Electricity theft detection and localisation in unknown radial low voltage network

IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308
_______________________________________________________________________________________
Volume: 04 Issue: 07 | July-2015, Available @ http://www.ijret.org 72
ELECTRICITY THEFT DETECTION AND LOCALISATION IN
UNKNOWN RADIAL LOW VOLTAGE NETWORK
Heman Shamachurn1
, Perenjordee Poollay Auroomoogum2
1
Lecturer, Electrical and Electronic Engineering, University of Mauritius, Reduit, Mauritius
2
Undergraduate Student, Electrical and Electronic Engineering, University of Mauritius, Reduit, Mauritius
Abstract
The distribution of electricity involves both technical and non-technical losses. One major cause of non-technical loss is the
illegal abstraction of electricity which is also known as ‘Electricity Theft’. The illegal usage of electricity has many associated
problems, both for utilities and consumers of electricity, implying that there is a pressing need for theft detection and localisation.
Traditional methods of identifying illegal electricity consumers are time consuming and ineffective as measurements have to be
performed at a large number of suspected locations. Smart metering in future electricity networks will lead to a more efficient
automated system for the detection and localisation of electricity theft. This will enable immediate action to be taken by
distribution network operators against the offenders and will help to improve the quality, reliability and security of electricity
supply systems. The aim of this study was to analyse the performance of an electricity theft detection and localisation technique in
an unknown grid. The method assumed the availability of measured voltages, currents, and powers from installed smart meters.
The detection step was a power comparison process and the localisation step was a voltage comparison process. The
investigation involved analysis in the presence of single and simultaneous multiple thefts. To better represent future networks with
increasing penetration of renewable energy generators, distributed generation was added to the system and the capability of the
detection and localisation technique was further explored. All the simulations were performed in Matlab/Simulink. It was found
that the method performed satisfactorily, with a minimum stolen power of 450 W successfully detected and localised.
Keywords: Electricity Theft, Smart Meters, Double Feeding, Low Voltage Radial Network, Matlab/Simulink
--------------------------------------------------------------------***----------------------------------------------------------------------
1. INTRODUCTION
Significant operational losses are involved in the generation,
transmission and distribution of electricity. The losses can
be classified as technical and non-technical. The technical
losses are associated with the components of the power
system and the non-technical losses (NTL) are associated
with external factors which do not directly involve the
power system. In some countries the illegal abstraction of
electricity takes a major proportion of the NTL. For
instance, 1200 GWh of electrical energy is illegally
consumed from the distribution grid each year in the
Netherlands, representing about 1% of the annual electricity
generation [1]. Electricity theft is a main concern for
utilities as the percentage of theft might be small overall, but
the associated financial loss is significant [2].
Several methods are employed to steal electricity including
tampering with the energy meter, bypassing the meter
through double feeding and evading bill payments.
Electricity theft can overload generator units as distribution
network operators (DNOs) cannot forecast the illegal
consumption, and if significant, the electricity supply can be
interrupted due to demand-supply mismatch. Moreover, the
stolen electricity increases grid losses and represents huge
monetary losses both for the DNOs and the Government.
Losses are eventually reflected in the price of electricity,
which implies that genuine customers have a pay a higher
electricity price because of illegal consumers [3].
Currently, tampering attempts are mostly detected by
measurements of electrical parameters in suspected
locations followed by an analysis of the acquired data. The
whole process is time consuming and ineffective, especially
in densely populated areas where there are many houses
very close together and many branches in the distribution
network. Smart meters and state of the art measurement
systems in the future grid will make electricity theft harder
[4]. The installed secure meters at the consumers’ premises
and in the substations will enable automated, fast and
successful electricity theft detection and localisation while
preventing tampering with the meter itself.
2. METHOD DESCRIPTION
2.1 Investigated Network Configuration
A three-phase, four-wire radial distribution system with a
TT earthing arrangement for residential customers was
considered. The model comprised a substation represented
by a 11 kV/ 400 V three-phase transformer and 35 single-
phase household loads as shown in Fig-1. The
corresponding parameters are provided in Table-1. The
distance between the substation and the first connected
house downstream was 100 m.
Each house was represented as a single-phase load
consuming both real and reactive powers. Each load was
modeled by voltage and power controlled current sources
[5]. The following assumptions were considered:

_______________________________________________________________________________________
 Each house was connected to the network via a smart
meter which could record the RMS voltage, the RMS
current, the real and reactive power flows.
 The same measurement capabilities were present at
the substation.
 The active power consumption of each house had a
normal distribution with a mean and a variance of 1
kW.
 The reactive power consumption of each house had a
normal distribution with a mean of 0 VAr and a
variance of 200 VAr.
Fig-1: Simulated network
Electricity theft at a house was simulated by the connection
of an illegal load in parallel to the house, but bypassing the
smart meter as shown in Fig-2.
Fig-2: Illegal load connection
Table-1: Network parameters
Properties Value
Average length between houses 9 m
Cross section of feeder sections 150 mm2
Resistance of feeder 0.206 Ω/km
Inductance of feeder 0.318 mH/km
Length of connection cable between
house and feeder
5 m
Cross section of connection cables 10 mm2
Resistance of connection cable 1.83 Ω/km
Inductance of connection cable 0.402 mH/km
2.2 Electricity Theft Detection and Localization
During a given time step of the simulation, a probable
electricity pilfering attempt was identified by a large
difference between the total active power consumptions of
all customers and the total active power measured at the
substation, taking into account the technical losses
throughout the feeder.
If a theft was detected, the localisation process would
involve comparing the estimated grid voltage at each house
to the actual measured voltage.
The following assumptions were made:
 Order of each house on the feeder was known
 Cable impedances were unknown
 Phase of each house was unknown
The detection and localisation process is summarized in Fig-
3 [6].
Fig-3: Theft detection and localisation for unknown grid
The non linear load flow problem [6] can be approximated
by a linear model to obtain equation (1) where Vh,k is the
voltage at house h at time step k; Vk
0
is the voltage
magnitude at the distribution transformer for the
corresponding phase at time step k; Ph`,k is the active power
of house h` at time step k; Qh`,k is the reactive power of
house h` at time step k; ah,h` is the influence of the active
power of house h` on house h; bh,h` is the influence of the
reactive power of house h` on house h.
Vh,k = Vk
0
+ ah,h`Ph`,k + bh,h`Qh`,k
N
h`=1
N
h`=1
(1)

_______________________________________________________________________________________
The influences ah,h` and bh,h` have to be determined by
solving a linear least squares problem. Considering the
technical losses, the influences are calculated for the time
steps without theft in order to identify the phase of each
house. After identifying the network, the voltage at each
house can be calculated by using equation (1). A large error
between the measured and the estimated voltages for a given
phase implies that a theft has been attempted on that phase.
Normally, for the theft location, the difference is
comparatively greater than other locations.
2.3 Simulations
Different cases were considered for the simulations as
follows:
Case 1: Without Theft
The model was simulated without any theft in the network
so as to compare the measured and expected voltage
profiles. The percentage voltage errors were calculated for
all the houses. The simulation was run for 1000 time steps
representing the different time frames during which data
was collected from the meters.
Case 2: Single Point of Theft on Feeder
An illegal load was connected at house 15, with the active
illegal power varied from 150 W to 3750 W in steps of 150
W for every 40 time steps starting at time step 0.
Case 3: Single Point of Theft on each Phase
Theft was simulated as shown in Table-2.
Table-2: Single point of theft on each phase
House
number
Phase Illegal active power
(kW)
Time steps of
theft
15 C 2 1 to 1000
16 A 1 401 to 1000
17 B 3 601 to 620
Case 4: Two Points of Theft on One Phase
Theft was simulated as shown in Table-3
Table-3: Two points of theft on one phase
House
number
Phase Illegal active power
(kW)
Time steps of
theft
16 A 2 1 to 1000
34 A 3 401 to 900
Case 5: Presence of Distributed Generator (DG) in
the Network
Several distributed generators were connected across the
network as per Table-4.
Table-4: DG connections
House number Phase DG active power (kW)
10 A 1
11 B 2
12 C 5
The effect of the generators on the voltage profiles were
initially analysed without theft. The algorithm was
subsequently tested for thefts of 2 kW and 3 kW at houses
10 and 17 respectively during all the time steps.
3. RESULTS AND DISCUSSIONS
3.1 Case 1: Without Theft
The maximum technical losses in the network was
determined from the power records of all meters and found
to be 3 % on each phase. This value was used as a
benchmark to detect illegal electricity consumption.
Moreover, the influences were determined for all the phases
to identify the houses on each of the three phases. Chart-1
shows the influence of active power of all houses on the
voltage magnitude of house 15. All the houses connected on
the same phase as house 15 can be identified and confirmed
from Fig-1. A similar plot was obtained for the influence of
reactive power.
Chart -1: Influence of active power of all houses on house
15
The percentage voltage errors obtained for all the houses
were very small (between 0 % and 0.008 %) and are
displayed in Chart-2. Expected and measured voltage
profiles for phase B at time step 500 are shown in Chart-3. It
can be observed that the voltage errors were very small.
Similar charts were obtained for the remaining two phases.

_______________________________________________________________________________________
Chart-2: Percentage voltage error for all houses at time step
500
Chart-3: Voltage profiles for phase B at time step 500
3.2 Case 2: Single Point of Theft on Feeder
Taking into account the 3 % technical losses, it was found
that the minimum stolen power detectable was 450 W based
on the increment of 150 W for every 40 time steps. The
voltage and percentage voltage error profiles for time step
81 are shown in Chart-4 and Chart-5 respectively.
Chart-4: Voltage profiles for phase C at time step 81
Chart-4 shows that a theft occurred on the system as the
measured and expected voltage profiles are no longer
superimposed on each other. Chart-5 depicts that the theft
occurred at house 15 as the corresponding percentage
voltage error is maximum. The same conclusions could be
drawn for profiles at time steps greater than 40, but with the
percentage voltage errors being higher due to increasing
stolen power at house 15.
3.3 Case 3: Single Point of Theft on Each Phase
The thefts at all the three houses were detected and located
successfully at the corresponding time steps. The charts used
for analysis were similar to Chart-4 and Chart-5.
Chart-5: Percentage voltage error profile for phase C at
time step 81
3.4 Case 4: Two Points of Theft on One Phase
The algorithm successfully detected and localised all points
of illegal abstraction as shown in Chart-6 and Chart-7.
Chart-6: Voltage profiles for phase A at time step 500

_______________________________________________________________________________________
Chart-7: Percentage voltage error profile for phase A at
time step500
3.5 Case 5: Presence of DG in the Network
The voltage profiles in the absence of electricity theft for
phase C at time step 500 are shown in Chart-8. The peak at
house 12 is due to the relatively high power of 5 kW
injected into the network by the DG of the house. The
voltage errors can be observed to be very small.
Chart-8: Voltage profiles for phase C at time step 500
In the presence of electricity thefts, Chart-9 and Chart-11
clearly point out the presence of theft due to relatively large
errors between the measured and estimated voltages. Chart-
10 and Chart-12 show that the theft locations were
successfully identified.
Chart-9: Voltage profiles for phase A at time step 500
Chart-10: Percentage voltage error profile for phase A at
time step 500
Chart-11: Voltage profiles for phase B at time step 500

_______________________________________________________________________________________
Chart-12: Percentage voltage error profile for phase B at
time step 500
4. CONCLUSION
Smart meters incorporating the tested algorithm will be
efficient in detecting single, multiple multi-phase and
multiple single-phase electricity thefts in a radial low
voltage network, both in the presence and in the absence of
distributed generators, even if the network design
parameters are unknown.
REFERENCES
[1] Frank Van Der Bergh, Petr Kadurek, Sjef Cobben,
and Wil Kling, "ELECTRICITY THEFT
LOCALIZATION BASED ON SMART
METERING," in 21st International Conference on
Electricity Distribution, Frankfurt, 2011, pp. 1-4.
[2] Thomas B Smith, "Electricity theft: a comparative
analysis," Energy Policy, vol. 32, no. 18, pp. 2067-
2076, August 2004.
[3] P. Kadurek, J. Blom, J. F.G. Cobben, and W. L.
Kling, "Theft detection and smart metering practices
and expectations in the Netherlands," in Proceedings
of the Innovative Smart Grid Technologies Europe
Conference, Gothenburg, 2010, pp. 1-6.
[4] R. Alves, P. Casanova, E. Quirogas, O. Ravelo, and
W. Gimenez, "Reduction of non-technical losses by
modernization and updating of measurement
systems," in 2006 IEEE/PES Transmission &
Distribution Conference & Exposition: Latin
America, Caracas, 2006, pp. 1-5.
[5] University of Texas at Austin - Center for
Electromechanics. (2011, September) Simulink
Smartgrid Simulation 1: The Basics. Video
(Youtube).
[6] Sam Weckx, Carlos Gonzalez, Jeroen Tant, Tom De
Rybel, and Johan Driesen, "Parameter identification
of unknown radial grids for theft detection," in
Innovative Smart Grid Technologies (ISGT Europe),
2012 3rd IEEE PES International Conference and
Exhibition on, Berlin, 2012, pp. 1-6.

Electricity theft detection and localisation in unknown radial low voltage network

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Electricity theft detection and localisation in unknown radial low voltage network (20)

More from eSAT Journals (20)

Recently uploaded (20)

Electricity theft detection and localisation in unknown radial low voltage network