A New Under-Frequency Load Shedding Method Using the Voltage Electrical Distance and Artificial Neural Networks

International Journal of Advanced Engineering, Management and Science (IJAEMS) [Vol-5, Issue-3, Mar-2019]
https://dx.doi.org/10.22161/ijaems.5.3.3 ISSN: 2454-1311
www.ijaems.com Page | 171
A New Under-Frequency Load Shedding Method
Using the Voltage Electrical Distance and
Artificial Neural Networks
Le. Trong. Nghia1,*, Quyen. Huy. Anh2, P.T.T. Binh3, Phung. Trieu. Tan4
1,2,4 Faculty of Electrical and Electronics Engineering, HCMC University of Technology and Education, Hochiminh city,
Vietnam
3,*Faculty of Electrical and Electronics Engineering, HCMC University of Technology, Hochiminh city, Vietnam
Abstract—This paper proposes a method for determining
location to shed the load in order to recover the
frequency back to the allowable range. Prioritize
distribution of the load shedding at load bus positions
based on the voltage electrical distance between the
outage generator and the loads. The nearer the load bus
from the outage generator is, the sooner the load bus will
shed and vice versa. Finally, by selecting the rate of
change of generation active power, rate of change of
active power of load, rate of change of frequency, rate of
change of branches active power and rate of change of
voltage in the system as the input to an Artificial Neural
Network, the generators outage, the load shedding bus
are determined in a short period of time to maintain the
stability of the system. With this technique, a large
amount of load shedding could be avoided, hence, saved
from economic losses. The effectiveness of the proposed
method tested on the IEEE 39 Bus New England has
demonstrated the effectiveness of this method.
Keywords—load shedding, Voltage Electrical Distance,
Artificial Neural Network, under-frequency load
shedding.
I. INTRODUCTION
The imbalance active power between the generation and
the load demand causes a decrease the frequency in the
power system. When the balance with the active power
occurs, it’ll give the serious for the frequency instability
of the system. The methods such as primary frequency
control, secondary frequency control, reserve power of
the generation units in the system are only effectiveness
when the system is slightly overloaded and the frequency
of the system is less decrease. But in cases of serious
power imbalance and lead to the blackout completely, the
system need be used load shedding program to recovery
the frequency.
Jianfeng and et al [1] have developed a method with risk
indicators to determine the bus should be targeted for load
shedding to maintain stable voltages. Buses with the
highest voltages risk are prioritized for load shedding.
This is estimated from the probability of the collapse of
the voltage occurring. Risk indicators are the products of
these probabilities and the effects of voltage collapse.
In [2], Hsu and et al presented a strategy of load shedding
by performing artificial neural network (ANN) and
transient stability analysis for an electrical system. To
prepare the training data for ANN, transient stability
analysis of a real power systemhas been made to address
the minimized load with different operating scenarios.
The Levenberg-Marquardt algorithm was combined with
the back propagation algorithm for neural network
training. By choosing the total generating capacity, total
load demand and decay frequency are neural inputs of
ANN, the output is the minimum number of load
shedding that are identified to maintain the stability of the
power system.
In the paper [3], a new approach based on hybrid Particle
Swarm-Based-Simulated Annealing Optimization
technique (PSO-B-SA) is proposed for solving under-
voltage load shedding (UVLS) problem. Under voltage
load shedding (UVLS) is one of the most important tools
for avoiding voltage instability. In this paper, the UVLS
problem is formulated based on the concept of the static
voltage stability margin and its sensitivity at the
maximum loading point or the collapse point. The voltage
stability criterion is modeled
directly into the load-shedding scheme. In any UVLS
scheme finding the global point is very important for
having cost effective economy. The proposed PSO-B-SA
methodology is implemented in the under voltage load
shedding scheme for IEEE 14 and 118 bus test systems.
Simulation
results show the efficacy and advantage of the proposed
scheme.
A good load shedding program must shed the minimum
number of loads as quickly as possible, it also meets all
technical constraints to ensure a stable system.
Conventional load shedding techniques are limited by

load shedding overloading required and slow
performance. Intelligent load shedding methods have also
been studied and developed such as genetic algorithm
(GA) [4] and particle swarm optimization (PSO)
algorithm [5]. These
focus on determining when and how much load should be
disconnected.Studies on the location of load shedding are
very limited.
This paper proposed the method of load shedding based
on voltage electrical distance method and neural
networks. ANN is used to identify and classify load
shedding control strategies based on the designed rules.
The effectiveness of the proposed method has been tested
on the IEEE 39 Bus New England.
II. THE VOLTAGE ELECTRICAL DISTANCE
The voltage electrical distance between the two nodes i
and j is the following formula [6]:
D(i,j) = D(j,i) = −Log(αij ∗ 𝛼𝑗𝑖) (1)
Where:
/
/
i j
ij
j j
V Q
a
V Q
 
 
 ,
/
/
j i
ji
i i
V Q
a
V Q
 
 
 (2)
i
j
V
Q


,
j
j
V
Q


,
j
i
V
Q


, i
i
V
Q


are components extracted
from the inverse matrix of the Jacobian matrix.
The formula ∆Vi = αij∗∆Vj [6] represents the voltage drop
at node i when disturbance occurs at node j. Fromformula
(1), the distance near found, relative with D(i,j) small or αij
as large. On the other hand, the larger the αij, the lower
the voltage drop at node i when the disturbance at j as
large. Thus, when the generator outage, the voltage
fluctuation range near the fault node is large, resulting in
the voltage drop at near nodes also increases, then load
shedding will be at the nearest distance load buses or the
largest voltage drop.
Fig. 1: The block diagram of the relationship between
the generator k and the loads
With: DV(k,1)< DV(k,2) < DV(k,3) < …< DV(k,n)
Prioritized load shedding: Load 1  Load 2  Load 3
 …  Load n
III. PROPOSED METHODS
3.1. Set up load shedding program
Load shedding program are based on three main factors:
the timing, the amount of load to be shed and the location
of load shedding.
The timing: The system data is sent to the control center
for continuous measurement, when the system frequency
is within the allowable range of 59.7Hz<f, the load
shedding program will start, the neuron function will be
activated to identify the generator outage and the load
shedding sequence. The operating time of the UFLS relay
is about 0.1s [7] after the frequency falls below the
allowable threshold and the process is carried out until the
recovery frequency reaches the allowable value. In some
emergency cases, (short circuit, loss of generator) this
method cannot maintain system stability or restore the
frequency with quite long time. Using smart computing
technology,the proposed effective load shedding intervals
require less than 500ms [8]. Here, the proposed load
shedding period is 200ms. This time period includes:
measurement of data acquisition, data transfer, data
processing and tripping trip. However, to ensure safe
margin in real time as well as allowable errors, a period of
100ms [9] is added. So, in the simulation, here the
proposed load shedding time is 300ms.
The amount of load to be shed: After obtaining the load
shedding sequence list for each generator failure, use the
offline PowerWorld simulation software to shed for each
generator in trouble at different load levels from 80% to
100% full load. Dismissed until the frequency of the
buses are within the allowable range of stopping, so that
for each case the incident will have the number of load
shedding corresponding to that case. The incident data
collected would correspond to a number of load shedding
from the trained neuron function.
Location of load shedding: Use voltage electrical distance
for calculating distance between nodes. The load
shedding position will be based on the distances fromthe
generator outage (generator bus) to the remaining load
buses to the load shedding order, or in other words the
priority of the nodes closer to the generator will be first
off, because these load nodes directly affect the generator
is the most trouble. The flowchart load shedding process
is shown in Fig. 2.
3.2. Application ANN to identify the load shedding
Due to the complexity of the power system, the above-
mentioned traditional methods take a lot of time to clear
up, thus causing delays in decision making, The ANN
method is used to solve difficult problems that traditional
methods do not solve in terms of speed and performance.

However, ANN needs to be trained on the basis of initial
data. Therefore, it is necessary to build the set to leran,
including various outages. Data samples representing in
each outage are the change in generator power ΔPG, the
change of frequency in buses Δfbus, the voltage drop of
buses ΔVbus, the change of load ΔPload and the change of
power distribution across the transmission lines ΔPbranch.
During the simulation, various load levels were
considered to cover the operating modes.
Fig. 2: Flowchart load shedding online
During the ANN model identification process, the
creation database of the generator outage is considered
the most important. Reliable databases not only determine
the accuracy of the assessment,but also have a significant
impact on the robustness of the model.
There are two elements that need to be clearly
demonstrated during the simulation:
- The database must cover the operational status and
must adequately represent the various incident scenarios.
- The generated database must ensure the objectivity
of the parameters of the test power system.
The process of creating a database based on simulation
PowerWorld and it is done through the following 5
stages:
Fig. 3: Simulation steps for input, output sampling
The process of creating input database set is shown in
Fig. 4.
Fig. 4: The process of creating input database set
The process of creating output database set is shown in
Fig 5.

Fig. 5: The process of creating output database set
Fig. 6: Neural network training model with inputs and
outputs
IV. TESTING ON THE IEEE 39 BUS NEW
ENGLAND
The proposed method is tested on The IEEE 39 bus New
England, using Power world software to collect samples
and Matlab software for neural network training.
Fig. 7: The IEEE 39 bus New England
4.1. Load shedding programthe neural network
The process created a load shedding program use the
neural network shown in Fig. 8.

Fig. 8: Flowchart simulatorsampling and neural network
training
4.2. Creating database set
Data is available when each generator outage occurs at
different load levels (80% to 100%). This process is done
by off-line simulation using PowerWorld software.
Parameters representing in each case of generator faults
are the change of generator power (ΔPG), the change of
frequency at the bus (Δfbus), the voltage drop at the bus
(ΔVbus), the change in load capacity (ΔPload) and the
change in power distribution across transmission lines
(ΔPbranch).
4.3. Calculate the Voltage Electrical Distance
Voltage electrical distance is physical relationship
between two buses in power system. Voltage electrical
distance can be obtained by following step.
Step 1: Turning all PV generator buses into PQ load
buses.
Step 2: From the matrix Jacobian J, have a matrix J4
from Powerworld, in which J4 = [∂Q/∂V] (3)
Step 3: Inverse J4, call B = J4−1
. Each element of
matrix B is written:
[ ]i
ij
i
V
b
Q



(4)
Step 4: Take the αij decrease matrix, between nodes
i and j, as follows:
ij
ij
jj
b
a
b
 (5)
ij
ji
ii
b
a
b
 (6)
Step 5: Calculate the voltage electrical distance
between nodes i and j calculated according to the formula
(1):
log( )ij ji ij jiD D     
After completing step 5, we obtain table 1
Table.1: Table voltage electrical distance between the generator buses and the load buses.
Bus 30 Bus 32 Bus 33 Bus 34 Bus 35 Bus 36 Bus 37 Bus 38 Bus 39
Bus 3 0.2713 0.6145 0.5104 0.6583 0.4860 0.5908 0.3627 0.5117 0.4162
Bus 4 0.4117 0.4983 0.5665 0.7144 0.5421 0.6469 0.4949 0.6263 0.4479
Bus 7 0.5514 0.5471 0.7253 0.8732 0.7009 0.8057 0.6386 0.7783 0.4446
Bus 8 0.5339 0.5491 0.7153 0.8631 0.6909 0.7957 0.6221 0.7643 0.4053
Bus 12 0.7925 0.6192 0.9137 1.0615 0.8893 0.9940 0.8725 0.9968 0.7824
Bus 15 0.4517 0.6132 0.4053 0.5531 0.3809 0.4856 0.5106 0.5905 0.5427
Bus 16 0.4124 0.6393 0.3183 0.4662 0.2939 0.3987 0.4651 0.5323 0.5243
Bus 18 0.3432 0.6505 0.4561 0.6039 0.4317 0.5364 0.4074 0.4983 0.4790
Bus 20 0.7244 0.9513 0.2636 0.1542 0.6059 0.7107 0.7771 0.8443 0.8363
Bus 21 0.5188 0.7458 0.4248 0.5726 0.2317 0.3679 0.5716 0.6387 0.6308
Bus 23 0.5941 0.8210 0.5000 0.6479 0.2020 0.2170 0.6468 0.7140 0.7060
Bus 24 0.4657 0.6926 0.3716 0.5194 0.3064 0.3967 0.5184 0.5855 0.5776
Bus 25 0.2592 0.7422 0.6092 0.7571 0.5848 0.6896 0.1742 0.4549 0.4567
Bus 26 0.3793 0.7671 0.5769 0.7247 0.5525 0.6573 0.3553 0.2737 0.5449
Bus 27 0.3901 0.7334 0.5179 0.6658 0.4936 0.5983 0.3963 0.3713 0.5405
Bus 28 0.5605 0.9483 0.7581 0.9059 0.7337 0.8384 0.5365 0.1557 0.7261
Bus 29 0.5702 0.9580 0.7678 0.9156 0.7434 0.8482 0.5462 0.0828 0.7359
Bus 39 0.5220 0.8264 0.8426 0.9905 0.8182 0.9230 0.6309 0.8187 0.0000
From Table 1, we build the order of load shedding for each of the generator outage following in Table 2.

Table.2: Proposed load shedding strategy
Bus
Order Bus 30 Bus 32 Bus 33 Bus 34 Bus 35 Bus 36 Bus 37 Bus 38 Bus 39
1 Load 25 Load 4 Load 20 Load 20 Load 23 Load 23 Load 25 Load 29 Load 39
Explanation: According to the suggested strategy table
above, if there is a fault at generator 30, the load shedding
order will be Load 25  Load 3  Load 18  ... until
the system is stabilized again. Similarly, for the remaining
generators.
4.4. Build a learning sample set, identify the input
variable, output variable
The number of sample data received via off-line
simulation
total plus 189 samples (including 21 load from 80% to
100%, per each level have 9 case with 9 generators
outage).
Building a learning template is a file under [samples x
variables]. The samples are rows, and variables are
columns.
Samples data in the learning template under the vector
include the variables to the ∆PG (10), ∆fbus (39), ∆Vbus
(39), ∆Pload (19) và ∆Pbranch (46).
Total the input variables are 153 = (10 + 39 +39 + 19 +
46)
x = [∆PG ∆fbus ∆Vbus ∆Pload ∆Pbranch]
The output variables y is assigned as following: y = [y1
STR1 STR2 STR3 STR4… STRn] = 13
where: y1 is the name of generator outage, STRn is
strategic load shedding
Total results of the strategic load shedding on the Table 3
Table.3: Load shedding strategy for each generator fault
Strategic load shedding Load to be shed
Strategic 1 (STR1) L25
Strategic 2 (STR2) L25, L3
Strategic 4 (STR4) L4, L7, L8
Strategic 10 (STR10) L25, L26, L3, L27
Strategic 11 (STR11) L29, L28, L26, L27
Data consists of 189 samples and split into two subsets:
training data and test data. Training data covers all
generator faults at various load levels as well as covers all
strategies load shedding when the generator faults.
Training data is 85% of the samples (162 samples), test
data is 15% of the samples (27 samples).
4.5. Neural network training
The training of the neural network using the Back
Propagation Neural Network (BPNN) with the Scaled
Conjugate Gradient training algorithm was developed by
Moller [10]. It was designed to reduce computational time
at each step of the search, so the training time is faster and
more optimal. It only needs 4 hidden layers neural and
much less than the Levenberg-Marquardt algorithm.
Meanwhile, the accuracy and error of the network is
equivalent to the training method by Levenberg-

Marquardt algorithm. The structure of the neuron network
is shown in Fig. 9.
Fig. 9: Neural network training structure
Fig. 10: The relationship between the number of neurons
in the hidden layers and the accuracy
4.6. Test load shedding on the IEEE 39 bus New
England by Powerwold.
Process simulation sampling with the IEEE 10 generators
39 bus are made as follows:
Assuming the generator outage is 34, at load level 100%,
the frequency and deviation of the rotor angle become
unstable.
Fig. 11: Diagram of frequency of system at fault
generator 34 load level 100%
When implementing the proposed method, just load
shedding the load on bus 20, the frequency has stabilized.
V. COMPARISON OF METHODLOGY
SUGGESTED WITH OTHER METHODS.
5.1. Load shedding based on under frequency load
shedding relays [11].
Load shedding based on under frequency load shedding
relays is the most commonly used method, which is still
being used in many parts of the world, including
Vietnam. When the grid frequency falls below the
permitted threshold, the relay will be load shedding,
prevents the systemfrequency declines. Without this load
shedding control, the greatest possible consequence is the
blackout and the widespread outage.
For example, the FRCC load shedding program, has a
load shedding plan under frequency load shedding shown
in Table 4.
Table.4: The FRCC load shedding program
UFLS
Steps
Frequency
(Hz)
Time
delay
(s)
Amount
of load (%
of
member
system)
Cumulative
amount of
load (% )
A 59.7 0.28 9 9
B 59.4 0.28 7 16
C 59.1 0.28 7 23
D 58.8 0.28 6 29
E 58.5 0.28 5 34
F 58.2 0.28 7 41
L 59.4 10 5 46
In case of fault at generator 34 the load is 100%, when the
system frequency drops to 59.7Hz, it was load shedding
9% of total load. System frequency continues to fall, and
when dropped to 59.4Hz, load shedding 7% of total load.
The system frequency started to stabilize, the total load
shedding was 16% (9% + 7%).
5.2. Load shedding based on the AHP algorithm
(Analytic Hierarchy Process) [12]
AHP is the approach to making decisions. This method
presents balanced assessment options and criteria, and
integrates them into a final decision. AHP is particularly
suitable for cases involving analysis and quantification,
make decisions when there are multiple options
depending on the criteria with multiple interactions.
Strategic load shedding according to the AHP algorithm
is shown in Table 5.
Table.5: The order of load shedding according to the
AHP algorithm
Order Load
1 L31
2 L12
3 L18
4 L26
5 L23
6 L25
7 L21
8 L28
9 L24
10 L3
11 L16
12 L15
13 L29
14 L27

15 L7
16 L20
17 L8
18 L4
19 L39
In the case of generator 34 outage, the load is 100%, load
shedding according to AHP algorithm will be shed 31-12-
18-26-23-25, the systemis stability.
5.3. Simulation and results
Comparative results of the methods are presented in Table
6.
Table.6: Comparative results of the methods
Method
Amount of
load
shedding
(MW)
Frequency
recovery
time
(s)
Frequency
stability
(Hz)
UFLS relay 975,52 65 60,75
AHP 785,2 47 60,157
Proposed 628 35 60,030
In the case of generators 34 outage with a load level of
100%, the proposed method has many advantages over
the AHP method and the low frequency method, as
follows:
The proposed load shedding method reduced the amount
of load by 347.52 MW (55.3%) compared to the
traditional load shedding method and 157.2 MW (25%)
compared to the load shedding method based on the AHP
algorithm.
The proposed load shedding method had a frequency
recovery time of 30 seconds (85.7%) compared to
traditional load shedding and 12 seconds faster (34.3%)
than load shedding based on the AHP algorithm. The
simulation result is shown in Fig. 12.
0 10 20 30 40 50 60 70 80 90 100
59.4
59.6
59.8
60
60.2
60.4
60.6
60.8
Time (s)
Frequency(Hz)
UFLS
AHP
Voltage Electrical
Distance
Fig. 12: Diagram of system frequency after generator 34
outage of all three methods of load shedding
0 10 20 30 40 50 60 70 80 90 100
1.01
1.02
1.03
1.04
1.05
1.06
1.07
1.08
1.09
1.1
Time (s)
Volt(V) UFLS
AHP
VOLTAGE ELECTRICAL DISTANCE
Nomal
Fig. 13: Diagram of voltage after generator 34 outage of
all three methods of load shedding
Table.7: Comparison of recovery voltages at 19 bus loads when 34 generators were lost for all three methods

In the case of generators 34 outage with a load level of
100%, the proposed method has many advantages over
the AHP method and the low frequency method, as
follows:
The proposed load shedding method has a voltage
recovery time of 131.6% compared to conventional
method and 39.5% compared to the AHP algorithm.
The proposed load shedding method has better recovering
voltage values than traditional methods and AHP
algorithm methods as follows: the change in voltage
variation compared to the voltage before the fault of
proposed load shedding method is 0.17%, the traditional
method is 3.65% and the AHP algorithm method is
2.27%.
VI. CONCLUSION
The paper proposed a load shedding scheme with priority
based on the voltage electrical distance between the
generator outage and the load nodes to ensure system
stability in the event of a severe load imbalance in the
event loss a generator occurring in the electric system.
The effectiveness of the proposed load shedding program
has been demonstrated by analyzing the simulation results
of the IEEE 39 system, 10 generators. The results show
that: the method of load shedding proposed to reduce the
amount of load shedding, frequency recovery time, faster
voltage, and better recovery voltage values than
traditional methods and algorithm AHP.
ACKNOWLEDGEMENTS
This research was supported by Ho Chi Minh City
University of Technology and Education, and Ho Chi
Minh City University of Technology.
REFERENCES
[1] Jianfeng, D., Z. Shuangxi, and L. Zongxiang (2005),
“A New Risk Indices Based Under voltage load
shedding scheme” Department of Electrical
Engineering.
[2] Hsu, C.T., M.S. Kang, and C.S. Chen (2005),
“Design of adaptive load shedding by artificial neural
networks,” IEE Proceedings - Generation,
Transmission and Distribution, vol. 152, pp. 415.
[3] Sadati, N., T. Amraee, and A.M. Ranjbar (2009), “A
global Particle Swarm-Based-Simulated Annealing
Optimization technique for under-voltage load
shedding problem”. Applied Soft Computing, vol. 9,
pp. 652-657.
[4] Ying-Yi Hong, and Po-Hsuang Chen (2012),
“Genetic-Based Underfrequency Load Shedding in a
Stand-Alone Power System Considering Fuzzy
Loads,” IEEE transactions on power delivery, vol.
27, pp.87-95.
[5] Abbas Ketabi, and Masoud Hajiakbari Fini (2017),
“Adaptive underfrequency load shedding using
particle swarm optimization algorithm,” Journal of
Applied Research and Technology, vol. 15, pp.54-60.
[6] Lagonotte Patrick (August 1990), “The different
electrical distance”. In Proceedings of the Tenth
Power Systems Computation Conference, Graz,
Austria, pp. 542-550.
[7] Hooshmand, R. and M. Moazzami (2012), “Optimal
design of adaptive under frequency load shedding
using artificial neural networks in isolated power
system,” International Journal of Electrical Power &
Energy Systems, vol. 42, pp. 220-228.
[8] Bevrani, H., G. Ledwich, and J.J. Ford (2009), “On
the use of df/dt in power systememergency control,”
in 2009 IEEE/PES Power Systems Conference and
Exposition.
[9] Shekari, T., F. Aminifar, and M. Sanaye-Pasand
(2016), “An Analytical Adaptive Load Shedding
Scheme Against Severe Combinational
Disturbances”. IEEE Transactions on Power
Systems, vol. 31, pp. 4135-4143.
[10] Moller (1993), “M.F.: A Scaled Conjugate Gradient
Algorithm For Fast Supervised Learning,” Neural
Networks 6, pp. 525–533.
[11] FRCC (2008), FRCC System Disturbance and
Underfrequency Load Shedding Event Report.
[12] Y.C. Erensal, T. O¨ zcan, M.L. Demircan (2006),
“Determining key capabilities in technology
management using fuzzy analytic hierarchy process,
A case study of Turkey,” Information Sciences 176,
pp. 2755–2770.

A New Under-Frequency Load Shedding Method Using the Voltage Electrical Distance and Artificial Neural Networks

More Related Content

What's hot (19)

Similar to A New Under-Frequency Load Shedding Method Using the Voltage Electrical Distance and Artificial Neural Networks (20)

Recently uploaded (20)

A New Under-Frequency Load Shedding Method Using the Voltage Electrical Distance and Artificial Neural Networks