Integration of distributed generation to Grid

INV ITEDP A P E R

Integration Issues of DistributedGeneration in Distribution GridsThis paper considers the probable operating problems and challenges

in connecting distributed generation to low- and medium-voltage

electric power grids.

By Edward J. Coster, Student Member IEEE, Johanna M. A. Myrzik,

Bas Kruimer, Member IEEE, and Wil L. Kling, Member IEEE

ABSTRACT | In today’s distribution grids the number of

distributed generation (DG) units is increasing rapidly. Com-

bined heat and power (CHP) plants and wind turbines are most

often installed. Integration of these DG units into the distribu-

tion grid leads to planning as well as operational challenges.

Based on the experience of a Dutch distribution system ope-

rators (DSO), this paper addresses several possibilities to

handle grid planning issues. Effects on voltage control, grid

protection, and fault levels are investigated and described.

These aspects are illustrated with the aid of simulations on an

existing distribution grid. It is demonstrated that in compact

distribution grids voltage control problems and blinding of

protection are not likely to occur and that false tripping and

fault level have to be considered carefully.

KEYWORDS | Distributed generation (DG); distribution grids;

grid planning; protection; voltage control

I . INTRODUCTION

The nature of today’s distribution grids is changing from

passive to active. The penetration level of distributed

generation units (DG units) is increasing and it is expected

that this growth will continue over the next years. At the

moment not only the number of DG units is increasing, butalso, in specific areas, the size of the units. In this way, in

the near future, the DG may contribute in a substantialpart of the power generation.

DG technologies can be categorized in renewable and

nonrenewable [1]. Examples of renewable technologies

are, e.g., solar, wind, and geothermal, while examples of

nonrenewable technologies are internal combustion en-

gines, microturbines, and fuel cells. Two most popular and

widely used DG schemes are:

• wind turbines;• combined heat and power (CHP) plants.

Currently, wind turbines have a large share in the

generation of electricity by DG units. Originally, small

wind turbines were connected to low-voltage networks.

Wind turbines increased in size and therefore in gener-

ating power and have also been clustered in groups, thus

making it necessary to connect to medium-voltage (MV)

and even high-voltage (HV) grids.CHP plants are most often used to generate heat, with

the generation of electricity as an important side product.

These types of DG units are mostly installed in industrial

environments, where the heat is used in the production

processes and the generated electricity is either used for

own consumption or sold to the market. Small CHP plants

are also used for district heating.

In The Netherlands local authorities have designatedspecific areas for the development of greenhouses. Each

greenhouse may contain a CHP plant with a capacity in the

range of 1–3 megavoltsampere (MVA) and thus to be

connected to the local MV grid. Due to the high density of

greenhouses in such areas, the penetration level of CHP

plants in the MV grid is very high. This can amount to a

total generated power of 100 MW or even more, and with

loads far less than the generation power.In Fig. 1, an overview of the development of the small

CHP plants in The Netherlands during the past ten years is

given. Due to financial stimuli, such as a subsidy for saved

Manuscript received January 8, 2010; revised May 6, 2010; accepted May 12, 2010. Date

of publication September 7, 2010; date of current version December 17, 2010.

E. J. Coster is with Network Planning, STEDIN, Rotterdam 3015 EK, The Netherlands

(e-mail: [email protected]).

J. M. A. Myrzik and W. L. Kling are with the Electrical Power Systems,

Eindhoven University of Technology, Eindhoven 5612 AZ, The Netherlands

(e-mail: [email protected]; [email protected]).

B. Kruimer is with Quanta Technology Europe, Rotterdam 3085 BW,

The Netherlands (e-mail: [email protected]).

Digital Object Identifier: 10.1109/JPROC.2010.2052776

28 Proceedings of the IEEE | Vol. 99, No. 1, January 2011 0018-9219/$26.00 �2010 IEEE

CO2 emissions and exemption of energy tax on the pur-

chase of natural gas, the costs of installing a CHP plant

were reduced which led to an enormous increase of small

CHP plants, especially in the Dutch greenhouse areas. Itcan be seen that in the last couple of years the total

installed CHP power has more than doubled.

Next to the many CHP plants also many wind turbines

have been installed favoring specific subsidy schemes.

These wind turbines are clustered in small wind parks,

which are usually connected to the subtransmission or a

distribution grid. The development of the integration of

photovoltaics (PV) systems into the grid does not show asignificant increase. The effect of the PV systems on the

Dutch distribution grid is rather small and will not be

addressed in detail.

Integrating these large numbers of CHP plants and

wind parks into the power system within this short time

span leads to serious challenges for the grid operators in

order to stay in compliance with the regulatory and legal

framework. In addition, there are important technicalissues to be solved as well. This paper shows an overview of

a grid operator’s approach to connect a large number of DG

units to the network. The impact on the power system

operation is presented and will be illustrated with dynamic

simulations on the existing distribution grid in a green-

house area.

II . GRID PLANNING INCLUDING DG

From grid planning point of view some utilities may use

the DG as one of the ways to solve bottlenecks in the

distribution grids. In [2] and [3], it is discussed how the

DG can be incorporated in the grid planning process.However, utilities today face independent power produ-

cers (IPP) or nonutility generators (NUG). For those

DGs, the location and the technology are given factors

and the utility has an obligation to connect them to their

grid and to transmit the power. A connection of IPP can

create new bottlenecks rather than helping to solve

them.

In the Dutch greenhouse areas, it is foreseen that manyCHP plants will be installed, which need to be considered

as IPP and will have to be integrated into the distribution

grid. The Dutch legal and regulatory framework requires

the distribution system operators (DSO) to provide a con-

nection to the distribution grid within 18 weeks upon a

client’s request. The capacity and structure of today’s dis-

tribution grids in these areas does not support a connection

of a large number of CHP plants. A number of technicalmeasures need to be taken such as grid reinforcements,

extension of existing substations as well as building of new

substations. This may also affect the transmission grid as

several of such areas are being developed at the same time

in the same region. When congestion in the local trans-

mission grid occurs, the transmission system operator

(TSO) has to reinforce its grid as well. In these cases, grid

planning really becomes an issue. The time it takes toexecute projects in the grid to secure reinforcements does

not match the legal time it takes to make a grid connection

for a new DG unit. Fig. 2 gives an overview of the average

project time for the various parties and shows clearly the

discrepancies leading to major planning issues. To mitigate

these planning issues, proactive grid planning by both DSO

and TSO is necessary.

To plan the grid in a proactive way, several transmis-sion development scenarios have to be established [4].

These scenarios are based on the expected growth of the

load or the number of CHP plants in a certain area as well

as technical and economical developments which can in-

fluence this growth. With the aid of these scenarios,

bottlenecks in current MV grids and subtransmission grids

can be identified. In each scenario, alternative grid designs

have to be generated for each bottleneck. These alter-natives are to be generated in such a way that the expected

growth of CHP plants in each scenario is covered. The

Fig. 2. Planning period of grid reinforcements of the DSO and the

TSO as well as the legal period for connection.

Fig. 1. Development of the major DG technologies in

The Netherlands.

Coster et al. : Integration Issues of Distributed Generation in Distribution Grids

Vol. 99, No. 1, January 2011 | Proceedings of the IEEE 29

alternative grid designs are checked with boundaryconditions. The boundary conditions are conditions which

can restrict the possible solutions. Examples of boundary

conditions are, e.g., minimum realization time, grid

operator constraints, or limited investment funds for the

grid operator. In the grid planning stage, for each scenario,

this approach results in an action plan, which gives, per

scenario, an overview of the steps to be taken. The actual

development of the horticultural area determines whattransmission scenario has to be followed and what steps

have to be taken by the grid operator. This approach is

depicted in Fig. 3.

In this approach, the main challenge for the grid

operator is to formulate realistic scenarios and assess

which scenario will most likely be followed. The formu-

lation of scenarios is based on a market survey, technical

and economical developments, and expert knowledge.Applying scenario-based grid planning offers a possi-

bility to develop the grid in a proactive way. With the aid of

scenarios, future bottlenecks can be recognized early on in

the planning stage. Necessary grid expansions to overcome

bottlenecks can be linked to one another in order to gua-

rantee grid capacity on time. A drawback of this approach

is the long and intense study time needed to formulate

realistic scenarios. In addition, it is found that taking adecision based on scenarios and their chances of occur-

rence will lead to new uncertainties and possibly higher

investment costs, and probably taking into account too

many issues to be solved and dealt with.

III . EFFECT OF DG ON DISTRIBUTIONGRID OPERATION

Large scale integration of DG units in the distribution grid

not only affects the grid planning but also has an impact on

the operation of the distribution grid. Aspects which are

influenced by the connection of DG units are [5] asfollows:

• voltage control;

• power quality;

• protection system;

• fault level;

• grid losses.

The effect of DG units on these quantities strongly de-

pends on the type of DG unit and the type of the network.DG units can be either directly connected to the distri-

bution grid, such as synchronous and asynchronous gene-

rators, or via a power electronic converter. In all cases, the

power flow in the distribution grid as well as the grid losses

and the voltage control are affected. Synchronous gen-

erators contribute a large short-circuit current influencing

the protection scheme and the fault level. DG units

connected via power electronic converters hardly contrib-ute to the fault current making the effect on fault level and

protection system negligible [6].

The Dutch distribution grid consists of cables only and

covered distances are relatively short compared to many

other countries with overhead line feeders. Different con-

ductor type and much shorter distances cause the Dutch

distribution grids to have much lower impedances com-

pared to distribution grids consisting of long overheadlines. This difference strongly determines what impact DG

has on the distribution grid operation.

IV. VOLTAGE CONTROL

For the power system and customers’ equipment to func-

tion properly the voltage level of distribution grids must be

kept within specific range [5]. This voltage range is welldefined in international standards. Grids are confronted

with varying loads and voltage fluctuations will occur.

Varying loads create a changing current through the

resistance and reactance of the feeder causing a changing

voltage drop along the feeder.

The connection of the DG along the feeder may affect

proper voltage control of the distribution grid. The impact

of the DG on voltage control is dependent on the powerflow in the network. The voltage profile is not much in-

fluenced when the injected power by the DG is lower than

or equal to the load of the feeder. In this case, the energy

supplied by the grid as well as the current through the

feeder are decreasing, resulting in a reduction of voltage

drop. However, when the generated power exceeds the

load of the feeder, voltage rise will occur. This voltage rise

caused by the reversed power flow is a function of thepower generated by the DG and the short-circuit power of

the grid at the point of interconnection [7]. The effect of

reversed power flow gets stronger when the DG injects

reactive power as well.

The effect of the DG on the voltage profile of a

distribution feeder is demonstrated in the simulation of a

10-kV test feeder shown in Fig. 4.

Fig. 3. Overview of the grid planning stages.


30 Proceedings of the IEEE | Vol. 99, No. 1, January 2011

The total feeder length is 20 km and the substations are

equally distributed along the feeder. The DG unit is

connected to Sub 4 and generates 4 MVA. Table 1 gives an

overview of the applied network data.Different generation scenarios are applied to determine

the voltage profile along the feeder as shown in Fig. 5.

Without active DG-power injection, an expected

voltage profile occurs. The power flow through the feeder

results in a smooth voltage drop from the beginning to the

end of the feeder. A 1-MW DG-power injection helps to

supply the local load, which reduces the flow from the

source, and thus also reduces the voltage drop in the firstpart of the feeder. When the injected power starts to in-

crease, more loads are supplied locally further reducing

the flow from the source.

In Fig. 5, it can be seen that the voltage at Sub 4

significantly increases at an injected power of 3.2 MW.

The simulations demonstrate that in a DG connection

study all possible load/generation scenarios need to be

studied to assess if the voltage profile will stay within theoperational limits.

Keeping the voltage within the allowed limits can be

achieved in several ways. One simple option is to reduce

the primary substation voltage. This solution works well

for those feeders with the DG connected. In feeders con-

nected to the same substation, but without the DG, the

voltage now may become too low. In most cases, voltage

may be reduced to some extend without violating the lowervoltage limit. Another solution is to increase the conductor

diameter of the specific feeder to which the DG unit is

connected. This is only possible when the distribution grid

has to be extended due to the connection of the DG unit.

An operational approach can be to constrain the DG unit at

times of low demand and so expensive grid reinforcements

can be prevented. This can be an attractive option when

the DG unit is dispatched. New techniques are also avail-able which are based on influencing the reactive power

flow in the grid. DG units including a synchronous ma-

chine can absorb some reactive power and in this way can

help reduce the voltage rise. Controlling the reactive

power flow in the distribution feeder can also be done by a

FACTS device named STATCOM. Because of the fast re-

sponse time, it can provide a dynamic voltage control in

the distribution grids [8]. However, controlling voltage

level by controlling the reactive power flow only works fordistribution feeders with a sufficient X/R ratio.

V. POWER QUALITY

The effect of the integration of the DG on power quality

concerns three major aspects [5], [9]:

• dips and steady-state voltage rise;

• voltage flicker;

• harmonics.

A. Dips and Steady-State Voltage RiseThe DG can affect the quality of the supply voltage in

several ways. Connecting the DG to a lightly loaded feeder

the power flow can reverse and the voltage at the DGconnection point starts to rise. This means that the supply

voltage for customers connected nearby DG units starts to

rise as well. This voltage rise is a steady-state effect and it

strongly depends on the X/R ratio, feeder load, and in-

jected power by the DG unit. However, the DG can also

have a transient effect on the voltage level. A rapid load

current variation of a DG unit causes a sudden increase or

Fig. 4. Overview of a test feeder including DG.

Table 1 Network Data of the Test Feeder of Fig. 4

Fig. 5. Voltage profile of the test feeder for different generating

scenarios.



decrease of the feeder current and hence an effect on thefeeder voltage. For instance, when the wind starts to blow,

the wind turbine output rapidly increases until the rated

power of the wind turbine is reached. The rapid output

change of the wind turbine changes the power flow in the

feeder and can cause a voltage transient. A sudden change

in the output power can also occur when the wind exceeds

a certain upper limit (25 m/s). At that point, the wind

turbines have to be protected against overload and strongmechanical forces, and are disconnected and shut down

[10]. This disconnection can cause an increase in the

feeder current and hence a dip or a drop in the supply

voltage.

B. Voltage FlickerIn the distribution grids, the most common cause of a

voltage flicker is a rapid and regular variation of the loadcurrent. An example of a cause of regular load current

variation is the tower effect of a fixed speed wind turbine.

This effect is due to the wind shielding effect of each blade

of a three-blade turbine as it passes the tower. When the

blade passes the tower the injected electrical power of the

wind turbines reduces, which has an effect on the grid

voltage [10]. The tower effect causes a power oscillation

with a frequency of three times the blade turning speedand hence voltage flicker with the same frequency. Con-

versely, some DG schemes contribute significantly to the

fault level which gives a reduction of the network im-

pedance. As a result, the changing load current caused by a

changing load will lead to a smaller voltage variation and

hence improved power quality [9].

C. Harmonics and ResonancesInverter connected DG units might cause harmonics.

The magnitude and the order of the harmonic currents

depend on the technology of the converter and the mode of

operation [5]. The injection of harmonic currents can dis-

tort the voltage waveform which can propagate throughout

the distribution grid. Furthermore, apparently small volt-

age distortions can cause large harmonic currents at a

series of resonance conditions of the cable capacitance andthe supply inductance (transformer leakage and cable in-

ductance) and have to be prevented. Especially for inverter

interfaced DG, such as PV systems and micro CHP, parallel

resonance of the parallel network capacitance (output

capacitance of the inverter and the cable capacitance) and

the supply inductance can lead to high-voltage distortion at

the connection point [11].

These types of resonances are also found in [11] and[12] where practical measurements were performed on

existing low-voltage networks including a large number of

PV systems. The measurements especially show that pa-

rallel resonances occur which can trip the inverters due to

a distorted supply voltage.

An ability to reduce the harmonic current injection is

filtering of the output current. Modern power electronic

converters are able to filter the injected current and reducethe injected harmonics. In [13], various harmonic reduc-

tion techniques are discussed and it is stated that harmonic

reduction can be an ancillary service of grid connected

converters. An overview and classification of all power

quality phenomena and the effect of the DG on power

quality can be found in [14].

VI. GRID PROTECTION

Distribution grid protection consists normally of a simple

overcurrent protection scheme since there is only one

source of supply and the power flow is defined. The con-

nection of the DG to the distribution grid leads to multiple

sources of the fault current which can affect the detection

of disturbances. The contribution of the DG to the fault

current strongly depends on the type of the DG and theway the DG unit is connected to the distribution grid. As

stated earlier, the converter interfaced DG hardly contri-

butes to the fault current. To study the effect on distri-

bution grid protection, the directly coupled DGs are

considered here.

In [15]–[17], potential problems to the distribution

grid protection are discussed. The main problems identi-

fied are:• prohibition of automatic reclosing;

• unsynchronized reclosing;

• fuse-recloser coordination;

• islanding problems;

• blinding of protection;

• false tripping.

Reclosure problems, blinding of protection, and false

tripping are further analyzed.

A. Recloser ProblemsAn automatic recloser is a protection device typically

applied in distribution grids consisting of overhead lines.

Most faults on these lines only last a short period of time,

therefore it is not necessary to switch the line off perma-

nently. The automatic recloser switches off the line for a

short period of time to allow the arc to extinguish. After abrief time delay the line is energized again by the auto-

matic recloser. When the fault is removed the line can stay

in service otherwise the automatic recloser switches off the

line again. In case of a permanent fault the line is switched

off permanently after three or four unsuccessful reclosing

actions.

The DG can disturb the automatic reclosing process

significantly. During the open time of the recloser, the DGunit still energizes the feeder, thus energizing the arc and

making the temporary fault a permanent one. The line will

be unnecessary, often switched off permanently. Also,

during the open time, a small island is created and the DG

units tend to accelerate or decelerate, which results in

unsynchronized automatic reclosing [16]. This can lead to

a serious equipment damage. Because the automatic



recloser is a powerful means to protect overhead line

feeders, the DG units will have to disconnect before the

recloser action takes place.

Besides unsynchronized reclosing the coordinationbetween fuse and recloser can also be lost. Temporary

faults will then be cleared permanently by the fuse instead

of temporarily by the recloser. To restore the coordination of

the fuse and the recloser a microprocessor-based recloser

may be applied [18]. In the microprocessor, several trip

curves can be programmed and the microprocessor keeps

track of which curve is in use. The recloser is equipped with

fast and slow curves. The fast curve should be programmedin such a way that this curve is selective with the lateral

fuses, especially in the presence of the DG. To prevent un-

synchronized reclosing, the DG first has to be disconnected,

which brings the grid back in the situation without the DG.

Hence, the fast curve has to be active only during the first

reclosing action. In the second reclosing cycle, the slow

curve, which is selective with the lateral fuses, is activated

and the fault can be selectively cleared.Overhead line feeders are widely used in the world.

However, the Dutch distribution grid consists of cables

only and all faults are permanent, therefore automatic

reclosers are not applied. Blinding of protection and false

tripping are independent of the type of feeder and will be

discussed in the next sections.

B. Blinding of ProtectionIn Fig. 6, a distribution feeder including a DG is shown.

When a short-circuit occurs at the indicated location, both

the grid and the DG unit contribute to the fault current.The division of the current contribution depends on the

network configuration, grid impedance, and power gener-

ated by the DG unit. Due to the contribution of the DG

unit, the total fault current will increase.

However, the grid contribution decreases. This can

lead to a poor fault current detection. It is even possible

that the short-circuit stays undetected because the grid

contribution to the short-circuit current never reaches thepickup current of the feeder protection relay. This mecha-

nism is called blinding of protection and is also known as

protection underreach.

To determine the impact of a DG on the short-circuit

current a test grid, consisting of an external grid and three

cable connected MV nodes, is defined and modeled in the

simulation software. Busbar 2 connects a synchronous

generator (see Fig. 7). The grid parameters are shown

in Table 2.

In the test grid, the location and the size of the gene-

rator are modified in repetitive calculations. The setup is

explained by the flowchart in Fig. 8.

The flowchart contains two loops: one loop to modifythe location and the other loop to modify the size of the

generator. The location of busbar 2 is adjusted by 10% of

the total feeder length, which is kept as a constant. In this

way, node 2 shifts from node 1 towards node 3 and the

effect of the location of the generator can be observed.

After the modification of the location (the distance to the

substation and the distance to the feeder end), the second

loop is entered performing a three-phase fault calculationat busbar 3 with increasing generator size in 2-MW steps to

a maximum of 10 MW. The short-circuit calculation is

based on the well-known IEC 909 method. For each loop

run, the grid contribution specific for that short-circuit

current is stored. When the power generation loop is

completed, the first loop is entered again moving the loca-

tion 10% and the process starts all over.

The grid contribution determination tests are run forthree-phase and two-phase faults. The results are shown in

Fig. 9(a) and (b).

In both situations with two-phase as well as three-

phase faults, it can be clearly seen that the generators have

an effect on the short-circuit contribution by the grid itself.

As expected, large size generators influence the grid con-

tribution more than small size generators. The graphs also

show that the effect of the generator is most significantnear the center of the feeder.

To determine if blinding of protection occurs, the

minimum grid contribution has to be compared with the

pickup current of the overcurrent protection. Normally,

the pickup value is set to 50% of the minimum end of the

line two-phase fault current [19]. In the test feeder, the

minimum end of the line two-phase fault current without a

DG contribution is 2.1 kA, and the pickup current of theovercurrent protection is set to 1.05 kA. In Fig. 9(a) and

(b), the minimum grid contributions are 1.75 and 1.58 kA.

It can be concluded that in this case blinding of protection

will not occur.

So, integrating DG into a Bsufficiently strong[ distri-

bution grid composed of cable feeders of a Bmoderate

length[ does not normally lead to blinding of protection.

Fig. 6. Principle of blinding of protection.

Fig. 7. Test feeder blinding of protection.



In [20], this is discussed in detail, and in [21], simple

equations are derived to determine what the grid contri-

bution to the total short-circuit current will be for feeders

connecting the DG.

C. False TrippingFalse tripping (also known as sympathetic tripping)

may occur when a DG unit contributes to the fault in an

adjacent feeder connected to the same substation. The

generator contribution to the fault current can exceed the

pickup level of the overcurrent protection in the DG feeder

causing a possible trip of the healthy feeder before the

actual fault is cleared in the disturbed feeder. This is

shown schematically in Fig. 10.The DG unit provides a major contribution to the fault

current when the DG unit and/or the fault are located near

the substation. Especially in weak grids with long feeders

protected by definite-time overcurrent relays, false trip-

ping can easily occur. In this case, the settings of the

protection relays have to ensure that faults at the end of

the feeder are also detected which leads to a relatively low

pickup current.

In [15], it is discussed that in some cases false trippingcan be prevented by finding another suitable relay setting.

In practice, this means that the fault clearing time has to

Fig. 9. (a) Three-phase fault grid contribution varying generator

size and location. (b) Two-phase fault grid contribution varying

generator size and location.

Fig. 10. Principle of false tripping.

Table 2 Network Data of the Test Feeder of Fig. 7

Fig. 8. Flowchart to illustrate the repetitive calculations.



be increased rather than increasing the pickup current.Increasing the pickup current results in a less sensitive

feeder protection which probably may not clear all faults

anymore. Changing the fault clearing time leads to

disconnecting the faulted feeder first and preventing the

healthy feeder from false tripping.

When selectivity cannot be reached by changing the

protection settings, applying directional overcurrent pro-

tection can solve the problem [15]. However, directionalprotection is slower, more expensive, and usually not a

standard solution for grid operators.

Trying to solve blinding of protection may introduce

false tripping. This example is discussed in [22] in which

a small wind farm is connected to a weak existing net-

work causing blinding of protection. This is solved by

reducing the pickup current but it causes false tripping

for faults in some specific locations. The proposed solu-tion is to install protection devices with an additional

time delay giving the feeder as well as the wind farm a

longer fault clearing time.

VII. GRID SIMULATIONS

In this section, the simulation results of an existing Dutch

distribution grid including a DG is discussed. The grid is

located in a greenhouse area and has 21 CHP plants

equipped with synchronous generators connected. A single

line diagram of the grid as well as the network data aregiven in Fig. 15 and Tables 3–7. The 10-kV distribution grid

is connected via three transformers to the 25-kV sub-

transmission grid. The 10-kV substation consists of three

sections coupled via coupling breakers and eight feeders

connecting the 21 DGs. Each feeder is equipped with a

definite-time overcurrent protection set to 1650A-0.4 s.

The test system is used to investigate in what ways the

CHP plants affect voltage control, grid protection, andfault level.

A. Voltage ControlTo study the effect of the CHP plants on the voltage

profile of the feeders, Bfeeder 8[ is taken as an example.

Based on stationary power flow calculations, the voltage

profiles of Bfeeder 8[ are determined for the cases includ-

ing and excluding CHP plants, and are depicted in Fig. 11.

In this area, the distribution grid is dominantly loaded

with CHP plants and the connected load is limited. Hence,

the voltage drop along the feeder excluding the CHP plants

is also limited. An important factor herein is the relativelyshort feeder length and the low cable impedance. As is

expected, a voltage rise occurs when the CHP plants are

connected. However, due to the mentioned factors, this

voltage rise is modest and does not exceed the maximum

allowable voltage deviation of �10%. It can be concluded

that in this distribution grid connection of CHP plants does

not lead to unacceptable voltage profiles.

B. Grid ProtectionAs discussed in Section VI, false tripping and blinding

of protection are mechanisms which are independent of

the type of feeder (cable or overhead line). To determine if

false tripping occurs, the contributions of the CHP plantsper feeder have to be known. The feeder contributions are

highest during a three-phase fault at the busbar. When the

feeder contributions do not exceed the protection pickup

current, then false tripping will not occur. Fig. 12 gives an

overview of the contribution to the faulted busbar per

feeder as well as the pickup current of the overcurrent

protection.

In six feeders, the feeder contribution exceeds thepickup current of the overcurrent protection and false

tripping can occur. However, the clearing time of the

overcurrent protections is 0.4 s and all CHP plants are

equipped with an undervoltage protection clearing the

CHP plants within 100–200 ms. After disconnection of a

CHP plant, the specific feeder contribution drops below

the pickup current and all relays reset. In this way, false

tripping is prevented, however it is dependent on the CHP

Fig. 11. Voltage profile along ‘‘feeder 8’’ including and excluding

CHP plants.

Fig. 12. Overview of feeder contributions to a three-phase

busbar fault.



plant undervoltage protection. To achieve independence of

the CHP plants, the overprotection pickup current needs

to be increased in such a way that the worst case fault can

be still detected. Due to the high fault level, it is possible tofind a suitable setting which fulfills both requirements. In

weaker systems, the overcurrent protections have to be

equipped with the directional protections.

Blinding of protection is studied using feeders 1 and

8 by illustrating the effect of the connected CHP plants

on the grid contribution to the fault current. Along the

feeders, a short circuit is simulated which means that the

location of the fault is equidistant taken along the feeders.During the fault calculations, the grid contribution is

stored. Fig. 13 shows the grid contribution to the fault

current for the varying fault locations. It demonstrates that

the CHP plants do not affect the grid contribution when

the location of the fault in the feeders is between the

substation and the first CHP plant. The effect of the

contribution of the CHP plants becomes noticeable for

faults further along the feeder passed the first CHP plant.However, in this grid, the effect is modest and does not

lead to blinding of protection.

C. Fault LevelDuring the development of the greenhouse area, CHP

plants have been installed and connected to the local dis-

tribution grid, which prior to that time did not have any

local power generation but loads only. For the fault cur-

rent, there was only one source of supply and fault levels of

the main substation and the ring main units (RMUs) were

dimensioned to withstand the maximum fault current of

that single source. With integrated CHP plants additional

sources have been introduced leading to an increased faultlevel.

To study the effect of the CHP plants on the fault level,

short-circuit calculations are applied to the test system.

Disturbances are applied to the substation busbar and to

three feeder locations: RMU1, RMU2, and RMU3. The

10-kV substation fault level is 25 kA and for the RMUs it

is 14.5 kA. The results are shown in Fig. 14.

Normally, the substation is operated with closed

coupling breakers. The short-circuit calculations show

that the fault level of the main substation and RMUs are

exceeded. Operating the substation with open couplingbreakers reduces the fault level significantly.

In the region, other greenhouse areas are being dev-

eloped causing reduction of the fault level in an upstream

subtransmission substation. This also affects the fault level

of the test system. However, this is not sufficient to ope-

rate the substation with closed coupling breakers as indi-

cated in Fig. 14. Therefore, the grid needs to be operated

with open coupling breakers both in the local substation aswell as in the upstream subtransmission substation. So,

operational measures need to be taken to manage fault

levels because of the integration of CHP plants in the

distribution grid, which influences grid availability and

thus the customer reliability.

VIII . CONCLUSION

This paper deals with integration issues of DG units in

medium voltage grids. The significant increase of

integration of DG units in distribution grids can lead tomajor planning issues and conflicts with the legal and

regulatory framework. Some approaches to overcome

these planning issues are discussed. When integration of

DG units in the distribution grid also causes expansion of

the transmission grid, a prediction-based grid planning is

recommended.

There may be some operational challenges as well.

Major aspects which are influenced by the connection ofDG units are voltage control, grid protection, and fault

level. Simulating different scenarios in an existing distri-

bution grid with a significant number of CHP plants im-

plemented shows that voltage control problems in compact

distribution grids consisting of cables are not to be ex-

pected. It is also demonstrated that in such a grid the effect

of the CHP plants on the grid contribution to the fault

Fig. 13. Grid contribution including/excluding CHP plants as a

function of fault location.

Fig. 14. Effect of CHP plants on the fault level in the distribution grid.



Fig. 15. Overview of the simulated distribution grid.

Table 3 Line Data of the Simulated Distribution Grid

Table 4 Generator Data of the Simulated Distribution Grid

Table 5 Cable Data

Table 6 Transformer Data



current is limited. Therefore, blinding of protection is not

likely. Because of the short feeder length, false tripping

can occur and has to be considered during the design of the

protection system. At specific locations in the grid, shortfeeder length also leads to exceeding the allowed fault

level. It is shown that fault level can be managed to stay

within the limits by operational measures. This is a cost-

effective way to control the fault level, however, these

measures do affect customer reliability which can be seenas a drawback. h

REF ERENCE S

[1] H. B. Puttgen, P. R. Macgregor, andF. C. Lambert, BDistributed generation:Semantic hype or the dawn of a new era,[IEEE Power Energy Mag., vol. 3, no. 1,pp. 22–29, Jan./Feb. 2003.

[2] R. C. Dugan, T. E. McDermott, and G. J. Ball,BPlanning for distributed generation,[ IEEEIndustry Appl. Mag., vol. 7, no. 2, pp. 80–88,Mar./Apr. 2001.

[3] R. C. Dugan and S. K. Price, BIncludingdistributed resources in distributionplanning,[ in Proc. IEEE PES Power Syst.Conf. Expo., 2004, pp. 1694–1698.

[4] E. J. Coster and D. van Houwelingen,BIntegration of DG in MV-grids: Challengesencountered by the grip operator,[ in Proc.CICRE/IEEE Joint Symp. Integr. Wide-ScaleRenewable Resources into the Power DeliverySyst., Calgary, AB, Canada, 2009.

[5] T. Ackermann and V. Knyazkin, BInteractionof distributed generation and the distributionnetwork: Operation aspects,[ in Proc. PEST&D Conf., Asia Pacific, 2002, vol. 2,pp. 1357–1362.

[6] J. Morren and S. W. H. de Haan, BImpactof distributed generation units with powerelectronic converters on distribution networkprotection,[ in Proc. IET 9th Conf. Develop.Power Syst. Protection, Glasgow, Scotland,2008, pp. 664–669.

[7] C. L. Masters, BVoltage rise the big issue whenconnecting embedded generation to long11 kV overhead lines,[ Power Eng. J., vol. 16,no. 1, pp. 5–12, Feb. 2002.

[8] R. O’ Gorman and M. A. Redfern, BVoltagecontrol problems on modern distribution

systems,[ in Proc. Power Eng. Soc. GeneralMeeting, 2004, vol. 1, pp. 662–667.

[9] N. Jenkins, R. Allen, P. Crossley, D. Kirchen,and G. Strbac, Embedded Generation,vol. 31, 1st ed. London, U.K.: IET, 2000,ser. Power & Energy.

[10] B. Fox, D. Flynn, L. Bryans, N. Jenkins,D. Milborrow, M. O’Malley, R. Watson, andO. Anaya-Lara, Wind Power integration,Connection and System Operational Aspects,vol. 50, 1st ed. London, U.K.: IET, 2007,ser. Power & Energy.

[11] J. H. R. Enslin and P. J. M. Heskes,BHarmonic interaction between largenumbers of Photovoltaic inverters and thedistribution network,[ IEEE Trans. PowerElectron., vol. 19, no. 6, pp. 1586–1593,Nov. 2004.

[12] P. M. J. Heskes, J. F. G. Cobben, andH. H. C. de Moor, BHarmonic distortionin residential areas due to large scalePV implementation is predicable,[J. Distrib. Energy Resources, pp. 17–32,vol. 1, no. 1, Jan.–Mar. 2005.

[13] P. M. J. Heskes, J. M. A. Myrzik, andW. L. Kling, BSurvey of harmonic reductiontechniques applicable as ancillary service ofdispersed generation,[ in Proc. IEEE YoungRes. Symp., Eindhoven, The Netherlands,Feb. 7–8, pp. 1–4.

[14] J. F. G. Cobben, BPower quality, implicationsat the point of connection,[ Ph.D. thesis,Dept. Electr. Eng., Univ. TechnologyEindhoven, Eindhoven, The Netherlands,2008.

[15] K. Kauhaniemi and L. Kumpulainen, BImpactof distributed generation on the protectionof distribution networks,[ in Proc. 8th IEE

Conf. Develop. Power Syst. Protection, 2004,pp. 315–318.

[16] L. K. Kumpulainen and K. T. Kauhaniemi,BAnalysis of the impact of distributedgeneration on automatic reclosing,[ inProc. IEEE Power Eng. Soc. Power Syst. Conf.Exhib., 2004, pp. 603–608.

[17] S. M. Brahma and A. A. Girgis,BDevelopment of adaptive protectionscheme for distribution systems with highpenetration of distributed generation,[IEEE Trans. Power Delivery, vol. 19, no. 1,pp. 56–63, Jan. 2004.

[18] S. M. Brahma and A. A. Girgis,BMicroprocessor-based reclosing tocoordinate fuse and recloser in a systemwith high penetration of distributedgeneration,[ in Proc. IEEE Power Eng. Soc.Winter Meeting, 2002, vol. 1, pp. 453–458.

[19] P. M. Anderson, Power System Protection.New York: IEEE, 1998.

[20] E. J. Coster, J. M. A. Myrzik, and W. L. Kling,BEffect of distributed generation on protectionof medium voltage cable grids,[ in Proc.CIRED 19th Int. Conf. Electricity Distrib.,2007, pp. 1–6.

[21] E. J. Coster, J. M. A. Myrzik, and W. L. Kling,BIntegration of distributed generation inmedium voltage gridsVProtection issues,[Int. J. Distrib. Energy Resources, vol. 5, no. 3,pp. 167–186, Jul. 2009.

[22] K. Maki, S. Repo, and P. Jarventausta, BEffectof wind power based distributed generationon protection of distribution networks,[ inProc. 8th IEE Int. Conf. Develop. Power Syst.Protection, 2004, vol. 1, pp. 327–330.

ABOUT THE AUT HORS

Edward J. Coster (Student Member, IEEE) was born in Leiden, The

Netherlands, in 1972. He received the B.Eng. degree in electrical engi-

neering from Technische Hogeschool Rijswijk (TH Rijswijk), Rijswijk, The

Netherlands, in 1997 and the M.Sc. degree in electrical engineering

from Delft University of Technology, Delft, The Netherlands, in 2000. In

April 2006, he joined part-time the Electrical Power System group,

Eindhoven University of Technology, Eindhoven, The Netherlands, to start

a Ph.D. research project.

Since 2000, he has been a Senior Specialist for Network Planning with

STEDIN, Rotterdam, The Netherlands. His fields of interest are: distrib-

uted generation, power system protection, dynamic behavior, and

stability of power systems.

Johanna M. A. Myrzik was born in Darmstadt, Germany, in 1966. She

received the M.Sc. degree in electrical engineering from Darmstadt

University of Technology, Darmstadt, Germany, in 1992 and the Ph.D.

degree in the field of solar inverter topologies from Kassel University,

Kassel, Germany, in 2000.

From 1993 to 1995, she was a Researcher at the Institute for Solar

Energy Supply Technology (ISET e.V.), Kassel, Germany. Since 2000, she

has been with the Eindhoven University of Technology, Eindhoven, The

Netherlands. In 2002, she became an Assistant Professor and in 2008 an

Associate Professor in the field of residential electrical infrastructures.

Her fields of interests are: power electronics, renewable energy,

distributed generation, and electrical power supply.

Table 7 Load Data



Bas Kruimer (Member, IEEE) was born in Curaçao, The Netherlands

Antilles, in 1963. He studied power engineering at the Delft Technical

University, Delft, The Netherlands, graduating in 1988.

He worked for ABB in substation automation, protection, and network

control nationally and internationally. In 2002, he joined KEMA T&D

Consulting leading the design and engineering team. In 2006, he led the

Quality Management Systems business of KEMA Quality, and in early

2008, he joined the Infra company of Dutch Utility Eneco, today named

Joulz. In August 2010, he joined Quanta Technology, Rotterdam, The

Netherlands. During his professional years, he has contributed to Dutch

Cigre and to IEEE and has been an active player in the IEC 61850

developments for many years.

Wil L. Kling (Member, IEEE) was born in Heesch, The Netherlands, in

1950. He received the M.Sc. degree in electrical engineering from the

Technical University of Eindhoven, Eindhoven, The Netherlands, in 1978.

From 1978 to 1983, he worked with KEMA and from 1983 to 1998 with

Sep. Then, he was with TenneT, the Dutch Transmission System Operator,

as a Senior Engineer for Network Planning and Network Strategy. In 1993,

he joined the Delft University of Technology, Delft, The Netherlands, as a

part-time Professor, and in 2000, he joined the Electrical Power Systems

(EPS) group at the Eindhoven University of Technology, Eindhoven, The

Netherlands, as a part-time Professor, where at the end of 2008, he was

appointed a Full-Time Professor and Chair of the EPS group. He leads

research programs on distributed generation, integration of wind power,

network concepts, and reliability.

Mr. Kling is involved in scientific organizations such as Cigre and IEEE.

He is the Dutch representative in the Cigre Study Committee C6

Distribution Systems and Dispersed Generation.



Integration of distributed generation to Grid

Documents