See

discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/268801009

Design and Implementation of a Feasible

Microgrid Model in Brazil
CONFERENCE PAPER OCTOBER 2014

READS

33

1 AUTHOR:
Marcos Aurelio Izumida Martins
Federal University of Santa Catarina
5 PUBLICATIONS 0 CITATIONS
SEE PROFILE

Available from: Marcos Aurelio Izumida Martins

Retrieved on: 25 September 2015

Design and Implementation of a Feasible Microgrid Model in

Brazil
Felipe Cassias, Celesc S.A., Avenida Itamarati, Brazil, [email protected]
Cesare Pica, Fundao CERTI, Campus da UFSC, Setor C, Brazil, [email protected]
Marcos Izumida, Fundao CERTI, Campus da UFSC, Setor C, Brazil, [email protected]
Luciano Gentilini, Fundao CERTI, Campus da UFSC, Setor C, Brazil, [email protected]
Victor Maryama, Fundao CERTI, Campus da UFSC, Setor C, Brazil, [email protected]
Daniel Makohin, Fundao CERTI, Campus da UFSC, Setor C, Brazil, [email protected]

The Power Point Presentation will be available after the conference.

Abstract
This paper aims to present the topology and energy management solutions implemented in a
20kW microgrid project in Brazil, integrating renewable resources, backup power, controllable
loads and the main grid. The objective of this project has been to develop a microgrid model
for neighborhood/district application that may become feasible in the short term. This microgrid
project has been designed using control techniques that allow it to operate in grid-connected
and island mode and to participate in Virtual Power Plant schemes. Preliminary results have
shown feasibility of the project, which should guide local utilities regarding the adoption of
microgrid technologies.

1.

Introduction

In order to promote the deployment of small-scale distributed generation (DG), in 2012, the
Brazilian Agency of Electric Energy (ANEEL) approved the Net Metering act [1], allowing
bidirectional energy flow between consumers/producers and the main grid. Due to this act,
along with a steady decline in the prices of DG, a massive connection of small-scale DG to the
main grid is expected in the near future. On the one hand, this picture is very motivating,
however this sudden allowance of small-scale DG may escape beyond the control of the
utilities and distribution operators, bringing new concerns such as impacts in power quality and
commercial issues between utilities and prosumers.
A promising solution to these concerns arises with the concept of microgrids [2] [3]. In this
context, the objective of the work presented in this paper is to develop, implement and evaluate
a feasible microgrid model, helping power utilities to understand better the available
technologies and solutions for this kind of system and their impact and benefits for the whole
power grid. In this work, a 20kW pilot microgrid project has been designed and implemented,
establishing a test bed where DG technologies, power quality solutions, control strategies
systems, islanded operation, etc. could be tested and validated from the utilities point of view.
Firstly, the paper shows the topology design and energy management system adopted and,
then, simulation and preliminary experimental results are presented.

2.
2.1.

The Pilot Microgrid

Topology

The pilot microgrid, named District Power Plant, has a radial topology for electricity distribution,
mirroring the most common distribution nets present around the country. The core of the
system is the Multicluster Box 6, a metering/switch panel with the capability to act under the
control of the Sunny Island inverters for operations like resynchronization, islanding and
protection.
On the connection side, there is an automatic transfer switch (ATS) responsible for switch to
the diesel generator when the main grid suffers a fault event; although it does not activate the
generator automatically, (the supervisory system starts the generator).
Deeper into the system, there is a generation branch for photovoltaic (PV) and wind power, a
load branch for alternate current (AC) loads and a storage branch where three inverters work
as an interface between the AC bus and the direct current (DC) bus for batteries and load.

Figure 1 - Simplified topology

Figure 2 - Wind generation unit and the photovoltaic panels

2.2.

Generation Units

Three power sources fall under the class of generation units: photovoltaic (PV), wind and diesel
generation, which totalize 13.6kVA of installed capacity. Out of this amount, 5.5kVA are used
only in case of islanded operation, as the diesel generator is not allowed to work under
connected mode because of the ATS mutual exclusion and it is intended for the diesel
generator to only server as a battery charger.
The 8.1kVA remaining generation belongs to the wind power generation (2.4kVA) and PV
generation (5.7kVA). These two sources help mitigate the power drained from the utility during
connected mode, or even feeding the utility when the loads are not at full capacity. During
islanded operation, it helps the battery to feed the loads or even feed the loads alone in
situations of high power generation und low load usage.

2.3.

Storage

Four 12V serial connected lead-acid batteries compose energy storage of the microgrid. As
they have a DC connection, there was a need for a DC bus of 48V that is connected to the
three battery inverters because of the inverters operational parameters. Each of the inverters
is responsible for one phase of the AC side. Because of the three inverters used, it is possible
for them to balance the phases of the system by using the DC bus.
In terms of capacity, the battery bank is able to yield 220Ah to the DC bus for 10,6kW of power
during an hour (10,6kWh of energy), but to avoid deep discharges that damage or reduce the
batteries life time, the research team placed limits for the discharge/recharge times and
current. The maximum current can only operate for 15 minutes while the minimum capacity
allowed is 30%, when operating in islanded mode. For connected mode, unless there is a
dispatch order made by the supervisory system, the batteries stay almost fully charged (around
90%) in order to absorb energy from the renewable sources in case of change to islanded
mode.

2.4.

Consumption

There are two kind of load fed by the district power plant: AC loads and a DC load. Four loads
are connected to the AC load branch of the multicluster box while one load gets power from
the DC bus. The loads are all controllable by the supervisory system that may turn them off
when there is need for power under islanded mode or when it is economically feasible to send
energy to the utility. The four AC loads are comprised of two resistor banks of 4.5kVA each
(both of low priority), the cooling system of 1.6kVA (highest priority) and 1.9kVA of illumination
(high priority). The sole DC load is a resistor bank of 2.5kW (low priority).

3.

Dispatch Simulation

Before finishing the executive project of the microgrid, the research team simulated several
scenarios to attest the feasibility of the project. These simulations were made in a software
built by the research team itself and the data for wind and PV power were obtained from a
weather station at the site of the microgrid for one year. The Figure 3, Figure 4 and Figure 5
show the curves used for the simulation of load profile and renewable generation. The power
sources and storage present in the section 2 are also present in this simulation. For the loads,
scenarios of 5, 10 and 20kW were tested. The simulations are valid only for active power.

Figure 3 - Typical commercial demand curve

Figure 4 typical wind speed for a week

3.1.

Operation under connected mode

While operating under connected mode, the batteries and the diesel generator play no active
role, as the main grid feeds the loads with the necessary amount of power. This leaves the PV
and wind generation as the power source from the microgrid side. Scenarios of 5, 10 and 20
kW of peak load were tested and they resulted in 7.77MWh of energy consumption and
2.23MWh of energy exportation for the 5kW load scenario, 19MWh and 1.53MWh for the 10kW
load scenario, and 31.58MWh and 1.14MWh for the 20kW load scenario. The Figure 6 shows
the power flow between utility and microgrid when operating with 10kW of peak load.

Figure 5 - typical solar irradiation for a week

Figure 6 - Power flow between microgrid and utility during two weeks under connected mode at 10kW of peak
load

3.2.

Islanded mode simulation

In this scenario, the main grid is offline, for reasons unknown, and the microgrid own generation
and storage must supply the loads. As the fuel is limited, the diesel generation usage is limited
to the cases where the battery is near empty and the loads must operate. In addition, because
of the lack of a utility grid, there is no power exchange in cases of excessive
generation/consumption, which leads the central controller to activate the diesel generation or
limit the renewable generation through power electronics.

Figure 7 - Simulation results for a week operation under islanded mode

The battery inverters together with the central controller control the diesel generation dispatch
according to the battery charge level. When the charge level of the battery reaches a preestablished level, the control system activates the diesel generation, unless a load shedding
strategy is running. The diesel generation then works until the battery is fully charged. For this
control strategy, the boundaries for charge and discharge are 90 % and 60 % of the battery
level to avoid deep discharges, which compromises the batteries, as well as to allow some
energy absorption when there is excessive renewable generation. Figure 7 shows an example
of operation under islanded mode for a load profile of 5kW of peak load.

4.

Microgrid Automation and Control

In order to achieve better results in the stead state operation, a control system was created to
run in the background, managing the energy usage and dispatch. By using the communication
and sensing infrastructure together with a SCADA, the central controller reads the current state
of the grid and take measures to events such as main grid failure and retailers reactive or
active power demands [4], acting as a virtual power plant [5].
The decision process for load shedding and power dispatch takes into account the available
energy, the price given by the utility and the loads priority, configuring a type of centralized
tertiary control level [6] responsible for the power flow between utility and microgrid.

Figure 8 - Microgrid supervisory system

For an easier integration process, the research team choose to acquire Modbus compliant
devices only, as it is a commonly used communication protocol in the power industry. The
information flows through a process bus via an Ethernet switch that connects all the microgrid
instrumentation (meters, PLC, power electronics) to the SCADA carrying only essential data
for stead state operation and control. An oscilloscope gathers transient state data during
experimental tests only.
As the inner control loops are dealt by the local devices, mainly the inverters, the central
controller needs to take no action to control variables like frequency and voltage. Its tasks lay
on the supervisory level, monitoring the power flow between microgrid and utility, pre-setting
contingence plans for the loads, deciding when to turn the diesel generator on and maintaining
procedure sequence for operational mode changes, all dynamically slower tasks if compared
to the transient operations.
One of the most important tasks of the central controller is the maintenance of the power flow
between grids at time of automated demand response (ADR) operations. During these events,
the utility needs a warranty that the virtual power plant (VPP) will not consume more power

than what was accorded between the two parts (feeding less power than it was accorded also
falls under this consideration). The equation 1, where P_Load stands for the internal
consumption, P_Gen stands for internal generation and Putility stands for the accorded ADR
power plus 10% of hysteresis for safety purposes as seen in [7]. Therefore, the SCADA is
programmed to measure the available power and the consumption and to give a dispatch order
to the battery or a disconnection order to any internal load should the equation 1 become
unbalanced.

= _ + _ + = 0
=0

(1)

=0

The central controller is also responsible to set a price for an ADR action. It solves the
optimization problem of the equation 2, minimizing the energy cost on the microgrid side FC in
function of the active power. The model used for optimization takes the generation forecast
and price for the renewable sources as well as the cost to turn the loads off or to dispatch the
batteries. The result is an optimal price for multiple ADR scenarios. Each ADR event lasts only
for 15 minutes because demand response operations are short term actions for balance in the
utility side as it can be treated as an ancillary service [8].
=

() = ( () + ) + ( () + ) () ( )

=1
=1

5.

(2)

Conclusion

This work leads the research team to conclude that microgrids are feasible in the short term
even in countries under development like Brazil regarding technical issues, although these
technologies are not as available as it is in economically developed countries, which make the
process of building slower and more expensive as it was expected.
As of the end of this paper, the whole system installation and commissioning is at its end. In
addition, the individual tests of the devices shows that they are able to follow the local rules for
electrical operation. Preliminary tests of the devices working as an integrated system are about
to be started and experimental results of the operational modes are expected to be obtained
very soon.
The next steps regarding this system are the tests for full compliance of the systems
operational modes according to national laws of the energy sector, the creation of an interface
for interaction between the utility and the microgrid and the development of an economically
feasible business model that brings benefits for both utility and costumer.

6.

References

[1] AGNCIA NACIONAL DE ENERGIA ELTRICA ANEEL, "RESOLUO NORMATIVA

N 482," 2012.
[2] N. Hatziargyriou, H. Asano, R. Iravani and C. Marnay, "Microgrids," 2007.
[3] S. Chowdhury, S. Chowdhury and P. Crossley, "Microgrids and Active Distribution
Networks," 2009.
[4] H. Nunna and S. Doolla, "Intelligent demand side management in smart-microgrids," 2013.

[5] L. Toma, B. Otomega, C. Bulac and I. Tristiu, "Coordination of Distributed Generators

Through the Virtual Power Plant Concept," 2012.
[6] J. M. Guerrero, J. C. Vsquez and R. Teodorescu, "Hierarchical Control of DroopControlled DC and AC Microgrids - A General Approach Towards Standardization,"
Industrial Electronics, 2009. IECON '09. 35th Annual Conference of IEEE, Novembro
2011.
[7] I. PVPS, "The role of energy storage for mini-grid stabilization," 2011.
[8] U. D. o. Energy, "Benefits of Demand Response in Electricity Markets and
Recommendations for Achieving Them," 2006.