EE5712 Power System Reliability

:: Outage Model & Power Flow Analysis

Panida Jirutitijaroen
 Announcement
• 2nd project: October 14th due November 11th.
• Two more homework # 5 and # 6
• Final exam on November 24th, 1 PM
Independent outage models
Dependent outage models
Power system component failure model
Review of network modeling
Power system modeling

OUTLINE
Forced outage
Planned outage
Semi-forced outage
Partial failure
Multiple failure

INDEPENDENT OUTAGE MODELS

 Forced Outage Model
• Random occurrence
– Use failure rate to model the occurrance
• Two types, repairable and non-repairable
• Repairable
• Non-repairable
– Aging failure
– Chance failure, generally the failure rate of the
chance failure is quite small and negligible in most
cases
 Repairable Forced Outage Model
• Two-stage Markov model
• λ = 1/MTTF and μ = 1/MTTR
• f (Frequency of failure) = 1/ (MTTF+MTTR)
λ

UP DOWN
μ
UP

nUD nUD nUD

DOWN

Tup
Tdown
 Non-Repairable Aging Model
• Aging failure is a conditional failure depending
on how long the component has survived.
 t 

Burn-in period Useful life Wear-out period

 Recall From Lecture 2
• We can use the following distributions to
model the hazard function of an aged
component
– Normal distribution
– Weibull distribution
– Gamma distribution
Characteristics of Normal Dist.
Characteristics of Weibull Dist.
Characteristics of Gamma Dist.
 Planned Outage Model
• Scheduled events
– Maintenance/Replacem
ent λ

– Operation requirement UP Forced Down

• Three-stage Markov μp
μ

model λp

Planned Outage
 Planned Outage Model Approximation
• Three-stage model • Two 2-stage models

λ
λ
UP Forced Down
UP Forced Down
μ
μp μ
λp

Planned Outage UP

μp
λp

Planned Outage
 Semi-forced Outage Model
• Physical problems cause time-
delay in outage
• Oil leakage example
– Do not cause immediate failure
UP Forced Down
– Outage occur within limited μs
time
• This model is usually included λs
in short-term operation μso
problem. Pre-Outage
• Note that Markov model does
not represent time delay
• The time delay can be
modeled using sequential
Monte Carlo simulation.
 Partial Failure Model
• Minor failure
• Component can still operate
λ at lower capability called
UP μ Down “Derated State”
• Three-state Markov model
μd
λd
μf • May be difficult to have
λf
data on transition rates
Derated State between derated and down
states since repair is usually
performed to bring
component from down to
up states
Simplified Partial Failure Model

λ λ
UP μ Down UP μ Down

μd μf μd
λd λf λd

Derated State Derated State

• Ignore transition rates

between down and
derated states
 Multiple Outage Model
• Multiple independent outages
• Use union of events concept.
• The probability of a union of events is found
from addition rules.
– We have seen this before in series network and
min-cut set method
Common-cause outage
Component-group outage
Station-originated outage
Cascading outage
Environment-dependent failure

DEPENDENT OUTAGE MODEL

 Common-Cause Outage Model
• Multiple components
fail due to the same
external cause 1U µ1 λ1 1D
2U 2U
• i, i, are the failure and
µ2 µ2
repair rates of
λ2 λ2
components i and cis λC
1U µ1 λ1 1D
the common mode 2D 2D
failure rate.
 Component-Group Outage Model
• One component failure Busbar 3
leads to other failures in Supply load
the same group.
– One component fails
Transformer 4
– Others do not fail
Breaker 1 and 2
– System goes to protective
stage--failure
• Simultaneous outage All UP
μ1
• Component can not fail μ2 μ3 μ4
individually. λ1 λ2 λ3 λ4
• Recall, dependent failure 1D 2D 3D 4D
in series network.
 Station-Originated Outage
Busbar 3
• Terminal outage of primary
Supply load
components in transmission
or distribution systems
Transformer 4
• Consider when there is a Breaker 1 and 2 Switch 5
fault on a transformer.
– All fails at the beginning
λto
– Switch 5 open UP Forced Down
– 1,2 and 3 operates μtr
– 4 still fails
λts
μsw
• Need to analyze failure
effects of each components Switching
in the system
Transformer fault model
 Cascading Outage Model
• Sequential failure of components in the system
• Example: Overloading in transmission lines
λ1 λ2 λ3 λn
Normal 1D 2D N down

μc

• Actual application, cascading failure is fast, we

can further simplify the cascading outage by the
group outage model
 Environment-Dependent Model
Normal Weather Adverse Weather • Failure and repair rate of
a component is weather-
Up Up dependent
1/S 1/N • Example: Transmission
U U lines has higher failure
µ µ’ rate in adverse weather
λ λ’
condition
• S is the duration for
1/S 1/N
D D adverse weather
Down Down
• N is the duration for
normal weather
Generating unit
Transmission line
Transformer
Circuit breaker
Data requirement

POWER SYSTEM COMPONENT

FAILURE MODEL
 Basic Components in Power Systems
• Generators
• Transmission lines
• Feeders
• Circuit breakers
• Transformers
• Reactors
• Capacitors
 Generating Unit Model
• Generators are modeled as
three state devices.
Up
– State 1: Up state with full
capacity.
– State 3: Down state with zero 1
capacity.
– State 2: Derated state with
partial capacity. Derated Down
• A two state model is a
special case of this model 2 3
where there is no transition
to state 2.
• Unit unavailability is known
as Forced Outage Rate
(FOR)
 Transmission Line Model
Normal Weather Adverse Weather
• Assume to be either in the
up state or failed state.
Up Up • Failure and repair rates are
further assumed to be
1/S 1/N dependent on weather.
U U – λ = Failure rate in the normal
weather.
µ µ’
– µ = Failure rate inthe
λ λ’ normalweather.
– λ’ = Repair rate in
1/S 1/N theadverseweather.
D D – µ’ = Repair rate in the adverse
weather.
Down Down – N, S = Mean durations of
normal and adverse weather.
 Transformer and Bus Model
• Like the transmission
lines, transformer and
bus are treated as two Up Down

stage devices. µ λ
U D
• Failure and repair rates
are assumed to be
independent of the
weather.
 Circuit Breaker Model
• Fault models classified by three modes of failure
• Ground faults
– Two-stage model as transformer or bus
• Failure to open
– Healthy components may be isolated due to the operation of
secondary zone protection.
– May caused by latent faults in the breaker or the associated
protection system
– Characterized by a probability p, a probability that it may not
respond
• Undesired tripping
– Characterized as a failure rate; its effect will be an open line.
Node voltage equation
Bus admittance matrix
Bus admittance matrix by inspection
Example

REVIEW OF NETWORK MODELING

 Node Voltage Equation
Ip z Iq
+ +
Vp y1 y2 Vq
- -

I p  V p y1

V p Vq   1 1
  y1  V p  Vq  1 1 
y
I p   1 z 
z  z z
 z  V p 
I q  Vq y2 
Vq  V p  1  1
  V p   y2  Vq



Iq 
 
1 1  Vq 
y2  
z z  z  z z

ELEN 460:Network Matrices 31

 Network Component Modeling
• Transmission line • Transformer (except phase
(Π-equivalent circuit) shifter transformer)
I1 z I2 I1 y I2
p.u.

+ + + +
y y
V1 V2 V1 V2
2 2
- - - -

y 1 1 
 I1   2  z 
z  V1   I1   yp.u.  yp.u.  V1 
I    1 y 1  V2   I    y yp.u.  V2 
 2      2   p.u.
 z 2 z

ELEN 460:Network Matrices 32

 Bus Admittance Matrix
• Ybus: Bus admittance matrix
• Matrix form of node voltage equation,
I  YbusV
Ybus
where I = Vector of injected node current
V = Vector of node voltage
Ip z Iq
 1 1 
 I p   y1  z   V 
+ +
z
I    1
p
1  Vq 
Vp y1 y2 Vq
 q   y2  
- -  z z

ELEN 460:Network Matrices 33

 Bus Admittance Matrix by Inspection
• Symmetric matrix:
Ybusk , m  Ybusm, k 
• Diagonal entries:
Ybusk, k  = Sum of the admittance of all components
connected to node i.

• Off-diagonal entries:
Ybusk, m = Negative of the admittance of all components
connected between node i and j.

ELEN 460:Network Matrices 34

Bus Admittance Matrix: Example
1 2
y12
zG1
G1 Load
yload
y13 y 24
z G2
y34 G2
3 4

ELEN 460:Network Matrices 35

Power flow equation
Complex power equation
Power flow variables
Power flow formulation
Solution methods
Decoupled power flow

POWER SYSTEM MODELING

 Purpose of Power Flow Calculations
• Planning and operational planning are done to
ensure continued operation
• Power flow calculations for
– Voltage limits
– Avoid line and generator overload
– Stability margin (angle between two generators)
• In our reliability analysis
– To determine whether a state is loss of load state
or not.

ELEN 460:Power Flow Analysis:

37
Introduction
 Power Flow Equations
• Assumptions
G1 G2 – Three phase balanced system.
– Transformer reactance
SG1 SG 2
V1 V2 absorbed into generator or
load reactance.
S D1 • Key: Conservation of
complex power in the
system.

V3
S D3
 Bus Current Equation at Node i
• Net current into bus i: n
I i  I Gi  I Di   I ik
k 1
• From I  YbusV ,
 I1  Y11 Y12  Y1i  Y1n  V1 
 I  Y  Y2i  Y2 n  V2 
 2   21 Y22
        
     
 I i   Yi1 Yi 2  Yii  Yin  Vi 
        
    
 I n  Yn1 Yn 2  Yni  Ynn  Vn 
n

we have I i   YikVk i  1,2,, n

k 1
Complex Power Equation at Node i
• Net complex power (generation –
load) into bus i is: G1 G2
Si  SGi  SDi
SG1 SG 2
• For n bus system, this S i can be V1 V2
found from complex power S D1 S12
transmitted from bus i to all other S13
buses:
n
Si   Sik i  1,2,, n
k 1
V3
S D3
Where Sik = complex power from bus i to bus k connected to i.
• Using bus current equation at bus i,
n
I i   YikVk
k 1

along with complex power equation,

Si  Vi I i*
we have, n
Si  Vi  Yik*Vk*
k 1

Vi  Vi e jVi  Vi e ji
where
 Power Flow Variables
• Consider power flow equation at bus i,
n
Si  Pi  jQi  Vi  Yik*Vk*
k 1

where Vi  Vi e jV  Vi e j is complex value.

i i

• Normal operation, Ybus is known, i.e., Yik is known.

• This means that there are 4 types of variable in
this equation; namely, Pi , Qi , Vi , and, i .
• Typically, 2 out of 4 will be given depending on
bus type as will be seen in the next section.
 Power Flow Formulation
• Types of buses in the system
– Generator bus: given Pi Vi
– Load bus: given Pi Qi
– Swing or Slack bus: given Vi i
• Specifying 2 out of 4 variables, the remaining
2 are asked to be found.
• Slack bus accounts for I 2 R loss which can not
be known in advance.
 Generator Buses (PV Bus)
• Specified Pi and Vi
Pi : Net power (real) injection
Vi : Voltage magnitude

• Governor controls valve which, as a result,

controls constant power.
• Generator field current controls constant
voltage magnitude.
PGi  jQGi Pi  PGi  PDi
G
Vi  Vi spec
PDi  jQDi
 Load Bus (PQ Bus)
• Specified Pi and Qi
Pi : Net real power injection
Qi : Net reactive power injection

• Bus without generators.

• Net power injection is negative value.

Si  SGi  SDi Pi  Qi  PDi  QDi

0
Pi   PDi
PDi  jQDi
Qi  QDi
 Slack Bus (Swing Bus)
• Only ONE such bus in the system, normally,
assigned as bus #1.
• Specified V1 and 1
V1 : Voltage magnitude
1 : Voltage angle, assigned to zero for reference
angle
PG1  jQG1
G
V1  V1spec
i  0
PD1  jQD1
 Variables Summary

Type of Buses Specify Unknown

Slack Bus
Vi i Pi Qi
Generator Bus
(PV Bus)
Pi Vi Qi i
Load Bus
Pi Qi Vi i
(PQ Bus)
 Net current injection = (+)
Net current injection = (-)

G Load

Net current injection = 0 G

Net current injection = (+)

ELEN 460:Network Matrices 48

 Complex Power Flow Equation
• Power flow equation at node i,
n
Pi   Vi Vk ReYik cos ik  ImYik sin  ik 
k 1
n
Qi   Vi Vk ReYik sin  ik  ImYik cos ik 
k 1

• Basically, we need to solve the set of

equations, f(x) = 0
n
Pi   Vi Vk ReYik cos ik  ImYik sin  ik   0
k 1
n
Qi   Vi Vk ReYik sin  ik  ImYik cos ik   0
k 1
 Solution Methods
• Solve f(x) = 0 by iterations
• Gauss iteration, rewrite
x = h(x)
x(k+1) = h(x(k))
• Newton Raphson, from Taylor’s expansion
f(x+Δx) = f(x) + f’(x)Δx + higher order term
 Newton-Raphson Method
• Matrix form,
f(x+Δx) = f(x) + J(x)Δx + h.o.t
• Let h.o.t. ≈ 0, and f(x+Δx) = 0,
Δx = J(x)-1f(x)
• We have,
-1
x(k+1) = x(k) + J(x(k)) f(x(k))
Application to Power Flow
 P2 x   P2  J J12 
  J   11
   J 21 J 22 
 P2 x   P2 
f x    0 Px 
Q x   Q Px 
 2 2 J11  J12 
   θ V
 
 n
Q x   Q n
 Qx  Qx 
J 21  J 22 
θ V

J12   Px 
1 1
 θ   J11 J12   J11
x      J   
J 22  J J 22  Qx 
f x
  V   21  21
 Decoupled Power Flow Model
• From,  θ   J11 J12   Px 
1

 V   
  J 21 J 22  Qx 

• Active power depends mostly on voltage angles

• Reactive power depends mostly on voltage magnitude.
• Approximate the Jacobian matrix, J12 ≈ J21 ≈ 0.
• Further simplify Jacobian matrix to,
J11  V BV  J 22  V B B  ImYbus
 

• Y’bus is the truncated Ybus by deleting row and column

associated with the reference bus
• [V] is diagonal{|V2|,|V3|,…,|Vn|}.
Fast-Decoupled Power Flow Equations
• The power mismatch,
 V BV θ  Px 

 V B V  Qx 

• Assume flat voltage profile, replace [V] by

identity matrix then,
 Bθ  V Px  Px
1 ~

 B V  V  Qx   Qx 
1 ~
Advantages of Decoupled Power Flow

 Bθ  V Px  Px

1 ~

 B V  V  Qx   Qx 
1 ~

• B matrix is fixed.
• Use LU decomposition to decompose B into
lower and upper triangular matrix.
• Ease the computation
Power flow equations
Data requirement
Flow calculation in reliability analysis
Problem statement
Optimization Problems

NETWORK MODELING IN POWER

SYSTEMS RELIABILITY ANALYSIS
 Power Flow Equations
• Use concept of ‘Power • Find the values of state
balance’. variables at any bus in the
• Given known variables, network
– Voltage magnitude
Type of Buses Specify Unknown
– Voltage angle
Slack Bus
Vi i Pi Qi – Real power
Generator Bus – Reactive power
(PV Bus) Pi Vi Qi i • Based on these values,
Load Bus
Pi Qi Vi i we can compute for the
(PQ Bus)
followings
n
Pi   Vi Vk ReYik cos ik  ImYik sin  ik   0 – Power flow in each
k 1 transmission lines
n
Qi   Vi Vk ReYik sin  ik  ImYik cos ik   0
k 1
 Data Requirement
Deterministic data Stochastic data
• System data • System data
– System configuration – Multiple failures from
– Remedial action after outages common effects
• Component data • Component data
– Transmission line impedances – Failure and repair parameters
, susceptances, current- of individual elements
carrying capacities
– Generating unit parameters,
impedances
Flow Calculation in Reliability Analysis
• Random variables in the problem
– Power Generation
– Transmission line capacity due to thermal limits
– Load
• The system has large state space .
• Each state represent the status of generation,
transmission line capacity, and load
 Problem Statement
• Given a system state,
– Does the state suffer from loss of load?
– If so, how much? what is the minimum loss of load?
• Constraints in the problem
– Kirchhoff's laws, which are power flow equations
– Thermal limits of each transmission lines
– Other operation limits such as voltage magnitude
limit, reactive power limit.
• How to use power flow to analyze these
possibilities?
 Optimization Problem
• Given the followings
– System state, generation, transmission line capacity, and load
– Generation may not be equal to load!!
• Need to know
– If a state is loss of load state
– The minimum loss of load
• Formulate the optimization problem
Minimize load curtailment (C)
Subject to
– Kirchhoff’s laws
– Thermal limits of transmission line capacity
– Voltage limits
– Reactive power limits
 AC Power Flow Model
• Handles both active and reactive power flow
model
• Power flow equations
• Reactive power constraint
• Voltage magnitude constraint
• Transmission line constraint
 AC Power Flow
n
• Denote f(x) = 0 as, Pi   Vi Vk ReYik cos ik  ImYik sin  ik   CiP  0
k 1
n
Qi   Vi Vk ReYik sin  ik  ImYik cos ik   CiQ  0
k 1

• The constraints are as follows

P ≤ Pmax
Qmin ≤ Q ≤ Qmax
Fij = gij(x)
F ≤ Ff
-F ≤ Fr
Where
x = (|V|, θ, P, Q)
gij(x) = function describe circuit flow in the network
F ij = Circuit flow between node i and j
Note that in this formulation the demand is also included in “P” and “Q” vector.
 DC Load Flow
• Also called linearized power flow model
Bθ + G + C = D
• The constraints are as follows
G ≤ Gmax
C≤D
Fij = (θj - θi)bij
F ≤ Ff
-F ≤ Fr
Where
G = Input capacity from a generator (MW)
C = Load curtailment (MW)
D = Demand (MW)
B = Susceptance matrix
bij = Susceptance between node i and j
 = Node voltage angle vector
F ij = Circuit flow between node i and j (MW)
 Linear Network Flow Model
• Also called transportation flow model
AF+G+C=D
• The constraints are as follows
G ≤ Gmax
C≤D
F ≤ Ff
-F ≤ Fr
where
A = Node-branch incidence matrix
F = Circuit flow vector
Ff = Maximum forward flow vector
Fb = Maximum backward flow vector
G = Generation vector
Gmax = Maximum generation vector
D = Demand vector

Adequate for multi-area reliability analysis, not accurate enough for

composite system reliability
 Comparisons
DC Flow Linear Network Flow
• Flow characterized by • Flow characterized by
– KCL and KVL – KCL
• More restricted • Less restricted
• Solution method is linear • Use simple algorithm
programming – Ford-fulkerson min-cut max-
– Simplex method flow
– Interior point method • Fast computation
• Time-consuming • Easy implementation
computation
 Summary
• Outage model of each components
• Independent and dependent model described
by Markov model
• Power Flow model
– Power flow equations
– Solution methods
– Fast decoupled power flow
– DC load flow
– Linear network flow
 Reading Materials
• Risk Assessment of Power Systems, Wenyuan
Li, Chapter 2
• Power Flow Analysis, Arthur R. Bergen and
Vijay Vittal, Chapter 9 and 10
 Next lectures
• Power system reliability anlaysis
• Simple assessment steps
• Levels of analysis
– Single-area reliability analysis
– Multi-area power systems reliability analysis
– Composite system reliability analysis
– Distribution system reliability analysis