Question 4 of 7: Bus Classification and a Three-Bus Power Flow
Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)
Notes on this paper
Paper format. National Exams, May 2013 —
07-Elec-B7 Power Systems Engineering. Open-book, three hours, non-communicating
calculator permitted. Seven problems of equal value (20 points each); any five
constitute a complete paper and only the first five appearing in the answer book
are marked. All seven are solved here, because the set is a study
resource.
Reference texts.
M. E. El-Hawary, Electrical Power Systems: Design and Analysis,
rev. ed., IEEE Press/Wiley — the source of this paper's notation
(ABCD two-port, the “cantilever” transformer equivalent, and the
equal-area treatment of transient stability).
J. D. Glover, T. J. Overbye and M. S. Sarma, Power System Analysis and
Design, 6th ed., Cengage — transmission-line parameters (Ch. 4–5),
power flow (Ch. 6), symmetrical faults (Ch. 7), symmetrical components and
unsymmetrical faults (Ch. 8–9), transient stability (Ch. 11).
J. J. Grainger and W. D. Stevenson, Power System Analysis,
McGraw-Hill — sequence networks and fault calculations.
S. J. Chapman, Electric Machinery Fundamentals, 5th ed.,
McGraw-Hill — transformers (Ch. 2) and synchronous machines (Ch. 4–5).
Canadian practice: CSA C22.3 No. 1 Overhead Systems (conductor
temperature, sag and clearance), CAN/CSA-C88 and CSA C50 (power transformers and
insulating oil), and CSA C22.3 No. 1 Annex on fault-current duty.
Problem 4: Bus Classification and a Three-Bus Power Flow (20 points)
Part (a) — the three bus types. Each bus of a power-flow
problem carries four quantities: the voltage magnitude $|V|$, the voltage angle
$\delta$, the net real power injection $P$ and the net reactive power injection
$Q$. Two of the four are specified as data and two are computed; which two are
specified is what defines the bus type.
Bus type
Also called
Specified
Computed
Physical meaning
Slack
swing, reference
$|V|,\ \delta$ (usually $1.0\angle 0^\circ$)
$P,\ Q$
the one machine that takes up the mismatch between generation, load and the as-yet-unknown network losses, and supplies the angle reference for the whole solution. Exactly one per island.
Voltage-controlled (PV)
generator bus
$P,\ |V|$
$Q,\ \delta$
a generator whose governor sets its output power and whose exciter/AVR holds its terminal voltage; the reactive output is whatever it takes, subject to the machine's Q limits.
Load (PQ)
demand bus
$P,\ Q$
$|V|,\ \delta$
a bus with a known demand and no voltage control — the great majority of buses. Injections are negative for a load.
Two practical riders complete the answer. A PV bus reverts to a PQ bus
during the solution if its computed $Q$ hits a machine limit: $Q$ is then fixed at
the limit and $|V|$ is released. And a bus with neither generation nor load
(a pure junction or a switching station) is simply a PQ bus with
$P = Q = 0$.
In the system of Figure 2 bus 1 is the slack bus
($V_1 = 1.0\angle 0^\circ$, $P_1$ and $Q_1$ unknown), bus 3 is a
voltage-controlled generator bus whose solved state
$V_3 = 1.05\angle 32^\circ$ is given here, and bus 2 is a load bus carrying
$P_2 = -4$ pu with $|V_2|$ and $\delta_2$ to be found. The question hands us
$\delta_2 = -8.5^\circ$, which converts the non-linear power-flow solution into a
direct calculation.
Figure 4.1 — the three-bus system of Figure (2). Lines are pure reactances, so all real power transfer is governed by the angle differences and all reactive transfer by the magnitude differences.
Given.
Quantity
Value
Meaning
$x_{12}$
$0.08\ \text{pu}$
bus 1 to bus 2 series reactance
$x_{23}$
$0.20\ \text{pu}$
bus 2 to bus 3 series reactance
$V_1$
$1.00\angle 0^\circ\ \text{pu}$
slack bus
$V_3$
$1.05\angle 32^\circ\ \text{pu}$
generator bus
$P_2$
$-4.0\ \text{pu}$
net real injection at bus 2 (a 4 pu load)
$\delta_2$
$-8.5^\circ$
given voltage angle at bus 2
Find. $|V_2|$ and $Q_2$, then the generation
$P_1 + jQ_1$ and $P_3 + jQ_3$.
Approach. Write the standard polar power-flow injection
equations for a purely reactive network; the $P_2$ equation is then linear in
$|V_2|$ and solves in one line, after which every remaining quantity is an explicit
evaluation.
Specialise the injection equations to a lossless network. For
a line of pure series reactance $x_{ik}$ the general polar forms
$P_i=\sum_k |V_i||V_k|\left(G_{ik}\cos\delta_{ik}+B_{ik}\sin\delta_{ik}\right)$ and
$Q_i=\sum_k |V_i||V_k|\left(G_{ik}\sin\delta_{ik}-B_{ik}\cos\delta_{ik}\right)$
collapse to
$$P_i=\sum_{k\neq i}\frac{|V_i||V_k|}{x_{ik}}\sin\delta_{ik},
\qquad
Q_i=\sum_{k\neq i}\left(\frac{|V_i|^{2}}{x_{ik}}
-\frac{|V_i||V_k|}{x_{ik}}\cos\delta_{ik}\right)$$
with $\delta_{ik}=\delta_i-\delta_k$. Real power is carried by angle difference,
reactive power by magnitude difference — the two decouple, which is the whole
reason a fast-decoupled solver works.
Solve the bus-2 real-power equation for the magnitude. Bus 2
connects only to buses 1 and 3, and $|V_2|$ factors out of both terms:
$$P_2=|V_2|\left[\frac{|V_1|\sin(\delta_2-\delta_1)}{x_{12}}
+\frac{|V_3|\sin(\delta_2-\delta_3)}{x_{23}}\right]$$
Evaluating the bracket with $\delta_2-\delta_1 = -8.5^\circ$ and
$\delta_2-\delta_3 = -40.5^\circ$,
$$\frac{1.00\sin(-8.5^\circ)}{0.08}
+\frac{1.05\sin(-40.5^\circ)}{0.20}
=-1.8459-3.4113=-5.2572$$
so
$$|V_2|=\frac{-4.0}{-5.2572}
\;\Longrightarrow\;\boxed{|V_2|=0.7609\ \text{pu}}$$
Reactive injection at bus 2. Substituting the same angles and
the magnitude just found into the $Q$ equation,
$$Q_2=\frac{|V_2|^{2}}{x_{12}}+\frac{|V_2|^{2}}{x_{23}}
-\frac{|V_2||V_1|\cos(\delta_2-\delta_1)}{x_{12}}
-\frac{|V_2||V_3|\cos(\delta_2-\delta_3)}{x_{23}}$$
$$\boxed{Q_2=-2.3129\ \text{pu}}$$
The sign convention is injection, so a negative value means bus 2 absorbs
2.3129 pu of reactive power: the load is drawing about
0.58 pu of lagging vars for every 1 pu of real power, i.e. a power factor of
roughly 0.87 lagging at that bus.
Part (c) — generation at the slack bus. Bus 1 sees only
bus 2, so
$$P_1=\frac{|V_1||V_2|}{x_{12}}\sin(\delta_1-\delta_2)
=\frac{1.00\times0.7609}{0.08}\sin(8.5^\circ)
\;\Longrightarrow\;\boxed{P_1=1.4058\ \text{pu}}$$
$$Q_1=\frac{|V_1|^{2}}{x_{12}}-\frac{|V_1||V_2|}{x_{12}}\cos(\delta_1-\delta_2)
=\frac{1}{0.08}-\frac{0.7609}{0.08}\cos(8.5^\circ)
\;\Longrightarrow\;\boxed{Q_1=3.0937\ \text{pu}}$$
The slack machine is delivering nearly 3.1 pu of reactive power for 1.4 pu of real
power — a consequence of the depressed voltage at bus 2, which forces the
reactive support to come from both ends.
Part (d) — generation at bus 3. Identically, with
$\delta_3-\delta_2 = 40.5^\circ$ and $x_{23} = 0.20$,
$$P_3=\frac{1.05\times0.7609}{0.20}\sin(40.5^\circ)
\;\Longrightarrow\;\boxed{P_3=2.5942\ \text{pu}}$$
$$Q_3=\frac{1.05^{2}}{0.20}-\frac{1.05\times0.7609}{0.20}
\cos(40.5^\circ)
\;\Longrightarrow\;\boxed{Q_3=2.4751\ \text{pu}}$$
Balance checks. The lines are lossless, so the real powers
must sum to zero:
$P_1+P_2+P_3 = 1.4058 - 4 + 2.5942
= 0.000000$ pu. The reactive powers must sum to the
$I^{2}X$ absorbed by the two reactances:
$$\sum Q=3.2559\ \text{pu}
\quad\text{versus}\quad
|I_{12}|^{2}x_{12}+|I_{32}|^{2}x_{23}
=0.9238+2.3321
=3.2559\ \text{pu},$$
with $|I_{12}| = 3.3981$ pu and
$|I_{32}| = 3.4148$ pu. Both balances close, which
validates every number above.
Check: the solved $|V_2| =
0.7609$ pu is far outside any operating standard (a CSA/IEEE
service voltage would be held within roughly 0.95–1.05 pu). That is a
property of the data the question supplies, not an arithmetic error: a 4 pu load
delivered over $j0.08$ and $j0.2$ pu with a fixed 40.5° angle spread simply
cannot be supported at a healthy voltage. The balance checks in step 6 confirm the
solution is internally exact. In a design study this bus would be flagged for
reactive compensation.