Student Poster Competition
4 minute ‘elevator pitches’
Project 1.1
Study of Dynamic Characteristics of
Autonomous Droop-controlled Microgrids
ABOUTALEB HADDADI (McGill University)
AMIRNASER YAZDANI (Ryerson University)
BENOIT BOULET (McGill University)
GEZA JOOS (McGill University)
www.smart-microgrid.ca
Problem statement
• A challenging task is to
regulate the voltage
amplitude and frequency of
an autonomous microgrid:
– Multi-input multi-output
(MIMO),
– Time varying,
– Non-linear,
– Considerable transient and
steady-state impacts
imposed by loads.
Problem statement
• Most commonly, the goal is achieved by coordinated control of
multiple Distributed Energy Resource (DER) units through droopbased control.
• Advantages:
–
–
–
–
Simple structure (use of SISO controllers),
Plug-and-play feature,
No need for communication between the DERs,
Enables power sharing.
• Challenges:
– Sensitivity of control to droop gains,
– Sensitivity of control to the steady-state power flow (real- and reactivepower outputs of DGs),
– Dependence of stability on the network and loads,
– No control over transient performance.
Challenges of droop control ̶ Example
Example: Poor transient performance of droop-based control due to dynamic coupling
between the DER units and loads in a droop-controlled autonomous microgrid ̶ connecting
an induction machine load to a test microgrid
•
•
•
•
A 12.47-kV North american Distribution network, Autonomous
mode of operation,
Droop-based voltage and frequency control,
An induction machine load is
connected to Bus 6 at 𝑡 = 6.5
s,
The network becomes unstable
following the connection of the
induction machine load.
Figure 2: Response of the DERs to connection of
an induction machine to the test microgrid
showing poor stability of droop-based control due
to dynamic coupling between the DER units and
loads.
Figure 1: The test microgrid.
A Robust Hierarchical Control Structure for
Virtual Power Plants
Project 1.2. Distributed and Hybrid Control
Fahimeh Kazempour and Reza Iravani
University of Toronto
Problem Statement
• Unlike the conventional grid-connected mode, a microgrid in the Virtual
Power Plant (VPP) mode of operation is obliged to:
 Exchange a specified active and reactive power at the PCC,
 Maintain PCC's voltage and frequency within specified values
 Provide dynamic power balance within the microgrid.
• A three-level hierarchical control structure needs to be employed to
realize the VPP mode of operation.
• A VPP is subject to:
 Operating point changes,
 Parametric and topological uncertainties, and unmodeled dynamics
in both the local DERs and their interactions,
 Exogenous disturbances.
– At the primary control level, a robust control strategy is used to
synthesize DER’s local controllers.
The Proposed Solution
• The method of decentralized robust H∞ control with integral quadratic
constraint (IQC) uncertainty description is employed to develop
dynamic reference-tracking controllers.
• Using this method, necessary and sufficient conditions are derived to:
 Guarantee the stability of the overall VPP system
 Minimize the H∞-norm bound on a map from the disturbance input to the
controlled output
• Our method only requires solving a set of game type algebraic Riccati
equations
• Algebraic Riccati equations are presented in terms of rank constrained
LMIs
• A suitable test system is employed, to investigate the robust dynamic
performance of the VPP system under the proposed control structure.
Project 1.3
Current State Estimation for Microgrid
Ahda P. Grilo, Pengfei Gao and Wilsun Xu (University of Alberta)
www.smart-microgrid.ca
Current State Estimation for Microgrid
• Smart grids will increase load management
opportunities for customers.
• For microgrids the real-time power consumption
of loads is a useful information for energy
management systems.
• In commercial facilities it is difficult to measure
currents.
 Idea: Voltages at the load terminals are often
accessible and can be measured by distributed
voltage sensors.
Current State Estimation for Microgrid
Proposed method:
1) To measure voltages at the load terminals by
distributed voltage sensors.
2) Use these voltages, combined with the
network topologies and parameters, to
estimate the load currents.
Traditional:
Microgrid:
Current State Estimation for Microgrid
 The proposed technique represents a good solution for
microgrid facilities where the load conductors are inaccessible
for current sensing.
 The idea of using voltages to estimate currents as presented in
this paper has some other applications. For example, it could be
used to monitor home appliance behavior by using distributed
voltage sensors installed at various locations of a home.
A Stochastic Simulation Tool for Studying the Electric
Features of Micro-grid
Qingxin Shi and Ricardo Torquato, Department of Electrical and Computer
Engineering, University of Alberta
The development of smart micro-grid requires a lot of planning and
monitoring studies. Therefore we need to model the real-time performance of
the micro-grid. The model should take into consideration all kinds of random
factors, such as random behavior of home appliance and random power
generation of photovoltaic (PV) panel. This poster presents a simulation
platform to handle this work.
The application of this platform is: to simulate the real-time network
response of the random loads and random PV generation, helping the engineers
to manage the micro-grid.
Project 1.3
Contribution of DGs to fault current and
their impacts on overcurrent Protection
H. Yazdanpanahi
PDS Lab, Department of Electrical & Computer Engineering, University of
Alberta
www.smart-microgrid.ca
Problem
• In spite of their undoubted advantages,
DG units impact on over-current
protection by contributing to fault
current.
DG
P2
Net
Net
R
F
P1
DG
Fault
Nuisance (sympathetic) tripping
Failure in fuse-saving scheme
Time (s)
A
B
Net.
Back-up
Back-up
Main
Main
DG
Fault
If1
Miscoordination between main and back-up
If2
Current
Research
Four types of DGs have been investigated.
Inverter-based
Synchronous Machine
Mechanical torque
Grid
DC link
Rotor-side
converter
PMSM
Grid-side
converter
Induction Machine
For each type, the magnitude and duration of the
current is assessed by analysis and simulation from
the perspective of relay coordination.
Project 1.4
Operational Strategies and Storage Technologies
to Address Barriers for a Very High Penetration
of DG Units in Intelligent Microgrids
Michael Ross (McGill University)
Dr. Chad Abbey (Hydro-Québec)
Professor Géza Joós
www.smart-microgrid.ca
Problem: Optimized Microgrid Dispatch with a Very High Penetration of
Renewable Energy
Energy
Storage
Demand
SystemResponse
Point of Common Coupling
Renewable
Diesel Generator
Energy
Distributed
Generation
- Time-dependent
- Ability
resource
to curtail non-critical load.
- Limit power fluctuations
- Fluctuating
- High greenhouse
and
variable
gas emissions
power
output
-fuel
Limited
by power
- achieve
Critical
and energy
load must
rating
be supplied with
- Reduce the peak power
- Power
- Expensive
flowbalance
through
isPCC
cost
difficult
to
energy (high reliability)
Electric
Power
System
Wind
Generation
(50%
penetration)
Solar
Generation
(50%
penetration)
Energy
Storage
System
Diesel
Generator
Non-Critical
Loads
Critical Loads
Proposed Solution: Multi-Objective Optimization Dispatch for Microgrids
 A Microgrid controller must be developed to coordinate the
control of available Distributed Energy Resources (DER) to:
•
•
•
•
•
Minimize cost of energy
Reduce peak power through PCC
Minimize power fluctuations through PCC
Improve reliability
Reduce GHG emissions
 The multiple benefits can be optimized while mitigating the
adverse effects of renewable energy integration.
Project 2.1
Cost-Benefit Framework: Secondary
Benefits and Ancillary Services
MIKE QUASHIE AND GEZA JOOS (MCGILL UNIVERSITY)
www.smart-microgrid.ca
A Methodology to Optimize Benefits
of Microgrids
• Objective:
This paper proposes a generalized methodology to determine
the optimal configuration of microgrids that maximizes its
benefits.
The technique proposed herein, incorporates an advanced
modeling of variable loads and distributed generators into a
power flow problem to solve the benefits’ optimization problem:
𝑝
min 𝐹 =
ML t, n, µ, σ + GCC t, n + GHC(t, n)
(1)
𝑛=1
Subject to:
𝑃𝐺,𝑗 − 𝑃𝐿𝑜𝑠𝑠,𝑗 t, n, µ, σ + 𝑃𝑊,𝑗 t, n, µ, σ + 𝑃𝑆,𝑗 t, n, µ, σ = 𝑃𝐿𝑜𝑎𝑑,𝑗 t, n (2)
• 𝑃𝑊,𝑗 t, n, µ, σ + 𝑃𝑆,𝑗 t, n, µ, σ ≤ 𝑃𝑝𝑒𝑛,𝑗 × 𝑃𝐿𝑜𝑎𝑑,𝑗 t, n …..(3)
The Methodology is applied to real
feeder in north America.
The study shows significant
reduction in cost of energy which
seek to advance the business case
of microgrids . It also provides
potential investors and
stakeholders a planning strategy to
maximize the benefits accrued
from microgrids.
Figure 1. Cigre's North American medium Voltage Distribution
Network Benchmark with DG connected to operate as Microgrid
A Novel Affine Arithmetic Method to Solve
OPF Problems with Uncertainties in
Microgrids
Mehrdad Pirnia
Claudio Cañizares
Kankar Bhattacharya
Alfredo Vaccaro
Department of Electrical & Computer Engineering
Motivation
• Increased focus on renewable generation has brought forth
many concerns in planning and operation of microgrids.
• Margins of operation for thermal generators are needed to
provide system reliability and efficiency in view of the
variability brought about by DR and DG technologies.
• Common methods to consider uncertainties from
renewable sources integration (e.g., Monte Carlo
Simulation) rely on pdfs of random variables and are not
efficient.
• Self Validated Computation Methods (SVC) do not need
pdfs and are efficient:
– Interval Arithmetic (IA)
– Affine Arithmetic (AA)
24
Methodology and Results
• Develop an accurate and efficient AA-based OPF
model to incorporate uncertainties in microgrids.
• Validate the AA-based operation system models
with the MCS based method.
• Use the resulting AA based intervals to estimate the
spinning reserve requirements in the presence of DR
and variable DG penetration in microgrids.
• Test and validate the model on a benchmark
microgrid.
25
EFFECT OF PRICE-RESPONSIVE DEMAND
ON DISPATCH AND COSTS IN
MICROGRIDS
Felipe Ramos-Gaete
Claudio Cañizares
Kankar Bhattacharya
Department of Electrical & Computer Engineering
Motivation
• With smart microgrids, loads in the grid are required to
react to stressful conditions of the system.
• Depending on load reaction, cost inefficiencies, line
congestions and even energy shortage may occur.
• It is necessary to examine the effect of price-responsive
demand on smart microgrid unit commitment of
distributed generators and storage.
• In the long run, is also relevant to study the interrelationship between demand elasticity and electricity
prices.
27
Methodology and Results
•
•
•
•
Develop models to represent demand response in dispatch.
Develop a microgrid unit commitment and dispatch model.
Establish a link between prices and customers’ response.
Some interesting findings:
– More elastic demand can make the system less stable, increasing or
decreasing the demand beyond feasible operating points.
– Multi-period price-responsiveness presents a similar behaviour as
energy storage systems, soothing system variability while minimizing
total operation cost.
– A corrected real-time pricing scheme can be derived from this work,
which would allow more controllability over demand response.
28
Stability and Control of Unbalanced
Synchronous Machine Based Distributed
Generators
Ehsan Nasr Azadani, Claudio Canizares, and Kankar Bhattacharya
Dept. Electrical and Computer Engineering, University of Waterloo
www.smart-microgrid.ca
Motivation
• Rapid development and increase in penetration of
decentralized or distributed generation (DG).
• Transition from a passive grid containing only loads to an
active grid, including loads and “small” generation.
• The dynamics of both transmission and distribution system
are affected.
• Lack of knowledge of the dynamics of DGs under unbalanced
conditions.
• A full characterization of the unbalanced system in stability
analyses would allow a better understanding of dynamic
behaviour of DGs.
• Most DGs nowadays are equipped with small synchronous
generators (e.g., diesel generators, microturbines)
Objectives
• Develop both static and dynamic models of synchronousmachine DG under unbalanced conditions.
• Perform:
– Voltage stability studies based on P-V and P-L curves.
– Small perturbation stability studies using a model
identification approach to compute the eigenvalues.
– Transient stability studies based on time domain
simulations to study contingencies.
• Propose a control strategy to improve the stability of
distribution systems with synchronous-machine DG units.
Project 2.4
Stochastic evaluation of transient stability
of Micro-grids
Mayssam Amiri, Ani Gole, Tomás Yebra Vega (University of Manitoba)
www.smart-microgrid.ca
Stochastic evaluation of the stability in a Micro-Grid is more
realistic and less conservative.
Model
Statistical approach
µ- Grid
• 100’s or 1000’s of variables
• Closer to reality
• Less conservative than
traditional methods
Monte Carlo Method
• Time consuming
Mathetical Generator
Results
Evaluation
Wind turbine
DG
Parallel Computing
Probability of failure
caused by unstability
• Discrimination type of Fault
Model
SM 0.0
SM 2.1
Prob.
LLL &
LLLG
0.0367
0.0254
Prob.
LL &
LLG
0.1047
0.025
Prob.
LG
0.1554
0.0013
PDF’s
PV
µ-turbine
Small SM
SM 0.0
SM 2.1
• Influence of the
mathematical model
• Influence of µ-Grid design.
• Influence of the degree of
penetration of DG’s
Project 3.1 Universal Communication Infrastructure
Are Omnidirectional Terminal Station Antennas a
Better Choice for Point to Multipoint Deployments
in NLOS Environments?
Prof. David G. Michelson (University Of British Columbia)
Sina Mashayekhi (PhD student)
www.smart-microgrid.ca
Problem: Directional Antennas Performance in NLOS Multipath
Environment
• Minimum requirement for P2MP radio systems for management of the
electricity supply in Canada:
– Minimum 12 dBi gain , maximum beamwidth of 30o
– Appropriate for high capacity systems operating under LOS
– What about for relatively low capacity systems in NLOS condition?
• Directional antennas do not have 100% of expected performance in NLOS
–
–
–
–
Broadening the pattern, Gain reduction
Reduction in Average Area Spectral
Increase of Co-channel Interference
Reduction in Coverage
Solution: Revising SRSP 301.7 to allow using Omni antennas for critical
applications in SG
• Impacts of Using Omni for P2MP deployments:
– Co-channel Interference, Coverage, ASE
– Increased second best server redundancy
– Cost effective deployments and Maintenances
THEME 3, PROJECT 3.2
Throughput analysis of Narrow-band Power Line
Communications in Advanced Distribution Automation
Chon Wang Chao (MEng Student)
Quang-Dung Ho (Research Associate)
Tho Le-Ngoc (McGill University)
www.smart-microgrid.ca
Overview of the project
Fig. 1 – Studied communications architecture of PLC for ADA
NSMG-Net Project 3.2: Chon Wang Chao
Contribution of the project
• Data rate estimation
– Packet structure specified by IEC 61850 clientserver communications
– Expected data requirements for advanced
distribution automation
• Impact of communication channel competition
– Throughput variation
– Bandwidth requirement
– Improvement from using Clear to Send/ Request
to Send mechanism
NSMG-Net Project 3.2: Chon Wang Chao
THEME 3, PROJECT 3.2
Frequency Regulation by Aggregator-based Electric
Vehicles Charging Control via Wireless CommunicationsChon
Chon Wang Chao (MEng Student)
Quang-Dung Ho (Research Associate)
Tho Le-Ngoc (McGill University)
www.smart-microgrid.ca
Overview
of
the
Project
• Frequency regulation with the battery capacity of electric vehicles (EV)
• Intelligence resides at the aggregators which are used to coordinate the
responses of the EVs
Fig. 2 – the proposed control and communications architecture of FR with EV
NSMG-Net Project 3.2: Chon Wang Chao
Contribution of the project
• Index system for selecting EV to participate in FR
– Map the EV private information (e.g. SOC,
departure time) to an index
– Aim to reduce the privacy concerns
• Impact of communications on FR
– Frequency regulation performance (e.g. stability)
under non-ideal communications
– Communication delay
– Packet Loss
NSMG-Net Project 3.2: Chon Wang Chao
THEME 3, PROJECT 3.2
EFFICIENT COMMUNICATION ARCHITECTURE FOR
INTELLIGENT MICROGRIDS
Tho Le-Ngoc (McGill University)
Quang-Dung Ho (Research Associate)
Yue Gao (MEng Student)
Gowdemy Rajalingham (MEng Student)
www.smart-microgrid.ca
Proposed System Architecture
1026
Endpoints
Router
Collector
LTE
UTILITY
Wired Backhaul
xxx
xxx
xxx
xxx
Command
Center
Fig. 1 – Neighbor Area Network
Performance Evaluation
Objective
• Determine capabilities and limitations of NAN with GPSR
• Investigate NAN clusters performance with various system parameters
Fig. 2 – Simulation Scenario, sweep of cluster size
TABLE 1 – SIMULATION PARAMETERS
Channel Model
Simple pathloss – pathloss exponent 𝜂, and Lognormal shadowing – variance 𝜎
MAC layer
IEEE 802.11
Routing Protocol
Greedy Perimeter Stateless Routing (GPSR)
Performance Metrics Packet Transmission Delay & Packet Delivery Ratio (PDR)
Traffic
Per-node data rate - 𝑟
Topology
Clusters of size - 𝑛
System Parameters
Sweeps of
• Variance 𝜎, default value of 𝝈 = 𝟒 [dB]
• Per-node data rate 𝑟, default value of 𝒓 = 𝟎. 𝟏 𝐩𝐚𝐜𝐤𝐞𝐭𝐬 𝐬 𝐧𝐨𝐝𝐞
• Clusters of size 𝑛, default value of 𝒏 = 𝟏𝟎𝟎𝟎 𝐧𝐨𝐝𝐞𝐬
NSMG-Net Project 3.2: Gowdemy Rajalingham
Project 3.3 Distribution Automation Communications:
① What is the problem for substation communication systems:
I.
The ZigBee wireless platform is a cost-effective wireless sensor networking system that can be
used to monitor substation components in electric substations.
II.
Impulsive noise with a short duration and strong energy content caused by partial discharge of a
dielectric breakdown can degrade the communication quality of ZigBee nodes.
② What are we doing for the above problem:
I.
Modelling the sequence of PD impulsive noise.
II.
Analysis of the impact of impulsive noise on ZigBee systems operated at 915 MHz and 2.4 GHz
bands.
③ What is the contribution of our research:
I.
Assessment of ZigBee operational bands that are more resistant to impulsive noise in electricity
substations.
II.
Possibility of a ZigBee sesnor network that can be utilized for the detection of partial discharge
events.
Project 3.4
Integrated Data Management and Portals
Student:
Moein Manbachi (Simon Fraser University)
Supervisors:
Dr. Hassan Farhangi (British Columbia Institute of Technology)
Dr. Ali Palizban (British Columbia Institute of Technology)
Dr. Siamak Arzanpour (Simon Fraser University)
www.smart-microgrid.ca
Canada Electric Power Generation, Consumption and T&D Losses
*Ref. International Energy Agency (http://www.iea.org)
 In 2010, Canada T&D Loss value: 65.661 Billion-kWh.
 Average Transmission & Distribution Losses: % 8.5388
 About %40 of T&D Losses occurs in Distribution Network
Ref. http://www.electricenergyonline.com
 Importance of Energy Conservation &
Distribution Network Optimization
Volt/VAR Optimization
(VVO)
 VVO Technique can be improved by new Smart Grid Technologies
More efficient solution with better results
8. Future
Plans &
Targets
7. Case Study
Results
6. What will
be the next
generation of
VVO?
1. How Smart
Grid new
features can
help VVO?
Project 3.4:
Real-time
Smart Grid
Adaptive
Volt/VAR
Optimization
5. Where we
are now?
How far we
can go?
2. What is our
proposed
solution?
3. What are
the key
benefits of
our solution?
4. Centralized
Control Vs.
Decentralized
Control