Adaptive Load Management:
Adaptive Load Management: Possible Implementation of Demand Response According to Well‐Understood
According to Well
Understood Value and Choice
Value and Choice
Jhi‐Young Joo and MarijaIlić
The 8th Annual Carnegie Mellon Conference On The Electricity Industry
Carnegie Mellon University, March 14, 2012
Contents Contents
 Background
B k
d
 Adaptive load management (ALM)
 Multi‐temporal aspect
 Multi‐layer aspect
 Understanding values and choices
 Objectives and constraints
 Balancing demand with the system
 One‐shot vs. moving‐horizon optimization
 Conclusion 2
Background Background
 Problems of price‐responsive loads based on baselines P bl
f i
i l d b d b li
[ ]
[1]
 Information asymmetry between end‐users and operators
 The value of the “baseline” changes over time
g
 The availability and value of the supply also changesover time
 Possible solution
P ibl
l ti
 Schedule more “certain” supply and demand in a longer time scale
 Schedule the difference when the information becomes more accurate and closer to the actual consumption
 Customers define their own Customers define their own “baselines”
baselines and become and become
responsible for them
[1] Hung‐po Chao, “Price‐Responsive Demand for a Smart Grid World,” The Electricity Journal, 2010, Vol.23, No.1
3
Overview of adaptive load management
‐‐ multi‐temporal aspect
year
month m = 1
h=1
time
…
…
load
load
hour
…
n=2
n=1
time
Risk prone LSE
p
Risk averse LSE
tim
e
4
Overview of adaptive load management
‐‐ multi‐layered aspect
 Why aggregation? Because small users
Wh
ti ? B
ll
 Are usually risk averse
 Cannot participate in the market directly
 Have higher uncertainty in demand quantity
: aggregated demand curve more predictable
demand
Aggregate demand
demand
demand
...
Σ
≈
5
Overview of adaptive load management
‐‐ multi‐layered and multi‐temporal aspects
Tertiary layer
Purchase bids
Market price
Secondary layer
Secondary layer
L
Long‐term contract
Load serving entity I
D
Demand function
d f ti
LSE II
LSE III
End‐user rate
Primary layer
End‐user
Power plant drawing by Catherine Collier, Integration and Application Network, University of Maryland Center for Environmental Science (ian.umces.edu/imagelibrary/)
6
Values and choices of energy service
Values and choices of energy service
Every user/customer has different values and gy
needs for energy service
minimize (energy cost) + (utility) + (risk)
minimize
(energy cost) + (utility) + (risk)
subject to (state dynamics)
(state energy cost constraints)
(state, energy, cost constraints)
7
Balancing demand with the system
Balancing demand with the system
O
One‐shot optimization
h t ti i ti
Expected price
Optimal energy
price
price
Exogenous parameters
Energy optimizer
hour
p
12pm
hour
Demand function at 12pm
pricce
12pm
p
Repeat for all hours
8
energy
Examples of balancing with the system
‐‐ Island of Flores, Portugal
One‐shot optimization
Internal temperature of the refrigerators
9
Balancing demand with the system
Balancing demand with the system
Moving‐horizon optimization
Moving‐horizon optimization
Current time step
Current time step
Expected price
Current user state
(e g indoor temperature)
(e.g. indoor temperature)
Energy optimizer
Demand bid with quantity ih
i
limits
System operator
SSupply l
bids
Cleared price
p
Demand/supply dispatch
Suppliers 10
Examples of balancing with the system
‐‐ Island of Flores, Portugal
Moving‐horizon optimization
Internal temperature of the refrigerators
f h
fi
11
Conclusion
 Multi‐temporal and multi‐layered optimization
M lti temporal and m lti la ered optimi ation
 Multi‐temporal aspect
 Contracts based on the best forecast/information of the system and the users at different time steps
at different time steps
 As the forecast and the information changes, additional contracts in a finer time step
 Dependent on the risk aversion of the customer
 Multi‐layered aspect
: reducing high uncertainty of small‐users’ demand through aggregation
 Balancing with the system
: Information exchange at the right time between demand entities and system operator is crucial for a physically implementable d
i
i lf
h i ll i l
bl
demand response scheme.
12
Related publications and patents
Related publications and patents  Chapters
Chapters 8 and 9, 8 and 9 “Engineering
Engineering IT‐Enabled Electricity Services; The Case Of Low‐Cost Green Azores IT Enabled Electricity Services; The Case Of Low Cost Green Azores
Islands”, Co‐edited by M. Ilić and L. Xie, to be published
 J.‐Y. Joo and M. Ilić, “Distributed Multi‐Temporal Risk Management Approach To Designing Dynamic Pricing”, IEEE Power and Energy Society General Meeting, July 2012, accepted
 J.
J ‐YY. Joo and M. Ilić, Multi
Joo and M Ilić Multi‐Temporal
Temporal Risk Minimization Of Adaptive Load Management In Electricity Risk Minimization Of Adaptive Load Management In Electricity
Spot Markets, IEEE PES Innovative Smart Grid Technologies, Europe, Dec 2011
 M. Ilić, J.‐Y. Joo, L. Xie, M. Prica and N. Rotering, A Decision Making Framework and Simulator for Sustainable Electric Energy Systems, IEEE Transactions on Sustainable Energy, Jan 2011
M. Ilić, L. Xie
Ilić, . Xie and
and J.‐Y. Joo, Efficient Coordination of Wind Power and Price‐Responsive Demand Part J. Y. Joo, fficient Coordination of Wind Power and Price Responsive emand Part
 M.
I: Theoretical Foundations, Part II: Case Studies, IEEE Transactions on Power Systems, to appear
 J.‐Y. Joo and M.D. Ilić, Adaptive Load Management (ALM) in Electric Power Systems, IEEE International Conference on Networking, Sensing and Control, Apr 2010
 J.‐Y. Joo and M.D. Ilić, A multi‐layered adaptive load management system: information exchange between market participants for efficient and reliable energy use, IEEE PES Transmission and Distribution Conference, Apr 2010
 L. Xie, J.‐Y. Joo and M.D. Ilić, Integration of intermittent resources with price‐responsive loads, 41st North American Power Symposium, Sep 2009
13