Intelligent MicroGrid Communication Networks - NSMG-Net

Report
T HEME 3, P ROJECT 3.2
T HO L E -N GOC (M C G ILL U NIVERSITY )
Q UANG -D UNG H O (R ESEARCH A SSOCIATE )
G OWDEMY R AJALINGHAM (ME NG S TUDENT )
C HON -W ANG C HAO (ME NG S TUDENT )
Y UE G AO (ME NG S TUDENT )
2
SUMMARY
Technology Integration and Network Architecture Design
• Performed an evaluation of promising wired and wireless technologies
• Conducted a comparative study of potential radio access technology interconnections
• Proposed the integration of Wi-Fi Mesh and LTE for the network architecture
Feasibility of wireless mesh for the NAN
• Determined promising routing protocols (GPSR, RPL) for the NAN
• Evaluated performance of GPSR in wireless mesh NAN
Applicability of PLC for Advanced Distribution Automation
• Estimated data rate requirements with IEC 61850 based messaging in ADA scenarios
• Examined impact of channel contention at MAC layer on achievable PLC throughput
Frequency Regulation Using EV Charging Control over Wireless Communications
• Proposing a new aggregator based EV charging control scheme with priority indices
• Proposing joint simulation platform to study the effects of communication on FR
NSMG-Net Project 3.2: Gowdemy Rajalingham
AGM 2013, Vancouver
3
PROPOSED SYSTEM ARCHITECTURE & EVALUATION
1026
Endpoints
Router
Collector
LTE
xxx
Wired Backhaul
UTILITY
xxx
xxx
xxx
Command
Center
Fig. 1 – Neighbor Area Network
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
NSMG-Net Project 3.2: Gowdemy Rajalingham
TABLE 1 – SIMULATION PARAMETERS
Channel Model
Simple pathloss – pathloss exponent
Lognormal shadowing – variance
MAC layer
IEEE 802.11
Routing Protocol
Greedy Perimeter Stateless Routing (GPSR)
Performance Metrics
Packet Transmission Delay, Packet Delivery Ratio
Traffic
Per-node data rate -
Topology
Clusters of size -
System Parameters
Sweeps of per-node data rate & clusters of size for [dB]
AGM 2013, Vancouver
4
FEASIBILITY OF CANDIDATE ROUTING FOR NAN
Performance Versus Cluster Size and Data Rate
Current Estimates
• Base Rate: NIST Data Rate of 0.00195 pps (based on simple meter readings)
• Typical AMI deployment NAN size: A few 1000s of smart meters
Results
• Can maintain latency < 100 ms for up to 6000 nodes for data rates up to 10x base data rate
• Can maintain PDR > 95% for up to 6000 nodes for data rates up to 10x base data rate
• At 100x base data rate, to maintain latency < 100 ms and PDR > 95%, cannot exceed a cluster size of roughly 1500
10000
0.9
0.8
Packet delivery ratio PD
Delay [ms]
1000
1.0
pps = 0.001
pps = 0.01
pps = 0.1
100
10
1
0.7
0.6
0.5
0.4
0.3
0.2
pps = 0.001
pps = 0.01
pps = 0.1
0.1
0.1
10
100
1000
Cluster size n [node]
Fig. 3 –  = , 95% Percentile of delay vs. Cluster size
NSMG-Net Project 3.2: Gowdemy Rajalingham
6000
0.0
10
100
1000
6000
Cluster size n [node]
Fig. 4 –  = , Packet Delivery ratio vs. Cluster size
AGM 2013, Vancouver
5
PUBLICATIONS
[1] Quang-Dung Ho, Gao Yue and Tho Le-Ngoc, “Challenges and Research Opportunities in Wireless Communication Networks for Smart
Grid”, IEEE Wireless Communications Magazine, June 2013.
[2] Chon-Wang Chao, Quang-Dung Ho and Tho Le-Ngoc, ”Challenges of Power Line Communications for Advanced Distribution
Automation in Smart Grid”, 2013 IEEE Power and Energy Society General Meeting, Vancouver-Canada, July 21-25 2013.
[3] Gowdemy Rajalingham, Quang-Dung Ho and Tho Le-Ngoc, “Attainable Throughput, Delay and Scalability for Geographic Routing on
Smart Grid Neighbor Area Networks”, 2013 IEEE Wireless Communications and Networking Conference (WCNC 2013), Shanghai-China, 710 April 2013.
[4] Gowdemy Rajalingham and Quang-Dung Ho, “LTE HetNets: Challenges and Opportunities for Integration of Smart Grid Networks”,
Technical Report, McGill, April 2013.
[5] Chon-Wang Chao and Quang-Dung Ho, “Communication Standard and Network Infrastructure Considerations for Smart
Grid”, Technical Report, McGill, 2012.
[6]Yue Gao and Quang-Dung Ho, “OMNET Implementation of RPL for Smart Grid Neighbor Area Networks”, Technical Report, McGill,
December 2012.
NSMG-Net Project 3.2: Gowdemy Rajalingham
AGM 2013, Vancouver
6
7
INTRODUCTION
Objective
Project 3.2 aims to study and to develop relevant transmission, information processing, and networking techniques
for an efficient and reliable IMG Communication Network (IMGCN)
Issues
The successful implementation of the Intelligent MicroGrids (IMGs) requires an efficient communications
infrastructure that is cost-effective, scalable, fault-tolerant, secure & satisfies the QoS requirements (data rate,
delay, reliability)
Microgrid
Customer
Substation
Non-Renewable
Energy
Smart
Meter
Substation
Electric Vehicle
Microgrid
Wind Enegy
Solar Enegy
Power System Layer
Control Center
Wired Backhaul
Network
3G/4G Cellular, Ethernet,
Leased Line, Fiber Optics,
Satellite...
 10-100 Mbps
Smart
Home
Devi ce
802.15.4, 802.11,
Wireless Mesh,
 10-100 Kbps
 Coverage of up to several 1000 km 2  Coverage of up to several km 2
802.15.4, Zigbee, 802.11,
 10-100 Kbps ...
 Coverage of up to 100 m2
Wide Area Network (WAN)
Neighbor Area Network (NAN)
Home Area Network (HAN)
Communications Layer
Fig. x – Full Abstract System Architecture Model
NSMG-Net Project 3.2: Gowdemy Rajalingham
AGM 2013, Vancouver
8
9
APPLICABILITY OF POWER-LINE COMMUNICATIONS
Key Contributions
•
•
•
Calculated the expected data rate requirements with IEC 61850 message architecture and power network
parameters
Examined the impacts of channel competition with Carrier Sense Multi-Access/Collision Avoidance (CSMA/CA)
algorithm on saturation throughput (T) and bandwidth requirement
Further details in poster “Throughput Analysis of Narrowband PLC in Advanced Distribution Automation”
Fig. x – Communication PLC Network
NSMG-Net Project 3.2: Gowdemy Rajalingham
AGM 2013, Vancouver
10
APPLICABILITY OF POWER-LINE COMMUNICATIONS
Summary of Results
•
•
Expected data rate with only PLC supporting advanced distribution automation is 310.69 kbps
Throughput and bandwidth requirement variation
• T decreases as the number of nodes increases (higher probability of collision)
• The optional Request to Send/Clear to Send mechanism can increase the T with same number of nodes and reduce to
growth rate of bandwidth requirement
•
•
Existing field tested PLC technology may not be able to provide enough data rate
Further details in poster “Throughput Analysis of Narrowband PLC in Advanced Distribution Automation”
Fig. x – Communication PLC Network
NSMG-Net Project 3.2: Gowdemy Rajalingham
AGM 2013, Vancouver
11
12
FREQUENCY REGULATION USING EV CHARGING CONTROL
Key Contributions
•
•
•
Proposed a new aggregator based electric vehicle charging control scheme with priority indices
Proposed to use the joint simulation platform to study the effects of communication delays and
packet loss
Further details in poster “Cost-Effective Frequency Regulation by Aggregator-based EV Charging
Control via Wireless Communications”
Fig. x – Proposed Control Structure and Neighborhood Mapping
NSMG-Net Project 3.2: Gowdemy Rajalingham
AGM 2013, Vancouver
13
FREQUENCY REGULATION USING EV CHARGING CONTROL
Control System Model
•
Further details in poster “Cost-Effective Frequency Regulation by Aggregator-based EV Charging Control via
Wireless Communications”
Fig. x – Proposed Control Block Diagram and Joint Simulation Setup
NSMG-Net Project 3.2: Gowdemy Rajalingham
AGM 2013, Vancouver
14
FREQUENCY REGULATION USING EV CHARGING CONTROL
Communications Model
•
Further details in poster “Cost-Effective Frequency Regulation by Aggregator-based EV Charging Control via
Wireless Communications”
Fig. x – EV Selection Algorithm
NSMG-Net Project 3.2: Gowdemy Rajalingham
AGM 2013, Vancouver
15
FREQUENCY REGULATION USING EV CHARGING CONTROL
Illustrative Example
•
Further details in poster “Cost-Effective Frequency Regulation by Aggregator-based EV Charging Control via
Wireless Communications”
Number of EV = 120
Ithreshold
= 98
Fig. x – Illustrative Example
NSMG-Net Project 3.2: Gowdemy Rajalingham
AGM 2013, Vancouver
16
17
SURVEY OF TECHNOLOGIES
Wired Technologies
• Economically feasible when network cables and related facilities are pre-existing and readily available at
acceptable low costs
• More suitable for back-haul links for large volume of traffic
• Example: Digital subscriber line (DSL), leased line, power line communications (PLC), fiber optics …
Wireless Technologies
• Home Area Network
• 10-100 kbps
• Coverage area of up to 100 m2
• Example: ZigBee, WirelessHART,
6LowPan, Bluetooth, …
• Neighbor Area Network
• 10-100 kbps
• Coverage area of up to several km2
• Example: Wi-Fi, Wi-Fi Mesh, …
• Wide Area Network
• 10-100 Mbps
• Coverage area of up to several 100 km2
• Example: WiMax, LTE, …
Fig. x – Potential Technologies
NSMG-Net Project 3.2: Gowdemy Rajalingham
AGM 2013, Vancouver
18
CANDIDATE WIRELESS ARCHITECTURES
LTE
TABLE X – LTE PERFORMANCE CHARACTERISTICS [1]
LTE Small Cell
Latency
• Best case latency
• 6 ms for short packets (<40 bytes)
• 11 ms for longer packets (>40 bytes)
Max
Users
• Max users is less than the number of LTE
Resource Blocks available
• LTE Control Channels are bottlenecks
• Data aggregation is necessary
(c) Multi-hop Transmission with LTE Small cells
eNB
Wi-Fi
Smart Meter
M2M Gateway
UE
(a) Direct Transmission
(b) Multi-hop Transmission with Wi-Fi clusters
Fig. X – Potential NAN/WAN Interconnections
TABLE X
INTERFERENCE
LATENCY
THROUGHPUT
SCHEDULING
SELF-ORGANIZING
NETWORKS
DIRECT
TRANSMISSION
Interference
LTE uplink
latency
• No data aggregation
• Wasted capacity
High complexity
No need for SONs
MULTI-HOP WITH
WI-FI CLUSTERS
• Out of band w.r.t.
LTE
• Interference
Extra tiering delay
• Data aggregation at GW &
in mesh network
• Potential for network coding
Lower
complexity
(aggregation)
SONs are Wi-Fi
mesh networks
MULTI-HOP WITH
LTE SMALL CELLS
• Interference
• Coverage gaps
• Power control
needed
Extra tiering delay
Data aggregation at small cell
BS
Lower
complexity
(aggregation)
Need for SONs for
LTE femto-cells
LEGEND: Excellent, Adequate, Deficient
NSMG-Net Project 3.2: Gowdemy Rajalingham
AGM 2013, Vancouver
19
PROPOSED SYSTEM ARCHITECTURE
TABLE X – LTE PERFORMANCE CHARACTERISTICS [1]
LTE
Latency
• Best case latency
• 6 ms for short packets (<40 bytes)
• 11 ms for longer packets (>40 bytes)
Max
Users
• Max users is less than the number of LTE
Resource Blocks available
• LTE Control Channels are bottlenecks
• Data aggregation is necessary
LTE Small Cell
(c) Multi-hop Transmission with LTE Small cells
eNB
Wi-Fi
UE
(a) Direct
Transmission
M2M Gateway
Smart Meter
(b) Multi-hop Transmission with Wi-Fi clusters
1026
Endpoints
Fig. X – Potential NAN/WAN Interconnections
Router
Data Aggregation Point (DAP)
UTILITY
Base Station
Backhaul for Cellular
Network
Smart
Meter
xxx
xxx
xxx
Collector
3G/4G Cellular
Wide Area Network (WAN)
802.11 Wireless Mesh,
Neighbor Area Network (NAN)
Fig. X – Proposed NAN/WAN Interconnections
NSMG-Net Project 3.2: Gowdemy Rajalingham
xxx
Command
Center
Fig. x – Neighbor Area Network
AGM 2013, Vancouver
20
21
NEIGHBOR AREA NETWORK
Microgrid
Customer
Substation
Non-Renewable
Energy
Smart
Meter
Substation
Electric Vehicle
Microgrid
Wind Enegy
Solar Enegy
Power System Layer
Control Center
Wired Backhaul
Network
3G/4G Cellular, Ethernet,
Leased Line, Fiber Optics,
Satellite...
 10-100 Mbps
Smart
Home
Devi ce
802.15.4, 802.11,
Wireless Mesh, …
 10-100 Kbps
802.15.4, Zigbee, 802.11,
 Coverage of up to several 1000 km 2  Coverage of up to several km 2
 10-100 Kbps ...
 Coverage of up to 100 m2
Wide Area Network (WAN)
Neighbor Area Network (NAN)
Home Area Network (HAN)
Communications Layer
Fig. x – Potential Technologies
Fig. x – Full Abstract System Architecture Model
1026
Endpoints
Network Characteristics
• Network of smart meters,
repeaters, collectors
• Static, line powered,
heterogeneous multi-tiered
network
• Communications protocols
must be robust, scalable,
self-configurable and selfhealing
Traffic Characteristics
• Multiple-Point-to-Point
• Point-to-Multiple-Point
• Point-to-Point
• Large volume of devices
• Short bursty packets
• Quality of Service (QoS)
differentiation
• Mix of real-time ( < 10
ms) and non-real-time
traffic (seconds - min)
Router
UTILITY
xxx
xxx
xxx
Collector
Command
Center
NSMG-Net Project 3.2: Gowdemy Rajalingham
xxx
Fig. x – Neighbor Area Network
AGM 2013, Vancouver
22
NEIGHBOR AREA NETWORK
1026
Endpoints
Router
UTILITY
xxx
xxx
xxx
Collector
xxx
Command
Center
Network Characteristics
• Network of smart meters, repeaters, collectors
• Static, line powered, heterogeneous multi-tiered network
• Communications protocols must be robust, scalable, self-configurable
and self-healing
NSMG-Net Project 3.2: Gowdemy Rajalingham
Fig. x – Neighbor Area Network
Traffic Characteristics
• Multi-point-to-point, point-to-multi-point, point-to-point
• Large volume of devices with short bursty packets
• Quality of Service (QoS) differentiation
• Real-time (<10ms) & non-real-time traffic (sec/min)
AGM 2013, Vancouver
23
FEASIBILITY OF CANDIDATE ROUTING FOR NAN
Objective
• Determine the capabilities and limitations of NAN
• With respect to ability to host MicroGrid
applications
• With GPSR routing protocol
• Thus, performance of NAN clusters with various
system parameters is investigated
Expected Results
• As channel conditions worsen, performance
degrades due to more likely packet corruption and
retransmissions
• As data rate increases, higher chance for channel
contention, back-offs and packet retransmissions
lead to increased delay and reduced reliability
• As cluster size increases,
• Increase in network load and average hop
count
• Significant increase in network delay with
decreasing PDR
NSMG-Net Project 3.2: Gowdemy Rajalingham
Fig. x – Simulation Scenario, sweep of cluster size
TABLE X – SIMULATION PARAMETERS
Channel Model
Simple pathloss – pathloss exponent
Lognormal shadowing – variance
Radio Access Technology
IEEE 802.11
Network Routing Protocol
Greedy Perimeter Stateless Routing (GPSR)
Performance Metrics
Packet Transmission Delay
Packet Delivery Ratio
Traffic
Per-node data rate -
Topology
Clusters of size -
System Parameters
Investigated
Sweeps of variance , per-node data rate
and clusters of size
Default values: , , [dB]
AGM 2013, Vancouver

similar documents