File preview

  CPS
 Energy
 Grand
 Challenges
  1.
 Achieve
 20%
 peak
 reduction
 via
 demand
 side
 management
 (equivalent
 188
  GWatts)—beyond
 demand
 response
  Improving
 capacity
 utilization:
 how
 to
 shape
 demand
 (saving
 capital
 costs
  later)
  Engagement
 of
 human
 decision
 makers/behavior
 in
 the
 process
  Economic
 analysis/studies
  Human
 interface
 –
 feedback
 from
 users
  Hope
 versus
 management:
 virtual
 power
 plants,
 manage
 the
 generation
 side
  to
 match
 the
 expected
 reduction
  How
 to
 make
 it
 reliable,
 agile?
  Integration
 of
 renewable
 sources
 –
 supply-­‐following
 loads
 via
 demand
  management
  Pervasive
 health
 monitoring
 of
 the
 energy
 system
  Are
 market-­‐based
 mechanisms
 sufficient?
 What
 are
 the
 right
 kind
 of
 market
  designs/interfaces/APIs/etc.
 More
 than
 price
 signals?
 Time
 scales,
 especially
 as
  they
 get
 more
 fine-­‐grained,
 is
 a
 major
 factor
 in
 understanding
 the
 system
 
  2.
 Achieve
 50%
 of
 energy
 generation
 portfolio
 from
 renewable
 sources
 
  High
 fluctuation/non-­‐dispatch
 
  New
 forecasting
 methods—new
 approaches
 are
 needed,
 not
 safety
 but
  usage-­‐based
 (worst
 case
 weather
 events
 vs.
 worst
 case
 energy
 demand)
 
  Real
 time
 weather
 information/cloud
 cover/wind
 intensity
 
  Challenges
 in
 connecting
 and
 processing
 vast
 instrumentation
 for
 rapid,
 agile
  response
 
  Agility
 of
 back-­‐up
 generation
 plans
 
  Implications
 for
 control
 loops
 in
 highly
 variable
 environment
 
  Don’t
 forget
 the
 natural
 gas
 turbines
 in
 the
 generation
 portfolio
 
  Distributed
 generation
 (renewables
 embedded
 in
 the
 distribution
 system)–
  distributed
 techniques
 needed,
 solar
 scales
 to
 backyards,
 but
 maybe
 not
 wind
 
 
  Solar
 hard
 to
 forecast
 
  Security
 and
 privacy
 issues
 
 
  3.
 Removing
 human
 operators
 from
 making
 routine
 decisions
 (by
 2050?)
 
  Operators
 in
 the
 loop
 for
 emergency
 conditions
 
  Autopilots
 for
 the
 energy
 system
 
  Better
 information
 collection
 and
 display
 for
 strategic
 energy
 decision
  making
 
  AC
 optimal
 power
 control
 solved
 in
 real
 time
 
  Evolution
 towards
 a
 new
 model
 of
 energy
 buying
 and
 selling/aggregation
 vs.
  operating
 a
 power
 plant
 or
 portfolio
 of
 generation
 capabilities,
 etc.
 
  Capture
 expertise/intuition
 of
 existing
 operators
 as
 they
 retire
 and
 replace
  with
 informed
 automated
 model-­‐based
 systems
 

Energy
 Birds
 of
 a
 Feather
 Session
  Third
 CPS
 PI
 Meeting,
 National
 Harbor,
 Maryland
 


  4.
 Making
 Energy
 Industry
 as
 Innovative
 and
 Agile
 as
 any
 High
 Technology
 Industry
 
  Economic
 and
 market
 models,
 financial
 incentives
 
  New
 technology
 development,
 efficiency
 of
 operational
 
  “CAD
 for
 the
 utility
 industry”—testbeds,
 models,
 ways
 to
 understand
 the
  benefits
 of
 new
 technology
 before
 the
 buy
 decision
 
  Introduction
 of
 competition
 is
 essential
 
  Models
 exist,
 how
 to
 integrate
 them
 across
 scales
 and
 scenarios
 
  Smart
 grid
 conceptual
 model
 
  Competition
 in
 the
 “local
 loop”/distribution
 net
 
  Understanding
 the
 implications
 of
 restructuring
 the
 markets
 
  What
 are
 the
 right
 constraints
 to
 allow
 the
 deconstructed
 grid
 
  5.
 Keeping
 the
 lights
 on,
 no
 matter
 what—at
 reasonable
 cost
 
  Allow
 innovation
 without
 sacrificing
 safety
 and
 service
 
  Leverage
 new
 IT
 technologies
 –
 sensors,
 information
 processing,
 etc.
 
  Big
 data
 meets
 energy
 for
 actionable
 control
 
  RFC
 like
 Internet
 Protocols
 for
 evolving
 energy
 technology
 plug-­‐and-­‐play
 
  architecture
 
  CPS
 Energy
 Research
 Community
 Shared
 Infrastructure
 for
 Scaling
 
  Data—economic
 data
 (smart
 data,
 results
 of
 demand
 response
 experiments)
 
  Synchrophasor
 data
 from
 utilities
 
  Load
 flow
 from
 utility
 
  Real-­‐time
 dynamic
 data
 
  Failure/fault
 data
 
  End-­‐to-­‐end
 data
 from
 generator
 to
 load
 (consistent/clean)
 
  Standardized
 NDA
 agreements
 to
 get
 data
 to
 researchers
 
  Security
 Compliant
 Cloud-­‐based
 Datacenter
 
  Publication
 of
 analysis
 results
 –
 move
 community
 to
 a
 scientific
 basis
 
  Shared
 software
 
 
  Data
 collection
 (decoupled
 from
 specifics
 of
 the
 environment,
  commercial
 santization)
 
 
  Analysis
 
 
  Visualization
 
 
  Large
 scale
 realistic
 test
 models
 for
 a
 variety
 of
 purposes
 
 
  Testbeds,
 simulators,
 emulators;
 hardware
 and
 software
 components
 
 
  Different
 testbeds
 for
 different
 purposes—no
 single
 testbed
 
 
 
  Smart
 meters
 
 
 
  Renewables
 
 
 
  PMUs
 
 
 
  How
 to
 test
 for
 disasters,
 like
 blackouts
 
 
 
  For
 control
 systems
 
 
  Exchange
 standards
 
 
  Ask
 the
 meta
 question
 of
 what
 the
 open
 testbed
 should
 be
 
 
  Policy—no
 money
 for
 industry
 projects
 without
 publication
 of
 data?
 


 


 

But
 must
 be
 flexible