1

Abstract The issue of controlled and reliable integration of

distributed energy resources into microgrids and large power
grids has recently gained considerable attention. The microgrid
concept, which basically corresponds to the coordinated
operation of a cluster of loads, distributed generators and energy
storage systems, is quite appealing due to its flexibility,
controllability and energy management capabilities. In order to
provide uninterruptible power supply to the loads, microgrids
are expected to operate in both grid-connected and stand-alone
modes, and economically meet the demand on an instantaneous
basis. The problem of optimal management of the resources in a
microgrid is being widely investigated and recent studies have
proposed the application of both centralized and distributed
control schemes by using multi-agent systems, heuristic methods
and optimization algorithms. This paper elaborates on the
conceptual design of a centralized energy management system
(EMS) and its desirable attributes for a microgrid in stand-alone
mode of operation. A number of test protocols are proposed to
analyze the performance of the system, as well as the impacts of
relevant parameters.


Index TermsMicrogrids, distributed generation, energy
management systems, multi-stage optimization.
I. INTRODUCTION
elped by deregulation processes in the energy sector, and
encouraged by the implementation of additional incentive
policies, the small-scale distributed energy resources
(DER), many of them based on renewable energies, have
experimented an unprecedented growth in the last couple of
decades. Because of this, important efforts have been made in
the energy sector in order to develop appropriate technologies
and techniques for the reliable and economic exploitation of
the renewable energy sources, and its integration into power
systems.
One of the main technical issues to be addressed in the
process of integration is the control and management of the
non-dispatchable renewable energy sources, like wind and
solar energy, which feature somewhat unpredictable
behaviour. Another major concern is the assessment of the
impact of such sources on the overall grid security and
reliability. Also, there is a the need for re-engineering the
protection schemes at the distribution level, to cope with the
bidirectional power flows and take advantage of the fast
response of power electronic devices commonly used for the
grid connection of distributed energy sources.

The authors are with the University of Waterloo, Waterloo, Ontario,
Canada (e-mail: dolivare@engmail.uwaterloo.ca).

Although small autonomic grids have existed for many
decades, the concept of microgrid was first introduced in [1],
[2] as a solution for the reliable integration of DER and for
harnessing their multiple advantages. A microgrid can be
described as a cluster of micro-sources, energy storage
systems (ESS) and loads that is perceived by the main grid as
a single element that can respond to centralized control
signals. Special protection, control and energy management
systems must be designed for the microgrid operation in order
to ensure reliable, secure and economical operation in either
grid-connected or stand-alone mode. In particular, the problem
of energy management in a microgrid gains more relevance
with the presence of highly-variable energy sources, where the
update rate of the unit dispatch command should be fast
enough to follow the sudden changes of load and non-
dispatchable generators, with time constants close to those of
the control system.
The problem of energy management in microgrids consists
on finding the optimal (or near optimal) unit commitment
(UC) and dispatch of the available generators so that certain
selected objectives are achieved. A commonly pursued
objective for stand-alone mode of operation is to economically
supply the local load [3], whereas under grid-connected mode
of operation the maximization of profits is typically sought
[4]. Additional objectives such as the minimization of
greenhouse gas emissions of the microgrid have also been
proposed in [3] and [5] by applying heuristic and multi-
objective optimization techniques.
With regard to the architecture of the energy management
system (EMS), two main approaches have been proposed to
date in the technical literature: (i) Centralized EMS (CEMS)
and (ii) Distributed EMS (DEMS). The CEMS architecture
consists of a central controller provided with the relevant
information of every distributed energy resource within the
microgrid and the microgrid itself (e.g., cost functions,
technical characteristics/limitations, network parameters and
mode of operation), as well as the information from
forecasting systems (e.g., local load, wind speed, solar
radiation) in order to determine an appropriate UC and
dispatch of the resources according to the selected objective.
On the other hand, DEMS provides a market environment
through the use of Multi-Agent Systems (MAS) where each
microgrid agent sends buying and/or selling bids to a Central
Microgrid Operator (CMO) according to their particular needs
and cost structures; the CMO then performs a binding process
to determine the operation of the microgrid for the next period.
In this case, a separated UC process must be realized to
determine the agents that will operate in each particular
period. In this paper, the characteristics of the existing EMS
A Centralized Optimal Energy Management
System for Microgrids
Daniel E. Olivares, Claudio A. Caizares, Fellow, IEEE, Mehrdad Kazerani, Senior Member, IEEE.
H
2
approaches are reviewed, to synthesize the conceptual design
of a CEMS suitable for microgrids in stand-alone operation.
The rest of the paper is presented as follows: Section II
describes the characteristics of the current developments in
EMS and identifies the challenges that need to be addressed in
this area. A typical microgrid configuration and its
components are presented in Section III. Section IV discusses
the proposed architecture and framework for the development
of an EMS for microgrids in stand-alone mode of operation.
Section V describes specific performance evaluation
procedures for comparison and calibration of EMS models.
Finally, relevant conclusions are presented in Section VI.
II. MICROGRID EMS
As mentioned in the previous section, there are 2 main
approaches to the development of EMS for microgrids, each
having its own advantages and drawbacks. These two
approaches are described and analyzed next.
A. Centralized EMS
A few cases of CEMS have been discussed in the literature,
taking into account varied microgrid characteristics and
configurations. A general framework for the development of
CEMS is proposed in [6], while a CEMS for a microgrid
composed of hydrogen storage and wind power, utilizing a
dynamic linear programming (LP) formulation is presented in
[7]. An LP solution technique together with heuristics is
proposed in [8] for the implementation of a CEMS for a PV-
storage microgrid, while [3] proposes a purely heuristic
optimization approach. Different evolutive algorithms for
optimization are also applied to the CEMS problem in [9] and
[10].
A typical CEMS architecture is shown in Fig. 1, where a
central agent collects all the relevant information from the
different microgrid actors to perform an optimization and
determine the inputs of the control system for the next period.
Depending on the particular resources present in the
microgrid, the input variables of the CEMS can be:
Forecasted power output of the non-dispatchable
generators for the following N consecutive periods.
Forecasted local load for the following N consecutive
periods.
State of charge of the ESS.
Operational limits of dispatchable generators and ESS.
Security and reliability constraints of the microgrid.
Interconnection status.
Main grid energy price forecasting.
Once all the input variables are gathered in the CEMS, a
multi-stage optimization is performed in order to determine
the optimal dispatch of units according to a defined cost
function, over a pre-specified time frame. Output variables of
the EMS are the reference values of the control system (e.g.,
output power and/or terminal voltage) for each dispatchable
DER, together with binary decision variables for connecting or
disconnecting loads for load shifting. An additional output
variable is the UC decision of the dispatchable generators (if
required); however, this problem can be solved at a lower
frequency than the dispatch, and separately.
The main advantages of the centralized approach are:
allowing for a broad observability of the microgrid, and
suitability for application of optimization techniques. Some of
its disadvantages are: reduced flexibility, as it needs to be
modified to incorporate additional generators, and extensive
computational requirements to perform the optimization.
B. Distributed (Multi-Agent) EMS
A DEMS based on MAS for microgrids was first proposed
in [11] as an alternative for coordinated operation of
microgrids in a competitive market environment and with
multiple generator owners. The relevant microgrid actors are
grouped and represented by different agents that interact in a
market environment in order to determine the operation of the
microgrid. In this way, consumers, generators, ESS and the
main grid participate in the market by sending buying and
selling bids to the CMO based on their particular needs,
availability, cost functions, technical limitations, expectations
and forecasts. The CMO is responsible for the settlement of
the microgrid market by matching buying and selling bids
maximizing the social welfare, while ensuring the feasibility
of the resulting operation plan. A similar MAS approach is
also proposed in [12]. Additional agents assigned to different
tasks such as load shifting and load curtailment to allow
demand side management are proposed in [13] as well.
The MAS-DEMS approach allows almost autonomous
operation of the generating units in a microgrid, and reduces
the need for manipulation of large amounts of data, thus
reducing computation time. Another important advantage of
DEMS is its flexibility, as it provides the plug-and-play
feature, facilitating the installation and coordination of
additional DER in the microgrid. On the other hand, DEMS
based on MAS shows disadvantages compared to CEMS when
applied to microgrids that require strong cooperation between
the different DER in order to operate the system in a secure
and reliable way. A typical DEMS model for a microgrid
operating in grid-connected mode is shown in Fig. 2.
In the case of isolated microgrids operating in stand-alone
mode, the small number of generators in the microgrid and the
uneven share of installed power, as well as the lack of a strong


Fig. 1. CEMS.


N-Period
Forecasting of
Non-Dispatchable
Generation
Microgrid
Centralized
EMS

State-of-Charge
of the ESS
Microgrid Model
and Policies
N-Period
Forecasting of
Electrical/Thermal
Load
Operational Limits,
Security and
Reliability
Constraints
Command to
Controllable Loads
(DSM) On/Off/Shift
Set-points
for dispatchable DER
(droop) controllers for
Next Period
Unit Commitment
Decisions for future
Periods
Main Grid
Interconnection
Status and Energy
Price Forecasting
3
price signal from the main grid, make a DEMS more difficult
to implement.
III. TYPICAL MICROGRID CONFIGURATION
Although microgrids can be located close to or within the
main grid and be operated in grid-connected mode, the
microgrids of interest for this work correspond to isolated,
locally-operated systems that can be found in remote locations
with accessibility problems, and where the connection to the
main grid of electricity is not feasible, either for technical or
economical reasons. Thus, the particular microgrid
configuration considered here, embodies multiple DER such
as wind generators, PV solar arrays, diesel generators,
microturbine CHP, Fuel Cell-Electrolyzer ESS, thermal
storage, and both electrical and thermal loads. A microgrid
with the aforementioned components connected in a typical
radial configuration is shown in Fig. 3.
Loads that have a defined energy consumption that can be
supplied with certain flexibility over a period of time are
referred to as shiftable loads, as they can be shifted by the
EMS to the most convenient period. A broader description of
various microgrid components is provided next.
A. Fuel Cell-Electrolyzer ESS
In order to fully embrace the benefits of renewable energy
sources, it is indispensable to have an ESS capable of dealing
with intra-day variations of generation. Fuel Cell-Electrolyzer
ESS offers interesting benefits for long term energy storage, as
it can consume electricity to produce hydrogen, store
hydrogen without time limitations, and consume hydrogen to
generate electricity, which makes the operation of the system
more flexible. Furthermore, these hydrogen conversion
processes generate significant amount of heat [16]. However,
one of the main drawbacks of this technology is its low round-
trip efficiency (energy input-hydrogen-energy output) which is
in the range of 30-40% [14], [15], and costs [16]. Constraints
on the operation of this ESS are typically the maximum and
minimum power output, minimum state-of-charge of the
hydrogen tank and ramp rates.
B. Wind Turbines
A wide range of levels of penetration of wind power can be
observed in existing microgrids; however, only wind turbines
with a significant share of the microgrid peak load are relevant
for the operation of the EMS. Different levels of penetration,
wind speed profiles and wind turbine characteristics should be
considered.
C. Diesel Generator
Mainly due to its flexibility, remote and isolated power
systems usually rely on the operation of diesel generators to
supply the local load. Although at a very expensive rate in this
case, this generator is typically able to supply the peak load on
its own, without the need for further energy sources. When
considering time frames in the order of minutes, the operation
of diesel generators is constrained by their power output
limits, power ramp-up and -down rates and start-up time.
D. PV Panels
Together with wind turbines, PV systems are one of the
fastest growing renewable energy sources. Although both are
based on non-dispatchable energy sources, the PV panels
usually have a more easily predicted power production.
Different levels of penetration, real solar radiation profiles and
PV panel characteristics should be considered.
E. CHP Microturbine
Gas microturbine technology has promising CHP
applications within microgrids, as it can achieve overall
efficiencies of around 65% when adding a heat recovery cycle


Fig. 2. DEMS (Multi-Agent).


Selling Price
Buying Price
Microgrid Operator
Matches Buying and Selling Bids, Maximizing
Welfare
Main Grid
Operation
Period N
Operation
Period N+1
Time
Announcement of
Main Grid Prices for
N+1
Bidding +
Market Settlement
for N+1
Selling Price


ESS
Distributed Generator Consumer / Load
Buying Price
Selling Price



Buying Price





Fig. 3. Microgrid configuration.

Electricity
Heat
PV
Critical
Load
Diesel
Electrolyzer
Fuel
Cell
Hydrogen
Storage
Critical
Load
Shiftable
Load
Wind
Turbines
Thermal
Storage
Shiftable
Load
Critical
Load
AC
AC
4
to the conventional power generation. In small sizes, this
technology shows a high operational flexibility, with start-up
times of around 3 minutes, and shut-down times of
approximately 10 minutes [17]. Its operation is constrained by
ramp-up and -down rates, start-up and shut-down times, and
minimum power output; additionally, auxiliary power needs to
be provided from the grid during the initial part of the turbine
warm-up [17].
F. Thermal Storage
Heat recovery and storage in form of hot water in isolated
accumulator tanks is a simple and efficient way to store
thermal energy from CHP plants for intra-day regulation [18],
as water offers a good heat recovery capacity per cubic meter,
and heat losses can be kept at minimum with a proper tank
design. The maximum and minimum storage capacities
constrain the operation of the tanks.
G. Shiftable Loads
Typically heating and cooling systems can be operated with
some flexibility over time, such that the power requirements
can be shifted to different times without significant
performance degradation. The incorporation of shiftable loads
into the microgrid is expected to produce an impact similar to
the addition of more energy storage capacity, allowing a more
flexible operation of the microgrid.
IV. PROPOSED EMS ARCHITECTURE
Depending on whether a microgrid is to be operated in grid-
connected or stand-alone mode, the most suitable EMS
architecture and objectives may be completely different. As
discussed in Section II, the particular characteristics of CEMS
make them more suitable for implementation in microgrids
operating in stand-alone mode, as they are more likely to
provide continuous operation of the system while dealing with
variable generation and load, without main grid support.
Hence, a in order to perform the coordinated operation of the
isolated microgrid described in the previous section, a CEMS
architecture is proposed here.
The interaction between the different control modules
within the microgrid is illustrated in Fig. 4. The control system
block represents the individual control stages of each DER
that keep the balance of active power and regulates the voltage
in the microgrid. The control system of power-electronics-
interfaced (no-inertia) microgrids is typically based on the
implementation of droop controllers that emulate the
behaviour of synchronous generators against changes in the
frequency and voltage of the grid [20], [21].
The proposed EMS embodies two main blocks: a multi-
stage Economic Load Dispatch (ELD) block and a UC block.
The EMS also requires input from a load and generation
forecasting system. The different blocks of the proposed
architecture are discussed in some details next.
A. Generation/Load Forecasting
The generation and load forecasting system must provide,
with two different resolutions, forecasted values of output
powers from renewable energy sources and total microgrid
load. A high resolution in the order of a few minutes over a
horizon of hours is required by the multi-stage ELD block,
while a lower resolution in the order of 30-60 minutes over a
horizon of tens of hours is required by the UC block.
B. Multi-Stage Economic Load Dispatch
The multi-stage ELD performs the calculation of optimal
dispatch for DER over a horizon H
ELD
, with a time step T
ELD
,
considering a fixed group of on-service units that will be
provided by the UC block. The UC block also provides the
boundary conditions for the two ESSs at the final time step of
the ELD, as it has a more extended visibility of load and
generation profile patterns. The ELD calculation is carried out
for each time step T
ELD
so that, although it is calculated over a
horizon H
ELD
, only the dispatch calculated for the immediately
following time step is used as reference for the control system,
and the dispatch for next time steps is updated during the next
iteration. Even when the UC and ELD problems can be solved
together, they are separated in order to speed-up the ELD
calculation and achieve faster update rates for dispatch. As the
UC block delivers the ESS state-of-charge boundary
conditions with a different time resolution, the multi-stage
ELD calculation considers a linear interpolation of the values
provided by the UC.
The off-line calculation of the optimal dispatch for each
possible demand scenario of a microgrid is discussed in [19],
where the optimal dispatch for each condition is stored in a
static look-up table or operating chart. However, several
drawbacks are identified in this approach regarding the
inability to handle variable conditions. In particular, the
presence of ESS in the microgrid introduces a time-
dependence in the calculation of the optimal dispatch, as the
state-of-charge of the ESS at a given time-step will depend on
the state-of-charge and dispatch at the previous step; therefore,
the optimal dispatch is not solely determined by a particular
demand scenario. To manage this time-dependence, a dynamic
optimization of the microgrid operation is required.
C. Unit Commitment
The UC block calculates the optimal schedule of the DERs
over a horizon H
UC
, with a time step T
UC
, and at least H
ELD
in
advance of the next period, such that there is always a defined
UC for the ELD calculations. The outputs of this block are the
start-up and shut-down decisions for each dispatchable DER
in the microgrid, together with the optimal dispatch of units
and ESS state-of-charge for each time step, according to the
forecasted values.


Fig. 4. CEMS block diagram.


Unit
Commitment
Multi-Stage
ELD
Control
System
Plant
(Microgrid)
Generation/Load
Forecasting
Other Inputs
Low Resolution
Long Term
5
The coordinated operation of UC and multi-stage ELD
blocks, and the overlapping of consecutive ELD calculations
is illustrated in Fig. 5, where each vertical line (long lines for
UC calculation and short lines for ELD calculation) represents
the instant at which calculations take place, and the dotted
lines indicate the operating period for which the calculation is
being performed. A more detailed time scale of the
calculations is depicted in Fig. 6.
V. PERFORMANCE AND CALIBRATION PROCEDURES
A number of procedures to comparatively evaluate the
performance and fine-tune the proposed EMS architecture are
proposed in this section.
A. Multi-stage ELD horizon vs ESS capacity
The only reason to perform a multi-stage ELD calculation
and not a single-stage ELD after the UC is the presence of
ESS in the microgrid. As the ESS capacity of the microgrid
increases, the link between the operations in consecutive time
steps becomes stronger; therefore, a more extended horizon
for multi-stage ELD (H
ELD
) must be chosen, slowing-down the
calculation process and imposing a lower limit on the time
step between consecutive ELD calculations (T
ELD
).
In order to determine the most suitable H
ELD
for a particular
microgrid, a study must be performed to find an optimal
balance between the dispatch accuracy and calculation speed.
Thus, a low H
ELD
values would not allow a correct use of the
ESS capacity, while high values would increase the
calculation times and introduce increasing errors from the
forecasting system. This would lead the microgrid operation to
deviate from the optimal condition, and therefore would
increase the total operating costs. The trade-off in the selection
of H
ELD
is illustrated in Fig. 7.

B. Impact of Forecasting Accuracy
The accuracy of the load and generation forecasting system
plays a very important role in the performance of an EMS, as
poor forecasting accuracy would lead to deviations from
optimal operation of the microgrid. An EMS that strongly
relies on the forecasting system outputs is not desirable;
moreover, an EMS design should be robust against extended
forecasting errors.
To evaluate the robustness of the EMS against forecasting
errors, the following three different forecasting scenarios were
proposed in [7], which will be adopted in this work:
persistence model (worst case), perfect forecasting (best case)
and uncertain forecast (realistic case). Although a lower
performance is expected from the scenarios without perfect
forecasting, the EMS should be able to reduce the loss-of-load
risk to acceptable levels. Similarly to the case of large power
systems, a trade-off will exist between the acceptable loss-of-
load probability of the microgrid, the accuracy of the
forecasting method, and the operating costs.
VI. CONCLUSIONS
This paper has presented a brief review of the existing
energy management system (EMS) architectures for
microgrids, identifying the main advantages of each approach,
and has proposed a centralized EMS architecture for
implementation on isolated microgrids in stand-alone mode of
operation. Some relevant considerations and procedures for
the model fine-tuning and performance evaluation have also
been presented. Future work will concentrate on the
implementation and testing of the proposed architecture.
VII. REFERENCES
[1] B. Lasseter, Microgrids [distributed power generation], in Proc. IEEE-
PES Winter Meeting, vol. 1, Jan./Feb. 2001, pp. 146-149.
[2] R. H. Lasseter, MicroGrids, in Proc. IEEE-PES Winter Meeting, vol.
1, Jan. 2002, pp. 305-308.
[3] E. Alvarez, A. C. Lopez, J . Gmez-Aleixandre, and N. de Abajo, On-
line minimization of running costs, greenhouse gas emissions and the
impact of distributed generation using microgrids on the electrical
system, in Proc. IEEE PES/IAS Conference on Sustainable Alternative
Energy (SAE), Sept. 2009, pp. 1-10.
[4] M. Y. Nguyen, Y. T. Yoon, and N. H. Choi, Dynamic programming
formulation of micro-grid operation with heat and electricity

Fig. 5. UC-ELD coordinated operation and overlapping.


ELD
t
UC UC
. . . . . .
H
ELD
H
UC
T
ELD


Fig. 6. UC-ELD operation time scales.


k
ELD for Period k ELD for Period k+1
Given the UC, calculate
ELD for (k+2)T
ELD


Given the UC, calculate
ELD for (k+1)T
ELD


kT
ELD
(k+1)T
ELD
(k+2)T
ELD
(N+2)H
UC
NH
UC
(N+1)H
UC
UC period N UC for period N+1
Calculate
UC for N+1
Calculate
UC for N+2
H
ELD
k+4

k+2



Fig. 7. Trade-off in selection of ELD horizon.


Improved dispatch
of ESS
Faster ELD
calculation and
lower TELD
Operation
Cost [$]
ESS-1
ESS-2
ESS-1 Capacity < ESS-2 Capacity
H
ELD
[t] H
ELD-1
* H
ELD-2
*
6
constraints, in Transmission & Distribution Conference & Exposition:
Asia and Pacific, Oct. 2009, pp. 1-4.
[5] C. M. Colson, M. H. Nehrir, and S. A. Pourmousavi, Towards real-time
microgrid power management using computational intelligence
methods, in Proc. IEEE-PES General Meeting, Jul. 2010, pp. 1-8.
[6] N. Hatziargyriou, G. Contaxis, M. Matos, J . A. P. Lopes, G.
Kariniotakis, D. Mayer, J . Halliday, G. Dutton, P. Dokopoulos, A.
Bakirtzis, J . Stefanakis, A. Gigantidou, P. O'Donnell, D. McCoy, M. J .
Fernandes, J . M. S. Cotrim, and A. P. Figueira, Energy management
and control of island power systems with increased penetration from
renewable sources, in IEEE-PES Winter Meeting, vol. 1, J an. 2002, pp.
335-339.
[7] M. Korpas and A. T. Holen, Operation planning of hydrogen storage
connected to wind power operating in a power market, IEEE
Transactions on Energy Conversion, vol. 21, no. 3, Sept. 2006, pp. 742-
749.
[8] S. Chakraborty and M. G. Simoes, PV-microgrid operational cost
minimization by neural forecasting and heuristic optimization, in Proc.
IEEE Industry Applications Society Annual Meeting IAS '08, Oct. 2008,
pp. 1-8.
[9] R. Noroozian and H. Vahedi, Optimal management of MicroGrid using
bacterial foraging algorithm, in Proc. 18th Iranian Conference on
Electrical Engineering (ICEE), May 2010, pp. 895-900.
[10] H. Vahedi, R. Noroozian, and S. H. Hosseini, Optimal management of
MicroGrid using differential evolution approach, in Proc. 7th
International Conference on the European Energy Market (EEM), J un.
2010, pp. 1-6.
[11] N. D. Hatziargyriou, A. Dimeas, A. G. Tsikalakis, J . A. P. Lopes, G.
Karniotakis, and J . Oyarzabal, Management of microgrids in market
environment, in Proc. International Conference on Future Power
Systems, Nov. 2005, pp. 7 pp.-7.
[12] T. Logenthiran, D. Srinivasan, and D. Wong, Multi-agent coordination
for DER in MicroGrid, in Proc. IEEE International Conference on
Sustainable Energy Technologies (ICSET), Nov. 2008, pp. 77-82.
[13] J . Oyarzabal, J . J imeno, J . Ruela, A. Engler, and C. Hardt, Agent based
micro grid management system, in Proc. International Conference on
Future Power Systems, Nov. 2005, pp. 6 pp.-6.
[14] D. B. Nelson, M. H. Nehrir, and C. Wang, Unit sizing and cost analysis
of stand-alone hybrid wind/PV/fuel cell power generation systems, in
Renewable Energy, vol. 31, Aug. 2006, pp. 1641-1656.
[15] K. Agbossou, M. Kolhe, J . Hamelin, and T. K. Bose, Performance of a
stand-alone renewable energy system based on energy storage as
hydrogen, in IEEE Transactions on Energy Conversion, vol. 19, Sept.
2004, pp. 633-640.
[16] G. Taljan, M. Fowler, C. Caizares, and G. Verbi, Hydrogen storage
for mixed windnuclear power plants in the context of a Hydrogen
Economy, Int J Hydrogen Energy, vol. 33, Sept. 2008, pp. 4463-4475.
[17] F. E. Pierce, Summary of Results fromTesting a 30kW Microturbine
and Combined Heat and Power (CHP) System, Federal Energy
Management Program - U.S. Department of Energy, DOE/EE-0316,
May 2007.
[18] T. Tveit, T. Savola, A. Gebremedhin, and C. Fogelholm, Multi-period
MINLP model for optimising operation and structural changes to CHP
plants in district heating networks with long-termthermal storage, in
Energy Conversion and Management, vol. 50, Mar. 2009, pp. 639-647.
[19] C. A. Hernandez-Aramburo, T. C. Green, and N. Mugniot, Fuel
consumption minimization of a microgrid, in IEEE Transactions on
Industry Applications, vol. 41, May/Jun. 2005, pp. 673-681.
[20] M. C. Chandorkar, D. M. Divan, and R. Adapa, Control of parallel
connected inverters in standalone AC supply systems, in IEEE
Transactions on Industry Applications, vol. 29, J an./Feb. 1993, pp. 136-
143.
[21] Il-Yop Chung, Wenxin Liu, D. A. Cartes, E. G. Collins, and Seung-Il
Moon, Control Methods of Inverter-Interfaced Distributed Generators
in a Microgrid System, in IEEE Transactions on Industry Applications,
vol. 46, May/J un. 2010, pp. 1078-1088.
VIII. BIOGRAPHIES
Daniel E. Olivares was born in Santiago of Chile, received the B.Sc. and the
Engineer degree in electrical engineering from the University of Chile,
Santiago, Chile, in 2006 and 2008 respectively. He is currently working
toward the Ph.D. degree in electrical and computer engineering at the
University of Waterloo, Waterloo, ON, Canada.
Claudio A. Caizares (S85, M91, SM00, F07) received in April 1984 the
Electrical Engineer diploma fromthe Escuela Politecnica Nacional (EPN),
Quito-Ecuador, where he held different teaching and administrative positions
from1983 to 1993. His MS (1988) and PhD (1991) degrees in Electrical
Engineering are fromthe University of Wisconsin-Madison. Dr. Caizares has
held various academic and administrative positions at the E&CE Department
of the University of Waterloo since 1993, where he is currently a Full
Professor. His research activities concentrate on the study of modeling,
simulation, control, stability and computational issues in power systems in the
context of competitive electricity markets.

Mehrdad Kazerani (S88M96SM02) received the B.Sc. degree from
Shiraz University, Shiraz, Iran, in 1980, the M.Eng. degree fromConcordia
University, Montreal, QC, Canada, in 1990, and the Ph.D. degree fromMcGill
University, Montreal, in 1995. From1982 to 1987, he was with the Energy
Ministry of Iran. He is currently a Full Professor with the Department of
Electrical and Computer Engineering, University of Waterloo, Waterloo, ON,
Canada. His research interests are in the areas of power electronic circuits and
systems design, active power filters, matrix converters, distributed power
generation, utility interface of alternative energy sources, battery electric,
hybrid electric and fuel cell vehicles, and FACTS. Dr. Kazerani is a
Registered Professional Engineer in the Province of Ontario.