EE 455

Introduction to Energy Distribution Systems

Lecture Notes # 11

Prof. S. S. Venkata

Copy Right 2001 S. S. Venkata. No part of the notes may be

reproduced in any form without the consent of the author.

Load Modeling
Under steady-state, complex power S at any location in a distribution
system is a function of voltage, V
S = VI* = P + jQ = f ( V )

Model I
It is usual to represent both P & Q as a polynomial function of |V|. That
is
P = ao + a1|V| + a2|V|2 + a-1|V|-1 +

(1)

Q = bo + b1|V| +b2|V|2 +

(2)

Such a representation is valid for individual type of loads or aggregate

(composite) type of loads.
In equations (1) & (2)
(i) if only ao and bo are only non-zero, then with all other a i s and b i s
being zero,
P = a0

Q = b0

CONSTANT POWER

MODEL

(ii) if only a1 and b1 are non-zero,

P = a1 | V|

Q = b1 | V|

CONSTANT CURRENT MODEL

(iii) if only a2 and b2 are non-zero,

P = a 2 | V|2

2
Q = b 2 | V|

CONSTANT IMPEDANCE MODEL

Example [1]
(a) Air-conditioning load demand (per unit values)
P = 2.97

2
4.00 |V| + 2.02 |V|

26.8 |V| + 14.90 |V2

|

Q = 12.90

(b) Flourescent lighting

P = 2.18 + 0.286 |V| - 1.45 |V|-1
Q = 6.31

2
15.60 |V| + 10.3 |V|

(c) Induction motor

P = 0.720 + 0.109 |V| + 0.172 |V|-1
Q = 2.80 + 1.63 |V| - 7.6 |V|2 + 4.89 |V|3
These models are usually arrived at by conducting a series of tests and
performing suitable regression analyses:

Model II
Composite loads which are assumed to be mixtures of the types
discussed above can be represented as
| V|
P = Pn

| Vn |

| V|
Q = Qn

| Vn |

, k varies between 0 to 3

(3)

, l varies between 0 to 3

(4)

Examples [2]
(a) If k = 1, l = 0, it implies the load is a constant current type, with
power factor unity.
(b) If k = l = 2, then the load is constant impedance type.
(c) If k = 2.5 and k = 2.7, then it represents an aluminum reduction
plant. This model is simpler and can be determined empirically knowing
|Vn| = initial or base value of voltage
Pn = initial or base value of real power
Qn = initial or base value of reactive power
Note: In all these models, one should be aware of the range of |V| for
which the equations (1) to (4) are valid.

Model III
Loads are seldom modeled to include frequency effects. If loads (or
demands) are sensitive to frequency effects, then these effects should be
included: In this case

| V|
P( f ,| V| ) = Pn

n | Vn |

| V|
Q( f ,| V| ) = Q n

n | Vn |
where,

(5)

= 2( f )
n = 2( f n )
f n = base frequency

(6)

, = constant exponents.
With this type of model, it is often a practice to determine four sensitivity
coefficients defined below:
dP
df

dP
d| V|

dQ
df

, and

dQ
d| V|

Knowing these coefficients, the new values for P and Q can be

determined from f and |V|. All the changes are assumed to be
small: Thus
P
Q =

(dP / df ) (dp / d| V| ) f
( dQ / df ) ( dQ / d| V| ) | V|

Then: P = Pn + P
Q = Qn + Q
for

f = fn + f

and

|V| = |Vn| + |Vn|

The following tables show the results of a data survey conducted by EPRI
[3].

Model IV [4]
This is particularly suitable for modeling aggregate loads at a node
knowing the demand profiles for a day, a season or a year.
Assuming Gaussian Distribution
Pp = Pn + k p

(7)

Q p = Q n + k p

(8)

where, Pp = Power having an excess probability by p %

kp = Coefficients related to p
= Standard Deviation
The power values in excess probability around 10 % are used in volt-drop
calculations; The mean values (50% excess probability) are used for loss
calculations. Smaller values of (p%) are used for overload and
emergency conditions determination.

References
[1] R. B. Adler, C. C. Mosher,
Power Loads

Steady-state Power Characteristics for

, Stability of Large Electric Power Systems, IEEE Press,

1974, (This book has a series of papers on modeling.) pp. 147-153.

[2] M. H. Kent et-al.,

Dynamic Modeling of Loads in the Stability

Studies, ibid, pp. 139-146.

[3] EPRI Project Report,

Long Term Power System Dynamics

, Vol. I,

EPRI RP. 90-7-0, June 1974.

[4] E. Lakervi & E. J. Holmes,
(Book)

Electricity Distribution network Design

, Peter Peregrinus td., 1989.

L

Demand:

Demand is the load averaged over a specified interval of time and is

expressed in kilowatts, kilovoltamperes, amperes or other suitable units applicable to the

load commodity. Demand interval is the period over which the load is averaged.
Figure 1 shows how demand varies with time and demand interval at it occurs.
This is a daily demand curve. Note, the demand interval could be 15, 30, or 60 minutes.
Maximum demand is the greatest of all demands which have occurred during the specified
period of time. It must state demand interval and period such as daily, monthly, or annual.
Load Duration Curve: Figure 2 shows the demand of a particular value cumulatively
occurring during the interval versus duration. It is a monotonically decreasing function as
to be naturally expected and is derived from Figure 1. Diversified or coincident demand is
the demand of a composite group of loads. The maximum diversified demand or
Maximum Coincident Demand, Dm (henceforth called maximum demand) is usually the
value of which is of most importance for related loads. Non-coincident demand is the sum
of the demands of a group of loads with no restrictions on the interval to which each
demand is applicable. Here also the maximum non-coincident demand, DM is of interest
for unrelated loads.
Demand factor, FD, is the ratio of the maximum demand of a system to the total
connected load, LC (or sum of continuous ratings) of the systems.
Utilization factor is the ratio of maximum demand (DM) of a system to the rated
capacity of the system, which is the smaller of thermal or voltage drop capacity.
Load factor, FLD is the ratio of average load, Davg over a designated period of time
to the peak load (or maximum demand) occurring during the period. Load factor varies
between 0 and 1.

Diversity factor, FDV is the ratio of the sum of individual maximum demands of
various subdivisions of a system to the maximum demand of the whole system. It is
greater than or equal to 1. This means it is the ratio of maximum non-coincident demand
to maximum coincident demand.
Coincidence factor, FC is the ratio of the maximum coincident total demand of a
group of consumers to the sum of the maximum power demands of individual consumers
comprising the group both taken at the same point of supply for the same times.
Coincidence factor is the reciprocal of the diversity factor.
Load diversity, LD is the difference between the sum of the peaks of two or more
individual loads and the peak of the combined load.
(1)

Diversity factor = FDV =

D m1 + D m 2 + + D mN =

D mi = D m

Dm

Fc

i =1

D m1 + D m 2 + + D mN =

D mi = Dm

Dm

Fc

i =1

(22)
where N = Number of loads connected
1
1
FDV FDV

(2)

Coincidence factor = FC =

(3)

Load diversity = LD = DM - Dm

(23)
(24)

The maximum demand of the group of (1+2+3+. . . . +N) loads can also be expressed as:
Dm = c1D1 + c2D2 + c3D3 + . . . . . . .+ cNDN

(25)

where cN =

the contributing factor of the Nth load which is defined as the contribution
of the nth load to the group maximum demand. It is equivalent to

hourly

listed in Table 2.

variation factor

With the above definition, the coincidence factor can be expressed as

D + c D +L L + c N DN
Fc = c1 1 2 2

(26)

D1 + D2 +L L + DN

If D1 = D2 = . . . . . . = DN then
Fc =

D1 ( c1 + c2 +L + c N )
ND1

c1 + c2 +L + c N
N

(True for ranges)

(27)

(True for air conditioners)

(28)

If c1 = c2 = . . . . . . = cN then
Fc =

c1 ( D1 + D2 +L + DN )
= c1
D1 + D2 +L + DN

Loss Factor:

Ratio of average power loss to peak load power loss during a specified

period of time.
Referring to Figure 3
x = peak load of duration t
y = minimum load of duration (T - t)
Average load =

xt + y (T t )
T

Load factor = FLD =

If y = 0, FLD =

xt + y (T t )
Tx

(a)
(b)

(29)

t
T

Peak loss = x2R

for duration t

(a)

Minimum loss = y2R for duration (T-t) where R is a constant

(b)

x 2 Rt + y 2 R ( T t )
T

(c)

Average loss =

(30)

Rx 2 t + y 2 R ( T t ) t y T t
= +
= FLS =
RTx 2
T x T
2

Loss factor

(d)

If y = 0 and x 0, FLS = FLD

If

T t
1. 0
T

t
2
y
0 and x 0 FLS (FLD)
T

This means loss factor varies between FLD and (FLD)2

Figure 4 shows load and loss factors as functions of peak load duration for load cycle of
Figure 2.
Figure 5 has curves of FLS as a function of FLD.
Empirical formula: FLS = 0.3 FLD + 0.7 F2LD for 30-minute monthly demand curve. In
general, this can be expressed as:
FLS = (1-A)FLD + AF2LD
The loss factor is primarily used in loss evaluation studies.
Power factor: Power factor of a load

= Fp
= cos ( - ) = cos

(33)

= Active Power, P
r
Apparent Power, S
In equation (33), = arc tan

Q
. Here again this is true only for pure sinusoidal qualities.
P

The above definition is not generally applicable to a distributed load or a group of

individual loads which are continually changing. In this case, such as light load or peak
load in which case the loads are expressed in demand. The concept of group power factor
is a complex subject. It may be more desirable to determine the average power factor,
rather than the power factor at a particular load condition. Such is the case in considering

industrial and commercial services, the contracts of which might contain power factor
clauses.
Balanced load: A balanced polyphase load is that which draws symmetrical currents when
symmetrical voltages are applied to it.
Balanced voltages = symmetrical polyphase voltages.
Voltage unbalance = maximum deviation from the average base phase voltage/average
phase voltage.
(34)

or V2 =
V1

negative-sequence voltage
positive-sequence voltage

(35)

This could easily be achieved more easily with the aid of computers.
The second definition is widely used. We will get more into it when we discuss the
subject of symmetrical components and overcurrent protection.
Load Distribution and Load Density
2 methods of representing distributed load
a) uniform distribution
b) loads are all same fraction of distribution transformer rating
2 methods of load density representation
a) linear density - kVA/1000
b) area density - MVA/mi2
Determining Load Characteristics
1) Measurements
2) Statistics (using a few spot measurements)
The greatest interest is in residential loads as opposed to commercial or industrial.

Methods of Making Field Measurements

Instruments include

ammeter (rms & thermal lag)

wattmeter
demand meter
recording - disk chart, strip chart, paper tape, magnetic
tape.

Means of measurement: spot check, maximum demand test, recording demand.

Estimating Load Characteristics
1) Diversified Demand Method (covered before)
2) Relationship between maximum demand and KWH of compensation.

The second method uses factors to estimate maximum demand for each consumer.
The factors are:
0.0039 for 0-500 kWh
0.0035 for 500-1500 kWh
0.0030 for > 1500 kWh
Coincidence or Diversity Factor
1 Fc

FcN

Fc

FcN

coincidence factor for

Fc

coincidence factor for

FcN

DN
D1

DN

maximum diversified demand for load

(51)

loads
loads
(52)

D1

average individual maximum demand

Fc

nFcn1
n 1

(53)

DN

D1FcN

(54)

Motor Starting Characteristics

Motor starting causes voltage dips. Starting currents are roughly 6 to 8 times
corresponding to full-load components.

EXAMPLE:

DISTRIBUTION TRANSFORMER SERVING RESIDENTIAL LOADS:

(REF: Electric Power Distribution Engineering by Gonen)

Let

s consider the concept of Maximum Diversified Demand (Coincident Demand) for

residential loads. The objective is to estimate distribution transformer loads serving

residential areas. Figure 7 shows typical maximum diversified demands for various loads.
Table 1 shows hourly variation factors for these load types on an hourly basis.
Let

s also assume that a given distribution tran

sformer is serving 6 houses.

Usually such transformers serve 4 to 14 houses depending upon the type of heating and
transformer sizes. Further, let us consider that each house has a clothes dryer, a range, a
refrigerator, lighting and miscellaneous loads. We want to estimate the 30-minute
maximum diversified demand as seen by the transformer. We are also given that the
maximum coincident demand seen by each of these load types for each house is (see
Figure 7)
Dryer

1.6 kW

Range:

0.8 kW

distribution transformer

Refrigerator: 0.0666 kW
Lighting:

6 similar houses

0.61 kW

Hence, the 30-min. maximum coincident demand

=

6(1.6 + 0.8 + 0.66 + 0.61)

(2) Suppose the primary feeder has 150 DT

18.5 kW

s serving 900 houses. Again from the above

curve,

1.2 kW/house
0.53 kW/ house

range

0.52 kW/house

refrigerator

0.044 kW/house

Pav, max =

dryer

lighting, etc.

hence, 30-min max diversified demand

=

(1.2 + 0.53 + 0.52 + 0.044) x 900 = 2064.6 kW

(3) Hourly variation factor of a particular load

=

demand (of a particular load) coincident with group demand

max diversified demand (of the particular type of load)
At 4pm,

clothes dryer
12am
1 am
2 am
3 am
4 am
5 am
6 am
7 am
8 am
9 am

0.03
0.02
0.00
0.00
0.00
0.00
0.00
0.00
0.08
0.20

Dryer (0.38 x 1.6 x 6) kW

Dryer 0.38

1.6 x 6 kW

Range 0.24

0.8 x 6 kW

Refrigerator 0.90 x

0.066 x 6 kW

lighting 0.32

0.61 x 6 kW

10 am
11 am
12 noon
1 pm
2 pm
3 pm
4 pm

0.65
1.00
0.98
0.70
0.65
0.63
0.38

Table 1
Hourly variation factors

Refer to (1), 6 house case

6.344 kW = total hourly diversified demand

(contribution of the particular type of load to group max demand)

Rate Structures
1.

Flat Demand Rate: Does not require meters. Used for street lighting.

2.

Straight-Line Meter Method: Single Price based on kWh consumed. Most

residential customers are charged on this basis.

3.

Blocked-Meter Rate: Graduated structure, in which higher the consumption lower

the rate applied. Not suitable for conservation.

4.

Demand Rate: Recognizes load factor and involves a two-part tariff, one based on
maximum (monthly) demand and the second one based on actual energy
consumed.

5.

Seasonal Rate: Summer versus Winter Rate. Here in the Northwest we pay
higher rate for winter and lower rate for summer. (Why?)

6.

Time-Of-Day (TOD) Rate: Recognizes on-peak versus off-peak duration

consumption. The philosophy used here is similar to that for seasonal rates.

Load Growth
P0(1 + g)n

Pn

Pn:

load forecast at nth year

(load forecasting)

g:

annual growth rate

n:

# of years

P0 :

initial load

Pn

P0
n
0

5. Rate Structure
Sample:
Min change (first 20 kWh)
Next 80 kWh

$0.0355/kWh

Next 100 kWh

$0.0321/kWh

Next 200 kWh

$0.0296/kWh

Next 400 kWh

$0.0265/kWh

Consumption in excess of 800 kWh

Ex:

$2.25/mo.

$0.0220/kWh

Consumption 2200kWh
2.25 + 80 x 0.0355 + 100 x 0.0321 + 200 x 0.0296 + 400 x 0.0265 + 1400 x 0.022
=

55.62

Environmental surcharges:
County tax, fuel cost adjustment, state sales tax
Total: $ 85.42
*rate schedule may be seasonal.
Northwest higher rate in Winter higher cost due to power import from California.
Fuel Cost Adjustment (factor FCAF)
FCAF =

$/kWh

Btu
kWh

B
10 6

1
1 D

$
Btu

A:

weighted average Btu/kWh

B:

amount by which average cost of fuel per million Btu exceeds or is

less than $1/million Btu.

C:

ratio of total net thermal generation to total net generation

KWH x

KWH x

1
:
total generation to meet the given demand
1 D
loss factor
1
C :
1 D

total thermal generated to meet the given

demand