Modelling and Simulation of Reverse Power Relay for Loss of Mains Protection of Distributed Generation in Microgrids

C. Buque1, S. Chowdhury2, S.P. Chowdhury3 Department of Electrical Engineering

University of Cape Town Cape Town, South Africa claudio.buque@mottmac.com1, sunetra.chowdhury@uct.ac.za2, sp.chowdhury@uct.ac.za3
Abstract Integration of distributed generation (DG) into the utility grid has led to a renewed emphasis on looking into novel power system control and protection issues pertaining to DG units. This paper focuses on loss of mains (LOM) detection and protection for DG. Commonly used methods of detection fail to effectively detect the loss of mains scenario when the local area network load and generation are closely matched. The proposed method of detection and protection is highly efficient when the demand and supply are similar. For this reason it can be used together with the present methods to provide a complete solution to LOM protection. Index TermsReverse Power Relaying, Loss of Mains, Distributed Generation, Point of Common Coupling.

The developed relay is installed on the utility side of the point of common coupling (PCC). This enables it to detect the LOM occurrence and signal the static switch (SS), which is a power electronics device at the PCC. Once the SS is open the microgrid can safely operate in isolated mode. In this mode the DG has enough capacity to support the local load with minimal negative effect on voltage levels and frequency stability. II. LOSS OF MAINS Loss of mains can occur due to faults in the utility grid, maintenance in the system or even circuit breaker nuisance trip. This event may have many negative effects on the distributed generator and microgrid. The real challenge faced by the system operators comes in when the microgrid has a small distributed generator. This can cause voltage dips and system instability. As a systems engineer it is important to determine at design stage whether the DG is suitable to support the utility load or not. In some areas for example, the UK, the DG is by no means allowed to support the utility load under a loss of mains scenario [2]. This is due to the lack of the reliability of the grid when operating with a DG as its only source and the difficulties encountered when attempting to reconnect the system. Loss of mains, for whatever reason it may occur is a severe situation which design engineers must take into account, especially in terms of system control and protection during the occurrence. A. Loss of Mains Detection and Protection Over the years there have been many detection techniques developed, each of these have their own advantages and disadvantages. Local detection means that long distance communication is not required. The detection device is located close to the

I.

INTRODUCTION

Loss of mains occurs when the microgrid, formed by the distributed generator and local load, become disconnected to the utility source of electrical power but remain connected to part of the utility load [1][9]. Loss of mains is less apparent in terms of system stability when the DG is large enough to support the utility load to which it is connected, however it can become a big challenge to system operators if the DG does not have enough capacity to support the utility load. Consequences related to LOM include frequency instability and voltage dips. There is no industry standard method for LOM detection and protection, various literatures present different methods of detection ranging from active methods, passive methods and remote techniques. In this paper a passive method (reverse power relaying) is explored in order to provide a single and complete solution for loss of mains protection for microgrids. This relaying system demonstrates the benefits of active methods including speed of operation as well as avoiding the difficulty presented in passive methods when establishing detection thresholds.
The authors are grateful to the authority of the Electrical Engineering Department, University of Cape Town for providing the support and infrastructure necessary for carrying out this research.

978-1-4799-1303-9/13/$31.00 2013 IEEE

switch. There are two types of local detection methods: Active and Passive method. In the active method the detection device functions by directly interacting with the system under consideration. Perturbations are purposely injected to the system. The systems response to the perturbations determines whether loss of mains has occurred or not. Examples of active techniques are [3]: Reactive power export error detection Impedance measurement method Slip-mode frequency shift method Active Frequency Drift method Automatic phase-shift method Active methods of detection have the disadvantage of directly affecting the system to which they are applied. If the injected signal is not adjusted correctly it may have a significant effect on the magnitude of the frequency of the voltage, current and power output of the DG, it can break the power balance between the DG and the local loads. However, it has a significant advantage in that it is cheaper than passive means of detection. In passive techniques there is no added signal to the system. This technique relies on the detection of certain distinct patterns at the DG output when islanding occurs. A difficulty faced when dealing with passive techniques is the proper selection of detection thresholds. This technique depends vastly on load condition. When there are local balanced loads LOM detection becomes difficult. Examples of these techniques are [3]: Rate of change of output power Rate of change of frequency (ROCOF) Rate of change of frequency over power Harmonics Detection Reverse Power Relaying Lastly there are remote techniques for LOM detection. These are based on communication systems between the utility and the DG. These are expensive to implement, especially when small DGs are involved. Examples of remote detection include: Power Line carrier communications Supervisory Control and Data Acquisition systems When implementing Passive techniques of detection, the DG protection against LOM is provided for by the relays used to detect LOM. Once they detect an occurrence they send a signal to open the appropriate Circuit Breaker (CB) and from those pre-determined actions, the DG and Microgrids are guaranteed protection. When implementing remote techniques, the communication links can be used to communicate with the PCC or the Static Switch to disconnect the microgrid from the utility grid when LOM occurs. The communications systems are usually reliable enough to provide effective signaling. However, when active detection techniques are implemented depending on the infrastructure, not always are appropriate protection mechanisms activated. Detectors will

merely warn that LOM has occurred, they will not necessarily open any breakers or disconnect the microgrid at the point of common coupling. An effective way of making use of the information provided by the detector is to send it to relays located at the microgrid. Based on pre-determined settings, the relays can activate the SS at the PCC to open and leave the microgrid to operate in isolated mode. When the generation and load in a network area are closely matched, it becomes difficult to detect a loss of grid supply at the generator [4]. III. POWER SYSTEM MODEL

The power system model, as shown in Fig. 1, consists of a utility source with impedance representing the estimated impedance for a utility grid of that capacity. There are two busbars with a reverse power relay between them. Once the relay detects an abnormality the Static Switch is opened and the microgrid becomes isolated. A normal installation would have different types of relays triggering the various circuit breakers due to certain events. For the purposes of this paper only 1 relay is used so that its capabilities can be clearly demonstrated. The System characteristics are shown in Table 1. Due to the configuration of the power system, the sizes of the generators and loads, it is expected that active power will only flow from busbar A to busbar B. If active power flows from busbar B to busbar A it is considered abnormal.

Fig. 1. Power System Model TABLE 1 POWER SYSTEM CHARACTERISTICS System Voltage (ph-ph) 132kV Nominal Frequency 50 Hz Utility Source Capacity 100 MVA DG Capacity 10MVA/ 10KVA Utility Load 100MVA Local Load 8MVA

IV.

REVERSE POWER RELAY

Reverse power protection is generally applied to prevent damage to mechanical plant items in the event of failure of the prime mover. Common generator damage includes gearbox damage and mechanical damage to shafts. Power relays currently used in industry are capable of measuring system voltage and current. They also measure the angle between these two signals, the angle . With this information they can calculate real power which is as shown in Equation 1.

P = V I cos

(1)

Under normal operation, real power flow is in a determined forward direction and -90 < < 90, in the case of reverse power flow 90 < < 270. The relay allows for unidirectional flow of active power from A to B. V. REVERSE POWER RELAY MODEL

generator that is primarily being protected. In this case -0.01 was chosen as the lower limit because this is the lowest integral value during normal system operation. For other systems this value would have to be chosen based on modelling results as it depends on the size of the generation capacity and the system loads. B. Definition of Trip Time Reverse power relays with either built in timers or external timers must be used to avoid spurious isolation under transient reversal of power, which may arise following synchronisation or in the event of a power transmission system disturbance. The bigger the generating capacity, the lower the time delay should be. For loss of mains detection the delay time will essentially be defined by breaker reclosing times. When using high speed auto-reclosing, it is important to know the time for which the line must be de-energised in order to allow complete de-ionization of the arc, so that it will not strike when the voltage is re-applied. The de-ionization time depends on various factors, of these factors circuit voltage is the most important. Utility reclosers usually reclose after 0.17s for a 132kV system as shown in Table 2. For this reason loss of mains relays must activate before the circuit breakers attempt to reclose and reconnect the utility power source, this is to avoid reclosing on to unsynchronised systems.
TABLE 2 ARC DE-ENERGISATION TIMES FOR DIFFERENT VOLTAGE LEVELS [8]

The reverse element of the relay is the most important in identifying an abnormal condition, therefore it is the most important in modelling the relay. In this paper the directional component of the relay is adopted from [6], where a similar relay is used for generator protection. In this model, as depicted in Fig. 2, voltage and current signals are modified to square waves, with maximum and minimum values of 1 and -1. When the signal is positive, the value is 1 and when the signal is negative it is represented by -1. The two signals are then multiplied to give an output of 1 when the signals overlap and -1 when they dont. The product is then integrated from 0 to L. The upper limit of the integrator is 0 so that under normal power flow conditions the integral remains less than 0. However, under reversed power flow conditions the integral output tends to fall until it reaches a threshold value of L. The value of L varies according to the allowable reverse power, the higher the value of L, the higher the amount of reverse power. Once there is a reverse power flow in the location where the relay is situated the relay identifies the abnormal condition and can immediately react to it. A time delay component is incorporated in the relay. This ensures that the circuit breaker only trips if a prolonged fault or abnormal event occurs. The relays shouldnt trip for transient power swings. The last 4 elements in the model are used to ensure a logic 0 is sent to the CB when an abnormality is detected so that the CB can open. A. Definition of Reverse Power Level When reverse power relaying is used to protect a generator a relay setting is chosen based on the type of generator it is. An example of this is protection of a Diesel Engine the allowable motoring power is 5 to 25% of the rated generation capacity. The relay setting is chosen based on Equation 2 [8]:
Setting = Motoring Power(%) x Generating Capacity (MVA) CTratio x VTratio
(2)

Transmission Line Voltage (kV)

66 110 132 220 275

Minimum deenergisation time

0.1 0.15 0.17 0.28 0.3

For Loss of mains detection the method of determining allowable reverse power flow is different because it is not the
1 Constant 1 I_in -1 Constant1 1 2 V_in Constant2 -1 Constant3 Switch1 Product Switch 1 s Integrator
if(u1 < -0.01 ) u1 else

The trip signal time delay has to be long enough to not trip for transient cases but fast enough to open the circuit breaker before the recloser activates an unsynchronised circuit to be reconnected. It is assumed that [7], at medium level voltage it takes a relay 2 cycles to pick up the fault and the circuit breaker another 3 to 5 cycles to open. Therefore, in addition to the set trip time, it takes from 5 to 7 cycles (0.1s to 0.14s in a 50Hz system) from the time the fault occurs to the time the fault is cleared. Based on the preceding argument the trip time delay should be: 0.1s < t < 0.17s. The value chosen for t in this paper is 0.14s, which corresponds to 7 cycles in a 50Hz system.

1 Constant6
In1 if { } Out 1

If 1 Constant4

If Action Subsystem

To

-1 Add1 Gain

1 Trip

0.06 Variable Constant5 Time Delay

Fig. 2. Reverse Power Relay Matlab Model

VI.

TEST CASES

A. Test Case 1: Loss of Mains due to open utility switch, large DG Capacity. In Test Case 1, CB1 is opened after 2 seconds. This represents a scenario where there is maloperation of a switch or nuisance tripping. It causes reverse power flow into the utility load from the 10MVA microgrid generator. It is important that the reverse power relay detects this scenario because it will prevent unsynchronized reconnection of the utility power source with the microgrid. In this case the reverse power relay is expected to cause the static switch at the point of common coupling to open. B. Test Case 2: Loss of Mains due to a 3-phase Fault,with large DG Capacity. 3-phase faults are the most severe faults in terms of the levels of fault current. It is important to provide a protection system that can handle such high currents. In this case the Utility Source is lost due to a 3-phase fault between CB1 and busbar A. After the fault has occurred, power will flow from the DG towards the fault. This power in the reverse direction should be detected by the reverse power relay and the fault is expected to be cleared by the relay and switch combination in between busbar A and busbar B. C. Test Case 3: Loss of Mains due to open utility switch, with Small DG Capacity. This case is similar to test Case 1 but here it is expected that there will be the least active power flow which might make it more difficult for the reverse power relay to detect the abnormality. This test is included in order to test the sensitivity of the reverse power relay. Again the switch CB1 will be opened at 2 seconds. This case will demonstrate the impact on the size of the DG with regards to reverse power relaying for loss of mains. D. Test Case 4: Loss of Mains due to a 3-phase Fault, with Small DG Capacity. When it comes to reverse power flow with small DG, detection is the main issue and sensitivity of the relay model is of paramount importance. The reverse current is so small that it makes it difficult to detect the reverse active power. In this case the three phase fault will cause a relatively high reverse power flow and the relay is expected to detect it. In all of the test cases the relay setting is the same. The relay plays the vital role of detecting reverse active power. For all the tests the setting is as described in Section V. VII. RESULTS A. Test Case 1 The results for this case indicate the relay detected the reverse power at 2 s. The SS was signalled to open at 2.14 s,

after the 0.14 s time delay, as shown in Fig. 3. This was as expected. The reverse current was 10A which was detected by the relay, as shown in Fig. 4. The reverse active power can be equated using the values in Table I, Equation 1 and the measured reverse current.
1 0.8 Trip Signal 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 Time (s) 3 3.5 4 4.5 5

Fig. 3. Relay status during Test Case 1

250 200 Current (A) 150 100 50 0 0

0.5

1.5

2.5 Time (s)

3.5

4.5

Fig. 4. Power System Current During Test Case 1

B. Test Case 2 The results for Test Case 2 indicate that the reverse current was 50A as shown by Fig. 6. This current was easily detected by the relay. The SS was opened at 2.14 s, as can be seen in Fig. 5.
1 0.8 Trip Signal 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 Time (s) 3 3.5 4 4.5 5

Fig. 5. Relay status during Test Case 2

250 200 Current (A) 150 100 50 0 0

0.5

1.5

2.5 Time (s)

3.5

4.5

Fig. 6. Power System Current during Test Case 2

C. Test Case 3 The lowest active reverse power was experienced in test case as shown in Fig. 8. For this reason the relay did not detect the reverse active power and the SS did not open, as suggested by the Relay Status signal in Fig. 7. This result

shows that reverse power relaying is ineffective when the installed DG capacity is small. For this reason it cannot be used as the only means of protection for loss of mains.
1 0.5 Trip Signal

-0.5

-1 0

0.5

1.5

2.5 Time (s)

3.5

4.5

Fig. 7. Relay status during Test Case 3

250

most methods of loss of mains protection being used at the moment are least effective when the DG capacity is large. This reverse active power relaying method can be used to compliment or as back-up to more traditional loss of mains protection methods, such as ROCOF relaying, to provide a complete solution which will detect all of the abnormalities and protect the microgrid from all of the harms of Loss of Mains. This scheme is useful when other methods such as ROCOF and frequency drop relays are not sensitive enough to detect an abnormal scenario. It is also better than active systems because it is not intrusive and better than most passive systems because it is reliable. REFERENCES
[1] P. O'Kane and B. Fox, "Loss of mains detection for embedded generation by system impedance monitoring," in Developments in Power System Protection, Sixth International Conference on (Conf. Publ. No. 434), 1997, pp. 95-98. P.D. Hopewell, N. Jenkins, A.D.Cross, Loss-of-mains detection for small generators, IEE Proc. Electr. Power Appl., Vol. 143, No. 3, May 1996. J. Yin, L.Chang and C. Diduch, Recent Developments in Islanding Detection for Distributed Power Generation, Power Engineering, 2004. LESCOPE-04. 2004 Large Engineering systems Conference, July 2004, pp. 124 128. P. Crolla, A.J. Roscoe, A Dysko and G.M. Burt, Methodology for testing loss of mains detection algorithms for microgrids and distributed generation using real-time power hardware-in-the-loop based technique, 8th International Conference on Power Electronics ECCE Asia, May 2011. Areva, Network Protection & Automation Guide, First Edition, July 2001, pp. M.M. Aman, G.B. Jasmon, Q.A. Khan, A.H.B. Abu Bakar, J.J. Jamian, Modeling and Simularion of Reverse Power Relay for Generator Protection, 2012 IEEE International Power Engineering and Optimization Conference (PEOCO2012), Malaysia, June 2012. S.Chowdhury, S.P.Chowdhury and P.Crossley, Microgrids and Active Distribution Networks, UK: The IET(UK), 2009. ABB, ABB Switchgear Manual, 10th Edition, 2001. S.Chowdhury, S.P.Chowdhury and P.Crossley. Microgrids and Active Distribution Networks, The IET(UK), July 2009.

200 Current (A) 150 100 50 0 0

[2] [3]
0.5 1 1.5 2 2.5 Time (s) 3 3.5 4 4.5 5

Fig. 8. Power System Current during Test Case 3

[4]

D. Test Case 4 In Test Case 4, where loss of mains occurs because of a fault, the reverse power flow is slightly higher than when loss of mains occurs due to a switch opening. The reverse current witnessed by the relay is 3A, this can be seen in Fig. 10. The relay detected this reverse in power flow and the SS was opened after a 0.14 s delay, as displayed in Fig. 9.
1 0.8 Trip Signal 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 Time (s) 3 3.5 4 4.5 5

[5] [6]

[7] [8] [9]

IX.

BIOGRAPHIES

Fig. 9. Relay Status during Test Case 4

250 200

C. Buque received his BSc in Electrical Engineering in 2011 and is currently pursuing his MSc in Electrical Engineering at the University of Cape Town, South Africa. Email: claudio.buque@mottmac.com S.Chowdhury is currently the Senior Lecturer in the Electrical Engineering Department of The University of Cape Town, South Africa. She became Member of IEEE in 2003 and Senior member of IEEE in 2011. She has published three books and over 150 papers in power systems modeling and simulation, power system protection, distributed generation and renewable energy systems. She is Member of SAIEE, Member of the IET (UK) and IE(India) and Senior Member of IEEE(USA). Email: sunetra.chowdhury@uct.ac.za S.P. Chowdhury is currently Associate Professor in Electrical Engineering Department of the University of Cape Town, South Africa. He became Member of IEEE in 2003 and Senior Member of IEEE in 2011. He has published three books and over 200 papers in power systems, distributed generation and renewable energy systems. He is a Member of SAIEE, Fellow of the IET (UK) with C.Eng. IE (India) and the IETE (India) and Senior Member of IEEE (USA). Email: sp.chowdhury@uct.ac.za

C u rre n t (A )

150 100 50 0 0

0.5

1.5

Fig. 10. Power System Current during Test Case 4

2.5 Time (s)

3.5

4.5

VIII. CONCLUSIONS It can be concluded from the test cases and results that reverse power relaying is most effective in detecting loss of mains when the DG capacity is large. This is useful because