4/10/2013

Load Flow Studies

Power Flow in Short Transmission Lines

The Load Flow Problem

Bus Admittance Matrix

Gauss-
Gauss-Seidel & Newton-
Newton-Raphson

Information from a Load Flow Study

Principles of Load Flow Control

1
 4/10/2013

Let us consider a short transmission line. The

single-phase equivalent circuit is shown
below:

We calculate for the

current I and its
conjugate I* :

We calculate the single-

phase complex power at
the sending and receiving
ends:

The direction of power flow will be inherent in the

direction of the current I, i.e., SS is the supplied
power when positive, and SR is the load power
when positive.

2
 4/10/2013

Looking at the
sending-end
complex power:

Getting the real and imaginary (reactive)

components:

If we assume the line reactance is much greater

than the line resistance, i.e., xL >> rL, then we can
neglect rL. This means θ = 90° and Z = X, which
when we substitute in the previous equations yield:

3
 4/10/2013

Looking at the
receiving-end
complex power:

Getting the real and imaginary (reactive)

components:

If we assume the line reactance is much greater

than the line resistance, i.e., xL >> rL, then we can
neglect rL. This means θ = 90° and Z = X, which
when we substitute in the previous equations yield:

4
 4/10/2013

Since we assumed that the transmission line

consists of pure reactance, real power is not
dissipated in the line and PS = PR.

If the transmission line resistance is non-
negligible, we will have to use the “unsimplified”
equations.

Maximum real power transfer occurs when α = 90°.

Real power transfer is more sensitive to the
difference between phase angles of the supply
voltage and the load voltage.

Reactive power transfer is more sensitive to the
difference between magnitudes of the supply
voltage and the load voltage.

The Load Flow Problem

How do you determine the voltage, current, power, and power factor
at various points in a power system?

5
 4/10/2013

Load Flow From the Real World

6
 4/10/2013

Load Flow of Distribution System

Sub-
Load Flow of Transmission and Sub -
transmission System

7
 4/10/2013

How do you determine the voltage, current, and

power flows, at various points in the power system,
under existing conditions of normal operations?

How do you determine the adequacy of the power
system in meeting the demand during
contingencies?

How about if there are contemplated changes in
the power system? How will you determine in
advance the effects of:

Growth of or addition of loads

Addition and decommissioning of generating plants

Expansion of the transmission and distribution systems

before the proposed changes are implemented?

ANSWER: THE LOAD FLOW STUDY!

Load Flow Analysis simulates (i.e.,

mathematically determines) the performance
of an electric power system under a given set
of conditions.

Load Flow (also called Power Flow) takes a
snapshot of the electric power system at a
given point in time.

8
 4/10/2013

The following power system components are

modelled in the load flow problem
formulation:

Loads

Generators

Transformers

Transmission/Distribution Lines

The static components of the power system

are modeled by the bus admittance matrix
[Ybus], bus impedance matrix [Zbus], etc.

The number of buses (excluding the neutral
bus) determines the dimension of the bus
admittance matrix [Ybus], bus impedance
matrix [Zbus], etc.

9
 4/10/2013

Voltage-controlled generating units to supply a

scheduled active power P at a specified voltage
magnitude V. The generators are equipped with
voltage regulators to adjust the field excitation so
that the units will supply or absorb a particular
reactive power Q in order to maintain the voltage.

Swing generating units to maintain the frequency
at 60Hz in addition to the specified voltage. The
generating unit is equipped with frequency
following controller (quick-responding speed
governor) and is assigned as the Swing Generator.

The power system is interconnected through

the busses. The busses must therefore be
identified in the load flow model.

Generators, shunt admittances, and loads are
connected from their corresponding bus to the neutral
bus.

Transmission lines, transformers, and series
impedances are connected from bus to bus.

10
 4/10/2013

To completely describe a particular bus, four

quantities must be specified:

Bus Voltage Magnitude, |VP|

Bus Voltage Phase Angle, δP

Bus Injected Active Power, PP

Bus Injected Reactive Power, QP

The difference between the total load demand

plus losses (both P and Q) and the scheduled
generations is supplied by the swing bus. The
voltage magnitude and phase angle are
specified for the Swing Bus, also called the
Slack Bus.

11
 4/10/2013

The total real power Pp injected into the

system through the bus is specified together
with the magnitude of the voltage Vp at the
bus. The bus voltage magnitude is
maintained through reactive power injection.

The total injected power Pp and the reactive

power Qp at Bus P are specified and are
assumed constant, independent of the small
variations in bus voltage.

12
4/10/2013

13
 4/10/2013

Numerical Methods

Direct Methods
a) Cramer’s Rule
b) Matrix Inversion
c) Gauss Method
d) Gauss-Jordan Method

Iterative Methods
a) Gauss Iterative Method
b) Gauss-Seidel Method
c) Newton-Raphson Method

An iterative method (root word: iterate) is a

repetitive process for obtaining the solution of an
equation or a system of equations.

The solutions start from arbitrarily chosen initial
estimates of the unknown variables from which a
new set of estimates is determined.

Convergence is achieved when the absolute
mismatch between the current and previous
estimates is less than some acceptable pre-
specified precision index (the convergence index)
for all variables.

14
 4/10/2013

Given the system of algebraic equations,

In the above equation, the x’s are unknown.

In general, the jth equation may be written as:

15
 4/10/2013

In general, the Gauss iterative estimates are:

where k is the iteration count.

From an initial estimate of the unknowns (x10,

x20,…xn0), updated values of the unknown
variables are computed using equation “a”.
This completes one iteration. The new
estimates replace the original estimates.
Mathematically, at the kth iteration,

16
 4/10/2013

A convergence check is conducted after each

iteration. The latest values are compared with
their values respectively.

The iteration process is terminated when:

Example:

Assume a convergence index of ε = 0.001

and use the following initial estimates:

17
 4/10/2013

Solution:

a) The system of equation must be expressed
in standard form.

Iteration 1
(k = 0):

18
 4/10/2013

Iteration 2
(k = 1):

Iteration 3
(k = 2):

19
 4/10/2013

Iteration 4
(k = 3):

Iteration 5
(k = 4):

20
 4/10/2013

Iteration 6
(k = 5):

Iteration 7
(k = 6):

21
 4/10/2013

Iteration 8
(k = 7):

The Gauss iterative method has converged at

iteration 8. The method yields the following
solution:

22
 4/10/2013

The Gauss-Seidel method is an improvement over the

Gauss iterative method. As presented in the previous
section, the standard form of the jth equation may be
written as follows.

From an initial estimates (x10, x20,…xn0), an updated

value is computed for x1 using the above equation with
j set to 1.This new value replaces x10 and is then used
together with the remaining initial estimates to compute
a new value for x2. The process is repeated until a new
estimate is obtained for xn. This completes one
iteration.

Within an iteration,
the most recent
computed values are
used in computing
for the remaining
unknowns. In
general, at iteration
k,

After each iteration, a convergence check is conducted.

The convergence criterion applied is the same with
Gauss Iterative Method.

23
 4/10/2013

An improvement to the Gauss Iterative Method

Example:

Solve the system of equations using the
Gauss-Seidel method. Used a convergence
index of ε = 0.001

24
 4/10/2013

Solution:

a) The system of equation must be expressed
in standard form.

Iteration 1
(k = 0):

25
 4/10/2013

Iteration 2
(k = 1):

Iteration 3
(k = 2):

26
 4/10/2013

Iteration 4
(k = 3):

Iteration 5
(k = 4):

27
 4/10/2013

The Gauss-Seidel Iterative Method has

converged after only 5 iterations with the
following solutions:

28
4/10/2013

29
4/10/2013

30
 4/10/2013

Gauss-Seidel Voltage Equations of the form

shown in (6) are written for all buses except
for the swing bus. The solution proceeds
iteratively from an estimate of all bus
voltages

For a Load Bus (Type 3) whose real power and
reactive power are specified, the G-S voltage
equation is used directly to compute the next
estimate of the bus voltage.

31
 4/10/2013

For a Generator Bus (Type 2) where the

voltage magnitude is specified, an estimate of
Qp must be determined first. This estimate is
then compared with the reactive power limits
of the generator. If it falls within the limits,
the specified voltage is maintained and the
computed Qp is inputted, in the Gauss-Seidel
equation. Otherwise, the reactive power is set
to an appropriate limit (Qmin or Qmax) and the
bus is treated as a load bus in the current
iteration.

32