Closed Loop Control of Soft Switched Forward Converter Using Intelligent Controller 157

IJCTA, 10(02), 2017, pp. 157-164

© International Science Press

Arduino based Autonomous Energy Mana-

gement of a Micro-Grid using Multi Agent
System
Leo Raju* Prithika Ravi M* Sajna. G* and Srinidhi. J*

Abstract : The objective of this paper is to develop Arduino based Multi Agent System (MAS) for advanced
demand side management of a solar micro-grid. High penetration of renewable energy resources needs new
coordination and control approaches to meet the stochastic nature of the environment and dynamic loadings.
We use Multi Agent System for advanced distributed, autonomous energy management of the micro-grid to
dynamically and flexibly adapt to the changes in the environment. We consider a micro-grid which contains
two solar Photo Voltaic (PV) system, local consumer, and a battery. We develop a simulation model in Java
Agent Development Environment (JADE) for dynamic energy management, which choose the best possible
action every hour to stabilize and optimize the solar micro-grid. Furthermore, environment variables are
sensed through Arduino microcontroller and given to agents of MAS. The resulting actions are reflected in
the LED out puts which can be readily deployed in the actual field.
Keywords : Autonomous Agent, Energy management, Multi Agent System, Micro-grid, JADE, Arduino.

1. INTRODUCTION
Environment consideration is leading proliferation of renewable resources in electrical industry.We are
moving towards a more decentralized, more sustainable, and smarter power system. Centralized approach
is used in most of the existing research on micro-grid operation problems. In order to reduce communication
overhead and improve robustness, distributed approach is used in energy management problem. One such
approach is Multi Agent System based modelling of micro-grid to provide a common communication
interface for all agents representing the autonomous physical elements. Furthermore, the distributed
nature and potential for modelling autonomous decision making entities in solving complex problems
motivates the use of multi-agent system for the operation of micro-grid. The design and implementation of
Multi Agent System in micro-grid energy management is discussed in detail in the paper [1]. Multi-agent
system for operation of an integrated micro-grid is discussed in [2]. Multi-agent based distributed energy
management for intelligent micro-grid is discussed in [3]. The complete review of micro-grids in multi-
agent system perspectives are discussed in [5]. Multi agent based micro-grid control is discussed [4].
Only in the very recent paper [5], the integration of micro-grid market operations and Distributed Energy
Resources (DER) is discussed in detail. Although many micro-grid research activities involving MAS have
been reported, they did not consider all the options available for optimal energy management of a micro-
grid, like, grid failure, no solar power in the night, peak load management, demand side management, etc.,
Also most of the references do MAS simulation in JADE framework. Real time implementation of MAS
is not discussed. So we propose a multi agent system based advanced distributed energy optimization of
solar micro-grid by comprehensively analysing and simulating all the possible options for the dynamic

* Department of Electrical and Electronics Engineering, SSN College of Engineering, Chennai-603110,Tamil Nadu, India. Corresponding
Author: [email protected]
158 Leo Raju, Prithika Ravi M, Sajna. G and Srinidhi. J
energy management in Java Agent Development Environment (JADE). Here the agent autonomously
choose the best option every hour, considering the intermittent nature of solar power, randomness of load,
dynamic pricing of grid and variation of critical loads, to stabilize and optimize the solar micro-grid.
Simulation results in this paper are capable of representing the dynamic behaviour of the micro-grid across
various possible solar-power and load values. All the features of smart grid are implemented and the
agents operations are verified by Arduino microcontroller for various scenarios.
The rest of the paper is organized as follows. A detailed discussion on multi agent system approach
is given in section 2. Micro-grid is explained in section 3. Arduino Micro controller is explained in
section 4. Problem formulation is given in section 5. Implementation of autonomous demand side energy
management and grid outage management in solar micro-grid is given in section 6. Simulation and results
are given in section 7. Conclusion is given in section 8.
2. MULTI AGENT SYSTEM
Distributed system with many on-going interactions and continuous communications are almost infeasible.
These considerations have motivated the development of approaches to distributed system based on agents,
which provide ways for adaptation and on-going interaction. A Multi Agent System (MAS) is a distributed
system consisting of multiple software agents, which form ‘a loosely coupled network’, to work together
to solve problems that are beyond their individual capabilities or knowledge of each entity. MAS is a
emerging sub-field of Distributed Artificial Intelligence (DAI). Multi-Agents overlay a way to elaborate
systems that are decentralized rather than centralized, emergent rather than planned, and concurrent
rather than sequential, with many advantages. MAS have inherent benefits such as flexibility, scalability,
autonomy and reduction in problem complexity [6]. In MAS, several autonomous and intelligent entities
called agents are working in collaboration to achieve the overall goal of a system. An agent receives
information about a state of its environment, takes actions which may alter that state and expresses
preferences among the various possible states. Agents have four behavioural attributes, autonomy, social,
proactive and reactive. Autonomy refers to the principle that agents can operate on their own to meet
their goals without the need for human guidance [7]. Agents are proactive, i.e., the ability to take the
initiative rather than acting simply in response to their environment. Agent can cooperate with other
agents for coordinated action. In order to cooperate, agents need to possess social ability, i.e., the ability
to interact with other agents with some communication language like Agent Communication Language
(ACL). Agents are reactive to changes in environment. Reasoning, optimizing, controlling and learning
are the inherent characteristics of an agent[8].
3. MICRO-GRIDS
Micro-grids are electricity distribution systems containing loads and distributed energy resources, (such
as distributed generators, storage devices, or controllable loads) that can be operated in a controlled,
coordinated way either while connected to the main power network or while islanded. Micro-grids can
coordinate all the assets and present them to the mega-grid in a manner and at a scale that is consistent with
current grid operations, thereby avoiding major new investments that are needed to integrate emerging
decentralized resources. Micro-grids have been proposed as a novel distribution network architecture
within the Smart-Grids concept, capable to exploit the full benefits from the integration of large numbers
of small scale distributed energy resources into low-voltage electricity distribution systems. The present
SCADA (Supervisory Control And Data Acquisition) is incapable of managing large numbers of renewable
energy resources due to its intermittent nature [9]. Multi Agent Systems are used to manage the dynamic
nature of the renewable resources..
4. ARDUINO SERIAL COMMUNICATION USING RXTX LIBRARY
This is the most suitable method for accessing the Arduino from the Java code. A library called RxTx is
used which facilitates serial communication with the Arduino UNO. Messages are sent to and received
 Arduino based Autonomous Energy Management of a Micro-Grid using Multi Agent System 159
from the Arduino serially. However, since the communication is serial, only one agent can access the
Microcontroller at a time. In our case, the control agent is responsible for this communication. The Arduino
board can, however, limit the number of values sensed from the environment due to the lack of many
analog pins. Potentiometers connected to analog pins of the Arduino board are used to input the sensed
values of the environment like Load Required and Solar Power Available,etc., Potentiometers connected
to analog pins of the Arduino board are used to input the sensed values. The Arduino IDE itself is written
in Java, and it can communicate to the serial port via the RxTx Java library. The required command for
turning on the LED is parsed from the message of the communication from Eclipse.
The potentiometer values can be obtained from the Arduino board through this serial communication.
The results of strategic operations of agents can be reflected in the microcontroller which is given to
the physical device for actuation. This information is again passed to the Arduino board through serial
communication and given to its digital pins to which LEDs are connected. These LEDs indicate the action
taken by the agents and the status of the LEDs trigger the action of the actuators. An Arduino program is
written and uploaded for the above mentioned purpose. The Loads will get power from a source based
on priorities for The first priority the lowest cost among all other sources and the second is the nature of
source. In this case renewable sources are preferred.The above two conditions not only gives the consumer
an advantage over hectic grid prices but also helps in the improving the environment leading to economic
and environmental optimization.
During Every instance, the Eclipse program will fetch the potentiometer details which in real time
correspond to the Load Required, Solar Power that can be Generated, etc.,. The Eclipse Communicates
with the Arduino at a standard baud rate of 9600 symbols/sec. At the end of every instance, ECLIPSE will
send information to ARDUINO in an asynchronous communication language that contains the pattern of
LED’s that has to glow for that instant. This will be parsed by the Arduino to issue the high command to
corresponding digital output pins. These LED’s in real time correspond to the command signal issued by
ECLIPSE MAS System to the Actuators that allows the power flow between various sources and loads.
All these actions take place parallel in Eclipse with the help of MAS concept for faster actions. During
a grid failure, the entire micro-grid will operate under islanding mode. The MAS System senses such
faults and initializes non critical load shedding. Loads such as Air Conditioner, Refrigerator and Washing
Machine are considered as non-critical loads.
5. PROBLEM FORMULATION
The challenges due to penetration of renewable energy resources like solar and wind have to be dynamically
managed in order to maintain stability, reliability and fault tolerance. In the solar micro-grid demand side
management and grid outage is implemented autonomously by using multi agent system and the agents
operations are practically verified using Arduino microcontroller.
6. IMPLEMENTATION OF DEMAND SIDE MANAGEMENT AND GRID OUTAGE
MANAGEMENT
The environment variables i.e. Load Demand, Solar 1 Generation, Solar 2 Generation, State of Charge of
the Battery, Information on grid availability and the NC Load are sensed through the potentiometer and
are given to the microcontroller. Using these values, the program calculates the optimal power distribution.
This is reflected in microcontroller which is given to the physical device for actuation (LEDs).In the flow
chart, initially the load gets power from solar1 and then if required it gets from solar 2. Then the remaining
excess solar power is used to charge the battery and finally the excess power is given to the grid. If there
is shortage of power for load then the load receives power from both the solar and then from the battery. If
grid is available the remaining required power is received from the grid. During the grid outage the micro-
grid gets isolated and it balances the power requirement by appropriate load shedding.
160 Leo Raju, Prithika Ravi M, Sajna. G and Srinidhi. J

Start

N
L>0
Y
N L=0
L>S1
S1 = S1 – L
Y
S1 = 0 Export to Bt
L = L – S1
Export to Grid
N
L>0

Y
N
L>S2 L=0
S2 = S2 – L

S2 = 0 Export to Bt
L = L – S2
Export to Grid
N
L>0
Y
N
L=0 L>Bt
Bt = Bt – L
Y
Bt = 0
L = L – Bt

N
L>0

Y
N NC LOAD
Grid! = 0 SHEDDING
Y
Get remaining L
from Grid

End

Figure 1: Flowchart

7. SIMULATIONS AND RESULTS

All the operations are considers as shown in the flow chart and for these scenarios, the java programming
is done in JADE [10] environment and executed in Eclipse Integrated Development Environment. Various
scenarios are considered and sniffer diagrams and the console output representing the interaction of the
agents and transaction details are studied. The Arduino board is connected to the bread board where all
the potentiometer and LEDs. This Arduino board is connected with computer serial port through ethernet
card. Arduino software is run to upload the program into the board. The potentiometers are fixed with the
 Arduino based Autonomous Energy Management of a Micro-Grid using Multi Agent System 161
environment variables of solar values, load values, battery values, grid value and Non Critical Load (NCL)
value. Four LEDs are fixed on the bread board representing two solar units, a battery unit, and grid.The
JADE agents are made to run in the Eclipse environment. The agents receive environment variables from
the pot through Arduino microcontroller and take strategic decisions by communicating and coordinating
with other agents. The resulting actions in the form of command signals are given to the various micro-
grid components represented in the form of LEDs through Arduino. In this case, all the LEDS glow as
all the agents are are involved in the process. The potentiometers can be varied for various environment
values and the LED actions are verified with environmental dynamics
Table 1
Ttest Case 1 – Sensed Values

Input Pins Agents Equivalent Power

A0 Load (kW) 100
A1 Solar 1 (kW) 50 50
A2 Solar 2 (kW) 30 30
A3 Battery (kW) 10 10
A4 Grid Zero (Outage)
A5 NC (%) 10

Process :
• Though the load needs 100 kW, due to the Grid outage and non-availability of enough power
sources, the NCL are shed to manage the deficiency.
• 90 kW is the demand now. This is satisfied by Solar 1, Solar 2 and the Battery.
• As Solar 1, Solar 2 and Battery give power to the Load after shedding, LEDs at pins 8, 9 and 10
glow.
• The further required 10Kw is managed by shedding Non Critical Load.

Battery Grid Load1 Solar1 Solar2

INFORM:-1 ( )

REFUSE:-1 ( )

PROPOSE:-1 ( )

REJECT-PROPOSAL:-1 ( )

REQUEST:-1 ( )

FAILURE:-1 ( )
SUBSCRIBE:-1 ( )

CONFIRM:-1 ( )

Figure 2: TEST CASE 1- Sniffer

Console Output : The console output the and JADE sniffer and Arduino output snapshot are shown
in Fig.4., Fig 5 and Fig 6.
162 Leo Raju, Prithika Ravi M, Sajna. G and Srinidhi. J

Figure 3: Test Case-1 console output

Hardware Output :
Table 2
Test Case-1 hardware output

Output Pins Pin 8 Pin 9 Pin 10 Pin 11

Agents Solar 1 Solar 2 Battery Grid

On/Off On On On Off
 Arduino based Autonomous Energy Management of a Micro-Grid using Multi Agent System 163

Battery Grid Load1 Solar1 Solar2

INFORM:-1 ( )

REFUSE:-1 ( )

PROPOSE:-1 ( )

REJECT-PROPOSAL:-1 ( )

REQUEST:-1 ( )

AGREE:-1 ( )

Figure 4: Sniffer Solar 1 giving power to the Load

Figure 5: Arduino with hardware

164 Leo Raju, Prithika Ravi M, Sajna. G and Srinidhi. J

Figure 6: Hardware Output

8. CONCLUSION
The demand side management and grid outage management is implemented in a solar micro-grid with
a Multi-Agent System approach. A MAS model was developed for the solar micro-grid by using JADE
and all the options available for the agents in the micro-grid are comprehensively analysed for optimal
energy management of a solar micro-grid.The operations of agents are practically verified by Arduino
microcontroller. Future work will focus on applying to IOT into our system to make it smarter.
9. REFERENCES
1. M.Pipattanasomporn, H.Feroze andS.Rahman,“Securing critical loads in a PV-based micro-grid with a multi-agent
system,” Renewable Energy, 39(1), 166-174,2012.
2. T. Logenthiran, D.Srinivasan, AM. Khambadkoneand H.NAung,“Multi agent system for real-time operation of a
micro-grid in real-time digital simulator,” IEEE Transaction on Smart Grid, 3(2), 925-933. 2012.
3. HSVSKNunna and S. Doolla, “Multi agent-based distributed-energy-resource management for intelligent micro-
grids,” IEEE Transactions on Industrial Electronics, 60(4),1678-87, 2013.
4. J Gomez-Sanz, S.Garcia, S. Rodriguez, N. Cuartero –Soler andL. Hernandez-Callejo,“Reviewing micro-grids from
a multi-agent systems perspective,” Energies. 7(5), 3355-3382,2014.
5. ND. Hatziargyriou, Micro-grids Architecture and Control. Wiley-IEEE Press; 2014.
6. YSF Eddy, HB. Gooi and SX Chen,“Multi-agent system for distributed management of micro-grids. IEEE Transactions
on Power Systems,”30(1),24-34,2015.
7. GH. Merabet, M. Essaaidi, MR.Talei,N.Khail,NM. Madkour and D Benhaddou,“Applications of Multi-Agent
Systems in Smart Grids: A survey,” Proceedings of 2014 International Conference on Multimedia Computing and
Systems (ICMCS), IEEE, 2014.
8. J. Lagorse, D. Paireand A. Miraoui,“A multi-agent system for energy management of distributed power sources,”
Renewable Energy,5(1), 174-182.2013.
9. S. Calderwood, W. Liu, J. Hong and M. Loughlin, “An architecture of Multi Agent System for SCADA-dealing with
uncertainty, plan and actions,” Proceedings of ICINCOPP,300-306,2013.
10. Java Agent Development Environment (JADE) development toolkit. Available from: http:// www.jade.tlab.com.