© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 1
PV and Storage in Communities: Challenges and Solutions
iiESI Workshop, Melbourne March, 2017
Prof Luis(Nando) Ochoa Professor of Smart Grids and Power Systems
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 2
Outline
Context: Flexibility
Challenges in PV-Rich LV Networks (Communities) – PV, Battery storage – Case Study, Control strategies – Effects on the LV Network and Customers
Future DSOs and Community-Based Flexibility
Conclusions
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 3
Flexibility Vision (UK Regulator)
TRANSMISSION NETWORK
DISTRIBUTION NETWORK
Renewable generators
conventional thermal generation
INDIVIDUAL PROVIDERS
Changing demand/supply in response to prices or direct signals from flexibility users.
LOCAL BALANCING
Community-based flexibility where consumers use local generation with DSR or storage, reducing the need to transport electricity.
THIRD PARTY AGREGGATION
Consumers, generators and storage can provide flexibility through an intermediary. Based on Ofgem’s
vision of flexibility
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 4
But before we get there…
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 5
Challenges in PV-Rich LV Networks
PV generation happens during the day, when many people are not at home Problems
V Max
Min
I I
V
Off-LTC
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 6
Challenges in PV-Rich MV Networks
Wide-spread PV adoption Challenges to HV networks
How can we address this?
Primary Sub OLTC
Off-LTC
V
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 7
… but people will buy storage.
Surely, this will sort it out (?)
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 8
Challenges in PV-Rich LV Networks
For Distribution Network Operators (DNOs) – PV impacts: voltage rise and thermal overload – Networks need to be reinforced
For householders
– Little or no incentive to export PV power – PV export limits are being adopted in some countries
(e.g., 70% of PV installed capacity is the limit in Germany) – Falling prices for household storage devices makes it more
viable
store the surplus during the day to use it later in the night!
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 9
Residential PV + Storage Systems
For householders – interested in their own benefits – want to save money with the PV + storage system – want to use as much low carbon energy as possible – do not care about the DNO’s technical problems!
For DNOs
– storage creates opportunities to mitigate technical issues and avoid/defer investments
– concerned with less grid dependency, lower revenue
… very different objectives
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 10
Storage – Basic Control and Sizing
Day (generation) Store the exports until full Night (no generation) Supplies the house load until empty Size of battery Enough to meet the average exports
Average energy production
High exports day
Storage becomes full network issues?
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 11
… Does residential storage
help mitigating PV impacts?
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 12
LV Network Impacts – Case Study
Real UK LV Network (NW England) – Residential, underground – Tx 800 kVA, 11/0.4 kV, 4 feeders – 200 single-phase customers – Voltage limit of 1.10 pu (253 V)
Summer, Weeklong Storage
– Similar to Tesla Powerwall – 2 kW, 10 kWh, 20 < SoC < 90% – Round trip efficiency of 87%
Feeder 1
Feeder 2
Feeder 3
Feeder 4
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 13
Probabilistic Approach 1/2
Monte Carlo Analysis – To cope with uncertainties (load/PV behaviour, location)
• Sets of 1,000 residential demand profiles (season/type of day) • Sets of 1,000 PV profiles (season)
– Unbalanced, time-series power flows (OpenDSS) – Different PV penetrations and storage control strategies – 100 week-long simulations per scenario
Key Metrics Voltage: EN50160 non-compliant customers Thermal: Feeder utilisation (1-hour moving avg) Effects on Customers: Grid Dependency, Self Consumption
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 14
Probabilistic Approach 2/2
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 24:00
Inso
latio
n [p
u]
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1Summer
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 24:00
Inso
latio
n [p
u]
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1Winter
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 24:00
Pow
er [k
W]
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0Load Profiles (Average)
Summer WDSummer WEWinter WDWinter WE
Tool developed by CREST (Loughborough Univ.)
1-min resolution
Number of Residents 1 2 3 4+ Probability (%) 29 35 16 20
PV Size (kWp) 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Probability (%) 1 1 8 13 14 14 12 37
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 15
Control Strategies
1. Optimal for the Customer – All PV surpluses (negative net demand) are stored
2. Delayed Charging – Surpluses are stored if the power export limit is reached
3. Store the Excess – Only PV generation beyond the export limit is stored
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 24:00-4
-3
-2
-1
0
1
2
3
Pow
er [k
W]
GenerationConsumption
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 24:00-4
-3
-2
-1
0
1
2
3
Net
Dem
and
[kW
]
PV Exports Limit
ExportsImports
70% limit
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 16
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 24:00-4
-3
-2
-1
0
1
2
3
Net
Dem
and
[kW
]
PV Exports Limit
ChargeDischarge
Mode: Optimal for the Customer
Even small surpluses early morning are stored
More energy for the customer
Battery is likely to be full before PV impacts
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 17
Mode: Delayed Charging
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 24:00-4
-3
-2
-1
0
1
2
3
Net
Dem
and
[kW
]
PV Exports Limit
ChargeDischargeExport
Charging starts closer to peak generation
Battery is likely to have room during PV impacts
Not necessarily making the most of PV generation for the customer
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 18
Mode: Store the Excess
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 24:00-4
-3
-2
-1
0
1
2
3
Net
Dem
and
[kW
]
PV Exports Limit
ChargeDischargeExport
Charging only PV generation beyond the export limit
Battery will significantly help reducing PV impacts
Does not make the most of PV generation for the customer
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 19
PV Penetration Level [%]
0 10 20 30 40 50 60 70 80 90 100Vol
tage
Non
-Com
plia
nt C
usto
mer
s [%
]
0
10
20
30
40
50
60
70
80
Feeder 1
No Storage
"Optimal" for the Customer"Delayed" ChargingStore the Excess
Results: Voltage Issues
Optimal for the customers Largest network impacts
Store the excess Significant mitigation of impacts
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 20
Results: Thermal Issues
PV Penetration Level [%]
0 10 20 30 40 50 60 70 80 90 100
Feed
er U
tiliz
atio
n Le
vel [
%]
50
60
70
80
90
100
110
120
130
140
150Feeder 1
No Storage
"Optimal" for the Customer"Delayed" ChargingStore the Excess
Optimal for the customers Largest network impacts
Store the excess No thermal issues
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 21
… Do all control strategies
benefit customers?
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 22
Index of Grid Dependence (IGD)
Effect of using storage on the electricity bill
𝐼𝐼𝐼𝐼𝐼𝐼 =𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖
𝑐𝑐𝑖𝑖𝑐𝑐𝑖𝑖𝑐𝑐𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑐𝑐
100% - household totally grid dependent (normal bill)
0% - household totally grid independent (no bill)
Grid
Household
consumption
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 23
Grid Dependency
No Storage Optimal for the Customer Delayed Charging Store the Excess0
102030405060708090
100
Inde
x of
Grid
Dep
ende
nce
(%)
PV w/o storage Produces savings but not large
Optimal for the Customer and Delayed Charging Greatest benefits, small electricity imports
Store the Excess Almost as not having storage!
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 24
… so, it is just a matter of
‘smart’ storage control
… not really
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 25
Smart & Low Carbon
Bulk Generation
Homes, Schools, Shops, Businesses,
Communities
Distributed Generation
New Service Markets $
Storage
Distribution System Operator
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 26
Current Control Philosophy (Distribution)
Grid Supply Point (GSP)
No control of LV and part of MV
Bulk Supply Point (BSP)
Primary Sub
DNO Control Room
OLT
C
Switc
hes
OLT
C
OLTCs Coordinated Decentralised Control
Secondary Sub
Centralised Monitoring and Control Reconfiguration Manual (Remote)
Voltage Decentralised (Fixed Targets)
OLT
C
Switc
hes
Reconfiguration Manual (Local) Voltage Off-load taps
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 27
Fully Centralised vs. Coordinated Decentralised
Fully Centralised – Opportunity for holistic ‘optimisation’ – Full flexibility – Increased reliance on monitoring/comms – Increased complexity due to scale – High deployment time
Coordinated Decentralised – More localised ‘optimisation’ (static areas) – Flexibility exists to centrally react to problems – Less centralised monitoring/comms – Less complexity due to reduced scale – Less deployment time
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 28
… but there are many other barriers
Defining Distribution Markets – Voltage management, congestion management – Near or post fault management (enabling
islanding/microgrids)
DSOs should be capable of
– Quantifying operational needs and constraints wide-scale observability, forecasting
– Exchanging data with the TSO and third parties (service providers)
… all this requires advanced management systems, investment, time, training, etc.
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 29
Conclusions 1/2
Residential PV + storage might become more viable – Lower capital cost, lower electricity bills less grid dependent
However, customer-focused storage control strategies do not reduce PV impacts – Storage gets full before critical periods (around midday)
New storage control strategies can mitigate network impacts from PV without affecting customers – Customers can achieve similar levels of grid dependency
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 30
Conclusions 2/2
Complexity will increase with more controllable elements and the need for more flexibility – Computational/data complexity shouldn’t be an issue in ~15 years
Regulators need to define the roles of potential future
players and technologies in the provision of flexibility – E.g.: aggregators, storage – This will prevent new business models being hindered by
regulatory barriers
This transition of DNOs to DSOs in the next few years will
bring exciting new regulatory environments where the envisioned smart grids are likely to finally emerge
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 31
Technical Reports and Publications
Technical Reports (most publicly available): https://sites.google.com/site/luisfochoa/publications/technical-reports
List of Publications (most publicly available):
Journal Papers https://sites.google.com/site/luisfochoa/publications/journals
Conference Papers https://sites.google.com/site/luisfochoa/publications/conferences
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 32
Thanks!
Questions?
Acknowledgements
• Dr. Tiago Ricciardi Post-Doctoral Research Associate • Mr. Kyriacos Petrou PhD Student • Dr. Sahban Alnaser Lecturer
© 2017 L. Ochoa - The University of Melbourne PV and Storage in Communities, Mar 2017 33
PV and Storage in Communities: Challenges and Solutions
iiESI Workshop, Melbourne March, 2017
Prof Luis(Nando) Ochoa Professor of Smart Grids and Power Systems
Top Related