Henggeler Antunes

The role of demand-responsive energy
management systems for smart cities
Carlos Henggeler Antunes

Dept. of Electrical Eng. and Computers, University of Coimbra
and R&D Unit INESC Coimbra
COST Exploratory Workshop on Smart Cities

Paris, September 2011
A multi-disciplinary project
Scientific areas and teams
Main area: Sustainable Energy Systems

Secondary area: Fundamentals of Engineering Systems
Involving expertise in:

•  electricity distribution networks
•  end-use energy efficiency
•  operational research models and algorithms
•  multi-criteria decision analysis
•  decision support systems
Research team: C. H. Antunes (DEEC, UC); R. Larson (MIT, ESD), A M.

Oliveira (FPCE, UC)
Industrial partner: ISA - Intelligent Sensing Anywhere
Energy Box - development and implementation of a
demand-responsive energy management system
Purpose
Further develop and
implement in practice the
concept of Energy Box as a
24/7 background processor
operating on a local computer
or in a remote location to
manage in an intelligent
manner (responding to price
signals, comfort
requirements, etc.), one’s
home or small business
electrical energy use
Motivation
•  Ongoing transformation of electric grids into smart grids provides the
technological basis to implement demand-sensitive pricing schemes aimed at
using the electric power infrastructure more efficiently.
•  This creates benefits

- for the end-users by lowering their electricity bill while satisfying their
energy service requirements namely regarding comfort levels,
- for the utility by contributing to manage the peak of the load diagram
and flattening the aggregate demand curve that may enable to meet
forecasted demand growth with the current portfolio of generation sources and
grid infrastructures.
Real-time energy management system

•  In the context of migration to smart grids, demand-sensitive pricing of
electricity will expectedly become the standard pricing mechanism.
•  The judicious use of short-term price signals, comfort requirements and user
preferences can be engineered to induce changes in the usage of end-use
loads and electricity consumer behaviour.
•  A real-time EMS is under development, which is aimed at further developing
and implementing in practice the concept of Energy Box.
•  It will operate as a 24/7 background processor to manage in an intelligent
manner electrical energy usage in homes or small businesses.
Real-time energy management system

Automated energy management system able to mimic the individual
consumer’s decision-making process.
The Energy Box will exploit the flexibility that consumers generally have in the
timing of their electricity usage to induce changes in their electricity-
consuming patterns through time-varying electricity pricing.
This becomes feasible with the smart grid infrastructure, including two-way
communication and short-interval meter reading, complemented with sensor
networks.
⇒ To achieve a total system optimal control.

The Energy Box

•  Energy management system (hardware + algorithms) coordinating in an
autonomous manner for the typical small consumer of electricity (residential
and small commerce/services):
- Management of electricity use
- Storage (inc. EV)
- Selling back to the grid
subject to conditions
- of the grid
- in-door comfort and air quality
- occupancy patterns
- weather
Tasks
•  Preference modeling and learning
•  Characterization of consumer behavior
•  Technical issues
•  The Energy Box as perceived by the electric utility

Tasks goals
Preference modeling and learning
A critical issue for the success of the Energy Box concept is the degree of
acceptance by the users  capturing the preferences of different user
profiles, namely concerning the attitudes towards short-term price signals and
comfort thresholds.
Learning distinct patterns of use (rates of occupancy, comfort requirements,
etc.)  dynamically adapting to them when responding to changes in prices.
 study how to model the preferences of different user profiles in a
way they can be used as an input to the optimization algorithms providing
optimal decisions
- to shed loads (which loads, when, how long),
- to manage storage (charging or discharging an EV),
- to sell electricity back to the grid if a local generator is installed.
✚ sensor information fusion and short-term decision making.
Tasks goals
Characterization of consumer behavior
Behavioral issues can emerge at distinct levels, ranging from market
positioning to product development to the fostering of attitude changes
towards large scale, overarching concerns (e.g., environmental awareness,
public policies).
" " to address relevant behavioral processes trough a cognitive
modeling strategy, resting operationally on recent variants of conjoint analysis
and best-worst scaling.
 discrete-choice variants of the experiments will allow contrasting the
Energy Box concept with alternative energy-saving investments.
" " assessing consumer’s reaction to different pricing strategies, as
well as induced changes in overall conservation patterns and consumer’s
standpoint regarding energy public policies.
Tasks goals
Technical issues
- electric and thermal storage
- characterization of the main end-use loads (physically-based models)
- battery load control and management with laboratorial experiments with
a prototype electric vehicle.
- battery technologies, inc. voltage, connectors and communication
between the vehicle and the grid
- multiple energy sources
Tasks goals
The Energy Box as perceived by the electric utility
- Study the effects of the Energy Box as perceived by the electric utility, as a
function of its market penetration.
- How the demand response automated by means of a multitude of Energy

Boxes with different parameterizations according to distinct user profiles and
preferences affect the grid, including instability issues?
- What should be the statistical characterization of these random delays to

achieve the desired effects of maintaining network stability and yet changing
the load status as soon as possible?
- Can traditional utility operated remote load control actions complement the
Energy Box decisions (dual mode)?
Tasks goals
Algorithms
- Determining the global optimum control of using, storing and selling
electricity  sequential decision-making process that balances the end-
user’s comfort, cost and lifestyle preferences in the face of uncertain
conditions regarding the price of electricity, weather conditions, and
electricity available from weather-dependent generation sources.
- Stochastic dynamic programming framing for determining optimal control

decisions (process of sequential decision-making under uncertainty), e.g.
automated real-time response by space conditioning appliances with
respect to spot pricing for electricity .
- Optimal policy  finding a sequence of decisions that maximizes the

expected value of
max d i E si [ ∑λ uj
i i
j
(si ,di ) ]
j
Tasks goals
Algorithms
-j = comfort,lifestyle,cos t
hom e hom e
si+1 = f i,i+1 (sihom e ,diuse ,distore ,disell )
⎧ sihom e. max Temp if diuse.temp > sihom e. max Temp
hom e.temp
⎪
si+1 = ⎨ diuse.temp sihom e. min Temp ≤ diuse.temp ≤ sihom e. max Temp
⎪ shom e. min Temp if diuse.temp < sihom e. min Temp
⎩ i
⎧ sihom e.store + 1 if distore.battery = charge and the battery is not full in stage i
hom e.store
⎪
si+1 = ⎨ sihom e.store - 1 if distore.battery = discharge and the battery is not empty in stage i
⎪ sihom e.store otherwise
⎩
€
Tasks goals
Algorithms
Utility measure regarding buying and selling electricity
Cost preference function uicost (si ,d i )
€
sihome.bill
€
Patterns of use vs. quantity of use
Pattern of use: influencing usage despite
appliances are efficient or not (e.g. deviate
consumption of an inefficient washing machine to
lower tariff periods)
Quantity of use: suggesting more efficient

appliances to lower consumption without
downgrading the energy service provided (e.g.
provide information about savings that could be
obtained with a more energy efficient washing
machine)

Henggeler Antunes

Uploaded by

Copyright:

Available Formats

Henggeler Antunes

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Henggeler Antunes

Uploaded by

Copyright:

Available Formats

The role of demand-responsive energy

management systems for smart cities

Carlos Henggeler Antunes

COST Exploratory Workshop on Smart Cities

Main area: Sustainable Energy Systems

Involving expertise in:

Research team: C. H. Antunes (DEEC, UC); R. Larson (MIT, ESD), A M.

• This creates benefits

Real-time energy management system

Real-time energy management system

⇒ To achieve a total system optimal control.

The Energy Box

• Characterization of consumer behavior

• The Energy Box as perceived by the electric utility

- How the demand response automated by means of a multitude of Energy

- What should be the statistical characterization of these random delays to

- Stochastic dynamic programming framing for determining optimal control

- Optimal policy  finding a sequence of decisions that maximizes the

Utility measure regarding buying and selling electricity

Cost preference function uicost (si ,d i )

Quantity of use: suggesting more efficient

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

•  This creates benefits

•  Characterization of consumer behavior

•  The Energy Box as perceived by the electric utility

- Optimal policy  finding a sequence of decisions that maximizes the