Abstract— Many recent blackouts might raise a concern on

how the introduction of distributed energy resources or distributed generation (DG) into the network influences its secure operation in terms of dynamic performance. In this paper, the maximum penetration level of DG is determined using a dynamic study. Test systems are studied with different DG technologies and control methods.

Index Terms—Distributed Generation, Penetration Level,

Transient Stability, Dynamic Performance, System Collapse

I. INTRODUCTION Since the early 1920’s, power system stability has been recognized as an important concern for secure system operation [1]. The complexity of large interconnected systems requires many control elements to yield a reliable operation. The secure operation of power system has always been a challenging task, due to their intrinsic dynamic nature as load, generation, topology and key operating parameters are constantly changing [2]. As many power systems are operated very close to their present limits, putting more DG on the system may give more stress on their secure operation. The tendency to increasingly use small generation technologies or distributed generation may raise many problems related to system operation stability and control. The introduction of DG might interfere with the control of large central generators and interact with the behavior of loads in the power system. DG units have not been designed to support system stability during power system disturbances and often do not have to satisfy severe grid codes. Most synchronous machine equipped DG units have no or very limited active and reactive power control capabilities while regular induction generators are not able to control the reactive power at all. The system inertia of DG units is normally low. The synchronous DG with an ability to supply reactive power might contribute to improving voltage stability, in contrast with induction machines that may reduce the voltage stability of the system [3].

This work was supported by Belgian “Fonds voor Wetenschappelijk Onderzoek Vlaanderen” and the IWT for granting a GBOU research project to support this research.

The authors are with Department of Electrical Engineering, research group ELECTA, Katholieke Universiteit Leuven, B-3001 Leuven, Belgium (e-mail: thong.vuvan@esat.kuleuven.ac.be, http://www.esat.kuleuven.ac.be/electa)

Furthermore, the power generated by DG leads to a reduction of power generation from central power units and the number of online generators, spinning reserve [4]. This may result in a larger uncertainty in terms of power system stability, when a major disturbance occurs. In order not to harm the stability of the power system, an acceptable penetration level of DG in a specific network should be known before allowing a large amount of DG to be connected. Most of the literature has been using steady state studies to determine an acceptable penetration level of DG based on voltage operation limits [5], harmonic limits [6], protection limits, transformer capacity limits, minimum active power loss, etc. Not much literature deals with using dynamic approaches. In [7] the maximum rotor speed deviation and oscillation duration of transient stability are calculated when N-1 safety at different DG penetration levels is studied. A system collapse criteria, accounting for frequency and voltage, is applied in this paper to find out the acceptable penetration level of DG. The stress is defined as DG penetration levels from 0% to 100%. The stress is gradually increased until the system collapses, and the maximum penetration is derived. The methodology is implemented using the Syscan (System Collapse Analysis) module of the industrial Eurostag software package [8]-[9]. An 8-bus and the 39-bus New England systems are used to test with different DG technologies and control methods [10].

II. SIMULATION METHODOLOGY The question is to know how much penetration level of DG is acceptable in terms of system security. The idea is to study whether the system can sustain under normal operation and after any disturbances with DG connected. Starting from an initial steady state of the base case without any DG connected, the system is subsequently put under stress with increasing penetration levels of DG from 0% to 100%. The penetration level can be calculated as a function of the total DG power generation PDG over the total load demand PL.

Using Dynamic Simulation to Study the Penetration level of Distributed Energy Resources

Vu Van Thong, Student Member, IEEE, Daniel Van Dommelen, Senior Member, IEEE, Johan Driesen, Member, IEEE, and Ronnie Belmans, Fellow Member, IEEE

Penetration level (%) = ∑∑

L

DG

PP

x 100 (1)

With the incentive to reduce greenhouse gases, many DG units using environmentally friendly technologies, such as wind turbines, photovoltaics and high efficient CHP units, are allowed to generate as much electrical energy as possible. Based on the assumption that the total load demand is constant, this results in a reduction of power generated by central power units, due to the fact that load and power generation must be kept balanced at all times. The order of turning down large generators follows their production costs. However, the output power of such distributed technologies depends on the weather and cannot be anticipated exactly. If the power system is not strong enough, this might cause many problems in terms of stability and security of operation. This uncertainty may require a larger spinning reserve. At least 10% of the demand is kept as spinning reserve in the simulation. Once the system is put under stress, a list of pre-determined incidents is applied being generic or individual. An individual incident can be a single event or either a group of events. The events can be opening or closing of a line, short circuit at a node or on a line, short circuit elimination, compensating capacitor bank modification, etc [9]. Voltage, frequency and angle stabilities are used as indicators to verify whether the system accepts that DG penetration. This step is repeated until the system collapses and the maximum penetration level of DG is determined. In this study, the system is considered unstable with a certain level of DG penetration if it remains in operation after applying the series of incidents, and the voltage at every node is lower than 0.8 p.u. or the frequency of the system is lower than 47.5 Hz during 20 seconds. Depending on the DG technologies, the unit is equipped with ancillary control of voltage and frequency or not. The DG synchronous machine is modeled using simplified model of generator and control method. For large central generators, full governor and excitation are modeled.

Fig. 1: Power generation by DG units reducing the output of central generators when demand is assumed constant.

III. CASE STUDY

A. Case Study 1: 8-bus system

The methodology is applied to study a simple 8-bus system. It contains 4 central generators connected to bus 01 with a capacity of 50 MW each. The voltage levels of the system are 15, 35 and 150 kV. DG units are connected to load buses 04, 06 and 08 via transformers (6/35 kV). The capacity of each DG unit is 10 MW. The system supplies a total load of 180 MW and 60 MVar. The load is kept unchanged when the level of DG increases.

02

05

01 03 04

06

07

08

Fig. 2: Single line diagram of tested 8-bus system

The DG units at nodes 04, 06 and 08 are gradually connected to the system while reducing the production output of central generators. The system is put under stress each time with one or three DG generators online. The penetration level increases from 0% to 16%, 33%,…, 100%. A short circuit is applied at every line at every stress level in order to test whether the system can sustain operation with N-1 safety or not. The short circuit is located at 50% of the line and it occurs at time t = 200 s and last 0.15 s during the simulation. All generators are equipped with under/over-frequency relays. But the settings are different between DG units and central generators. For all DG units the under-frequency is set at 49 Hz with a time delay of 1.5 s and the over-frequency is set at 51 Hz with a time delay of 0.5 s. When the frequency is lower than 48 Hz or higher than 52 Hz, the DG units are tripped instantaneously. For central generators, under/over-frequency relays are set up at 48/52 Hz with a time delay of 0.5 s. Fig. 3 shows the frequencies of central generators and DG units at 44% DG penetration level with short circuit at 02-05 line. This scenario corresponds to two central generators online and eight DG units operated (80 MW). After the short circuit at 02-05 line has occurred, the DG units connected at node 06 accelerate faster than other generators because they are closer to the short circuit. The generators are tripped by the instantaneous frequency relay when frequency exceeds 52 Hz. Other generators swing in different ways before reaching to new stable value depending on their inertial values and characteristics. After the short circuit is eliminated, the system is divided into two subsystems: one includes nodes 05, 06, load and DG units at 06; the other is the rest. The voltage of the system at nodes 05 and 06 are recovered to a certain level until the DG units at

Central Generation

Distributed Generation

Load Demand

150 kV

35 kV

35 kV 35 kV

Central Generator

DG

15 kV

DG units

Load

LoadDG units

Load

node 06 are tripped (Fig. 4). The voltages of other nodes oscillate with rather high first swing before reaching to a new stable value. If one more DG unit is connected to the system, the penetration level is 50 %, and total power generation of DG is 90 MW. The same N-1 criterion above is verified with this scenario. A short circuit located at 50% of the line and it occurs at time t = 200 s and lasts 0.15 s. It is applied to every line in the system. Some tests are not met and system collapses happen. Fig. 5 shows the frequency behavior of central generators and DG units at 50% DG penetration level with a short circuit at 03-07 line. The behavior is similar to the former case. The DG units at node 08 accelerate faster and are tripped by instantaneous frequency relays. The DG units at node 04 swing some cycles before going into unstable operation because of the small size and the simple control technique. They are also tripped by instantaneous frequency relays. When the six DG units are tripped, the power generation does not meet the load demand in the system, the frequencies of the remaining generators decrease. Under-frequency relays react and send signal to trip all these generators. The system is totally collapsed due to frequency instability. The voltage at every node for this collapsed scenario is shown in Fig. 6. Node 07 and 08 are blacked-out after the short circuit due to tripping of the DG units at 08. The voltages at other nodes recover to nearly the values before the disturbance, although the first swing is high.

200 201 202 203 204 205 206

49.0

49.5

50.0

50.5

51.0

51.5

52.0

s

Hz

Fig. 3: Frequencies of central generators and DG units at 44% DG penetration level with a short circuit at line 02-05.

200 201 202 203 204 205 206

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

1.1

1.2

1.3

s

p.u.

Fig. 4: Voltages of different nodes at 44% DG penetration level with a short circuit at line 02-05.

200.0 200.5 201.0 201.5 202.0 202.5 203.0

46.5

47.0

47.5

48.0

48.5

49.0

49.5

50.0

50.5

51.0

51.5

52.0

52.5

Hz

Fig. 5: Frequencies of central generators and DG units at 50% DG penetration level with a short circuit at line 03-07.

199.5 200.0 200.5 201.0 201.5 202.0 202.5 203.0

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

1.1

1.2

1.3

s

p.u.

Fig. 6: Voltages of different nodes at 50% DG penetration level with a short circuit at line 03-07.

B. Case Study 2: 39-bus New England system In order to better understand and to work with a larger network, a modification of the 39-bus New England test system is used for the other case study [10]. The system consists of 10 power plants with 15 central generators. The total load is 6097 MW and 1350 MVar. The DG units are gradually connected to 20 nodes where they supply directly the loads. The capacity of each DG unit is 50 MW. In this study, synchronous DG is used to test. Each machine is equipped with a simple governor and an excitation system.

Fig. 7: Single line diagram of the 39 bus New England test system [10]

All generators are also equipped with under/over-frequency relays. For all DG units, the under-frequency is set at 49.5 Hz with a time delay of 0.5 s and the over-frequency is set at 50.5 Hz with a time delay of 0.5 s. This is tighter than the previous

case study. When the frequency is lower than 48 Hz or higher than 52 Hz, the DG units are tripped instantaneously. For central generators, under/over frequency relays are set up at 48/52 Hz with a time delay of 0.2 s. The same methodology is applied for this case study. The penetration level of DG is gradually increased. Ten DG units are connected to the system each time. A short circuit located at 50% of the line and it occurs at time t = 200 s and lasts 0.15 s. It is applied to every line at every stress level. The system sustains operation and meets all criteria of voltage and frequency until the penetration level of DG is 49%. The scenario corresponds to 60 DG units online and 8 central generators offline. Fig. 8 shows the frequencies of DG units when a short circuit occurs at line 05-06. The DG units connected at nodes 05 and 06 are tripped by over frequency relays, other DG units are tripped by under frequency relays. The frequencies of central generators decrease after the short circuit as well. And the tripping of DG units puts more stress on the central generators. The spinning reserve is not large enough to cover the production lost of DG. This leads the frequency of the system to decrease faster, and the system is collapsed after nearly 3 seconds after the short circuit, although the voltage of the system seems to recover to its value before the disturbance (Fig. 10).

200.0 200.5 201.0 201.5 202.0 202.5 203.049.0

49.5

50.0

50.5

51.0

51.5

s

Hz

Fig. 8: Frequencies of DG units at 49% DG penetration level with a short circuit at line 05-06.

200.0 200.5 201.0 201.5 202.0 202.5 203.0

47.5

48.0

48.5

49.0

49.5

50.0

50.5

s

Hz

Fig. 9: Frequencies of central generators at 49% DG penetration level with a short circuit at line 05-06.

200.0 200.5 201.0 201.5 202.0 202.5 203.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

1.1

s

p.u.

Fig. 10: Voltages of different nodes at 49% DG penetration level with a short circuit at 05-06 line.

IV. CONCLUSIONS A methodology to determine the penetration level of distributed generation based on dynamic system studies is given in the paper. It shows the ability of dynamic study to investigate thoroughly and deeply the behavior of the system at different penetration level of DG under normal and abnormal operation. It is a tool to help the network operators to better understand what may happen with their system when

a large portion of DG is connected. Furthermore, they can give a good protection scheme or a solution to encourage the development of more environmentally friendly technologies.

V. ACKNOWLEDGEMENT The authors are grateful to the Belgian “Fonds voor Wetenschappelijk Onderzoek Vlaanderen” for its financial support of this work and the IWT for granting a GBOU research project to support this research. Special thanks go to Tractebel Engineering for providing the simulation software.

VI. REFERENCES [1] P. Kundur, “Power system stability and control,” McGraw-Hill, 1994.

[2] IEEE TP-138-0, “Techniques for power system stability limit search,” A Special Publication of the Power System Dynamic Performance Committee of the Power Engineering Society.

[3] T. Vu Van, D.M. Van Dommelen, J. Driesen, R. Belmans: “Impact of large scale distributed and unpredictable generation on voltage and angle stability of transmission system,” CIGRE General Meeting, Paris, France, 2004; 8 pages.

[4] T. Vu Van, E.Vandenbrande, J. Soens, D.M. Van Dommelen, J. Driesen, R. Belmans: “Influences of large penetration of distributed generation on N-1 safety operation,” PES General Meeting, IEEE, Denver, Colorado, USA, June 6-10, 2004; 5 pages.

[5] T.E. Kim, J.E. Kim: “A method for determining the introduction limit of distributed generation system in distribution system,” PES General Meeting, IEEE, July 2001; Vol. 1, pp. 456-461

[6] Bhowmik, A.; Maitra, A.; Halpin, S.M.; Schatz, J.E.: “Determination of allowable penetration levels of distributed generation resources based on harmonic limit considerations,” Power Delivery, IEEE Trans. on, Vol. 18 April 2003; pp. 619 – 624.

[7] M. Reza, P. H. Schavemaker, J. G. Slootweg, W. L. Kling, L. van der Sluis: “Impacts of distributed generation penetration levels on power systems transient stability,” PES General Meeting, IEEE, Denver, June 2004; 6 pages.

[8] http://www.eurostag.be

[9] J. Deuse, K. Karoui, A. Bihain, J. Dubois: “Comprehensive approach of power system contingency analysis,” PowerTech Conference Proceedings, 2003 IEEE Bologna, Volume: 3, June 23-26, 2003; pp. 352 – 357.

[10] M. A. Pai:” Enenrgy function analysis for power system stability,” Kluwer Academic Publishers, 1989.

VII. BIOGRAPHIES

Vu Van Thong (S’02) received the B.E. and the M.E. degrees in Electrical Power Systems at Hanoi University of Technology, Vietnam in 1997 and Asian Institute of Technology, Thailand in 2001 respectively. Before coming to Thailand for ME, he had worked as an electrical engineer at Electricity of Vietnam for nearly 3 years. Since 2001 he has been working towards a Ph.D. in the Electrical Energy research group, Department of

Electrical Engineering of the K.U.Leuven in Belgium. His special fields of interest include distributed generation, dynamic study, voltage stability and optimal power flow.

Daniel Van Dommelen (SM '78) is electrotechnical engineer from the K.U.Leuven, Belgium, has an M.Sc. in Electrical Engineering (U. Wisc.), and a Ph.D. from the K.U.Leuven. Since 1977 he is full professor at this university, teaching both undergraduate and graduate courses in Power Systems, High Voltage and Electroheat. He is author of a book on Production, Transmission and Distribution of Electric Power, presently in its fourth reprint, and author of numerous publications in both national and international journals. He is Belgian representative in the ERE Committee of the UIE. He has been chairman of the IEEE Benelux Section, and is presently SAC for the Section. He is distinguished member of CIGRE and SEE and member of national electrical engineering societies.

Johan Driesen (S’93–M’97) graduated as an Electrotechnical Engineer and received the Ph.D. degree in electrical engineering from the Katholieke Universiteit Leuven (KULeuven), Leuven, Belgium, in 1996 and 2000, respectively. In 1996, he became a Research Assistant of the Fonds voor Wetenschappelijk Onderzoek-Vlaanderen (Fund for Scientific Research of Flanders - F.W.O.-Vl.). From 2000 to 2001, he was a Visiting Lecturer with Imperial College, London, U.K.

In 2002, he was a Visiting Scholar with the Electrical Engineering Department, University of California at Berkeley. He is currently a Postdoctoral Research Fellow of the F.W.O.-Vl. at KULeuven. Dr. Driesen received the 1996 R&D Award of the Belgian Royal Society of Electrotechnical Engineers (KBVE) for his Master’s thesis on power quality problems. In 2002, he received the KBVE R. Sinave Award for his Ph.D. dissertation on coupled problems in electrical energy transducers.

Ronnie Belmans (S’77-M’84-SM’89-FM’05) received the M.S. degree in electrical engineering in 1979 and the Ph.D. degree in 1984, both from the K.U.Leuven, Belgium, the Special Doctorate in 1989 and the Habilitierung in 1993, both from the RWTH, Aachen, Germany. Currently, he is a full professor with the K.U.Leuven, teaching electric power and energy systems. His research interests include techno-economic aspects of power systems, power quality and distributed generation.

He is also guest professor at Imperial College of Science, Medicine and Technology, London-UK. Since June 2002 he is chairman of the board of directors of ELIA, de Belgian transmission grid operator.