NETWORK MATRIX &

POWER FLOW MODEL

Transmission Networks

2
 Distribution Network

69 70 71
S88 S67
S68 S69
Tie switch
Modified tie lines S86
67 S89

S83 S90
S66
S87
S85
61 62 63 64
S46 S49
38 S41 S79
S84 S56 S60 65
S61 S62
36 37 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58
S39 S63
S81
S53 66 60
S36 S37 S38 S40 S42 S43 S44 S45 S47 S48 S50 S51 S52 S54 S55 S57 S58 S82
59
S76 S64
S65
S59 S74
S71
S70 S73
S72
S77
2 3 4 68 5 6 7 8 9 S31
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
31 32 33 34 35
S1 S2 S4 S15 S78
S3 S6 S7 S8 S9 S11 S12 S14 S16 S18 S19 S20 S21 S23 S24 S25 S26 S28 S29 S30
S32 S33 S34 S35
S5
S10 S75
S13
S17
S22
S27
S80

3
 Why Do We Need Networks?

▪ Economic efficiency
o Fuel sources usually far from demand centres
o Generating plants can be located in the best places
o Always use the most efficient generating plants

▪ System reliability
o Interconnection of plants
- Increases reliability
- Reduces the need for reserve

4
 Problems with Networks

• Must operate satisfactorily under variety of conditions

- Voltage and flows must be acceptable at all times

• Must continue operating under adverse conditions (e.g.

N-1)
- Loss of lines, lost of generators (single outage = N-1)

• They are large and complex

We must be able to predict how the network behave

→ Power flow computation
→ Central calculation in power system analysis

This course will focus on techniques for computing

voltages and flows for large systems 5
 Review of Nodal Analysis

▪ Systematic method for solving circuits problems

o This will form the basis to understand how the admittance
matrix is created

▪ Algorithm of the technique

o Apply Kirchhoff's Current Law (KCL) at each node
- Sum of currents in = sum of currents out
o Express branch currents in terms of nodal voltages
o Solve for the nodal voltages

6
 Example of Nodal Analysis

→ Find the voltage at every node

→ Find the current flows in every branch
7
 Step 1 : Current sources

▪ Change all voltage sources into current sources

𝑉
𝐼𝑠 = 𝑠ൗ𝑅
𝑠
8
 Step 1 : Current sources

▪ Change all voltage sources into current sources

9
 Step 2 : Identify the variables

▪ Choose a reference node (if possible its voltage is 0)

▪ Identify voltages with respect to the reference node : 𝑉1 , 𝑉2 , 𝑉3 , 𝑉4
▪ Identify currents in all the branches

10
 Step 3 : Write KCL for each node

Convention
Current into the node : -ve
Current out from the node : + ve

Node 1: −𝐼21 − 𝐼31 − 𝐼41 = 0

Sum of currents in
Node 2: 𝐼21 − 𝐼32 − 𝐼42 = 0
equals to
sum of currents out Node 3: −𝐼𝑠3 + 𝐼30 + 𝐼31 + 𝐼32 = 0
Node 4: −𝐼𝑠4 + 𝐼40 + 𝐼41 + 𝐼42 = 0
11
 Step 4 : Express currents in terms
of voltages

𝑉𝑎− 𝑉𝑏
𝐼𝑎𝑏 =
𝑅𝑎𝑏

𝑉1 − 𝑉2 𝑉1 − 𝑉3 𝑉1 − 𝑉4
+ + =0
0.125 0.25 0.4

−𝐼21 − 𝐼31 − 𝐼41 = 0 𝑉2 − 𝑉1 𝑉2 − 𝑉3 𝑉2 − 𝑉4

+ + =0
0.125 0.25 0.2
𝐼21 − 𝐼32 − 𝐼42 = 0
𝑉3 𝑉3 − 𝑉1 𝑉3 − 𝑉2
−𝐼𝑠3 + 𝐼30 + 𝐼31 + 𝐼32 = 0 + + = 𝐼𝑠3
1.25 0.25 0.25
−𝐼𝑠4 + 𝐼40 + 𝐼41 + 𝐼42 = 0
𝑉4 𝑉4 − 𝑉1 𝑉4 − 𝑉2
+ + = 𝐼𝑠4 12
1.25 0.4 0.2
 Step 5 : Re-organize the equations
For Node 1
𝑉1 − 𝑉2 𝑉1 − 𝑉3 𝑉1 − 𝑉4 Simplify to whole number
+ + =0
0.125 0.25 0.4 8𝑉1 − 8𝑉2 + 4𝑉1 − 4𝑉3 + 2.5𝑉1 − 2.5𝑉4 = 0
Collect the same term
𝑉2 − 𝑉1 𝑉2 − 𝑉3 𝑉2 − 𝑉4 14.5𝑉1 − 8𝑉2 − 4𝑉3 − 2.5𝑉4 = 0
+ + =0
0.125 0.25 0.2
𝑉3 𝑉3 − 𝑉1 𝑉3 − 𝑉2
+ + = 𝐼𝑠3 Do the same for all nodal equations for each node
1.25 0.25 0.25
𝑉4 𝑉4 − 𝑉1 𝑉4 − 𝑉2
+ + = 𝐼𝑠4
1.25 0.4 0.2

Express the simplified equations in matrix form

14.5 −8 −4 −2.5 𝑉1 0
−8 17 −4 −5 𝑉2 0 Injected currents
=
−4 −4 8.8 0 𝑉3 1
−2.5 −5 0 8.3 𝑉4 0.68

Admittance matrix Y Unknown voltages 13

 Shortcut to Step 3-5

▪ Build the admittance matrix directly

o Diagonal element (i,j) :
- Sum of the admittances connected to node i including
reference nodes

o Off-diagonal element (i,j) :

- Minus the admittance of the branch between nodes i and j

o No branch between nodes i and j → element (i,j) is zero

14
 Shortcut (using previous example)

• Convert to branch admittance

• Add up the admittance for Admittance Matric

14.5 −8 −4 −2.5
−8 17 −4 −5
−4 −4 8.8 0
−2.5 −5 0 8.3

15
 Step 6 : Solving the matric

14.5 −8 −4 −2.5 𝑉1 0
−8 17 −4 −5 𝑉2 0
=
−4 −4 8.8 0 𝑉3 1
−2.5 −5 0 8.3 𝑉4 0.68
or

𝑌∙𝑉 =𝐼

𝑉 = 𝑌 −1 ∙ 𝐼
• Solving the nodal equations is straightforward because it involves only
the solution of a system of linear equations. There are calculator (for
small system) or good computer programs to do that!

16
 Notes on the Admittance Matrix Y

o Elements of 𝑌 depend only on admittances of branches of the

network
o If there is no branch between nodes i and j, element 𝑌𝑖𝑗 is zero
o For a large network such as a power system, most of the elements
of 𝑌 are zero – Y is a sparse matrix
o Special methods are needed for dealing with sparse matrices to
further speed up the computation.

17
 Sparse Matrix

o Example :
• Network with 1,000 nodes
• 1,000,000 elements in Y
• On average each node connected to about three other nodes
• On average each row of Y bus has 4 non-zero elements
➢ Diagonal element
➢ 3 non-diagonal elements corresponding to ‘neighbours’ of the
node
• There is thus only about 4,000 non-zero elements in Y
• Only about 0.4% of the elements of Y are non-zero
o Specialised computer code to take advantage of the sparsity of the
admittance matrix
18
 Exercise 3.1

Determine the bus voltages for the following power system network:
Given 𝑆1 = 1.1∠0° and 𝑆2 = 1.0∠0°

There is no connection
between bus 1-4 and bus 2-4,
thus a few terms are not exist
in the NV equations!

Solution:
The N-V equations of the system is of the form:
1.1∠0°
𝐼1 𝑌11 𝑌12 𝑌13 0 𝑉1 𝐼1 = = −𝑗1.1
𝑌 𝑌22 𝑌23 0 𝑉2 𝑗1.0
𝐼2 where
= 21 1.0∠0°
0 𝑌31 𝑌31 𝑌33 𝑌34 𝑉3
𝐼2 = = −𝑗1.25
0 0 0 𝑌43 𝑌44 𝑉4 𝑗0.8
19
 Solution 3.1

Build the admittance matrix :

𝑌11 = 𝑦10 + 𝑦12 + 𝑦13 𝑌12 = 𝑌21 = −𝑦12

𝑌22 = 𝑦20 + 𝑦12 + 𝑦23 𝑌13 = 𝑌31 = −𝑦13

𝑌33 = 𝑦13 + 𝑦23 + 𝑦34 𝑌23 = 𝑌32 = −𝑦23

𝑌44 = 𝑦34 𝑌34 = 𝑌43 = −𝑦34

Arranging in matrix form;

𝑌11 𝑌12 𝑌13 0 𝑉1 𝐼1 −𝑗8. 50 𝑗2. 50 𝑗5.00 0 𝑉1 −𝑗1.1

𝑌12 𝑌22 𝑌23 0 𝑉2 𝐼 𝑗2. 50 −𝑗8. 75 𝑗5.00 0 𝑉2 −𝑗1.25
= 2 =
𝑌13 𝑌23 𝑌33 𝑌34 𝑉3 𝐼3 𝑗5.00 𝑗5.00 −𝑗22.50 𝑗12.50 𝑉3 0
0 0 𝑌34 𝑌44 𝑉4 𝐼4 0 0 𝑗12.50 −𝑗12.50 𝑉4 0
Solve for 𝑉1 , 𝑉2 , 𝑉3 , 𝑉4
20
𝑽𝟏 = 1.0500 𝑽𝟐 = 1.0400 𝑽𝟑 = 1.0450 = 𝑽𝟒
 Summary of Nodal Analysis Method

o Write KCL for each node

o Express the currents as a function of the node voltages
o Solve matrix equation YV = I for the node voltages
- Shortcut : build the admittance matrix Y directly
o Nodal analysis method can be applied to AC circuits where there are
- Complex voltages
- Complex currents
- Complex impedance/admittances
o Solving complex equations is tedious by hand but not hard by a computer

21
 Nodal Analysis for Power Flow

o Write active power balance at each bus (power in = power out)

o Write reactive power balance at each bus
o Express power flowing in the branches in terms of the nodal
voltages
o Solve for the nodal voltages

22
 Power Balance

𝑁𝑘 : Set of buses connected to bus k by a branch

𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑
𝑃𝑘 = 𝑃𝑘𝐺 − 𝑃𝑘𝐿 = ෍ 𝑃𝑘𝑖 Total active power flowing on the
𝑖∈𝑁𝑘 branches connected to bus k

𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 Total reactive power flowing on

𝑄𝑘 = 𝑄𝑘𝐺 − 𝑄𝑘𝐿 = ෍ 𝑄𝑘𝑖
𝑖∈𝑁𝑘 the branches connected to bus k 23
 Power Balance using Nodal Analysis

From what we have learned, expressing injected currents in terms of the nodal voltages

𝐼𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 = 𝑌 ∙ 𝑉

For the power system in the example above, current injected at bus k is :

𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑
𝐼𝑘 = ෍ 𝑌𝑘𝑖 ∙ 𝑉𝑖 ෍ 𝑌𝑘𝑖 ∙ 𝑉𝑖 = 𝑌𝑘𝑘 ∙ 𝑉𝑘 + 𝑌𝑘1 ∙ 𝑉1 + 𝑌𝑘2 ∙ 𝑉2 + 𝑌𝑘3 ∙ 𝑉3
𝑖∈𝑁𝑘 𝑖∈𝑁𝑘

Note : 𝑖 ∈ 𝑁𝑘 means i is an element of

a set of N bus, including bus k
24
 Power Balance using Nodal Analysis

𝑁𝑘 : Set of buses
connected to bus
k by a branch
(including bus k
itself)

Relating the current equations to the injected power into

the bus using nodal voltages

𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑∗

𝐼𝑘 = ෍ 𝑌𝑘𝑖 ∙ 𝑉𝑖 𝑆𝑘 = 𝑃𝑘 + 𝑗𝑄𝑘 = 𝑉𝑘 ∙ 𝐼𝑘
𝑖∈𝑁𝑘

= 𝑉𝑘 ∙ ෍ 𝑌𝑘𝑖∗ ∙ 𝑉𝑖∗
𝑖∈𝑁𝑘

= ෍ 𝑌𝑘𝑖∗ ∙ 𝑉𝑘 ∙ 𝑉𝑖∗
𝑖∈𝑁𝑘
25
 A bit of Complex Algebra

𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑

𝑆𝑘 = 𝑃𝑘 + 𝑗𝑄𝑘 = ෍ 𝑌𝑘𝑖∗ ∙ 𝑉𝑘 ∙ 𝑉𝑖∗
𝑖∈𝑁𝑘

26
 The Power Flow Equations

o Expresses balance of active and reactive power at each bus

o Set of non-linear simultaneous algebraic equations
o “Usual” unknows : voltage magnitudes and angles – but more complex
o Elements of matrix Y used in the equations

27
 Types of Node

▪ Active and reactive power balances must be satisfied at all buses

and for the system as a whole
▪ Three different ways of satisfying this balance :

PQ Node – Load Bus

𝑠𝑝 𝑠𝑝
Slack Node – Ref Bus Specified Variables : 𝑃𝑖 , 𝑄𝑖
𝑠𝑝
Specified Variables : 𝜃𝑖 = 0, 𝑉𝑖
𝑠𝑝
Unknown Variables : 𝑉𝑖 , 𝜃𝑖
Unknown Variables : 𝑃𝑖 , 𝑄𝑖

PV Node – Generator Bus

Specified Variables : 𝑃𝑖𝑠𝑝 , 𝑉𝑖𝑠𝑝
Unknown Variables : 𝑄𝑖 , 𝜃𝑖
28
 PQ Bus

o At a PQ bus
✓ 𝑃𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 is specified
✓ 𝑄𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 is specified
✓ Voltage magnitude and angle are unknown
✓ Voltage magnitude and angle are such that the active and reactive
power flows in the lines connected to the PQ bus match the injections.
o Load buses are usually called PQ buses. 29
 PV Bus

o At a PV bus
✓ 𝑃𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 is specified
✓ Voltage magnitude is specified
✓ Voltage angle is unknown
✓ 𝑄𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 is unknown
o 𝑄𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 adjusted to keep voltage magnitude constant.
o Must have a variable var source at each PV bus
o Most generation buses are PV buses (in transmission system) because
they have a controller that keeps voltage constant. 30
 Balancing the Power Losses

o Power balances must also be satisfied at the system level

o Can’t specify injection at all busses because of losses

𝐺𝑒𝑛𝑒𝑟𝑎𝑡𝑖𝑜𝑛 𝐿𝑜𝑎𝑑 𝐿𝑜𝑠𝑠𝑒𝑠

𝑃𝑆𝑦𝑠𝑡𝑒𝑚 = 𝑃𝑆𝑦𝑠𝑡𝑒𝑚 + 𝑃𝑆𝑦𝑠𝑡𝑒𝑚
o Losses are a function of current flows, which are a function of voltages,
which are not yet known.

31
 Balancing the Power Losses: The Slack Bus

o Solution: leave the 𝑃𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 and 𝑄𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑 unspecified at one bus – the
slack bus
𝑖𝑛𝑗𝑒𝑐𝑡𝑒𝑑
o During the power flow calculation, 𝑃𝑆𝑙𝑎𝑐𝑘 is adjusted to keep the
system in balance.
o Also, need reference for the angles
- Set the voltage angle at slack bus arbitrarily (equal to 0 degree)
o Voltage magnitude and angle are specified at the slack bus
o A generator bus is usually chosen as slack bus because active and
reactive power of that generator can be adjusted by the power flow.

32
 Comments on the Slack Bus

o In real life, there is no slack

- Operator adjusts generation to meet load + losses
- All angle measurements are relative
o Doesn’t the choice of slack affect the solution?
- Yes, if adjustments are large
- Reduce adjustments
• Estimate losses
• Distribute adjustments over several generators.

33
 Equations and Variables

o N buses
o 2N equations (2 at each bus)
o 4N variables (P, Q, V, and 𝜃 at each bus)
o 2N unknown variables (2 at each bus) :
- V, 𝜃 at PQ bus
- Q, 𝜃 at PV bus
- P, Q at slack bus
o 2N known variables (2 at each bus)
- P, Q at PQ bus
- P, 𝑉 at PV bus
- V, 𝜃 at slack bus

34
 A Three-Bus Example

o All values are per unit impedances.

o Write the power flow equations for this system

35
 Y Matrix

o Calculate the branch/line admittances, then the Y matrix

−𝑗15 𝑗10 𝑗5
𝑌 = 𝐺 + 𝑗𝐵 = 𝑗10 −𝑗14 𝑗4
𝑗5 𝑗4 −𝑗9

36
 Power Flow Equations

o Two equations for each node

- First equation: balance of active power at that node
- Second equation : balance of reactive power at that node
o Left hand side : power injected at the node (generator or load)
o Right hand side : power flowing in the branches
37
 Power Flow Equations

−𝑗15 𝑗10 𝑗5
𝑌 = 𝐺 + 𝑗𝐵 = 𝑗10 −𝑗14 𝑗4
𝑗5 𝑗4 −𝑗9

𝑃1 = −15𝑉1 𝑉1 sin 𝜃11 + 10𝑉1 𝑉2 sin 𝜃12 + 5𝑉1 𝑉3 sin 𝜃13

𝑄1 = 15𝑉1 𝑉1 cos 𝜃11 − 10𝑉1 𝑉2 cos 𝜃12 − 5𝑉1 𝑉3 cos 𝜃13
𝑃2 = 10𝑉2 𝑉1 sin 𝜃21 − 14𝑉2 𝑉2 sin 𝜃22 + 4𝑉2 𝑉3 sin 𝜃23
𝑄2 = −10𝑉2 𝑉1 cos 𝜃21 + 14𝑉2 𝑉2 cos 𝜃22 − 4𝑉2 𝑉3 cos 𝜃23
𝑃3 = 5𝑉3 𝑉1 sin 𝜃31 + 4𝑉3 𝑉2 sin 𝜃32 − 9𝑉3 𝑉3 sin 𝜃33
𝑄3 = −5𝑉3 𝑉1 cos 𝜃31 − 4𝑉3 𝑉2 cos 𝜃32 + 9𝑉3 𝑉3 cos 𝜃33

o Note: All the “G” terms (conductance) are zero because we neglected the
resistance of the lines. 38
 Simplifying the Power Flow Equations

o Since we have :

sin 𝜃11 = sin 𝜃22 = sin 𝜃33 = sin 0° = 0

cos 𝜃11 = cos 𝜃22 = cos 𝜃33 = cos 0° = 1

o The power flow equations become :

𝑃1 = 10𝑉1 𝑉2 sin 𝜃12 + 5𝑉1 𝑉3 sin 𝜃13

𝑄1 = 15𝑉12 − 10𝑉1 𝑉2 cos 𝜃12 − 5𝑉1 𝑉3 cos 𝜃13
𝑃2 = 10𝑉2 𝑉1 sin 𝜃21 + 4𝑉2 𝑉3 sin 𝜃23
𝑄2 = −10𝑉2 𝑉1 cos 𝜃21 + 14𝑉22 − 4𝑉2 𝑉3 cos 𝜃23
𝑃3 = 5𝑉3 𝑉1 sin 𝜃31 + 4𝑉3 𝑉2 sin 𝜃32
𝑄3 = −5𝑉3 𝑉1 cos 𝜃31 − 4𝑉3 𝑉2 cos 𝜃32 + 9𝑉32

39
Bus Types

40
 Separating the Equations

o We know the injections : implicit equations (simultaneous)

𝑃2 = 10𝑉2 𝑉1 sin 𝜃21 + 4𝑉2 𝑉3 sin 𝜃23 P equation of PQ bus

𝑄2 = −10𝑉2 𝑉1 cos 𝜃21 + 14𝑉22 − 4𝑉2 𝑉3 cos 𝜃23 Q equation of PQ bus

𝑃3 = 5𝑉3 𝑉1 sin 𝜃31 + 4𝑉3 𝑉2 sin 𝜃32 P equation of PV bus

o We don’t know the injections : explicit equations

𝑃1 = 10𝑉1 𝑉2 sin 𝜃12 + 5𝑉1 𝑉3 sin 𝜃13 P equation of Slack bus

𝑄1 = 15𝑉12 − 10𝑉1 𝑉2 cos 𝜃12 − 5𝑉1 𝑉3 cos 𝜃13 Q equation of Slack bus
𝑄3 = −5𝑉3 𝑉1 cos 𝜃31 − 4𝑉3 𝑉2 cos 𝜃32 + 9𝑉32 Q equation of PV bus

41
 Implicit Equations

o Known variables : 𝑃2 , 𝑄2 , 𝑃3 , 𝑉1 , 𝑉3 , 𝜃1
o Unknown variables : 𝑉2 , 𝜃2 , 𝜃3

−1.5 = 10𝑉2 𝑉1 sin 𝜃21 + 4𝑉2 𝑉3 sin 𝜃23 P equation of bus 2

−0.8 = −10𝑉2 𝑉1 cos 𝜃21 + 14𝑉22 − 4𝑉2 𝑉3 cos 𝜃23 Q equation of bus 2
1.0 = 5𝑉3 𝑉1 sin 𝜃31 + 4𝑉3 𝑉2 sin 𝜃32 P equation of bus 3

o Set of simultaneous non-linear equations

42
 Explicit Equations

o Known variables : 𝑉1 , 𝑉3 , 𝜃1
o Solution of implicit equations : 𝑉2 ,𝜃2 , 𝜃3

o Unknown variables : 𝑃1 , 𝑄1 , 𝑄3

𝑃1 = 10𝑉1 𝑉2 sin 𝜃12 + 5𝑉1 𝑉3 sin 𝜃13 P equation of bus 1

𝑄1 = 15𝑉12 − 10𝑉1 𝑉2 cos 𝜃12 − 5𝑉1 𝑉3 cos 𝜃13 Q equation of bus 1
𝑄3 = −5𝑉3 𝑉1 cos 𝜃31 − 4𝑉3 𝑉2 cos 𝜃32 + 9𝑉32 Q equation of bus 3

43
Solving the Implicit Equations

o Implicit equations : set of non-

linear equations
o No direct solution
o Need an iterative solution
method
o Newton-Rhapson Method
o Example : one function of one
variable
o Solve f(x) = 0 where f(x) is a non-
linear function of one variable.

44