DISTRIBUTION NETWORK LOSS CALCULATION IN A DEREGULATED

ELECTRICITY SUPPLY INDUSTRY

Innocent E. Davidson Darin Naidoo

Department of Electrical and Electronic Engineering Eskom Distribution
University of Natal, Durban 4041, South Africa Eastern Region
Tel. (031) 260 1246, Fax (031) 260 1300 South Africa
E-mail: [email protected]. Email: [email protected]

Abstract Reform in the electricity supply industry is a global phenomena that can be attributed to some critical factors, such as:
private ownership as opposed to state ownership, consumers demand for price transparency and regulation of cross-subsidies,
the introduction of competition in the industry as a means of providing lower energy prices to end-users, the need for higher
energy efficiency in technical systems of utilities, significant growth in energy demand especially in developing countries, and
pressure on governments to increase funding for social services. A practical model is unbundling generation, transmission and
distribution into separate companies with an independent system operator. Economic dispatch is an important consideration in
power network performance and electricity pricing mechanisms. Lower technical losses will provide for cheaper electricity and
lower production costs, with a positive influence on economic growth. This paper presents an approach to technical loss
computation in distribution and reticulation systems, results of technical loss estimation in South Africa’s grid network, and
discusses loss minimization and power network performance under a deregulated electricity supply industry.

Key words: Technical losses, distribution networks.

I. INTRODUCTION the conductors and transformers along the delivery route. To

maximize revenue, these losses need to be minimized. Losses
Regulatory, economic and environmental challenges are are estimated to add 6 - 8 percent to the cost of electricity and
shaping a different world for electric utility operations in both some 25 percent to the cost of delivery [3]. The accurate
developed and developing nations. The market liberalization estimation of technical and non-technical losses will enable
and industry privatization initiative in the communications regional electricity distributors (RED) determine with greater
and energy sectors is a positive economic policy, which is not accuracy the operating costs for maintaining electricity
unique to South Africa, but a global phenomenon. supply to consumers. This in turn enables a more accurate
The structure of deregulated electricity industries across estimate of the system lifetime costs, over the expected life of
the world shows unbundling of generation, transmission and the installation [4].
distribution with an independent system operator (ISO) [1] as
the emerging standard. South Africa’s proposed structure will Hitherto, utilities operating as state monopoly institutions
see the ISO and transmission company function as a business in rigid structure have overlooked these losses since they do
entity [2]. Hitherto, electric utilities focused on construction not constitute major operational or power quality problems.
costs of new transmission and distribution (T&D) systems, or However, their impact is economic, and in a market-driven
providing grid extensions to emerging communities. deregulated environment, loss minimization has become
To maintain profitability, electric utilities have narrowed increasingly important. Distribution and reticulation system
down their strategic goals to: reducing cost of electricity to losses are both technical and non-technical.
customers, ensuring high environmental and occupational Non-technical losses (NTL) are more dominant in lower
health standards, improving service quality and better safety levels of distribution networks. They include unauthorized
measures for personnel and equipment. From the distribution line tapping, equipment vandalization, metering errors,
and reticulation systems perspective, the challenge is how to inadequacies/inaccuracies of meter reading, inaccurate
manage these at the lowest operating and maintenance costs; customer billing and inaccurate estimation of non-metered
maximize the life-span of existing plant and equipment; supplies, such as public lighting. Technical losses include line
invest in appropriate technologies, operate optimally with losses, copper resistance and iron losses of transformers. The
reduced manpower, how to locally implement deregulation of technical losses in the network constitute an economic loss.
the electricity industry both effectively and efficiently, and NTL are construed as a loss of revenue by the utility.
how to meet these constraints while enhancing system
reliability and efficiency. Technical losses include: copper, induction/radiation and
dielectric and losses [5]. Transformer losses include copper
II. LOSSES IN DISTRIBUTION SYSTEMS losses due to the internal impedance of transformer coils and
core loss. Distribution transformers have no-load losses to be
The supply of electricity to customers leads to significant considered. Some factors, which influence distribution losses
losses along lines. Some of the input energy is dissipated in include voltage regulation, magnitude of circulating currents,
power factor and phase balancing: From the perspective of where m and n indicate independent summations to include
losses, this is of significance when dealing with heavily all generators. B is the loss coefficients or B-coefficients.
loaded lines. The objective would to balance the phase loads
so that the maximum deviation from the average is below The B matrix takes into account the system parameters
10% [6]. such as impedances and var generation. Transmission and
distribution losses impact both the cost of production of
III. ECONOMIC DISPATCH & MODEL OF LOSS energy as well as the cost of electricity delivery. Reducing
MINIZATION IN DISTRIBUTED SYSTEMS this cost will lead to significant savings on the cost of energy
production. Lower technical losses will provide for cheaper
For specified system load, economic dispatch determines electricity, lower production costs with positive influence on
the power output of each plant that will minimize the overall economic growth. It will have the most impact during peak
fuel cost. Both the economic dispatch and minimum loss loads conditions, as it will reduce the peak demand thereby
problems are complex to compute manually and are usually reducing energy requirement of the system at the critical
solved with the aid of digital computers [7,8]. The first case stage when the cost of generation is usually the most
is the economic distribution of load between generating units expensive. A deregulated energy market pricing structure
within the same plant. The minimum cost of fuel is achieved will be expected to reflect this loss factor in spot prices.
when the incremental costs are minimal [9]. In a power system total generation equals total demand
plus line losses [10].
  aT PGT  bT …………….(1)
Pi
G   PD   PL …….….…(4)
k n
1
 K 1
where, aT   i1  From the above it can be assumed that total generation
 ai  revenue will equal total customer revenue plus operating cost
of the transmission and distribution network.
 K b 
bT   i 1 i 
 ai 
C
i
GR   PD * C SP   PL * C XL ….…(5)
k n

PgT  i 1 Pgi
K
where CGR is generation revenue, PD is consumer demand,
CSP is spot price paid by consumer, PL is line losses, and CXL
 is the total increment fuel cost of the plant, PGT and PGi is operational cost of transmission line [10].
represent the total plant output and the output of individual
generating units respectively. ai and bi are the constants that PL C XL T  CGR T  PD * CSP T ….....(6)
describe how the incremental fuel costs varies with unit
output. dC GR d PD * C SP 
  k ….…….(7)
Transmission losses are not considered in the above case, dPi dPi
because the load was shared by generating units in the same where  is marginal transmission cost, Pi is power at node I
plant. However when distributing load between various and  is the Lagrangian multiplier for maximum capacity of
power stations, transmission losses have to be factored in to line k.
determine most economical operation since power plants are Line loss is a function of power flow and system
in different locations. Using a penalty factor allocated to each constraints. From the above we can minimize the following
plant by the equation [9]: equation [9]:

PL   Pm Bmn Pn
1
Li  …………...……. (2)
PL
1 m n

PGi
subject to the constraints:
Li is the penalty factor of the plant, PGi is the power generated  min     max
by the plant, and PL is the system losses [9] defined as:
PD min   PD  PD max 
PL   Pm Bmn Pn ….………….(3) PGi min   PGi  PGi max 
m n
QGi min   QGi  QGi max 
 PGi is active power at node i, QGi is the reactive power at From these figures [6] it is easy to estimate the impact of
node i. losses on natural resources (coal, water) and the environment,
and conversely, the economic gains and pollution reduction.
IV. DISTRIBUTION SYSTEM LOSS ESTIMATION
Environmental Typical Impact due to
The total system loss is given by the difference between Measure Impact Losses
the energy generated or delivered and the energy sold. The (10 800 GWh)
energy used in power station or substation auxiliaries is Water use 1.25 KL/MWh 13 500 ML
deducted from the losses to obtain the system losses. Ash emission 0.37 Kg/MWh 4 Kt
Coal use 480 Kg/MWh 5.18 Mt
 100 …(8)
Energy Delivered - Energy Sold CO2 output 900 Kg/MWh 9.72 Mt
System loss % 
Energy Delivered SOx output 7.4 Kg/MWh 80 Kt
NOx output 3.7 Kg/MWh 40 Kt
The value obtained from (8) includes both technical losses
and a component of non-technical losses. NTL cannot be Table 2: Environmental Impact of Losses
computed easily, but can be estimated from preliminary
results, i.e. the result of technical losses are first computed The peak demand losses (MW) also has significant
and subtracted from the total losses, with the balance as NTL. financial implications, because peaking stations (for top
The technical losses can be calculated using appropriate load looping) are required to meet the increase in demand, which
flow software. Hitherto, many utilities arbitrarily select are usually more expensive to run than base load stations.
estimated values for losses, because of the accuracy of Therefore small reductions in network losses will amount to
available data and the simulation time required. However, significant financial savings to utilities as well as customers.
software tools that allows load profiles to be considered will
reduce simulation time and ensure accurate calculation of V. METHOD OF LOSS CALCULATION
technical losses for various conditions. For South Africa, the
losses are [6]: for transmission, 2.5%-4.5% of generation; In Eskom (South African Power Utility) operations, the
distribution: 3%-5% of energy supplied at main transmission distribution system is made up of subtransmission (132kV-
substation, MTS; others: 0.5%-1% of energy supplied at MTS 33kV) and reticulation (22kV - 11kV). The load is measured
Distribution losses include line and transformer losses in at each point of common coupling between transmission and
the 132 to 33kV networks. Other losses include line and distribution. The copper losses of the transformer are
transformer losses in networks with voltages  22kV and calculated assuming R = 4% of X. (R – copper resistance, X –
non-technical losses. By considering the above estimates the leakage reactance). The procedure is listed below [6]:
cost of distribution technical losses to Eskom Eastern region
for the month of January 2000 can be estimated [6].  Obtain loading data at MTS substations
 Calculate Load Factor (LF), where
Energy Purchases Energy Losses
Cost of
 100% ……..(9)
Total Energy Delivered
Distribution Losses
Peak LF 
Rands Technical 7.4 Peak Load x hours in period
GWH c/kWh Demand
(millions) Losses c/kWh
Loss
(GWh) (Rands
mil)  Apply the peak load to scale the base case in PSS/E
2860. 210.85 7.4 143.04 10.58 238MW (power system simulation load flow program)
99 (5% of (trans-  Run PSS/E to calculate the Peak load loss (Ppl) in
Purchases) mission
MW.
losses
included)  Determine the correct  and  factors
 These have been found (empirically) to be  = 0.3
Table 1: Distribution losses January 2000 and  = 0.7
 Calculate the approximate power loss (Apl)
The cost of the distribution losses in the region for the
month was estimated to be R10.58 million. This figure is
based on the average energy purchase price of 7.4c/kWh. In

Apl  Ppl *  * LF   * LF 
2

1998 Eskom’s gross energy sent out was 180 000 GWh and
the peak demand was 27.8 GW. Assuming 6% technical  Calculate the approximate energy loss in GWh
losses in the South African power system, the energy losses  Distribution technical loss (DTL) = Apl x hours in
can be estimated to be 10800 GWh. At peak demand, accounting period.
1668MW is used to meet network losses.  Calculate iron losses of transformers using
developed loss tables.
  Add all zones together for total technical loss. Electricity Regulator to the South African Institute of
 To obtain non-technical losses this figure is Electrical Engineers, 8th February 2001.
subtracted from the total system loss. [3] Tobin N and Sheil N, “Managing to reduce power
transmission system losses”, Transmission
Performance, Publication of Electricity Supply Board
VI. DISCUSSION AND CONCLUSION International, Dublin, Ireland 1987.
[4] Sellick R.L. and Gaunt C.T. “Load Data Preparation
To evaluate the impact of system losses, the magnitude of for Losses Estimation”. Proceedings of the 7th
these losses needs to be determined. As shown, the total Southern African Universities Power Engineering
system losses can easily be determined. The losses are then Conference (SAUPEC), Stellenbosch, South Africa,
divided into technical and non-technical losses. The next step Vol. 7, pp 117-120, 1998.
is to assess the continuing cost of losses, usually in annual [5] Dorf, R.C., `The Electrical Engineering Handbook’,
terms. The suitable method for costing losses should reflect CRC Press Inc., Boca Raton, 1993.
cost of production, in terms of generating capacity, fuel [6] Krishnaswamy N, “A holistic approach to practical
consumption and reflect the cost of capacity in the techniques to analysis and reduce technical losses on
transmission and distribution networks. These losses should Eskom power networks using the principles of
be recognized as real and substantial costs on the supply of management of technologies”, Technology Leadership
electricity and should be managed like other costs, and Program, Management of Technology Dissertation,
reflected in pricing methodologies on the spot market in a April 2000.
deregulated structure. [7] Paul C.R., “System loss in a Metropolitan utility
Loss management can be approached from supply side by network”, IEE Power Engineering Journal, pp. 305-
providing generation and upstream network capacity and 307, September 1987.
consuming fuel to supply them. Alternatively we can reduce [8] Miller R.H. and Malinowski J.H., “Power System
losses by demand side option. By using this option one not Operation”, 3rd Edition, McGraw-Hill Book Company,
only reduces the generation and system capacity but also New York, 1970.
reduces the cost of electricity delivery, which is a major [9] Grainger J.J. and Stevenson W.D (Jr.), “Power System
contributor to the final cost of electricity. This option also Analysis”, McGraw-Hill, New Jersey, 1994.
improves system efficiency. [10] Innocent E. Davidson, Abimbola Odubiyi, Michael O.
Kachienga and Brian Manhire, “Technical Loss
Some techniques of loss reduction include optimal Computation and Economic Dispatch Model in
positioning of normally open points to distribute load as Transmission & Distribution Systems in a Deregulated
evenly as possible and proper phase balancing. Losses ESI”, Research Paper, University of Natal, Durban,
generate substantial cost and are a significant issue in power October 2001.
system management. Loss evaluation is complex but critical
in assessing the overall performance of a power network.
Reducing electricity costs to the customer and minimizing
loss of high quality energy are worthy objectives for the
industry and this will be a challenge in the proposed
deregulated electricity supply industry.

VII. ACKNOWLEDGEMENTS

The University of Natal Research Fund (URF) and Dasela

Engineering Consultants have supported this work.

REFERENCES

[1] Davidson I.E. and Odubiyi A., “Power System

Operation in Developing Economies – The Nigeria
Experience (Part 1)”, `Energy supply and management
feature’, South Africa Electricity + Control Journal,
July 2001.
[2] Mkhwanazi, Xolani, ‘The Changing Regulatory
Environment during Electricity Supply Industry
Reform’, Speech given by the CEO of the National