1

Abstract-- This paper proposes a comprehensive approach to the planning of distribution networks and the control of microgrids. Firstly, a Modified Discrete Particle Swarm Optimization (MDPSO) method is used to optimally plan a distribution system upgrade over a 20 year planning period. The optimization is conducted at different load levels according to the anticipated load duration curve and integrated over the system lifetime in order to minimize its total lifetime cost.

Since the optimal solution contains Distributed Generators (DGs) to maximize reliability, the DG must be able to operate in islanded mode and this leads to the concept of microgrids. Thus the second part of the paper reviews some of the challenges of microgrid control in the presence of both inertial (rotating direct connected) and non-inertial (converter interfaced) DGs. More specifically enhanced control strategies based on frequency droop are proposed for DGs to improve the smooth synchronization and real power sharing minimizing transient oscillations in the microgrid. Simulation studies are presented to show the effectiveness of the control.

Index Terms — Microgrid, intelligent control, power system

planning & control.

I. INTRODUCTION HE planning of a reliable distribution network that satisfies the annual load growth for the planning period is

a significant issue for distribution network companies striving to survive in the competitive electricity market [1]. The resulting distribution network upgrade may consist of installation of new substations, the upgrade of existing power system infrastructure (substations & lines) and/or the installation of Distributed Generation (DG).

If DG forms part of the solution, then to maximize the system reliability, the DG should be able to operate in an islanded mode. In the islanded mode of operation the DG supplies the local load and is disconnected from the main supply grid. This leads to the concept of a microgrid [2] where a number of DGs can supply local loads in an autonomous manner.

This paper considers two main aspects: • The optimal planning of distribution systems. • The control of the resulting DG (microgrid) network.

The work reported in this paper was supported by the Australian Research Council and the CSIRO Energy Transformed flagship.

1 M. Wishart, M. Dewadasa, I. Ziari, G. Ledwich & A. Ghosh are with Queensland University of Technology, Brisbane, Australia (e-mails: mike.wishart@gmail.com, i.ziari@qut.edu.au, j.dewadasa@qut.edu.au, a.ghosh@qut.edu.au, g.ledwich@qut.edu.au)

Firstly, an integrated planning method is presented for determining the optimal solution for rural distribution systems given a certain load duration curve and growth parameters. The method considers DGs, capacitors and transformer and line upgrades in its solution to minimize the total lifetime cost of the system. Since all the solution technologies are available in discrete ratings, the planning problem is discrete in nature which causes the objective function to have a number of local minima. Heuristic methods are commonly used for solving this type of optimization problem. This paper employs Particle Swarm Optimization (PSO) as a heuristic method [3, 4]. Due to the discrete nature of the objective function, a Discrete PSO (DPSO) is used. To further mitigate the local minimum problem, a Modified DPSO (MDPSO) is presented in this paper. In this approach, the diversity of the optimizing variables is increased using the Genetic Algorithm (GA) mutation and crossover operators to avoid trapping in local minima. The objective is to minimize the lifetime cost over the planning period including line loss and the reliability costs along with the investment needed in DGs, capacitors, lines, and transformers. The bus voltage, feeder current, and the DG output power are incorporated in the optimization procedure as constraints.

Secondly, since the solution contains DG, issues for control of a microgrid need to be addressed. Power management strategies are vital for an autonomous microgrid in the presence of multiple DG units, where no single dominant energy source is present to supply the energy requirement [5]. Additionally, fast and flexible power control strategies are necessary to damp out transient power oscillations [6]. Many researchers have addressed the operational, control and protection issues in microgrids [7-11]. The real and reactive power output of a generator can be independently controlled by changing the voltage angle (based on frequency) and the magnitude respectively [12, 13]. Therefore, frequency and voltage droop controls are the most common methods used to share the real and reactive load power in a microgrid. In this paper, following issues are addressed under microgrid operation, control and protection.

1. Steady state control and power sharing. 2. Transients between converter interfaced DGs. 3. Dynamic power sharing amongst different DG types. 4. Incorporation of non-schedulable (renewable) DGs. 5. Protection of microgrid in grid connected & islanded

modes of operation.

Intelligent Distribution Planning and Control incorporating Microgrids

Michael T. Wishart, Member, IEEE, Manjula Dewadasa, Member, IEEE, Iman Ziari, Student Member, IEEE, Gerard Ledwich, Senior Member, IEEE, Arindam Ghosh, Fellow, IEEE1

T

978-1-4577-1002-5/11/$26.00 ©2011 IEEE

2

Decentralized control amongst DG sources is implemented as a simple and cost effective solution. Each DG has its own local control for connection and disconnection from microgrid, and controlling the real and reactive power. Both inertial & non-inertial (converter-interfaced) DGs are considered.

II. OPTIMAL DISTRIBUTION NETWORK PLANNING

A. Problem Formulation The main objective of the Optimal Distribution Network

Reinforcement (ODNR) problem is to maintain (or improve) voltage profile and avoid overloading subject to an increasing load, all at a minimum lifetime cost. The lifetime cost includes investment in DGs, capacitors, lines and transformers as well as operation & maintenance (O&M), losses and reliability cost. As constraints, the bus voltage should be kept within the standard range and the feeder current should be maintained lower than the rated current.

Given that all of the objective function elements are simply converted into a composite equivalent cost, this problem can be solved using a single-objective optimization method. This objective function is defined as follows:

DP)CCCCC()r1(

1OFY

0y

yES

yR

yL

yM&O

yCAPy +++++

+=∑

=

(1)

where OF is the objective function which is the net present value of the total cost, CCAP is the total capital cost, CO&M is the total operation and maintenance cost, CL is the loss cost, CR is the reliability cost, CES is the energy saving resulted by installing DGs, r is the discount rate, Y is the number of planning intervals in the study timeframe, and DP is the constraint penalty factor. The reliability cost is calculated in equation (2) below. This cost is based on the total unsupplied demand after an outage:

∑=

××=LL

0ll

yllllNS

yR DNSTkC (2)

where yRC is the reliability cost in the planning interval y, kNS

is the cost of energy not supplied ($/MWh), and yllDNS is the

demand not supplied in load level ll in planning interval y. The energy saving (CES) is calculated in a similar manner to the line loss cost.

B. Methodology A novel sequential technique is proposed in this paper to

solve the ODNR problem for a planning time framework as the load demand is growing. The proposed algorithm is shown in Fig. 1.

As observed in this flowchart, the algorithm is composed of two parts. In the first part, an initialization is performed. The second part is the main body of the procedure. Further details can be found in [15].

Fig. 1. Flowchart of the proposed technique.

C. Implementation of MDPSO

PSO was introduced originally by Kennedy and Eberhart in 1995 [14]. In this optimization method, the algorithm uses a population of individuals in parallel to search possible areas of a multi-dimensional space for the optimal solution. The individuals are called particles and the population is called a swarm [4].

As the discrete version of PSO, DPSO is applied to discrete problems like ODNR where part or all of the particles have an integer value. During this procedure, the optimal solution can be achieved by rounding off the actual particle value to the nearest integer value during the iterations. In [4], it is mentioned that this rounding off process does not influence the performance of DPSO. The particle which is composed of the optimizing variables was chosen as shown in Fig. 2.

The discrete nature of the planning problems like ODNR causes a number of local minima in the objective function. As a result of this, a modification is applied to the employed DPSO to avoid trapping in local minima. For this purpose, the diversity of the optimizing variables is increased by employing GA mutation and crossover operators [15] to 50% of the population. This reference [15] also indicates that this optimization method is more robust and accurate compared with some other optimization methods, such as pure DPSO, GA, Simulated Annealing, and Discrete Nonlinear Programming.

3

Fig. 2. The structure of a particle.

D. Results The 18-bus IEEE distribution system [16, 17] is used to

validate the proposed optimization method. A five load level characteristic is used to approximate the load duration curve in this paper, helping to decrease the computation time. It is assumed that the load is at the peak level for 0.05% of the total time. The load level is 81.25% of the peak level for 4.95% of the total time, 62.5% for 35% of the total time, 50% for 35% of the total time, and 37.5% for 25% of the total time. The characteristics of the test system are listed in Table I.

TABLE I CHARACTERISTICS OF THE TEST SYSTEM

Parameter Value DG Installation Cost $50000+$550/kVA DG O&M Cost ¢11.4/kWh DG Base Unit 300 kVA Capacitor Installation cost $3000+$35/kvar Capacitor O&M Cost $1/kvar Capacitor Base Unit 300 kvar Line Upgrading Cost (120000+30000×ΔLT)/km Line O&M Cost $2000/km Load Growth for Peak Load Level 40% per planning interval

Load Growth for Other Load Levels 13% per planning interval

r 0.07 Y 20 years (4 planning intervals) Failure Rate 0.01 (fault/km.yr) DG Time 30 minutes Switching Time 30 minutes Repair Time 180 minutes

The prices in this table are in the same range as [18-20] and

show a similar trend to the generating electricity cost from different types of technology in the UK [21]. The transformer upgrade cost is calculated based on a constant cost (e. g. labor cost) which is assumed to be $100000 and an additional cost based on the installation cost of the new and old HV/MV transformers and related facilities. It is assumed that the old transformer and associated facilities can be sold at half price.

It is assumed that the loads, located in the distribution network, are growing every five years during the planning period. Therefore, each planning interval is 5 years and four periods, each lasting 5 years, are defined (i.e. giving a 20 year lifetime). The results corresponding to this scenario are summarized in Figs. 3 & 4. Fig. 3 shows the test system configuration after planning in the last period. The red elements in this figure illustrate the optimal location of capacitors and DGs and the required line upgrades in the last

period of planning. Fig. 4 summarizes the incremental costs for each planning period. It can be seen that the DGs are added in the first planning period, while the transformer is upgraded in period 3, lines are upgraded in periods 3 & 4 and capacitors are added in all periods. The reliability cost indicates that the reliability and losses are maintained reasonably constant throughout the 20 year planning horizon.

Fig. 3. The test system configuration after planning in the last time interval.

Line Transformer Capacitor DG Loss Reliability Energy Saving10

3

104

105

106

107

Cost Elements

Cos

t ($)

Period 4 Period 3 Period 2 Period 1

Fig. 4. A summary of the optimization results.

The optimal solution clearly includes DGs. To maximize

reliability, these DGs must be able to operate in autonomously as a microgrid. Indeed, this has been implicit in the calculation of the reliability cost. The next section explores the issues related to autonomous microgrid operation, control and protection. Control strategies are proposed using frequency and voltage droop control to improve the dynamic power sharing in a hybrid microgrid.

III. OPERATION, CONTROL AND PROTECTION OF MICROGRIDS

A. Steady State Control and Power Sharing Real and reactive load power sharing amongst DGs in a

microgrid can be achieved using frequency and voltage droop control respectively. In a conventional frequency droop control, each DG uses the frequency at its point of connection (PC) to determine the amount of real power injection required. Thus, the system frequency will act as the communication signal amongst the DGs to share the real power appropriately.

4

The conventional frequency droop characteristic can be expressed as [12, 22-23]

)( ∗∗ −×+= PPmff rr (3)

where f* is the instantaneous frequency setting for a DG considered, fr is the rated frequency of the system, Pr is the rated real power output of the generator and P* is the measured actual real power output of the DG. The droop coefficient is denoted by m.

Moreover, the output voltage magnitude of a DG can be controlled to change the reactive power supplied to the system. However, in the presence of multiple DGs, maintaining a voltage to a pre-defined value can cause reactive power circulation amongst the sources. The best solution to solve this problem is to implement voltage droop control in DGs. This voltage droop results in reactive load power sharing in the microgrid. The conventional voltage droop control characteristic is given by [12, 22]

)( ∗∗ −×+= QQnVV rr (4) where V* is the instantaneous voltage magnitude setting, Vr is the rated voltage of the microgrid system, Qr is the rated reactive power output of the generator and Q* is the measured actual reactive power output. The voltage droop coefficient is denoted by n.

B. Transient between converter interfaced DGs High gain in droop control improves the power sharing

however can lead to transient oscillations. To minimize the transient oscillations, it is desirable to have low transient gains. Angle based droop control is proposed in [24] for a microgrid which consists of several converter interfaced DGs. The author compared the performance of angle droop to conventional frequency droop and it was shown that angle droop control can minimize the real power fluctuations during load changes in a microgrid. The proposed angle droop can be applied to a converter interfaced DG microgrid to share the real power amongst DGs. Power sharing accuracy can be increased by selecting the output inductance of converters to be inversely proportional to DG rating.

C. Dynamic Power Sharing Amongst Different DG Types The limitations of using conventional frequency droop

have been investigated by previous researchers. Several drawbacks can be identified using conventional frequency droop such as slow transient response, frequency and amplitude deviations, and high dependency on the converter output impedance [25]. System stability is one of the major concerns in microgrids with higher penetration levels of DGs, especially if high feedback gains are used to achieve proper power sharing [23]. Also, both inertial and non-inertial DGs can be present in a microgrid. However, the dynamic response

of inertial and non-inertial DGs is different. The non-inertial DGs, connected through converters, can have a significantly faster response than inertial DGs. This can lead to dynamic sharing problems and further oscillations. To minimize these, DGs are required to respond in a similar rate during a transient [23].

A modified droop control characteristic is proposed to improve the dynamic power sharing of a microgrid containing both inertial and non-inertial DGs. This ensures that the change of load is proportionally picked up by all the DGs. The proposed droop control is called as “integral to system droop line” and only implemented on converter interfaced DGs in the microgrid. In the proposed integral to system droop line control, steady state gain and transient gain are able to be set independently using an integral controller. Thus, system can respond with a medium gain during a transient event but reach a steady state point corresponding to a high gain. Once an appropriate time constant is selected for the integrator, converter interfaced DGs can respond in a similar manner to inertial DGs. This results in a smooth transition to system steady state. To implement the proposed droop, the frequency droop in (3) is modified by introducing an integration process to the DG to reach the steady state frequency droop point in the system. The error between calculated droop frequency in (3) and frequency at the PC is passed through an integrator to force the operating frequency of DG to reach the steady state droop point within a defined time period. The proposed method has the ability not only to minimize transient instability but also improve the proper power sharing amongst DG sources. Moreover, the proposed droop allows using high gain in steady state droop since it is providing sufficient flexibility avoiding transient instability. The proposed droop control is given by

∫ −+= ∗∗ dtffff pcd )( (5)

where fd is the modified droop frequency for the DG, f* is the droop given in (3) and fpc is the frequency at PC. The time constant of the integrator is selected according to the inertial DG dynamics (i.e., time constant of governor) to ensure a similar response from the non-inertial DGs in the system. However, it is to be noted that real power injection to the system can be controlled by changing the output voltage angle of a converter. Therefore, an angle (φ ) corresponding to the frequency deviation (i.e., the amount of real power required to inject into the system) given by (5) is calculated and used in reference generation to the converter. For example, if output feedback voltage control is used to control three phase converters, the reference voltages for three phases are generated using voltage magnitude obtained from voltage droop and calculated angle corresponding to droop frequency in (5). In this case, the reference for phase A can be generated as

)2( φπ += tfSinVV pcma (6)

5

where Vm is the voltage magnitude calculated from the voltage droop in (4), fpc is the PC frequency obtained from a phase locked loop (PLL) and φ is the angle corresponding to droop frequency fd in (5). All the converter interfaced DGs are controlled using the proposed modified droop control to enhance better dynamic power sharing amongst inertial and non-inertial sources in a microgrid during a transient event.

To demonstrate the concepts, consider the microgrid system shown in Fig. 5, which includes both inertial and non-inertial sources. The dynamic power sharing before and after implementing integral to system droop line control in converter interfaced DGs is investigated. It is assumed that DG1 is a diesel generator (i.e., an inertial), while DG2 and DG3 are converter interfaced DGs (i.e., non-inertial). Three impedance type loads, load1, load2 and load3 are connected at buses 2, 4 and 6 respectively. The system parameters are given in Table II.

Fig. 5. Diesel generator with two converter interfaced DGs

TABLE II SYSTEM PARAMETERS

System data Value System frequency 50 Hz System voltage 0.415 kV rms (L-L) DG1 (diesel) power rating (12 + j 8) kVA DG2 power rating (15 + j 10) kVA DG3 power rating (10 + j 6.7) kVA Feeder impedance (Z12=Z23) (0.025+ j 1.2566) Ω Load impedance load1=load2 (15+ j 11.781) Ω load3 (5+ j 3.92) Ω Frequency droop coefficient DG1 m1=0.0417 (Hz/kW) DG2 m2=0.0333 (Hz/kW) DG3 m3=0.050 (Hz/kW) Voltage droop coefficient DG1 n1=1.5 (V/kVAR) DG2 n2= 1.2 (V/kVAR) DG3 n3=1.79 (V/kVAR)

As a reference, the system response is investigated by implementing the conventional frequency droop in all DGs during synchronization and load changes. It is assumed that DG1 is connected to the system supplying load1 and load2 operating in conventional frequency and voltage droop in (3) and (4). DG2 and DG3 are then synchronized to the system at 3.5 s and 6.5 s respectively. Subsequently, load3 is connected to the system at 9 s. Finally, load1 and load3 are disconnected from the system at 12 s.

The real and reactive load power sharing amongst DGs is shown in Fig. 6. DG2 and DG3 start to inject real power after their connection. The results show frequency and real power

fluctuations during both DG synchronization and load changes.

Fig. 6. Real and reactive power sharing with conventional droop.

Fig. 7. The variation of PC frequency of each DG.

These oscillations are due to the different dynamic response

of inertial and non-inertial DGs. Thus, the system takes several cycles to reach steady state. The variation of the PC frequency of each DG is shown in Fig. 7.

Next, the same system is simulated by applying the proposed droop in (5) to converter interfaced DGs. The response of the DGs is shown in Fig. 8. It is evident that there are no power oscillations during the DG synchronization and load changes in the system. Moreover, the real power outputs of DGs are proportional to the generator ratings as expected from the droop coefficient selection. Thus, the power sharing amongst DGs is very accurate. The variation of system frequency during this simulation is also shown in Fig. 9. This shows that no large frequency oscillations are observed in the system.

It can therefore be concluded that in the presence of both inertial and non-inertial sources in a microgrid, conventional frequency droop can initiate frequency and real power oscillations during synchronization and load changes. However these oscillations can be avoided by implementing integral to system droop line control in converter interfaced DGs. This results in improved dynamic power sharing in the microgrid while at the same time maintaining accurate steady state sharing.

6

Fig. 8. Real and reactive power sharing with integral to system droop line.

Fig. 9. The variation of system frequency.

D. Incorporation of Non-schedulable DGs In this section, power management and control strategies

required to incorporate non-schedulable (renewable energy based) DGs and battery storage into a microgrid are discussed. The microgrid can include diesel generator(s), battery storage (BS), wind and solar PVs. The control of microgrid should enable the plug and play capability of DG sources, thus maximizing the benefits of renewable based energy sources. Decentralized control amongst DG sources is proposed as a simple and cost effective solution. Each DG has its own local control for connection and disconnection from the microgrid, and for controlling the real and reactive power output.

The BS is connected to the microgrid through a converter ensuring bidirectional power flow between microgrid and battery. Therefore, the BS can act as either a load or a source to absorb or inject real power into the microgrid. Also, the BS can assist in controlling the microgrid frequency. Moreover, the converter associated with BS has the ability to regulate the voltage at PC by injecting reactive power into the microgrid. The converter rating determines the maximum reactive power injection capacity into the system. The BS is employed with an intelligent control system (ICS) (or battery management system (BMS)) to manage the power effectively. The ICS in BS is continuously monitoring the state of charge (SOC) of the battery. If the battery is not fully charged and there is surplus power in the microgrid, the surplus power is used to charge the batteries. The battery storage can be controlled as

“operating reserve” to supply or absorb any transient power during changes in generation or loads within the energy limits. For example, when the load changes in the microgrid, the BS can react very quickly to match the load power change. The ICS is responsible for managing the operating reserve in the battery and controlling the battery charging and discharging.

The DGs connected through wind and PVs are controlled using maximum power point tracking (MPPT) to enhance the benefits of renewable energy sources. Therefore, any deficit in load power is supplied by other dispatchable sources (i.e., diesel, BS) operating in frequency and voltage droop control. Voltage control of each dispatchable DG and voltage droop amongst DGs ensures the voltage regulation, stability and proper reactive power sharing, avoiding reactive power circulation in the microgrid. The proposed frequency droop lines for BS and diesel generator are defined to ensure the battery is charged when there is excess power available in the microgrid.

The droop lines for the diesel generator and BS are shown by DE and AFG respectively in Fig. 10. The line segment AF represents the droop for battery charging while the droop for battery discharging is represented by line FG. According to the droop lines shown in Fig. 10, BS starts to supply the load power once the diesel generator reaches its maximum power output at rated frequency. However, it is to be noted that slope of the droop line is controlled by the ICS and it can be changed towards points H or K. The reason for proposing adaptive droop slope is to give ICS an opportunity to enhance the flexibility of control of the BS in the microgrid. For example, consider the power sharing in the presence of a few BS systems in a microgrid. In this circumstance, the slope of the droop line can be changed according to SOC of the batteries to enable the power sharing effectively since power sharing according to each BS converter rating is not viable. These control actions can be embedded in the ICS to respond whenever required. However, when there is a surplus of generated power (i.e. as determined using the system operating frequency), the BS can be charged. During this charging, the slope of the droop is selected appropriately by the ICS. The droop line is determined using the microgrid frequency, the required battery charging power (or current) and the time of day.

Fig. 10. Frequency droop characteristics for BS and diesel generator.

7

The main aim of proposing an adaptive droop for charging

is to manage the battery charging effectively maintaining the microgrid stability. The advantages of having adaptive droop for charging are listed below. (i) The output power of wind and PV fluctuates with time, is

difficult to predict and these intermittent sources are controlled in MPPT. If wind and PV start to inject more power into microgrid than the loads require, the system frequency (fsys) starts to rise. In this circumstance, the excess power can be used to charge the battery appropriately by selecting the appropriate slope for the droop. However, once the battery is fully charged, the ICS changes the droop line such that power absorption is zero. If the frequency increases further as a result of additional power generation, generation shedding is implemented to manage the frequency rise and stability of the microgrid.

(ii) The ICS of the BS continuously monitors the state of charge and the time of day. If the peak load demand is about to occur and the battery is not fully charged (to the operating reserve level) the excess power available from the diesel generator can be used to rapidly charge the battery. To increase the rate of battery charging, the droop line should be moved towards point B in Fig. 10. However, the maximum charging current will be limited by the converter rating. Alternatively, the rate of battery charging can be decreased by moving the droop line towards point C if required. It is to be noted that battery charging should not violate the maximum charging current given by battery manufacturer and charging should be carried out according to specifications (i.e., constant current and voltage). Also, the ICS changes the droop line if the SOC becomes low allowing system frequency to drop further triggering non-critical (prioritized) frequency based load shedding.

E. Protection of Microgrid in Grid Connected & Islanded Modes of Operation

Overcurrent (OC) protection is conventionally employed to protect radial distribution networks due to its simplicity and low cost [26, 27]. However, if DGs are added, several protection issues can arise and are well documented [28-32]. The current practice of automatic DG disconnection for every fault during loss of main grid supply drastically reduces the DG benefits [33]. Also, the islanded operation with DGs is usually not allowed since restoration by reclosing is difficult and power quality within the islanded section cannot be guaranteed [30]. However, to enhance the benefits of optimal distribution network planning as explained in Section II, protection issues should be solved, thus allowing both grid connected and islanded mode operations.

Several protection schemes have been proposed and most of them need a reliable communication medium. In [31], protection algorithms are proposed using a neural network for a DG connected network to locate a fault and isolate the faulty zone. A protection scheme based on communication to a multi-source distribution system has been proposed in [34] while the sign of wavelet coefficients of the fault current

transient is used to locate and isolate a faulted segment in [35]. A new inverse time admittance (ITA) relay characteristic is proposed for DG connected networks in [36]. The ITA relay is capable of operating without any communication between relays in the presence of different fault currents, thus providing the required protection for both grid connected and islanded mode of operation.

To enhance the proposed planning in Section II and proposed control strategies in Section III, a reliable protection scheme can be implemented considering the aforementioned solutions.

IV. CONCLUSION

This paper has presented an optimal approach for planning distribution networks including Distributed Generation (DG). A Modified Discrete Particle Swarm Optimization (MDPSO) algorithm has been utilized to optimize the network over its lifetime with consideration for the anticipated future load duration curve. The resulting optimal solution is shown to contain multiple DGs from the first 5 year planning period onwards. To maximize reliability, some DGs must be operated in islanded mode and this leads to the concept of a microgrid. The paper continues by investigating operational issues of microgrids. An “integral to system droop” control scheme is proposed to achieve desirable steady state and dynamic performance. Simulation results are presented to demonstrate the effectiveness of the scheme in improving the dynamic stability and damping system transients. A methodology for incorporating renewable energy sources is presented using the same form of droop control. Finally protection issues are briefly discussed and a new inverse time admittance relay is recommended.

Optimally designed distribution systems invariably contain microgrids. This paper has shown pathways for successful operation of microgrids including incorporation of renewable sources.

V. REFERENCES [1] W. El-Khattam, Y. G. Hegazy and M. M. A. Salama, “An Integrated

Distribution Generation Optimization Model for Distribution System Planning”, IEEE Trans. Power Systems, Vol. 20, No. 2, pp. 1158-1165, May 2005.

[2] R. H. Lasseter, "Certs Microgrid," in Proc. IEEE International Conf. on System of Systems Engineering, pp. 1-5, 2007.

[3] M. R. Alrashidi and M. E. El-Hawary, “A Survey of Particle Swarm Optimization Applications in Electric Power Systems”, IEEE Trans. Evolutionary Computation, Vol. 13, No. 4, pp. 913-918, August 2009.

[4] Y. Del Valle, G. K. Venayagamoorthy and S. Mohagheghi, “Particle Swarm Optimization: Basic Concepts, Variants and Applications in Power Systems”, IEEE Trans. Evolutionary Computation, Vol. 12, No. 2, pp. 171-195, April 2008.

[5] F. Katiraei and M. R. Iravani, "Power management strategies for a microgrid with multiple distributed generation units," IEEE Trans. Power Systems, vol. 21, pp. 1821-1831, Nov. 2006.

[6] M. Shahabi, M. R. Haghifam, M. Mohamadian, and S. A. Nabavi-Niaki, "Dynamic Behavior Improvement in a Microgrid with Multiple DG Units Using a Power Sharing Approach," in Proc. IEEE Bucharest Power Tech Conference, 2009.

[7] T. C. Green and M. Prodanović, "Control of Inverter-based Micro-grids," Electric Power Systems Research, vol. 77, pp. 1204-1213, July 2007.

8

[8] Z. H. Jiang and X. W. Yu, "Active Power - Voltage Control Scheme for Islanding Operation of Inverter-Interfaced Microgrids," in Proc. IEEE Power & Energy Society General Meeting, pp. 1738-1744, 2009.

[9] F. Katiraei, R. Iravani, N. Hatziargyriou, and A. Dimeas, "Microgrids management," IEEE Power & Energy Magazine, vol. 6, pp. 54-65, 2008.

[10] J. A. P. Lopes, C. L. Moreira and A. G. Madureira, "Defining Control Strategies for MicroGrids Islanded Operation," IEEE Trans. Power Systems, vol. 21, pp. 916-924, 2006.

[11] H. H. Zeineldin, E. F. El-Saadany and M. M. A. Salama, "Distributed Generation Micro-Grid Operation: Control and Protection," in Proc. IEEE Power Systems Conf.: Advanced Metering, Protection, Control, Communication, and Distributed Resources, 2006.

[12] K. D. Brabandere, B. Bolsens, J. V. d. Keybus, A. Woyte, J. Driesen, and R. Belmans, "A Voltage and Frequency Droop Control Method for Parallel Inverters," IEEE Trans. Power Electronics, vol. 22, pp. 1107 - 1115, 2007.

[13] L. Zhang, L. Harnefors and H.-P. Nee, "Power-Synchronization Control of Grid-Connected Voltage-Source Converters," IEEE Trans. on Power Systems, vol. 25, pp. 809 - 820, 2010.

[14] J. Kennedy and R. Eberhart, “Particle Swarm Optimization”, in Proc. IEEE International Conf. on Neural Networks, pp. 1942-1948, 1995.

[15] I. Ziari, G. Ledwich, A. Ghosh, D. Cornforth and M. Wishart, “Optimal allocation and sizing of capacitors to minimize the transmission line loss and to improve the voltage profile”, Computers & Mathematics with Applications, Vol. 60, No. 4, pp. 1003-1013, August 2010.

[16] M. A. S. Masoum, M. Ladjevardi, A. Jafarian and E. F. Fuchs, “Optimal Placement, Replacement and Sizing of Capacitor Banks in Distorted Distribution Networks by Genetic Algorithms”, IEEE Trans. Power Delivery, Vol. 19, No. 4, pp. 1794-1801, October 2004.

[17] M. Ladjevardi and M. A. S. Masoum, “Genetically Optimized Fuzzy Placement and Sizing of Capacitor Banks in Distorted Distribution Networks”, IEEE Trans. Power Delivery, Vol. 23, No. 1, pp. 449-456, January 2008.

[18] S. Segura, R. Romero and M. J. Rider, “Efficient Heuristic Algorithm Used for Optimal Capacitor Placement in Distribution Systems”, International Journal of Electrical Power and Energy Systems, Vol. 32, No. 1, pp. 71-78, January 2010.

[19] T. L. Huang, Y. T. Hsiao, C. H. Chang and J. A. Jiang, “Optimal Placement of Capacitors in Distribution Systems Using an Immune Multi-Objective Algorithm”, International Journal of Electrical Power and Energy Systems, Vol. 30, No. 3, pp. 184-192, March 2008.

[20] I. C. Da Silva, S. Carneiro and E. J. De Oliviera, “A Heuristic Constructive Algorithm for Capacitor Placement on Distribution Systems”, IEEE Trans. Power Delivery, Vol. 23, No. 4, pp. 1619-1626, November 2008.

[21]http://www.raeng.org.uk/news/publications/list/reports/Cost_Generation_Commentary.pdf, accessed January 2011.

[22] M. C. Chandorkar, D. M. Divan and R. Adapa, "Control of parallel connected inverters in standalone AC supply systems," IEEE Trans. Industry Applications, vol. 29, pp. 136-143, 1993.

[23] R. Majumder, F. Shahnia, A. Ghosh, G. Ledwich, M. Wishart, and F. Zare, "Operation and Control of a Microgrid Containing Inertial and Non-Inertial Micro Sources," in Proc. IEEE Region 10 Conference (Tencon), Singapore, 2009.

[24] R. Majumder, A. Ghosh, G. Ledwich, and F. Zare, "Angle Droop Versus Frequency Droop in a Voltage Source Converter Based Autonomous Microgrid," in Proc. IEEE Power & Energy Society General Meeting, 2009.

[25] J. M. Guerrero, N. Berbel, L. G. de Vicuna, J. Matas, J. Miret, and M. Castilla, "Droop control method for the parallel operation of online uninterruptible power systems using resistive output impedance," in Proc. IEEE Applied Power Electronics Conf. and Exposition, 2006.

[26] J. C. Gómez and M. M. Morcos, "Coordination of Voltage Sag and Overcurrent Protection in DG Systems," IEEE Trans. Power Delivery, vol. 20, pp. 214-218, 2005.

[27] J. Driesen, P. Vermeyen and R. Belmans, "Protection Issues in Microgrids with Multiple Distributed Generation Units," in Proc. Power Conversion Conf., pp. 646-653, 2007.

[28] A. Zamani, T. Sidhu and A. Yazdani, "A Strategy for Protection Coordination in Radial Distribution Networks with Distributed Generators," in Proc. IEEE Power & Energy Society General Meeting, 2010.

[29] S.-M. Cho, H.-S. Shin and J.-C. Kim, "Impact of Distribution System Quality on DG Interconnection Protection," in Proc. IEEE Power & Energy Society General Meeting, 2010.

[30] J. A. Martinez and J. Martin-Arnedo, “Impact of Distributed Generation on Distribution Protection and Power Quality,” in Proc. IEEE Power & Energy Society General Meeting, 2009.

[31] S. A. M. Javadian, M. R. Haghifam and N. Rezaei, "A Fault Location and Protection Scheme for Distribution Systems in Presence of DG Using MLP Neural Networks," in Proc. IEEE Power & Energy Society General Meeting, 2009.

[32] H. Cheung, A. Hamlyn, L. Wang, C. Yang, and R. Cheung, "Investigations of Impacts of Distributed Generations on Feeder Protections," in Proc. IEEE Power & Energy Society General Meeting, 2009.

[33] S. P. Chowdhury, S. Chowdhury, C. F. Ten, and P. A. Crossley, "Islanding Protection of Distribution Systems with Distributed Generators — A Comprehensive Survey Report," in Proc. IEEE Power and Energy Society General Meeting, 2008.

[34] X. Wu, Y. Lu and J. Du, "Wide-area Current Protection for Distribution Feeders with Distributed Generators," in Proc. Third International Conf. on Electric Utility Deregulation and Restructuring and Power Technologies, pp. 2558-2563, 2008.

[35] N. Perera, A. D. Rajapakse and T. E. Buchholzer, "Isolation of Faults in Distribution Networks with Distributed Generators," IEEE Trans. on Power Delivery, vol. 23, pp. 2347-2355, 2008.

[36] M. Dewadasa, A. Ghosh, and G. Ledwich, "An Inverse Time Admittance Relay for Fault Detection in Distribution Networks Containing DGs," in Proc. IEEE Region 10 Conference (TENCON), 2009.

VI. BIOGRAPHIES Michael T. Wishart (M’91) obtained a Ph.D. in Electrical Engineering from University of Natal, South Africa in 1994. Prior to joining Queensland University of Technology in 2009, he spent 15 years in industry in various Research and Development roles in the Power Electronics and Power Systems areas. This included periods working for Eskom in South Africa and for various Schneider Electric subsidiaries in the Asia Pacific region. His research interests are in Power Systems, Power Electronics and Intelligent Control. Manjula Dewadasa (S’08, M’10) is working as a research associate at Queensland University of Technology, Australia. He received the Ph.D in Electrical Engineering from Queensland University of Technology, Australia in 2010 and B.Sc. (Hons) in Electrical Engineering from the University of Peradeniya, Sri Lanka in 2005. He worked as an Electrical Engineer for two years in a leading manufacturing company before starting the PhD studies at QUT in 2007. His interests are in fault analysis, power system protection and distributed generation.

Iman Ziari (S’09) received his B.E. in Electrical Engineering from Sahand University of Technology, Iran in 2000 and his M.Sc. (Eng.) degree from Iran University of Science and Technology (IUST) in 2004. His interests are in Distribution Network Planning, Optimization and Power Quality.

Gerard Ledwich (M’73, SM’92) received the Ph.D. in electrical engineering from the University of Newcastle, Australia, in 1976. He has been Chair Professor in Power Engineering at Queensland University of Technology, Australia since 2000. His interests are in the areas of power systems, power electronics, and controls. He is a Fellow of I.E.Aust. Arindam Ghosh (S’80, M’83, SM’93, F’06) is Professor of Power Engineering at Queensland University of Technology, Brisbane, Australia. He obtained a Ph.D. in electrical engineering from University of Calgary, Canada in 1983. Prior to joining the QUT in 2006, he was with the Dept. of Electrical Engineering at IIT Kanpur, India, for 21 years. He is a fellow of Indian National Academy of Engineering (INAE) and the IEEE. His interests are in Control of Power Systems and Power Electronic devices.