An Adaptive Overcurrent Protection Scheme for Microgrids based

on Real Time Digital Simulation

Zhang Kailun, Dhivya Sampath Kumar, Dipti Srinivasan Anurag Sharma

State Grid Chengdu ECE department Electrical and Power Engineering,
Electric Power Company, China NUS, Singapore NU International Singapore
[email protected] [email protected] [email protected]

Abstract— A microgrid can operate in grid-connected as with distributed generation, has brought a huge change in
well as islanded mode using the power available from increasing the reliability of the system [3]. However, there are
distributed generation (DG) sources. As the mode of certain challenges as well which we need to address. As the
operation changes for the microgrid, the protection microgrid operates in grid-connected as well as islanded
settings, i.e. current settings and time settings, for the mode, the protection settings need to be adjusted according to
relays need to be altered according to the change in the the change in the network configuration [4]. It is also noted
direction and fault current level. Moreover, the protection that inverter interfaced distributed generation (IIDG) takes up
settings should also be updated considering the stochastic a huge proportion of DGs used in microgrids. The fault
nature of the power available from the DGs. Hence, in this current from these IIDGs is limited and minimal, which may
paper, an adaptive overcurrent protection relay scheme is not be sensed by the relays. Thus, will require a change in the
proposed for microgrid protection. The proposed scheme protection settings calculation. Mutual effect between high-
is tested on a modified CIGRE test micro-grid simulated frequency operation of power electronic devices and energy
in real-time digital simulator (RTDS) for various case storage element in microgrid also amplifies the harmonic
studies. Results obtained show that the proposed adaptive current, which may cause false operation of the protection
protection scheme can operate fast and accurately devices’ [5]. Moreover, IIDGs generating power using
according to the changes in the network. renewable energy sources like solar and wind are intermittent
and stochastic in nature. Thus, when they are connected to the
Index Terms—Overcurrent protection; distributed grid, the protection settings cannot be a fixed value, and need
generation; micro-grid; real-time digital simulator (RTDS） to be adaptive according to the change in the current
magnitude and direction.
I. INTRODUCTION There are different solutions proposed in the literature to

W ith the increasing demand of power and environmental

concerns, there arises an urgency to utilize renewable
sources of energy to generate power. With characteristics
address the aforementioned issues. Some of the research
efforts have suggested the use of Fault current limiters (FCLs)
[6-7], while in [8-10] distance protection is suggested as a
such as low pollution and high energy-use efficiency, solution to the protection issues. In [11] a scheme is proposed
distributed generation (DG) sources can solve this problem. where DG connection status and the level of fault are two
According to the Distributed Power Coalition of America deciding factors for relay settings of overcurrent protection.
(DPCA), DG can be defined as any small-scale power In [12], the authors have considered the islanded mode of
generation technology that provides electric power at a site operation for the microgrid in the presence of distributed
closer to customers in addition to the central generation energy resources. The authors have presented in [12] software
sources. In addition to operating as power supplement to main as well as hardware implementation in the laboratory to verify
grid, distributed generators can also work as an independent their proposed concept. The authors in [13] highlight the
power unit. When a fault happens and the main grid fails to assignment of two parameters, i.e. fault current coefficient
provide enough power, DG sources can supply power to and relay hierarchy, which are critical for proper operation of
important loads. Such an ability of DGs led to the evolution a microgrid protection system. In [14], the authors proposed
of microgrids in the form of a localized power grid that can an adaptive protection scheme that accommodates for the
operate independently or in conjunction with the utility grid varying fault current levels and network scenarios. However,
[1, 2]. Thus, a microgrid can be termed as a small-scale power in the aforementioned research works [6-14] real time
system, which consists of various distributed energy simulation is not implemented to validate the protection
resources and loads. schemes. In recent years, more and more relay manufacturers
This change in the structure of power network, from are using real time digital simulator (RTDS) as the main test
centralized generation sources to decentralized microgrids means of research and validation [15, 16, 17]. In this work,

This work is supported by Ministry of Education, Singapore (R-263-

000-C48-114)

978-1-5386-4950-3/17/$31.00 ©2017 IEEE

we extend our previous work [14] and complete real-time
modeling and simulation of a benchmark micro-grid in the
RTDS real-time simulation environment.
The organization of the paper is as follows: Section II will
give a brief introduction of the proposed overcurrent
protection scheme. In Section III, the modelling of the
microgrid in RTDS will be presented. Section IV will
highlight the case studies simulated in this work to
demonstrate the ability of the proposed approach. In Section
V, the conclusions are drawn.
II. OVERCURRENT PROTECTION SCHEME Fig. 1. Range of phase angle
A. Setting and Sampling
Basic overcurrent relays can operate with or without delay,
and it can operate for any given direction of current. General overcurrent relay to be selective when current measured by
operating characteristics of overcurrent relays are: definite instrument transformer is above the pickup level and has
time, moderately inverse, inverse, very inverse, extremely exceeded the time required to operate. There are various
inverse. It is noted that the various inverse-time characteristic methods for directional control for overcurrent relays, and in
curves provide one of the most basic forms of protection used this paper, 90-60 degree method for phase fault detection is
to protect power system components. In this paper, inverse adopted [19].
time overcurrent protection is adopted as a protection scheme Generally, during fault condition, the phase angle will be
for microgrids along with a directional element. In in the range of 5 to 15 degrees lagging for faults with low fault
conventional relays, the trip time is given as, voltage, and for faults with high voltage, it will be between 75
to 85 degrees lagging [20]. The protection strategy used in this
  paper assumes the angle of maximum torque to vary. The
 
 A  angle where maximum torque occurs is defined as 90 degrees.
t p = TMS ×  B + P  (1) From Fig. 1, it can be observed that when the operating
  I  
   −1  quantity is leading the polarizing quantity by 30 degrees, the
  IP   expected maximum torque is produced, and the polarizing
where I is the fault current, and IP is the pick-up current. TMS quantity being a voltage. Therefore, the actual current that
is a multiplier added to the overall time required to operate; A flows because of the applied voltage will be 60 to 70 degrees
and B are two constants depending on types of operating lagging.
characteristic; p is a variable exponent for the ratio of applied
C. Fuzzy Logic Decision Making (FLDM) Module
current to pick-up current. Reset time of the relay is given as,
Tr For conventional directional overcurrent relay protection,
tr = 2
(2) the pick-up current is a fixed value determined by calculating
 I  fault current values using short circuit analysis. However, due
  −1
 IP  to the inclusion of DGs, the level of fault current does not
where I is the fault current, and IP is the pick-up current, Tr is remain constant, and thus, the pick-up current needs to be
a time multiplier setting. updated according to the change in network situations. A
To calculate pick-up current, instrument transformers decision-making module is proposed to determine the value
measure voltage and current signals in the distribution system, of pick-up current whenever there is a change in power
and then signals are processed for fundamental phasor network, which is based on fuzzy logic in this paper.
estimation. The magnitude and phase angle is measured using The whole process of FLDM scheme [21] includes four
a technique [18] involving two samples of a waveforms taken main steps: fuzzification, inference, aggregation, and
at discrete time intervals. The root mean square value is defuzzification. The output power of DG “P(t)” and its
calculated using (3) while the phase angle between two terminal voltage “V(t)” signals obtained previously are taken
waveforms is determined using (4). as inputs to the fuzzy module. Then the two input values are
fuzzified using their corresponding triangular membership
V 2 + Vk2+1 − 2VkVk +1 cos ω0 Δt
V2 = k (3) functions μ ( P ) and μ (V ) , which is ideal for point-to-point
sin ω0 Δt 2
connection in the fuzzification process. The output is the pick-
I kVk + I k +1Vk +1 − ( I kVk +1 + I k +1Vk ) cos ω0 Δt up current in the network.
θ = cos −1 (4)
IV sin ω0 Δt 2 For each input and output, there are six membership
functions with variables set according to the value of the per
B. Directional Control unit value of input, namely, very low (VL), low (L), low-
medium (LM), medium (M), high (H) and very high (VH), as
The sensing of current flow direction allows the basic
 Fig. 2. Input and output membership functions

Fig. 4. Micro-grid network

TABLE I.
IMPEDANCE DATA OF LINES IN THE NETWORK
Rph Xph Rneutral R0 X0
Line Type (Ω/k (Ω/k (Ω/km (Ω/km (Ω/k
m) m) ) ) m)
Fig. 3. Closed-loop scheme of FLDM OL - Twisted
1 cable 4x120 0.28 0.08 1.14 0.42
mm2 Al
shown in Fig. 2. It is noted that the input-output relationship SC - 4x6 mm2
2 Cu
3.69 0.09 13.6 0.47
among the fuzzy inputs are mapped with the help of a defined
2
SC - 4x16 mm
rule base of 36 rules [14]. The rules are determined based on 3 Cu
1.38 0.08 5.52 0.42
the concept that the current setting decreases gradually with OL - Twisted
the decrease in power provided by the DG for the respective 4
cable 3x70 mm2
0.50 0.08 0.63 2.39 0.45
Al + 54.6 mm2
terminal voltage. The complete closed-loop scheme of the AAAC
proposed FLDM-based relay is as shown in Fig. 3. It is noted 5
SC - 3x50 mm2
0.82 0.08 0.52 2.04 0.42
Al + 35 mm2 Cu
that the adaptive protection scheme proposed by [14] was SC - 4x25 mm2
developed on MATLAB Simulink. In this work, it is migrated 6 Cu
0.87 0.08 3.48 0.41

to RSCAD before simulations on RTDS could be carried out. OL: Overhead line, UL: Underground line, SC: Service connection.
As, RSCAD only supports import of control components, thus
kV feeder and 13.2 kV feeder. The capacity of
only the control component of the adaptive protection scheme
transformer is 25 MVA and its type is Dy11. The
(FLDM module) can be migrated to RSCAD, and the rest of
neutral point of its secondary winding is grounded with
the adaptive protection scheme is developed from scratch.
a 0.1Ω resistance.
3. Transmission line: There are six types of transmission
III. MICROGRID MODELLING IN RTDS
lines in the micro-grid, all of which are in 3-phase 4-
RTDS provides a large number of simulation models for wire system. Detailed parameters are given in Table I.
power system, including transmission line, transformer, 4. Load: There are four loads in the micro-grid. Critical
generator, dynamic load and so on. Real time simulation of load (always connected) is load 4, located at bus 4.
micro-grid stresses on the transient process of various fast 5. DG resources: A PV system is assumed on bus 8,
changes in networks and the system response during a short with a capacity of 1.74 MW. Real time simulation of
time range. Thus, detailed modelling of power system micro-grid stresses on the transient process of
components is necessary. In this paper, an adaptation of various fast changes in networks and the system
CIGRE benchmark network, shown in Fig. 4, is simulated as response during a short time range. Thus detailed
the test microgrid in RTDS [22, 23]. The details of the modelling of PV is necessary. In this paper, VdcQ
parameters is shown as follows: control [24] is adopted for inverter, and the value of
1. Frequency: 60 Hz. reference voltage for DC voltage control is obtained
2. Transformer: A step-down transformer connects 138 from maximum power point tracking (MPPT)
 protection scheme are benchmarked with conventional
protection system for validation.
(a) (A) Case Study 1. Variation of Insolation
To demonstrate the adaptive nature of the proposed
protection scheme, in this case study the pick-up current
(b) (B) setting of the relay is determined with variation of insolation
of the sun. It is noted that insolation is proportional to the
output current of PV system, which affects the output power.
(c) (C)
Therefore, by changing the value of G, the pickup current
Fig. 5. Per unit value of terminal voltage, DG power, and pickup settings determined by FLDM control component can be
current setting: for Conventional protection system (a) G = observed. The results as shown in Fig. 5 highlight that as the
2000 W/m2. (b) G = 1500 W/m2 (c) G = 1000 W/m2 and isolation decreases from 2000 W/m2, the pick-up current
proposed adaptive protection settings (A) G = 2000 W/m2. (B)
settings remain constant for the conventional protection
G = 1500 W/m2 (C) G = 1000 W/m2
scheme. However, for the proposed protection scheme, the
pick-up current adapts to the change in isolation (power from
the PV) and decreases from 0.6504 A to 0.1961 A as the
isolation reduces from 2000 W/m2 to 1000 W/m2.
Case Study 2. Operation in grid-connected mode
Assuming that the PV system operates for 1000 W/m2
insolation and 25 degrees centigrade, a ‘Phase a-ground’ fault
is created at bus 8 that lasts for 0.2 seconds. In grid-connected
mode, the operation of two kinds of relays, conventional as
well as proposed adaptive, is observed.
(a) During the whole process, trip signal (TRIP) and reset
signal (KTRST) of relay can be observed as shown in Fig. 6.
For conventional protection system, because of improper
current setting, the relay does not operate when a fault
happens, so there is no trip signal but always a reset signal.
For the proposed fuzzy logic based adaptive protection, the
reset signal disappears once there is a fault, while the relay
trips after a short delay to prevent malfunction. After the fault
ends, reset signal can be detected again. It is noted that the
pick-up current setting for the grid-connected mode is 0.5453
(b) pu as shown in Fig. 7. The trip time and reset time is observed
Fig. 6. Operation signal grid-connected mode. (a) conventional to be 0.1374 seconds and 0.0437 seconds as shown in Fig. 8.
protection system. (b) adaptive overcurrent protection system Case Study 3. Operation in islanded mode
The same fault conditions as assumed in Case Study 2
are simulated in this case study to test the sensitivity of the
adaptive relay protection in islanded mode. However, the PV
system operates under the condition of 2000 W/m2 insolation
Fig. 7 Per unit value of terminal voltage, DG power, and pickup and 25 degrees centigrade. It is noted that in order to simulate
current setting: for grid-connected mode.
the islanding mode, the recloser switch, marked as main in fig.
4, is opened.
The results as shown in Fig. 9 demonstrate that the
proposed adaptive protection scheme successfully senses the
change of current level in the grid and a different pickup
Fig. 8. Operation time in grid-connected mode. (a) trip time. (b)
current setting is obtained. It is decreased to 0.1525 pu as
reset time.
compared to the 0.5453 pu for the grid-connected mode as
algorithm [25]. shown in Fig. 10 and Fig. 7, respectively. This is due to the
reduction in fault current level in islanded mode. The trip time
IV. SIMULATION RESULTS AND ANALYSIS and reset time is observed to be 0.1831 seconds and 0.0937
Extensive case studies are simulated in RTDS to seconds as shown in Fig. 11.
demonstrate the effectiveness of the proposed protection
scheme and show its operation process. However, due to
space constraint only three case studies are presented in this
paper. It is noted that the results of the proposed adaptive
 1, 2009. CIRED 2009. 20th International Conference and Exhibition
on. IET, 2009.
[6] A. Agheli, H. A. Abyaneh, R. M. Chabanloo, H. H. Dezaki, "Reducing
the impact of DG in distribution networks protection using fault current
limiters", Proc.4th Int. Power Eng. Optimization Conf., pp. 298-303,
2010.
[7] G. Tang, M. R. Iravani, "Application of a fault current limiter to
minimize distributed generation impact on coordinated relay
protection", Int. Conf. Power Syst. Transients, 2005.
[8] M. Baran, I. El-Markabi, "Adaptive over current protection for
distribution feeders with distributed generators", Proc. IEEE Power
Fig. 9. Operation signal in islanded mode. Eng. Soc. Power Syst. Conf. Expo., pp. 715-719, 2004.
[9] W. El-khattam, T. S. Sidhu, "Resolving the impact of distributed
renewable generation on directional overcurrent relay coordination: A
case study", IET Renew. Power Gen., vol. 3, pp. 415-425, 2009.
[10] V. P. Mahadanaarachchi and R. Ramakuma, "Impact of distributed
Fig. 10. Per unit value of terminal voltage, DG power, and
generation on distance protection performance - A review," 2008 IEEE
pickup current setting in islanded mode. Power and Energy Society General Meeting - Conversion and Delivery
of Electrical Energy in the 21st Century, Pittsburgh, PA, 2008, pp. 1-7.
[11] W. El-khattam, T. S. Sidhu, "Resolving the impact of distributed
renewable generation on directional overcurrent relay coordination: A
case study", IET Renew. Power Gen., vol. 3, pp. 415-425, 2009.
[12] N. Schaefer, T. Degner, A. Shustov, T. Keil, J. Jaeger, "Adaptive
Fig. 11. Operation time in islanded mode. (a) trip time. (b)
protection system for distribution networks with distributed energy
reset time.
resources", 10th IET Int. Conf. Develop. Power Syst. Protect., 2010.
[13] T. S. Ustun, C. Ozansoy and A. Ustun, "Fault current coefficient and
time delay assignment for microgrid protection system with central
V. CONCLUSION protection unit," in IEEE Transactions on Power Systems, vol. 28, no.
Due to increasing DG penetration level in micro-grids, 2, pp. 598-606, May 2013.
many difficulties for relay protection appear. The fault current [14] Kumar, Dhivya Sampath, D. Srinivasan, and T. Reindl." A Fast and
varies continuously because of the intermittent and Scalable Protection Scheme for Distribution Networks With
Distributed Generation." IEEE Transactions on Power Delivery
randomness brought by DG, which requires efficient tracking
31.1(2016):67-75.
of power system signals and logical decision making so that [15] S. R. Lee, J. y. Yoon, J. h. Kim, B. Yang and B. Lee, "Protective Relay
the pick-up current setting can be dynamic and effective. Real Tests of Hybrid SFCLs in a Korean Distribution Power System Using
time simulation shows its superiority in protection analysis. RTDS," in IEEE Transactions on Applied Superconductivity, vol. 21,
This paper models a micro-grid with an adaptive protection no. 3, pp. 2188-2192, June 2011.
scheme based on fuzzy logic decision-making in RTDS for [16] B. Wang, X. Dong, Z. Bo and A. Perks, "RTDS Environment
real time simulation. The protection scheme uses the output Development of Ultra-High-Voltage Power System and Relay
Protection Test," in IEEE Transactions on Power Delivery, vol. 23, no.
power of DG and its terminal voltage as inputs, to determine
2, pp. 618-623, April 2008.
the value of pick-up current. [17] H. Muda and P. Jena, "Superimposed Adaptive Sequence Current
Results of simulation have proved that this protection Based Microgrid Protection: A New Technique," in IEEE Transactions
scheme has the capability of fast signal estimations and online on Power Delivery, vol. 32, no. 2, pp. 757-767, April 2017.
protection settings in a distributed network with DGs i.e. [18] Johns, A.T., Digital Protection for Power Systems, IEE Power Series
microgrids. 15, Peter Peregrins Ltd., Six Hills Way, Stevenage, UK, 1995.
[19] Blackburn, J. Lewis, and Thomas J. Domin. Protective relaying:
principles and applications. CRC press, 2015.
References
[20] Mason, C. Russell. The art and science of protective relaying. Vol. 195.
[1] Thomas Ackermann, Göran Andersson, Lennart Söder, “Distributed Wiley, 1956.
generation: a definition,” Electric Power Systems Research, Volume [21] H. Cheung, A. Hamlyn, Y. Cungang, R. Cheung, "Network-based
57, Issue 3, 20 April 2001, Pages 195-204. adaptive protection strategy for feeders with distributed generations",
[2] D. Jin; Z. Li; C. Hannon; C. Chen; J. Wang; M. Shahidehpour; C. W. Proc. IEEE Canada Elect. Power Conf., pp. 514-519, 2007.
Lee, "Towards a Cyber Resilient and Secure Microgrid Using Software- [22] Papathanassiou S ， Hatziargyriou N ， Strunz k. A benchmark low
Defined Networking," in IEEE Transactions on Smart Grid , in press.
voltage microgrid network[C]//CIGRE Symposium. Greece：2005.
[3] S. Conti, S. A. Rizzo, E. F. El-Saadany, M. Essam and Y. M. Atwa,
[23] Rudion K，Styczynski Z A，Hatziargyriou N，et al. Development of
"Reliability Assessment of Distribution Systems Considering
Telecontrolled Switches and Microgrids," in IEEE Transactions on benchmarks for low and medium voltage distribution networks with
Power Systems, vol. 29, no. 2, pp. 598-607, March 2014. high penetration of dispersed generation[C]//International Symposium
[4] B. Kroposki et al., “Making microgrids work,” IEEE Power and Energy on Modern Electric Power Systems. Wroclaw，Poland：2006.
Mag., vol. 6, pp. 40–53, 2008. [24] R. Khezzar, M. Zereg and A. Khezzar, "Comparative Study of
[5] D'Adamo, Christian, Samuel Jupe, and Chad Abbey. "Global survey on Mathematical Methods for Parameters Calculation of Current-Voltage
planning and operation of active distribution networks-Update of Characteristics of Photovoltaic Module," in International Conference
CIGRE C6. 11 working group activities." Electricity Distribution-Part on Electrical and Electronics Engineering 2009.
[25] G. Farivar, B. Asaei, S. Mehrnami, “An Analytical Solution for
Tracking Photovoltaic Module MPP,” IEEE Journal of Photovoltaics,
vol.3, no.3, July 2013.