 230 317 64
4 L 138 74 15 16 G 230 100 20
5 L 138 71 14 17 L 230 0 0
6 L 138 136 28 18 G 230 333 68
7 G 138 125 25 19 L 230 181 37
8 L 138 171 35 20 L 230 128 26
9 L 138 175 36 21 G 230 0 0
10 L 138 195 40 22 G 230 0 0
11 L 230 0 0 23 G 230 0 0
12 L 230 0 0 24 L 230 0 0
(G: Generator Bus, L: Load Bus, S: Swing Bus)
2.1.2 Load Curve
The report [3] also oers a set of load from hourly to daily, weekly, and yearly
peak. The yearly peak load is 2850 MW, which is set as 100% of the total load in
the system. The hourly load data present the peak load in every hour of the year.
Multiplied by 24 hours per year, 7 days per week and 52 weeks every year, the total
number of hourly peak load data is
24 7 52 = 8736: (1)
14
The 8736 states of hourly peak load form a very detailed load curve throughout a
year, which is drawn in Figure 2.2. The dashed prole above the curve in the gure
is weekly peak load. It can be seen from the curve that load in RTS changes quickly
in everyday ranging from about 50% to 100%, and it is really hard to simulate the
PF of every condition in detail. Even the weekly peak load change is too complicated
to be analyzed in every aspect. Meanwhile, for snapshot analysis method, too many
cases would make the results vague and implicit. Therefore, a simplication in load
curve is implemented.
Figure 2.2: Hourly peak load through 364 days
The criteria for load curve simplication in dierent analyses are as followed:
1. PF analysis: A PF analysis assume generation buses to be fully loaded, and the
exceeded power would
ow out from the swing bus. In this way, a low power
demand requires shutdown of some generators which relies on rigorous power
dispatch guidelines. Meanwhile, the lower demand case may not be explicitly
useful if the generator units are shut down relatively in a geographically equal
way, similar to globally reducing the generator capacity in the same ratio.
15
For example, a 25% solar penetrated RTS with 50% load and 50% generator
capacity down is very likely to perform in the same pattern as a 50% penetrated
full-loaded RTS does. As a result, the thesis only consider nearly full load
condition in PF analysis, when all generating units are operating.
2. N-1 contingency analysis: The contingency analysis simulates all line and trans-
former outages in the system as PV penetration level increases. The major part
of it will be conducted under peak load condition, while some special cases will
be considered, like the peak and o-peak load hour with no solar insolation.
3. SC analysis: The SC analysis is used as an indicator of how the system behaves
before and after the installation of PV plants and other devices. So only peak
load SC condition is considered.
4. Reliability analysis: The reliability analysis considers the adequacy of power
supply utilities for dierent load. Therefore, the load curve should be taken
into account as completely as possible. In this thesis, the 8736 states of the
load are simplied to 12 states which represent dierent time period during a
year, as listed in Table 2.3. Each period peak load is calculated from averaging
the hourly peak loads during the corresponding period.
Table 2.3: Simplied load data for reliability analysis
Winter Summer Spring/Fall
Weekday Weekend Weekday Weekend Weekday Weekend
O-peak hours 21 8 0 9 22 9 22 9 21 8 0 9
Peak load 58.4 47.5 53.4 44.4 49.6 40.0
Peak hours 8 21 9 24 9 22 9 22 8 21 9 24
Peak load 82.4 62.2 76.6 58.3 67.4 51.3
16
2.2 Base Line Load Flow
The one-line diagram of RTS model built in ETAP platform is shown in Figure
2.3. A base case is set as the benchmark for comparison with future results. In this
paper, the base line refers to 100% load in Table 2.2, which actually is the annual
maximum load which takes place on Tuesday afternoon (5-6pm) of the 51st week [3].
The load
ow analysis shows the basic voltage prole on steady state.
The base line power
ow results are shown in Table 2.4. The total real power
losses are 53.04 MW. Given that the total load is 2850 MW, the percentage power
losses is:
Loss% =
P
loss
P
load
+P
loss
= 1:83% (2)
Table 2.4: Load Flow Result for Base Case
Bus # 1* 2* 3 4 5 6 7* 8
jV
bus
j (p.u.) 1.0350 1.0350 0.9720 0.9873 1.0123 1.0025 1.0250 0.9845
\V
bus
(

) -4.0 -4.1 -3.4 -7.1 -7.2 -9.9 -0.2 -6.1
Bus # 9 10 11 12 13** 14* 15* 16*
jV
bus
j (p.u.) 0.9826 1.0163 0.9946 1.0087 1.0200 0.9800 1.0140 1.0170
\V
bus
(

) -5.3 -7.2 -1.0 -0.5 0 3.4 12.8 11.5
Bus # 17 18* 19 20 21* 22* 23* 24
jV
bus
j (p.u.) 1.0385 1.0500 1.0232 1.0385 1.0500 1.0500 1.0500 0.9838
\V
bus
(

) 16.1 17.4 9.9 10.4 18.3 23.9 11.3 6.8
(*: Voltage control bus, **: Swing bus)
17
Figure 2.3: One-line diagram of the RTS model
18
2.3 Realization of PV penetration
From Table 2.1, the generation mix is calculated in Table 2.5 [3]:
Table 2.5: Generation Mix
Source MW %
Oil 951 28
Coal 1274 37
Nuclear 800 24
Gas 80 2
Hydro 300 9
In accordance with the purpose of achieving high PV penetration described in
Chapter 1, which is to save fossil fuels, reduce air pollution, and mitigate global
warming, an eective but also quite aggressive solution is the retirement of coal
power plants as well as the construction of PV plants in the same area. Generally
the land of retired coal plants could be polluted and far from city downtown so that
it is hard to be used for commercial or agricultural purpose. To build a solar plant
on this site has three benets:
1. It could partly save the capital cost for the PV plant in land rent;
2. It is able to take advantages of the former transmission towers and transformers;
3. It can prevent the light pollution caused by re
ection of solar panels.
The new solar power plant is assumed to have the same output voltage level as the
previous coal plant does. And it is supposed to generate as much real power as the
coal plant does so as to keep the supply capacity of power system in the same level.
Practically, smaller coal plants are more likely to be replaced before the larger ones,
because they are easier to retire, less eective to grid, and less costly. Following these
19
principles, the replacement starts from Bus 1, where the coal plants with smallest
capacity of 76 MW are at, and to Bus 2, 15, 16, and then 23, where the largest coal
plant with 350 MW locates. According to Table 2.5, the entire substitution of coal
power plants with PV plants could lift the penetration level of PV to 37% or more,
which builds a good model for testing the impacts of high PV penetration.
A solar power plant is built by the parallel connections of solar arrays to supply
real power, while a solar array is formed by series connections of solar panels to
support voltage. The standard test conditions (STC) species a testing environment
where the panel is operating in a cell temperature of 25

C under a solar irradiance
of 1000 W/m
2
with a spectrum of air mass 1.5 (AM1.5). Under STC, a solar panel
in this thesis has the following attributes:
Table 2.6: Parameters of a Solar Panel under STC
Type Multi-crystalline Silicon
P
max
265 W
V
mp
30.8 V
I
mp
8.61 A
V
oc
38.3 V
I
sc
9.10 A
Eciency 16.2 %
T
op
43 2

C

Pmax
-0.47%=

C

Isc
0.052%=

C

Voc
-0.344%=

C
Dimension 1650 992 35 mm
Weight 18.6 kg
20
Entering this parameters into the Solar Panel model in ETAP, it also generates
a P-V curve and a I-V curve as shown in Figure 2.4. ETAP assumes that the P-V
and I-V curves are linear between two setting points, which is not exactly practical.
However, it only changes the power output and panel resistance while the ambiance
is varying, in perspective of power systems, the result will not be aected..
Figure 2.4: Solar panel characteristics in ETAP
The solar plant is built following the rules below:
1. In 138 kV level, every 25 panels are connected in series, forming an array with
a voltage rating of 770 V;
2. In 230 kV level, every 21 panels are connected in series, forming an array with
a voltage rating of 646.8 V;
3. Every 6000 arrays are connected in parallel to form a larger array;
21
4. Due to software constrains, the solar arrays are connected with an inverter and
directly to the power grid. The eciency of inverter above 75% load is set to
be 95%, while the AC voltage is dened as the expected operating voltage of
each bus in Table 2.1.
Figure 2.5 shows the conguration of solar plants in aspects of array connections,
array ratings, and the output DC power under an irradiance of 1000 W/m
2
, an
ambient temperatureT
a
= 30

C, and a cell temperatureT
c
= 61:3

C. The conditions
above are the operation conditions of the solar array in this thesis. The inverter
settings are shown in Figure 2.6. As a result, the AC power output of each solar
power plant is shown below:
• 138 kV: 37.789 MW (p.f.=100%)
• 230 kV: 31.743 MW (p.f.=100%)
In this way, each coal power plant can be replaced by a certain number of the
PV power plant above without signicantly changing the net real power generation.
The number of PV plants needed to make up the capacity curtailed by coal plant
retirement is presented below:
Table 2.7: Number of PV plants needed on each bus
Bus #
Coal plant
MW
Number of
Coal Plants
Total MW
of coal
Number of PV
plants needed
Total MW
of PV
1 76 2 152 4 151.156
2 76 2 152 4 151.156
15 155 1 155 5 158.715
16 155 1 155 5 158.715
23
155 2 310 10 317.43
350 1 350 11 349.173
22
(a) 138 kV
(b) 230kV
Figure 2.5: Conguration and power generation of PV arrays
23
(a) Bus 1 - 138 kV
(b) Bus 16 - 230kV
Figure 2.6: Conguration and power output of inverters in solar plants
24
Substituting the 9 coal power generating units in the RTS from the smallest
capacity to the largest, the PV penetration level varies as in Table 2.8.
Table 2.8: PV Penetration Level Variation with Replacement of Coal Plants
Step # 0 1 2 3 4
Bus # 1 2
Total coal MW 1274 1198 1122 1046 970
Total PV MW 0 75.578 151.156 226.734 302.312
Penetration (%) 0 2.22 4.44 6.66 8.88
Step # 5 6 7 8 9
Bus # 15 16 23
Total coal MW 815 660 505 350 0
Total PV MW 461.027 619.742 778.457 937.172 1286.345
Penetration (%) 13.53 18.17 22.80 27.42 37.64
The maximum penetration level of 37.64% is derived from using total generation
capacity as denominator. However, during the operation, the exact ratio of PV
generation to total power generation does not equal the numbers in Table 2.8. Under
the generation conditions shown in Figure 2.5, the exact power output of a solar plant
is a bit less than its rating, as shown in Figure 2.7, which decreases the penetration
level in some extends. While another factor is that, during peak load operation
period, not all the all generating units are fully operated because there is only 2850
MW power demand in the system. In this way, the actual penetration level could be
calculated as:
P
PV
=
1286:345
2850
= 0:4513 (3)
As a result, during operation, the real power output of all PV plants is about 45%
of the total real power supply in the system with highest PV penetration level.
25
(a) Bus 1 solar plants
(b) Bus 16 solar plants
Figure 2.7: Normal real power output of solar plants under given conditions
26
CHAPTER 3: POWER FLOW ANALYSIS
The PF analysis consists of two parts. Firstly, without any outages, the coal
plants are replaced by PV plants following the rules in Section 2.3. A PF analysis is
performed to test the system stability changes. Then the system load is increased in
order to test the system strength of supporting the future load growth under steady
state. Then the N 1 contingency analysis is conducted following the variation of
PV penetrations.
3.1 Voltage Prole
3.1.1 Impact of PV penetration on bus voltage
Ignoring the resistance of transmission lines, a simple set of equations can be used
to describe the power
ow from bus 0 to bus 1:
8
>
>
<
>
>
:
P =
V
0
V
1
X
sin
Q =
V
1
X
(V
0
cosV
1
)
(4)
where bus voltages:V
0
=V
0
,V
1
=V
1
0, andX is the reactance of the line between
two buses.
Due to the fact that PV panels only generate real power, the greatest in
uence of
PV penetration in transmission system is the reduction of reactive power generation.
As a result, it would lead to two consequences:
• The power angle increases on load bus, and decreases on generation bus;
• The voltage drops on all buses.
27
3.1.2 Simulation result
With the increase of PV penetration, the voltage variation on each load bus is
shown in Figure 3.1.
(a) 138 kV level bus
(b) 230 kV level bus
Figure 3.1: Voltage prole of each bus due to PV penetration
28
According to Table 2.8, as penetration is below 10%, the coal replacement takes
place on 138 kV level, while the left happens on 230 kV. It can be seen from the
gure that the general trend of voltage bus is dropping. But since the inserted PV
plants are not evenly added to all buses, the changes of bus voltages are also not
smoothed. Mainly there are three features:
1. The impact of reactive power reduction is not too signicant during steady state
operation, since the voltage level is still higher than 96%. This is because of a
relatively low reactive power demand. At rst, the reactive power supply is still
sucient enough to support system voltages. So there is no obvious downturn
at the beginning. The rst sharp drop occurs at around 8% penetration, where
all the coal plants on 138 kV level is removed. With simulation of system
without PV, the total reactive power output from Bus 1 and Bus 2 is about
58 MW. Comparing the reactive power capacity of generators on Bus 1 and
2 from Table 2.1, retiring three coal plants, the reactive power capacity is 70
Mvar, indicating the demand can still be roughly satised. But without a coal
plant, the four oil plants can only handle 40 Mvar generation in total, which is
why the voltage drops remarkably. Similarly, the second sharp drop on higher
penetration of 30% can be explained by the same reason.
2. The system is divided into two areas by transformers, which blocked some
impacts from another voltage level. So the 2 voltage sharp drops are more
signicant in the voltage level where the corresponding reactive power shortage
occurs. Not only the transformers, but the transmission lines have a similar but
weaker eect,since the bus close to PV plants (such as Bus 4) is more aected
than the one far from PV plants and close to other non-coal-fueled plants (Bus
8). This makes the voltage impacts of PV penetration regional.
3. Since the voltage drop is caused by the reduction of reactive power supply, no
29
load bus is aected much less than the load bus. As shown in Figure 3.1b. The
four no load buses, Bus 11, 12, 17, and 24, have more
at voltage proles than
others.
3.2 Real Power Loss
The real power loss is an important index of the operation condition of a power
system. It can be derived using PF analysis and is calculated by simple Ohm's law.
A high real power loss could lead to overheating on transmission lines, overloading
of generation buses, and nancial loss of electric utilities. The total real power loss
curve is shown in Figure 3.2. The curve turns down in low penetration level (less
than 10%), while it increases quickly as penetration further grows. However, even
the maximum real power loss is still low enough to claim that it is acceptable and
the system is operating well. Meanwhile, since the real power loss decreases on low
penetration level, the implementation of PV plants in some extends might benet
the dispatch of power and hence save the energy.
Figure 3.2: Total real power loss of the system with PV penetration
30
3.3 N-1 Contingency
N-1 contingency analysis measures the number of violations on each single-line
outage considered. The violation is dened under a certain criterion. In this thesis,
a WECC standard is used that for all single-line outages, the post transient voltage
deviation should not exceed 5% at any bus [2]. While the thermal MVA deviation is
also limited at 5%. The entire table of WECC standard of disturbance performance
is listed in Table 3.1
Table 3.1: WECC Disturbance-Performance Table of Allowable Eects on Other
Systems [2]
Here the single-line outages include the outage of any single line or transformer in
RTS. The voltage deviation is dened as the dierence between the operation voltage
and 1 p.u. value. The MVA ratings are from report [3], and is presented in Table
3.2.
31
Table 3.2: Impedance and Rating Data of RTS [3]
3.3.1 Special action on L0610 open case
During the procedure of contingency analysis, not all cases can simply be calcu-
lated out through PF analysis. When the cable L0610, which connects Bus 6 and Bus
10, is disconnected from the system, the iteration of PF equations cannot converge,
which means the system cannot reach stability. The one line diagram of line L0610
is shown in Figure 3.3. It can be seen that except for the load, there is a 100 Mvar
32
reactor at Bus 6 which is designed to prevent overvoltage on Bus 6. Once L0610 is
out of service, the voltage on Bus 6 becomes extremely low and cannot support the
load anymore. Therefore, after a disturbance on L0610, the corresponding actions
must include both the opening of CBs between L0610, and the disconnection of the
reactor at Bus 6.
Figure 3.3: One line diagram of cable L0610
It can be indicated from this event that the RTS is not suciently stable and
resilient. As the PV penetration level rises, the reactive power shortage would fur-
ther aggravate the voltage collapse on Bus 6 hence even the disconnection of reactor
cannot solve the problem. In order to support the voltage on Bus 6 during contin-
gency, a static VAR compensator (SVC), or a capacitor bank, is added to Bus 6.
The specic plan is that, after the rst implementation of PV plant on Bus 1, a 75
Mvar SVC is also built at Bus 6. And once a transient occurs on line L0610, three
operations have to be triggered at once:
• the opening of CBs between L0610,
• the opening of CB of reactor on Bus 6,
• the closing of CB of the capacitor on Bus 6.
33
3.3.2 Contingency for base line
The system is tested under base line of 100% load, and the results are listed in
Table 3.3. A practically feasible system is expected to have no violations in simulation
of n 1 contingency. But the result actually shows that the RTS has 9 violations,
indicating a weak system. Such an attribute is mentioned in the original report of
RTS-79, saying that it \...was designed to have a lower reliability than is typically
considered acceptable in utility planning" [3].
Among all the violations in contingency analysis, some can be viewed as not
important. For example, when the transmission line L1718 is disconnected, a non-
load Bus 17 would go through an overvoltage of 1.051 p.u. Not only does the voltage
exceeds the criterion for only 0.001 p.u., but the overvoltage does not aect any
loads in the system so as not to reduce its reliability. One critical problem for the
system is that, since there is only one transmission line L0708 between Bus 7 and 8,
a serious islanding will occur if it is disconnected from the system. A study of this
islanding scenario is not in the scope of this thesis, because the related buses have
no PV plants connected, and all the operations needed all not dierent with normal
islanding treatments.
However, such a fragile system is actually better for theoretical study. It is more
sensitive to changes than the real ones, hence provides a more explicit marginal
information when PV plants are introduced. In this way, the 9 violations are chosen
to be the most allowable violations for modied RTS model with any changes. A PV
penetrated RTS will easily break this limitation because of reactive power reduction.
So some corresponding correction devices have to be installed in the system and try
to improve the system back to the one with 9 or less contingency violations.
34
Table 3.3: Base Line RTS Contingency Violations
Event
Violations
Event
Violations
Voltage Thermal Voltage Thermal
Normal 0 0 L1113 open 0 0
L0102 open 0 0 L1114 open 0 0
L0103 open 1 0 L1213 open 0 0
L0105 open 0 0 L1223 open 0 0
L0204 open 1 0 L1323 open 0 0
L0206 open 0 0 L1416 open 0 0
L0309 open 0 0 L1516 open 0 0
T0324 open 1 0 L1521 open 0 0
L0409 open 0 0 L1524 open 2 0
L0510 open 0 0 L1617 open 1 0
L0610 open 1 0 L1619 open 0 0
L0708 open 2 0 L1718 open 0 0
L0809 open 0 0 L1722 open 0 0
L0810 open 0 0 L1821 open 0 0
T0911 open 0 0 L1920 open 0 0
T0912 open 0 0 L2023 open 0 0
T1011 open 0 0 L2122 open 0 0
T1012 open 0 0 Total Amount 9 0
35
3.3.3 Contingency with PV penetration
The increase of voltage violations due to PV penetration is plotted in Figure 3.4.
The thermal violation is not taken into consideration because the MVA ratings of
branches are quite sucient and they only appear on the highest penetration level
of 38%. As a result, the voltage violation curve is much like the voltage proles in
Figure 3.1: one sharp rise on 8% penetration, and another after 25%. Therefore, it
can be concluded that the voltage violations in contingency analysis is also due to
the reactive power shortage introduced by PV penetrations.
In fact, at the highest 38% penetration level, the number of voltage violations
increases to 31, which means the system is extremely unstable, and any disturbance
might cause the collapse of the whole system. All violations of 38% PV penetrated
RTS are listed in Table 3.4. The most serious outage event is the opening of line
L0708, which is the only connection between Bus 7 and the system. Since Bus 7 has
three hydro power units rated as 100 MW and 60 Mvar, such a loss of reactive power
places the 138 kV area really in peril and incurs 6 violations.
Figure 3.4: Voltage violations due to PV penetration
36
Table 3.4: 38% Penetration RTS Contingency Violations
Event
Violations
Event
Violations
Voltage Thermal Voltage Thermal
Normal 0 0 L1113 open 0 0
L0102 open 2 0 L1114 open 1 0
L0103 open 1 0 L1213 open 1 0
L0105 open 0 0 L1223 open 0 0
L0204 open 1 0 L1323 open 1 0
L0206 open 0 0 L1416 open 4 1
L0309 open 0 0 L1516 open 0 0
T0324 open 3 1 L1521 open 0 0
L0409 open 0 0 L1524 open 4 0
L0510 open 0 0 L1617 open 3 1
L0610 open 1 0 L1619 open 0 0
L0708 open 6 0 L1718 open 0 0
L0809 open 0 0 L1722 open 0 0
L0810 open 0 0 L1821 open 0 0
T0911 open 1 0 L1920 open 0 0
T0912 open 2 0 L2023 open 0 0
T1011 open 0 0 L2122 open 0 0
T1012 open 0 0 Total Amount 31 3
37
Table 3.4 also presents 3 thermal violations which are all occurring on transmis-
sion line L0708. However, as shown in Table 3.2, there are a so-called \short term"
and a \long term" ratings for each branch, which is the maximum allowable MVA
within 15 minutes and 24 hours, respectively. According to the load data in [3],
the peak load (or near peak load) only lasts for several hours everyday. Notice that
all of the MVA violations mentioned are based on the normal rating, and none of
them have exceeded the long term or short term ratings. So precisely, no thermal
violations are observed, and that is why only voltage violations are considered in
previous part.
3.4 System Improvements
3.4.1 Problematic system conguration
Some of the problems of the RTS model have been discussed in previous sections
including the voltage unbalance on Bus 6, and some thermal limitations on trans-
mission line L0708. The one line diagram of transmission line L0708 is shown in
Figure 3.5. As previously said, L0708 has been the most problematic transmission
line in RTS. The fact it is the only line connecting a generation bus not only causes
islanding issues, but also brings it extra burden to transfer power out from genera-
tors and into system. During peak load normal condition, the power
ow through
L0708 has already reached its normal rating of 175 MVA. Such a conguration will
strongly hinder future expansion of the system, and denitely cannot adequately t
any future growth in system load.
38
Figure 3.5: One line diagram of transmission line L0708
Another problem is on Bus 3. As shown in Figure 3.6, totally there are four
connections at Bus 3, including one load, one transformer, and two transmission
lines, while half of them have violations during contingency analysis. Even in the base
line contingency case, line L0103 outage and transformer T0324 outage have already
caused voltage violations. Meanwhile, in many violation events, Bus 3 has voltage
regulation problem in either overvoltage or undervoltage. Such a great
uctuation
indicates that the voltage on Bus 3 needs to be regulated.
Figure 3.6: One line diagram of Bus 3
39
Lastly, there is a invisible violation on voltage regulation among all the buses with
PV plants. Since initially all the buses with coal plants are supposed to be voltage
control buses, this setting is greatly damaged by the introduction of solar plants
since a simple inverter cannot generate reactive power to support the bus voltage.
The reactive power shortage hence leads to a voltage drop on every generation bus
with PV plants, and also results in a full load in reactive power generation of every
generating units near PV plants. The reason such a violation is \invisible" is that
the voltage drop does not exceed the 5% line and will not cause a signicant impact
on loads at those buses. However, an undervoltage might aect the speed of rotors
or destroy the generators.
3.4.2 Improvement devices
In Section 3.3.1, a 75 Mvar SVC has already been placed on Bus 6 in order to
solve voltage collapse during L0610 outage. Based on the fact that Bus 3 also has
voltage drops, a similar treatment is performed to put a 100 Mvar capacitor bank
on Bus 3.
Besides, in order to solve the undervoltage on generation bus, a voltage-control
inverter has to be put into use. Basically, a voltage-control inverter is nothing but
an inverter and a synchronous compensator. But due to the introduction of a new
compensator, the power supply capacity is also aected: the increase of reactive
power supply can lead to a decrease in real power output. Also, the voltage control
inverters may not be deployed in all PV plants because of a huge increase in capital
cost and the low frequency of utilization. After completion of PV penetration, it
can be seen from the generation settings in Table 2.1 that on Bus 16 and 23, there
are only solar plants existing. And the voltage regulation problem is not important
anymore since no turbines will be aected. Therefore, it is reasonable to only change
40
the inverters on Bus 1, 2 and 15 into the voltage-control type so as to protect the
oil plants on them. The Mvar capacity of the voltage control inverter is determined
by the default minimum power factor setting of 98.5%. So the actual number of the
maximum Mvar of each inverter is
8
>
>
<
>
>
:
Q
max138
= 6:420 Mvar
Q
max230
= 5:395 Mvar
(5)
With all these modications, a contingency analysis is conducted again for 38%
penetration RTS, and the result is shown in Table 3.5. The new RTS solves all the
reactive power supply problem while even improves the voltage regulation perfor-
mance of Bus 3 and Bus 6. So as indicated in Table 3.5, the improved RTS with
38% solar generation has only 4 violations, 5 less than in the base case.
As for line L0708, since the goal of system improvement is to recover the modied
RTS to a state no worse than the original one, no further measures are needed on
L0708.
Table 3.5: Improved 38% Penetration RTS Contingency Violations
Event
Violations
Voltage Thermal
L0204 open 1 0
L0708 open 2 0
L1617 open 1 0
Total Amount 4 0
41
3.4.3 Real power curtailment due to voltage control inverters
As mentioned in Section 3.4.2, a voltage control inverter will consume real power
generated from the solar panel and output reactive power. In this way, an evalu-
ation of the real power curtailed by adoption of voltage control inverter has to be
performed. The results of power
ow analysis on the old and new base case RTS
(without and with voltage control inverters on Bus 1 and 2) are shown in Figure 3.7.
(a) Without voltage control inverters
(b) With voltage control inverters
Figure 3.7: Bus 1 and Bus 2 power
ows with 38% PV penetration RTS
The total real power generation of oil plants on each bus is 40 MW so that the
total real power output of solar plants is 148.959 MW. While the reactive power
generation of solar plants on each is 16.723 Mvar on Bus 1 and 9.487 on Bus 2. The
42
dierence of total MVA generated from solar plants on each bus is:
8
>
>
<
>
>
:
S
1new
S
1old
=
p
148:959
2
+ 16:723
2
148:959 = 0:936 MVA
S
2new
S
2old
=
p
148:959
2
+ 9:487
2
148:959 = 0:302 MVA
(6)
Even under the largest Mvar output with 98.5% power factor, the extra power needed
is only:
1
0:985
1 = 0:015 p.u. (7)
Therefore, in order to implement a voltage control inverter, only a negligible amount
of several KW real power is needed from a solar plant. The system can still operate
in a good condition without any change.
3.5 Early evening load peak without PV
3.5.1 Supply shortage after sunset
Although the system is now suciently supplied, it is only met under high inso-
lation conditions in Figure 2.5. So the system has to be studied more since the load
peak in RTS model does not always coincide with generation peak.
The hourly peak load is listed in Table 3.7 [3]. It is indicated in the table that
no matter when the 100 percent peak occurs, the load is still higher than 90% after
7 p.m. in any season. While the sunset is before 7 p.m., the PV plant generates no
power afterwards, and a power shortage will take place on RTS every night. The
MW value of power shortage can be simply derived by subtracting the maximum
43
real power output without PV plants from the peak load:
P
Gen
= 3405 MW 1274 MW = 2131 MW (8)
P
Short
=P
peakload
P
Gen
(9)
= 2850 MW 2131 MW = 719 MW (10)
Such a huge gap between supply and demand on early evening could lead to the
system collapse. In order to bridge the dierence, a gas plant can be installed.
Based on calculation, the capacity of the gas power plant should be 800 MW or
greater.
3.5.2 Capacity factor issue
Assume that the solar plants cannot generate power after 5 p.m. and will get
back to work after 6 a.m. The total nonrenewable generation can supply the load if
it is lower than:
2131
2850
100% = 74:77% (11)
Meanwhile, the weekend daily peak load in [3] is roughly 75% of the weekly peak,
while the weekday peak is from 93% to 100%. As a result, the approximation of
operation time of the gas turbine during 364 days in [3] is listed in Table 3.6.
Table 3.6: Approximation of Gas Turbine Operation Time
Winter Weeks Summer Weeks Spring/Fall Weeks
Wkdy Wknd Wkdy Wknd Wkdy Wknd
Weeks 17 17 13 13 22 22
Hours 5 3 6 1 6 2
44
In total, the approximate operation hours are
(5 + 3) 17 + (6 + 1) 13 + (6 + 2) 22 = 403 h (12)
The corresponding capacity factor (assuming full load during operation):
C:F: =
404
24 364
= 4:6% (13)
Such a low capacity factor is not recommended in practice. Therefore, a better plan
should be made, such as to use battery instead. However, cost problem will rise if
battery is used. The full design of real power compensation during the short time in
early evening is beyond the range of this thesis.
45
Table 3.7: Hourly Peak Load in Percent of Daily Peak [3]
46
CHAPTER 4: SHORT CIRCUIT ANALYSIS
The previous chapter tested the post-transient performance of system under
single-line outages. Practically, a SC may also take place on a bus which contains
substation or other related devices. Same as the relays and CBs on transmission
lines, an overcurrent protection device has to be implemented at each bus to detect
the current and operate correctly. Normally, the inverter in a solar plant is expected
to open when a SC occurs at the bus. A operation failure on CBs could lead to an
overcurrent or reverse directional power
ow into the solar plant. So the design and
settings of protection devices should be considered thoroughly. In this thesis, a SC
analysis is conducted in RTS to measure the magnitude of the SC current on each
bus, and determine the interruption capability of protection devices by compare their
ratings with the SC current.
4.1 Event Denition
The RTS SC analysis is conducted on ETAP software based on American Na-
tional Standards Institute (ANSI) C37 Standard [31]. Totally four types of SC are
dened including 3-phase, line-to-ground (LG), line-to-line (LL), and line-to-line-
ground (LLG) faults. Initially, the prefault voltage of each bus is 1.0 p.u. During
the simulation, each of the four faults will take place on a single bus in RTS, and the
magnitude of a SC current will be calculated based on ANSI.C37 Standard and the
system settings. The time period for testing is the rst half cycle, where the peak
current locates. All of the 24 buses will be tested, and the
uctuations of the SC
currents due to increase of PV penetrations will be recorded.
One more parameter that is needed for SC analysis on ETAP is the X/R ratio
of the generators. In this thesis, as shown in Figure 4.1, a default p.u. number of
47
sequence impedance in ETAP is used:
8
>
>
>
>
>
>
<
>
>
>
>
>
>
:
X
00
d
= 0:19 R
00
a
= 0:01
X
2
= 0:18 R
2
= 0:02
X
0
= 0:07 R
0
= 0:01
(14)
Figure 4.1: Default X/R ratio settings for the 76 MW coal power steam turbine in
ETAP
The PV model in ETAP does not have a X/R parameter. During the fault
transient, the inverter would act quickly to cut the connection from PV arrays and
the plant should open all the CBs at the same time. Therefore, a factor K is used
in SC analysis instead of X/R ratio. The K is dened as the SC contribution to the
48
AC system, as shown in Figure 2.6, and it is 150% of the full-load ampacity (FLA)
of the inverter. The SC analysis will be conducted based on this conguration, and
a PV penetration change will cause the variation of fault current on each bus.
4.2 Base Line SC
As can be seen in the base line SC analysis result presented in Table 4.1, the fault
currents on all the buses are within the range of 5 kA to 35 kA. The fault currents
on the designated PV generation Bus 1,2, 15, 16, and 23 are as great as 20 kA or
more, which indicate a potentially good performance in determining fault.
Meanwhile, most of the low fault current events (I
f
15 kA) occur at buses
without generating units except for Bus 22. If a fault cannot be cleared immediately
on a non-load bus or load bus, still many measures can be done later to make up.
But if a generation bus has a SC fault and cannot be cleared or even detected, the
damage to the generating units and the system would be devastating. It can be
roughly calculated that if the 6 hydro turbines are fully operating under normal
condition, a total of about 300 MVA power would generate a current output from
Bus 22 of:
I =
S
p
3V
=
300
p
3 230
kA = 1:304 kA (15)
As a result, a SC current of only 8 kA due to LL fault on Bus 22 may not be
able to trigger the interruption operation of protection devices and hence destroy
the turbine. A specic design of the protection device will not be discussed in this
thesis.
49
Table 4.1: Base Case SC Currents at Each Bus
Bus 3-Phase Fault LG Fault LL Fault LLG Fault
ID Voltage (kV) Current (kA) Current (kA) Current (kA) Current (kA)
1 138 19.121 20.814 16.714 20.125
2 138 18.701 20.395 16.347 19.706
3 138 12.344 11.174 10.692 11.908
4 138 7.845 7.042 6.797 7.573
5 138 9.234 8.484 8.004 8.968
6 138 10.366 8.586 8.979 9.776
7 138 13.962 15.810 12.245 15.131
8 138 12.951 11.310 11.235 12.464
9 138 18.577 17.441 16.094 18.191
10 138 19.952 18.035 17.286 19.246
11 230 12.924 10.933 11.204 12.249
12 230 10.795 8.975 9.360 10.196
13 230 28.616 28.893 24.979 29.011
14 230 20.210 17.922 17.565 19.421
15 230 33.592 29.133 29.194 32.062
16 230 31.312 29.122 27.196 30.508
17 230 18.976 18.751 16.476 18.926
18 230 33.941 32.752 29.578 33.664
19 230 22.534 18.404 19.539 21.112
20 230 21.289 19.754 18.517 20.737
21 230 25.376 27.540 22.147 26.732
22 230 9.185 10.843 8.051 10.316
23 230 21.474 24.233 18.800 23.264
24 230 6.132 5.189 5.312 5.799
50
4.3 SC Current Variation due to PV Penetration
The LG is selected to be studied rst because it is the most common fault type.
In the 138 kV area, the LG fault current curves due to PV penetrations are shown in
4.2a. Generally, due to the reactive power generation reduction, the total magnitude
of SC current will decrease, which is veried in the gure. By further investigating
it, three conclusions can be made:
1. The SC current is calculated on the basis of the power
ow on each bus, while
the dierence in current magnitudes by fault types are caused by the dierent
congurations of impedance. Therefore, since the power
ow is not changing
signicantly by PV penetration, the variation of fault current is also not so
obvious.
2. The two sharp drops of SC current are on the curves for Bus 1 and 2, where
the PV locates. Besides, there are four slight drops on Bus 4, 5, 6, and 10,
among which Bus 4, 6 and Bus 5 are the load buses directly connected to Bus
1 and 2, and Bus 10 is directly linked to Bus 5 and 6. So it is indicating that
the PV penetrations only impact SC currents in a very local area close to
the PV generation bus.
3. The SC current of Bus 1 and 2 only changes when the PV penetration is
happening on them or near them. And further penetrations on 230 kV area do
not aect the other area in aspect of SC fault current. It is not only because
of the local impact of PV penetration, but also due to the hindering eect
on fault current by transformers.
51
(a) All buses in 138 kV area
(b) Selected buses in 230 kV area
Figure 4.2: LG fault current variations due to PV penetration
52
Similar results can be found out on 230 kV buses. In order to make the gure
clearer, only 5 buses are posted in Figure 4.2b: 3 PV generation Bus 15, 16, and
23, and 2 load Bus 19 and 20. Because Bus 15 and 16 is every close to each other,
there is a coupling eect on power
ow as well as SC currents of both buses. It can
be seen from this gure that there is an obvious turning point on 20% penetration
level where fault currents on Bus 15 and 16 start to be constant and the one on Bus
23 is decreasing. It is exactly where the PV plants start to be installed on Bus 23.
From Figure 4.2b, again, it can be concluded that the impact of PV penetration on
SC current is localized around the PV bus.
The 3 phase and LL fault currents on 138 kV area buses are also shown in Figure
4.3 and 4.4. The same pattern is observed.
Figure 4.3: 3 phase fault current variations in 138 kV area due to PV penetration
53
Figure 4.4: LL fault current variations in 138 kV area due to PV penetration
4.4 SC Analysis for Improved RTS
It can be seen from Figure 4.2 that at the highest penetration of 38%, many PV
generation buses like Bus 1, 2, and 23 have decreased fault currents of only half the
base line values, which may lead to mis-operations on overcurrent protection devices.
Meanwhile, the modied RTS is enhanced in Section 3.4.2 because it cannot handle
all the n 1 contingencies in high PV penetration. The so-called improved RTS,
however, does not enhance the system in aspect of SC current, as shown in Figure
4.5. The reasons that the fault current in the new RTS is less the initial RTS include:
• The SC contribution factor K is still 150% during the simulation, which means
the SC from a PV plant does not change even with the implementation of
voltage control inverter. However, as discussed in Section 3.4.3, the reactive
power contribution is not too great. So the increase of K in the voltage control
inverter might not be too great. Meanwhile, most of the LG fault current
decreases are not on the PV generation buses. Since the PV penetration impact
54
on SC current is relatively regional, such a current drop on load and non-load
buses cannot be simply explained by the issue of K value.
• The newly-installed capacitor on Bus 3 reduces the reactive power demand from
generation bus. So the net total reactive power output from each generation
bus has decreased, causing the reduction in total magnitude of SC current.
Figure 4.5: Comparison of RTS and improved RTS in aspect of LG fault current
As a consequence, with the installation of PV plants and all the voltage correction
devices, the ratings of overcurrent protection devices must be redesigned to coincide
with the new SC current with smaller magnitude.
55
CHAPTER 5: RELIABILITY ANALYSIS
5.1 Reliability Indices
The reliability analysis is conducted on the RTS to determine its adequacy in
supplying loads. Generally, a reliability of a system have three aspects: the interrup-
tion frequency, the loss of energy or load (power), and the cost for interruption. In
this thesis, the nancial cost is not the major object to study. The interruption fre-
quency indices that ETAP is able to measure include System Average Interruption
Frequency Index (SAIFI), System Average Interruption Duration Index (SAIDI),
Customer Average Interruption Duration Index (CAIDI), Average Service Avail-
ability Index (ASAI), and Average Service Unavailability Index (ASUI) which are
calculated in the following ways:
SAIFI =
P
customer interruptions
P
customers
=
P

i
N
i
P
N
i
(16)
SAIDI =
P
customer interruption durations
P
customers
=
P
U
i
N
i
P
N
i
(17)
CAIDI =
P
customer interruption durations
P
customer interruptions
=
SAIDI
SAIFI
(18)
ASAI = 1
SAIDI
8760
(19)
ASUI = 1 ASAI =
SAIDI
8760
(20)
where
i
is the interruption rate on Bus i, N
i
is the total number of customers on
Bus i, and U
i
is the annual interruption duration of Bus i. Since RTS model is
a large-scale transmission system, no customer denition is done, and each bus is
simply viewed as one customer in the system. Therefore, the indices above is not
practically accurate for evaluation of RTS model and is usually used in distribution
systems. The loss of energy indices include Expected Energy Not Supplied (EENS)
56
and Average Energy Not Supplied (AENS). AENS is dened as the average EENS
per customer. Usually the EENS is calculated as below [32]:
EENS =
X
i
X
L
k
>C
i
(L
k
C
i
)R
i
T
i
(21)
where L
k
is the load during a time period k, C
i
is the total capacity of the system
during interruption event i,R
i
is the annual frequency of event i, andT
i
is the dura-
tion of event i. The rst summation is for all the interruption eventsfig, while the
second summation only counts in the dierence between the demand and generation
when the load cannot be fully supplied. Generally, the EENS in [32] measures the
annual energy shortage due to interruptions. However, ETAP software computes the
EENS in another method [33]:
EENS =
X
i
P
i
U
i
(22)
This algorithm denes the EENS as the production of annual outage duration at
Bus i U
i
and the average load at Bus i P
i
. This simpler denition makes the EENS
index in ETAP similar to the interruption frequency indices mentioned before. In
this thesis, the second EENS denition is used since it inherently exists in ETAP.
5.2 Reliability Settings
The major reliability parameters of a power system element is the Mean Time To
Failure (MTTF) and the Mean Time To Repair (MTTR). A device state is always

owing from the operation state to the failure state. In ETAP, the MTTF is the
average time duration between two failures, and the MTTR is the average time
duration when the device is out of service. The completed reliability settings of RTS
are in report [3]. The additional capacitors in the improved RTS have the same
57
reliability parameters as the reactor at Bus 6, which are an MTTF of 88.2 years
and an MTTR of 19 hours. As for the solar plants, in ETAP, only the inverters
have reliability parameters, and the solar plant does not take part in the reliability
analysis, as shown in Figure 5.1. Therefore, an overall interruption rate of 0.2 failure
per year is adopted, which corresponds to an MTTF of 5 years. The MTTR is set
as 50 hours.
Figure 5.1: The reliability conguration of a solar plant at Bus 1
58
5.3 Base Line Reliability
The reliability indices of RTS base line can be divided into two parts:
• SAIFI, SAIDI, CAIDI, ASAI and ASUI are not related with the load MVA ,
so they are plainly presented in Table 5.1.
Table 5.1: Part of the Reliability Indices for RTS Base Case
• EENS is a power-related index, which means a load curve has to be used to
provide a more accurate measurement. Using the load states determined in
Table 2.3, the EENS of each state is listed in Table 5.2
Table 5.2: EENS on Dierent Load States
Winter Summer Spring/Fall
Weekday Weekend Weekday Weekend Weekday Weekend
O-peak hours 21 8 0 9 22 9 22 9 21 8 0 9
EENS (MWh/y) 4885.25 3971.99 4463.12 3717.02 4143.84 3341.29
Peak hours 8 21 9 24 9 22 9 22 8 21 9 24
EENS (MWh/y) 6889.48 5198.26 6406.71 4875.47 5635.78 4286.08
As a result, the nal EENS for RTS base case is 5103.95 MWh/y, the AENS
is 283.55 MWh/costomer/y.
59
5.4 Change of Reliability Indices due to PV Penetration
Among all the indices provided by ETAP, the maximum number of independent
indices are 3, which are chosen as SAIDI, SAIFI, and EENS. The changes of the
three indices dut to PV penetration are plotted in Figure 5.2. It can be seen from
the gure that three curves are growing in a relatively linear manner. Since the
MTTF and MTTR parameters of all inverters are identical, the linear growth of
reliability indices could indicate a direct connection between reliability of the system
and the installation of PV plants. A further study requires a complete simulation of
RTS in tree structures so that some other tools have to be used. In this thesis, the
result could at least indicate that, the impact of PV penetration on reliability of the
system is not local, or quite global throughout the system.
5.5 Reliability Analysis for Improved RTS
With installations of more capacitors, the reliability indices will denitely in-
crease. The comparison of the indices of two RTS is shown in Table 5.3. The three
indices are increased a little by the capacitor added to RTS. However, the changes
are not signicant so the reliability issues brought by the improved RTS are worthy
compared with its benets in voltage regulations.
Table 5.3: Comparison of RTS and Improved RTS in Aspect of Reliability Indices
(38% PV penetration)
SAIFI SAIDI EENS
RTS 0.050567 3.0567 5706.64
Improved RTS 0.050585 3.0572 5758.845
60
(a) SAIDI
(b) SAIFI
(c) EENS
Figure 5.2: The variations of SAIDI, SAIFI, and EENS in RTS due to PV penetration
61
CHAPTER 6: CONCLUSIONS
In this thesis, a modied IEEE RTS model is built on ETAP platform. Many
snapshot analyses including power
ow, contingency, short circuit, and reliability
analysis are performed in the system. The model is proved to be relatively less
reliable and stable than practical systems. PV plants are built and replace the coal
plants on the same bus in order to achieve a high penetration level of 38%. Further
studies of the PV penetrated RTS indicate more serious problems caused by the
characteristics of PV plants, including voltage drops, system collapse after single-
line outages, short circuit current reduction, and reliability strength decrement.
Three suggestions on system improvement are provided. A capacitor bank is
added to Bus 6 of RTS in order to mitigate the voltage collapse caused by the outage
of cable L0610. Another capacitor is installed on Bus 3 to solve the voltage violations
at Bus 3 during some single-line-outage events. The voltage-control inverters are
adopted on Bus 1, 2 and 15 in order to regulate the operating voltages and protect
other generators on those buses. The improved RTS can be regarded as preliminarily
eligible to handle 38% PV penetrations.
Further improvements are recommended for RTS with high PV penetration level,
including addition of transmission lines connecting Bus 7, installation of backup
power supplies for condition of evening peak load without sunlight, and redesigning
of overcurrent protection devices on buses with PV plants. They are beyond the
scope of this thesis.
The limitations and further works of this thesis contain the following points:
1. The RTS is only tested by snapshot methods, which could not re
ect its dy-
namic performance. Relating problems unsolved include the transient stability
62
issues concerning line outages and short circuit faults and the dynamic behavior
of PV inserted systems following variation of solar radiation.
2. Not all the load states of RTS are simulated, which leads to inaccuracy to
the analysis. Some problems involving load changes are omitted, such as the
power dispatch issues with increase of solar power in low load RTS on early
morning. The backup power supply in early evening peak load RTS has not
been determined due to lack of optional sources.
3. Some parameters of components are not close to practice. The reliability pa-
rameters of inverters and capacitors, and the X/R ratios of generators are set
equally in the model, which could lead to a wrong result in reliability or short
circuit analysis. A better specied system with more given parameters should
be used.
4. The constrains in license of ETAP software limit the conguration of the model.
The ETAP used in this thesis only allows simulation for systems with less than
25 buses, which limited the modeling of PV plants. By using a software with
more allowable buses, a detailed model of PV plant with more inverters and
transformers must be used to reconrm the results derived in this thesis. A
distribution network can be added into the system to conduct a more accurate
reliability analysis.
5. System protection devices are not dened and added into the model. A thor-
ough simulation of RTS can be performed with more protection operations
being implemented.
63
References
[1] C. Whitaker, J. Newmiller, M. Ropp, and B. Norris, \Renewable Systems In-
terconnection Study: Distributed Photovoltaic Systems Design and Technology
Requirements," 2008.
[2] WECC Planning Coordination Committee, \System Performance Criterion Un-
der Normal Conditions, Following Loss of a Single BES element, and Following
Extreme BES Events," 2012.
[3] IEEE RTS Task Force of APM Subcommittee, \IEEE Reliability Test System,"
IEEE Transactions on Power Apparatus and Systems, vol. PAS-98, no. 6, pp.
2047{2054, 1979.
[4] Energy Information Administration (US), \Annual Energy Outlook 2017: With
Projections to 2050," 2017.
[5] J. Simitian, \California renewable energy resources act," 2011.
[6] K. de Le on, \Clean energy and pollution reduction act of 2015," 2015.
[7] J. Duncan Glover, M. S. Sarma, and T. J. Overbye, Power System Analysis and
Design. Cengage Learning, 2016.
[8] V. J. Mishra and M. D. Khardenvis, \Contingency analysis of power system,"
in 2012 IEEE Students' Conference on Electrical, Electronics and Computer
Science (SCEECS). IEEE, 2012, pp. 1{4.
[9] Touseef A. F. Mohammed, \Quasi-Static Time-Series Simulation using
OpenDSS in IEEE Distribution Feeder Model with High PV Penetration and
Its Impact on Solar Forecasting," 2012.
64
[10] M. Ebad, \Integration and Simulation Challenges of High-Penetration PV,"
2016.
[11] Y. Tang, \Distribution System Modeling, Analysis and Design with High Pen-
etration of Photovoltaic Generation," 2016.
[12] \2010 Phase 1 CTPG 2020 Study Report," 2010. [Online]. Available:
http://www.ctpg.us/archived-documents
[13] \2010 CTPG Final Study Report: Phase 2," 2010. [Online]. Available:
http://www.ctpg.us/archived-documents
[14] \2010 Final CTPG Study Report: Phase 3," 2010. [Online]. Available:
http://www.ctpg.us/archived-documents
[15] \2010 CTPG DRAFT Phase 4 Study Report," 2011. [Online]. Available:
http://www.ctpg.us/archived-documents
[16] N. W. Patapo and D. R. Mattijetz, \Utility Interconnection Experience with
an Operating Central Station MW-sized Photovoltaic Plant," IEEE Power En-
gineering Review, vol. PER-5, no. 8, pp. 31{31, Aug 1985.
[17] S. M. Chalmers, M. M. Hitt, J. T. Underhill, P. M. Anderson, P. L. Vogt, and
R. Ingersoll, \The Eect of Photovoltaic Power Generation on Utility Opera-
tion," IEEE Power Engineering Review, vol. PER-5, no. 3, pp. 28{29, 1985.
[18] W. T. Jewell, R. Ramakumar, and S. R. Hill, \A Study of Dispersed Photo-
voltaic Generation on the PSO System," IEEE Transactions on Energy Con-
version, vol. 3, no. 3, pp. 473{478, 1988.
[19] W. Jewell and R. Ramakumar, \The Eects of Moving Clouds on Electric Util-
ities with Dispersed Photovoltaic Generation," IEEE Transactions on Energy
Conversion, vol. EC-2, no. 4, pp. 570{576, Dec 1987.
65
[20] W. T. Jewell and T. D. Unruh, \Limits on Cloud-Induced Fluctuation in Pho-
tovoltaic Generation," IEEE Transactions on Energy Conversion, vol. 5, no. 1,
pp. 8{14, 1990.
[21] Desmond W. H. Cai, S. Adlakha, S. Low, K. Mani. Chandy, and P. de Martini,
\The Impact of Distributed Energy Resources on Utility Rate Structure," in 8th
Annual Carnegie Mellon Conference on The Electricity Industry, 2012.
[22] G. J. Shirek and B. A. Lassiter, \Solar plant modeling impacts on distribution
systems PV case study," in 2012 Rural Electric Power Conference, April 2012,
pp. B5{1{B5{10.
[23] D. Cyganski, J. A. Orr, A. K. Chakravorti, A. E. Emanuel, E. M. Gulachenski,
C. E. Root, and R. C. Bellemare, \Current and Voltage Harmonic Measurements
and Modeling at the Gardner Photovoltaic Project," IEEE Transactions on
Power Delivery, vol. 4, no. 1, pp. 800{809, 1989.
[24] Twenty-First Century PV Community and R. Program, \Photovoltaic Gener-
ation Eects on Distribution Feeders, Volume 1: Description of the Gardner,
Massachusetts," 1990.
[25] H. Asano, K. Yajima, and Y. Kaya, \In
uence of Photovoltaic Power Genera-
tion on Required Capacity for Load Frequency Control," IEEE Transactions on
Energy Conversion, vol. 11, pp. 188{193, 1996.
[26] B. Kroposki and A. Vaughn, \DG Power Quality, Protection, and Reliability
Case Studies Report," 2003. [Online]. Available: http://www.osti.gov/bridge
[27] A. F. POVLSEN, \Impacts of Power Penetration from Photovoltaic
Power Systems in Distribution Networks," 2002. [Online]. Available:
http://www.iea.org
66
[28] Union for the Coordination of Transmission of Electricity (UCTE), \Final Re-
port of the Investigation Committee on the 28 september 2003 Blackout in Italy,"
2004.
[29] T. Degner, J. Schmid, and P. Strauss, Dispower: Distributed Generation with
High Penetration of Renewable Energy Sources, Final Public Report. ISET,
2006.
[30] \The IEEE Reliability Test System - 1996," IEEE Transactions on Power Sys-
tems, vol. 14, no. 3, p. 1010, 1999.
[31] HVCB-WG C37.04 High Voltage Circuit Breaker Standard Working Group,
\IEEE C37.04-1999: IEEE Standard Rating Structure for AC High-Voltage
Circuit Breakers," 1999.
[32] R. Allan and R. Billinton, Reliability Evaluation of Power Systems. Springer
US, 2016.
[33] \ETAP 14.0 User Guide."
67

Abstract (if available)

Abstract The studies of system with high solar penetration are highly demanded due to the RPS requirement of rapid increase in renewable energy. In this thesis, a 24-bus transmission RTS model is built using ETAP software. Totally four analyses are conducted on the model including power flow, contingency, short circuit, and reliability analysis. The simulation result indicates that a high PV penetration level of 38% is achievable in RTS if two capacitors are installed on Bus 3 and 6, several voltage control inverters are implemented on Bus 1, 2, and 15, and the interruption ratings of protection devices at buses with PV plants are redesigned to coincide with the low fault current. This thesis is a preliminary evaluation of a power transmission system with high solar penetrations.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

System stability effect of large scale of EV and renewable energy deployment

PDF

The dynamic interaction of synchronous condensers, SVC, STATCOM and superconducting magnetic energy storage on electric vehicles

PDF

USC microgrid assessment study

PDF

Distribution system reliability analysis for smart grid applications

PDF

Electric vehicle integration into the distribution grid: impact, control and forecast

PDF

An adaptive distance protection scheme for power systems with SFCL

$Data-driven aggregated fractal dynamic load modeling for early warning of voltage collapse$

PDF

Data-driven aggregated fractal dynamic load modeling for early warning of voltage collapse

PDF

In-situ digital power measurement technique using circuit analysis

PDF

The system architecting process for a solar power satellite concept

PDF

Reinforcement learning in hybrid electric vehicles (HEVs) / electric vehicles (EVs)

PDF

High power, highly efficient millimeter-wave switching power amplifiers for watt-level high-speed silicon transmitters

PDF

Developing frameworks to quantify the operational and environmental performance of energy systems within the context of climate change

PDF

Extending systems architecting for human considerations through model-based systems engineering

PDF

A new hot-spot temperature model and methodology to determine the loss of insulation life in distribution transformers

PDF

Power efficient multimedia applications on embedded systems

PDF

Predictive real-time assessment for power grid monitoring and situational awareness using synchrophasors with partial PMU observability

PDF

High-accuracy adaptive vibrational control for uncertain systems

PDF

Investigation of health system performance: effects of integrated triple element method of high reliability, patient safety, and care coordination

PDF

Utility – customer interface strategy and user application for electric vehicles and distributed generation management

PDF

A joint framework of design, control, and applications of energy generation and energy storage systems

Asset Metadata

Creator Gao, Bingtao (author)

Core Title Impacts of high solar penetration on power transmission systems

School Viterbi School of Engineering

Degree Master of Science

Degree Program Electrical Engineering

Publication Date 04/13/2017

Defense Date 03/21/2017

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag IEEE RTS,OAI-PMH Harvest,power flow analysis,power transmission systems,PV penetration,reliability analysis,renewable portfolio standard

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Beshir, Mohammed (committee chair), Bogdan, Paul (committee member), Maby, Edward (committee member)

Creator Email bingtaog@usc.edu,ccccclllliiifff@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-353528

Unique identifier UC11258202

Identifier etd-GaoBingtao-5176.pdf (filename),usctheses-c40-353528 (legacy record id)

Legacy Identifier etd-GaoBingtao-5176.pdf

Dmrecord 353528

Document Type Thesis

Rights Gao, Bingtao

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA