17-06-26
1
Linking Storage to Renewable Energy Production: State of the Art and Applications
Catherine Rosenberg
Dept. of Electrical and Computer Engineering
ISS4E: MISSION
To use information systems and science to increase the efficiency reduce the carbon footprint
of energy systems
ISS4E.ca
17-06-26
2
Contexts
Smart homes and buildings
Distribution networks
Distributed generation
New technologies
Solar and wind
Storage
Electric vehicles
LED lighting
Pervasive computation and communication
Approaches
Internet-inspired
Energy system design
Data-driven analysis
Optimization
LATEST STUDIES
4
§ State Estimation in Power Distribution Systems Based on Ensemble Kalman Filtering, S. 2017
§ Can Flexible Solar Panel Orientation Help the Electrical Grid? S.2017 § Accurate Black-box Modelling of Lithium-Ion Batteries, S. 2017. § PV-Storage System Profitability in Multiple Jurisdictions, S. 2017. § On the Interaction between Personal Comfort Systems and Centralized
HVAC Systems in Office Buildings, S. 2017. § An Analysis on the Energy Consumption of Circulating Pumps of
Residential Swimming Pools for Peak Load Management, P. 2017. § Modelling Weather Effects for Impact Analysis of Pricing Policies:
Methodology and Case Study, P. 2017. § Practical Strategies for Storage Operation in Energy Systems: Design
and Evaluation, P. 2016. § Energy Storage and Regulation: An Analysis”, P. 2016. § Joint Optimal Design and Operation of Hybrid Energy Storage
Systems, P. 2016.
17-06-26
3
COLLABORATORS
5
Waterloo: S. Keshav L. Golab C. Canizares K. Battacharya
Canada: O. Ardakanian (U. of Alberta)
Europe: Y. Ghiassi-Farrokhfal (Erasmus U. Rotterdam) P. Jochem (KIT) K. Pettinger (Hochschule Landshut U.)
China: C. Song (Chinese Academy of Sciences)
India: R. Kalpana (IIT, Madras)
USA: S. Garg (NYU) A. Sangiovanni-Vincentelli (U.C. Berkeley) D. Callaway (U.C. Berkeley) D. Culler (U.C. Berkeley)
Many students!
TODAY’S ELECTRICAL GRID
6
17-06-26
4
MOSTLY DIRTY…
7
capacity
OVERPROVISIONED BY DESIGN
8
17-06-26
5
INEFFICIENT
5% better efficiency of US grid
= zero emission from 53 million cars
http://www.oe.energy.gov/ 9
OLD
Post-war distribution infrastructure is reaching EOL 10
17-06-26
6
UNEVENLY DISTRIBUTED
11
Electricity Consumption
https://yearbook.enerdata.net/#world-electricity-production-map-graph-and-data.html
China’s population > 4 X USA’s population
POORLY MEASURED
12
17-06-26
7
WITH LITTLE STORAGE
13
China leads the world in hydroelectric output followed by Canada, United States, Brazil, Russia and India. No European countries in the first 50 largest hydroelectric producers!
SMART GRID VISION
14
17-06-26
8
Source: European Technology Platform Vision Document
15
Source: European Technology Platform Vision Document
16
Renewable generation to reduce carbon footprint
17-06-26
9
Source: European Technology Platform Vision Document
17
Demand side management to reduce peak/average ratio
Source: European Technology Platform Vision Document
18
Storage to decouple supply and demand
17-06-26
10
Source: European Technology Platform Vision Document
19
Pervasive sensing, communication, control
Source: European Technology Platform Vision Document
20
Self-contained ‘microgrids’
17-06-26
11
Source: European Technology Platform Vision Document
21
Heavy investment for grid deployment/renewal
THE FUTURE IS (ALMOST) HERE!
§ Costa Rica actually ran on 100 percent renewable energy for 300 out of 365 days in 2015. Almost all of the country's infrastructure and utility energy is provided by hydroelectric and geothermal power (Iceland, Albania and Paraguay are also in the category of almost 100% renewable)
§ Portugal was 100% powered by renewables from May 7 to May 11, 2016!
§ Google to be powered 100% by renewable energy from 2017
22
17-06-26
12
THE FUTURE IS (ALMOST) HERE!
Image: Pitt and Sherry Consultants, Australia 23
GETTING THERE
24
17-06-26
13
PROGRESS Renewable/distributed generation (Wind) Solar
Storage Batteries (Electric vehicles)
Communication, computation, sensing, control
(Microgrids)
(Demand-side management)
25
TECHNOLOGY INFLECTION POINTS
Storage Solar Sensing and control
26
17-06-26
14
PERVASIVE CONTROL IS A NECESSITY
27
WHY?
28
Generation Load
17-06-26
15
CONVENTIONAL GRID
29
Generation Load
30
Generation Load
FUTURE GRID: DOUBLE NEED TO FORECAST
17-06-26
16
31
Generation Load
FUTURE GRID
Barnhart et al, Proc. Energy and Environment, 6:2804, 2013 34
MATCHING DEMAND AND SUPPLY
17-06-26
17
CONTROL OVER MANY TIME SCALES
Jeff Taft, Cisco 35
COMPLEX CONTROL ARCHITECTURE
Wid
e A
rea
Mon
itorin
g an
d C
ontro
l Sys
tem
(WA
MC
S) N
etw
orks
Conventional Wind Solar Storage
Internet
Distributed Energy Resources (DER)
(large scale)
Converged Voice, Data, and Video
Area ControlError (ACE)
Control
Virtual Power Plant (VPP)/Demand Response (DR)
Application
Reference Model
Trans-Regional Balancing Org.
Interchange Authority/Reliability Coordinator Grid Direct Current (DC) Inter-Ties Other Regions
Other Energy Traders
Utility EnergyTrading
UtilityNOC
• Customers• Suppliers• Retail Energy Providers (REP)• Third Parties
Third PartyStabilization
Services
Other TransmissionSystem Operators (TSO)/
Distribution System Operators (DSO) and Verticals
IndependentPower
Producers(IPP)
Other Region Phasor Measurement Unit
(PMU) Data
Other TSO PMU Data
Web Portals
Balancing Authority
VPP/DRApplication
ACEControl
Other Balancing Authorities
Trans-Regional Energy Markets
Wholesale Energy Markets
Trans-Regional/Trans-National TierTrans-Regional Networks
System Control TierInter-Substation Networks
Interchange TierInterchange Networks
Balancing TierControl Area Networks
Synchronous Grid Inter-Tie Control
WAMCS Network Operations Center (NOC)/
Data Center
Inter-Control Center NetworksEnterprise Networks Utility Tier
Intra-Control/Data Center Tier
AncillaryServices
VPP/DR Data/Signals
Control CenterNetwork
FieldDispatch
CallCenter
Control Center Control CenterControl CenterApplicationsand Users
EnterpriseApplicationsand Users
Control CenterNetwork
Data CenterData Center
Network
Transmission PMU Data
Substation Tier
D Level 1 Tier
D Level 2 Tier
Transmission Substation
Intra-Substation Networks
Sensor Networks
PhasorMeasurement
Unit (PMU)
Distribution Substation
Intra-Substation Networks
Sub-Transmission Substation
Intra-Substation Networks
Traction Substation
Intra-Substation Networks
Urban FANs Rural FANs
Sensor Networks
Electric Vehicle (EV)Sub-networks
Prosumer Tier
Third Party Aggregator of DR, Distributed
Generation (DG), etc.
FeederD-PMU
Nets
Sensors
Controls
Home Energy Controller
Controls
Displays
ChargerMeter
Inverter Control
Protection and Control
Inverter Control
Distribution PMU
Feeder
DistributeResourc
(large
Streetlight Systems
Line Sensor
Volt/ VAR Regulation
Fault Isolation and Service Restoration Nets
Sensor Subsystems
Inverter Control
DA Devices
DER Protection
Distribution Automation (DA) Sub-networks
Gateway
Transmission LevelStabilization
© 2011 Cisco and the Cisco Logo are trademarks of Cisco and/or its affiliates in the U.S. and other countries. A listing of Cisco’s trademarks can be found at www.cisco.com/go/trademarks. Third-party trademarks mentioned are the property of their respective owners. The use of the word partner does not imply a partnership relationship between Cisco and any other company. (1009R) C82-676734 6/11
Distribution LevelStabilization
(DSTATCOM)
Neighborhood Area Networks (NAN) (Residential, Commercial and Industrial)
Residential Networks Building Networks Private Microgrid Networks In-Vehicle Networks
In
D
Cisco 36
17-06-26
18
SOLAR
40
3 CHARACTERISTICS OF SOLAR GENERATION
1. Sunlight is free! § Near-zero OPEX, all cost is CAPEX
2. 20-25 year nearly maintenance-free lifetime 3. Amount of generation over lifetime depends on
geography
42
17-06-26
19
SOLAR
43
0.0
50.0
100.0
150.0
200.0
250.0
300.0
350.0
400.0
450.0
500.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Cell phone penetration (ITU) 1990=1
Cumulative Solar (EPIA) 2000=1
Solar PV is growing faster than cell phones
http://stats.areppim.com/stats/stats_mobile.htm http://www.epia.org/fileadmin/user_upload/Publications/EPIA_Global_Market_Outlook_for_Photovoltaics_2014-2018_-_Medium_Res.pdf
(c) S. Keshav [email protected] http://iss4e.ca
17-06-26
20
UNFORTUNATELY…
Problem 1: No sun at night…
SOLUTION: USE SOLAR BY DAY AND GRID BY NIGHT
46
17-06-26
21
PROBLEM 2: EXCESS SOLAR GENERATION
47 8 May 2016, Germany (from Agora Energiewende)
SOLUTION: USE ‘FREE ENERGY’ FOR SOMETHING
48
17-06-26
22
Barnhart et al, Proc. Energy and Environment, 6:2804, 2013
PROBLEM 3: VARIABILITY
SOLUTION: STORAGE
Storage decouples supply and demand
17-06-26
23
STORAGE
51
Global investment in energy storage technologies to reach $122 Billion by 2021
Source: Pike Research
A HOT AREA
52
17-06-26
24
Nature Climate Change: 2014 Tesla/Panasonic and GM/LG Chem battery costs are already (in 2016) down to the lowest projections for 2020! 53
BATTERY COSTS: CURRENT AND PROJECTIONS
54
17-06-26
25
STORAGE IS EXPENSIVE
Buying 1 KWh = 10c Storing 1 KWh = ~$250!
56
MANY STORAGE TECHNOLOGIES
57 www.iec.ch/whitepaper/pdf/iecWP-energystorage-LR-en.pdf
17-06-26
26
MANY STORAGE TECHNOLOGIES
58
In 2020: Build 500,000 EVs per year 35GWh/year of cell output From $600 to under$200 per KWh
TESLA GIGAFACTORY (IN NEVADA)
59
17-06-26
27
Storage stores energy (like bits)
measured in Joules or Watt-hours
What is drawn from a storage is power
Energy is the product of power and time
Power is measured in Watts (like bits/sec)
1 Joule = 1 Watt * 1 second
1 kWh = 1000 W * 3600 s = 3.6 million Joules
BASICS
60
“Bytes”
“Bits/s”
TYPES
61
17-06-26
28
Storage decouples supply and demand
Allows
Reliability for large scale renewable integration
Flexibility for energy management
10
WHY STORAGE?
APPLICATIONS
63
12
17-06-26
29
64
C(t)
C(t) = curtailment (waste of power).
A STORAGE MODEL (BASED ON FIRST PRINCIPLE)
The storage has some capacity B in Wh. At time t, it is charged with power c(t) or discharged with power d(t). Its content is b(t)
The ideal behavior is (Markovian behaviour):
0 ≤ b(t+∂) = b(t) + c(t) ∂ – d(t) ∂ ≤ B with c(t)d(t)=0, c(t), d(t) ≥ 0 and ∂ the time-slot duration
It is important to consider imperfections, such as:
• charging/discharging speed limits
• energy conversion/inversion efficiency
• capacity limits
• self-discharge (leaking) 65
c(t) d(t) b(t)
17-06-26
30
A STORAGE MODEL (2) § (Dis)charging limits: To avoid damaging storage, the controller might prevent charging or discharging too quickly
c(t) ≤ µc d(t) ≤ µd
§ (In) efficiency: losses in energy conversion
b(t+∂) = b(t) + ηcc(t) ∂ – ηdd(t) ∂
§ Capacity limits: Some storage (e.g., batteries) degrade quickly if they are nearly empty or nearly full for extended periods of time: a1B ≤ b(t) ≤ a2B
§ Self-discharge: Over time, storage naturally loses its energy
b(t+∂) = (1-γ1)b(t) + ηcc(t) ∂ – ηdd(t) ∂ - γ2B 66
HOW ACCURATE IS THIS MODEL?
§ Flywheels are not Markovian! § Batteries? Let’s see
67
17-06-26
31
Nissan Leaf chassis
68
LET’S TALK ABOUT BATTERIES
REVISITING THE STORAGE MODEL
69
Li-Titanate case (model 1 works better for LiFePO4)
17-06-26
32
BEYOND FIRST-PRINCIPLES Preceding slides have defined a basic storage model
Precise modelling of a specific storage technology requires more (information? detail?)
For electrochemical batteries, need to go beyond “power”
Power Injected
Power Drawn
Energy Content
Current(+)
Current(-)
Energy Content,Cell Voltage
ELECTROCHEMICAL BATTERY
Key parameters: Min and max Voltage Nominal voltage Charge capacity Max charging and discharging current (C-rate) Internal resistance (important for deducing the efficiency)
A
B
C
T
L
W
Technical DataNominal Capacity 30 Ah (measured at C/10 discharge rate, RT)
Nominal Voltage 2,3 V
Voltage range 1,7 V to 2,7 V
Impedance (1 kHz) < 2 mOhm
Dimensions Length (L) Width (W) Thickness (T)
287 mm ± 1 mm178,5 mm ± 1 mm (153 mm main body)12 mm +0,1/-0,5 mm
Weight 1100 g
Volume 475 ml
Housing Foil packaging
Tabs Length Distance Width Thickness
Aluminium (+ Pole), Ni-coated Copper (- Pole)33 mm ± 1 mm90 mm ± 0.25 mm50 mm ± 0,5 mm0,2 mm ± 0,02 mm
Expected lifetime Up to 15,000 cycles (at 1C charge/discharge full DoD and RT)
Expected calendar life 20 years (at RT)
ChargeCharging method CC/CV (constant Voltage with limited current)
Max. charge voltage 2,7 V (+0,05 V)
Recommended charge current 30 A (1C)
Max. charge current 120 A (4C)
End of charge U = 2,7 V and I < C/10
Max. temperature range -20°C to +55°C
DischargeRecommended discharge current 30 A (1C)
Max. discharge current 120 A (4C)
End of discharge Voltage 1,7 V
Max. temperature range -20°C to +55°C
Storage and transportMax. temperature range -20°C to +40°C (Capacity losses per year, at 50°C, 100% SOC: 0.9%)
17-06-26
33
REVISITING THE STORAGE MODEL (2)
72
§ To derive a more accurate model, we need to understand the inner workings of the battery
§ Go beyond simplistic power view (i.e., introduce voltage and current)
§ I will only explain one effect that is battery-specific.
REVISITING THE STORAGE MODEL (3)
73
Based on Peukert’s law (1897) Challenges the notion of constant energy limits a1 and a2.
17-06-26
34
VOLTAGE, CURRENT Similar effect when charging A function of the current (approximately linear)
Need to estimate the current..
u1
⇣ d(k)
Vnom
⌘+ v1 b(k) u2
⇣ c(k)
Vnom
⌘+ v2
Current ⇡ Power
Nominal Voltage
New constraint on energy limits:
The new model is still linear!
NOTION OF OPERATING RANGE
75
• We have developed even better explicit (non linear) models.
• We are evaluating what we gain by using them in different case studies.
17-06-26
35
A SIMULATION MODEL BETTER THAN SOA
76
§ The state of the art is Tremblay model (implemented in Simulink). § Our model does much better and does not require much more information than the Tremblay model (only information on the specification sheet). § We provide a Matlab system block that is compatible with Simulink simulation software: Matlab File Exchange link: https://www.mathworks.com/matlabcentral/fileexchange/63078-lithium-ionpi-model
Germany Trade and Invest 2014
IMPACT OF JURISDICTIONS ON PV/STORAGE PROFITABILITY
17-06-26
36
78
IMPACT OF JURISDICTIONS ON PV/STORAGE PROFITABILITY
IMPACT OF ToU PRICING ToU (time-of-use ) pricing started in Ontario on Nov. 2011. The price difference is meant to provide an incentive for residential customers to use less electricity during peak-demand times.
Is ToU pricing effective? Compared to flat pricing, has there been any change in electricity consumption during high-demand periods?
Many studies tried to answer these research questions in various contexts, but there was no consensus on the methodology and the results. We proposed a clear and objective study using a transparent, efficient and reproducible methodology.
Data: we obtained access to (smart meter) energy consumption data from 20,000 households in southwestern Ontario. The dataset covers nine months before the introduction of TOU and nine months after.
79
17-06-26
37
IMPACT OF ToU PRICING Measuring the effects of pricing is difficult because there are many other factors that influence consumption such as weather. To address this issue, in our study, we developed a deep statistical methodology to isolate the impact of pricing from the impact of weather and the number of working days.
Result: Our analysis shows that on-peak and mid-peak consumption has dropped by about 2.5 percent, while off-peak consumption has stayed about the same. Thus, it appears that electricity demand is not being shifted to off-peak periods, but is being conserved.
80
CONCLUSIONS
We’re well on our way to the Smart Grid But many challenges remain An exciting and complex research area!
81
