Summary and Next Steps-- short version

1
EPRI/NSF Workshop 02
GLOBAL DYNAMIC OPTIMIZATION OF THE ELECTRIC POWER
GRID

• Steering Committee
• Ronald G Harley, Georgia Institute of Technology
  (Chair)
• James Momoh, National Science Foundation
• Paul Werbos, National Science Foundation
• Massoud Amin, Electric Power Research Institute

2

• ?Stability depends on energy balance between
• turbine inputs to generators
• demands from loads

?Network contains capacitors, power electronic
converters, tap changers on transformers, phase
shifters.
2/14
3

• Control possibilities on a generator
• ?Inputs
• Steam power Pt into turbine
• Current into generator field winding If
• ?Outputs
• Speed or rotor angle
• Terminal voltage (power factor)
• Thus a two-input two-output nonlinear
  nonstationary system.

3/14
4

• Network Support Devices for Voltage Control
• FACTS
• Tap changers
• Phase shifters
• Switched capacitors

4/14
5

• ? Global Dynamic Optimization of the Electric
  Power Grid
• ? Transmission losses ? Earned revenue
• ? Restoration sequence after islanding ?
  Voltage deviations
• ? Stability after disturbance ? Security

5/14
6

• Optimized Operating Conditions
• ? Maintain constant frequency (energy, or balance
  of Watts)
• ? Maintain voltage profiles throughout network
• (balance of Vars)
• ? Respect thermal limits of equipment.
• ? Satisfy defined security criteria.
• ? Satisfy defined economic criteria.
• ? Maintain dynamic and voltage stability
• during disturbances.

6/14
7

• Present Operating Practices
• ? Real time control performed locally fast
  (equipment).
• ? Set points provided regionally slowly (humans).
• ? This is a form of decomposition in control as
  well as
• temporally.
• ? Faster disturbance rejection and stabilization
  function.
• ? Slower load tracking trajectory function.
• ? First stabilize after disturbance, then satisfy
  security,
• economic.

7/14
8

• Calculating Stability and Security (Present)
• ? Calculated off-line, updated typically energy
  hour or
• longer.
• ? Use mostly linearized models and eigenvalue
  analysis.
• ? Uses local information only.
• ? No on-line centralized coordination.
• ? Relatively large reserve generating capacity
  required.
• ? After disturbance should have enough margin
  within
• safe operating space to reach new stable
  point.

8/14
9

• Present Temporal Layers of Control
• ? Tertiary level (hourly updated or longer)
• Generator MW bids. Transmission contracts.
• Security. Economics.
• ? Secondary level (performed in seconds to
  minutes)
• Responds to system data. Often humans
  intervene
• to interpret data.
• ? Primary level (milli-seconds to seconds)
• Generator AVR and governor. FACTS. HVDC.
• Energy storage.
• ? Sensors, controller, actuator geographically
  close.

9/14
10

• For How Long Will Present Practices Work?
• ? Relies on sheer size, inertia of generator,
  enormity of
• total system which allows even the worst
  disturbance
• (not loss of network) to be proportionally
  small effects
• which are handled by sufficient safety
  margins.
• To survive loss of large network sections and
  generators
• more control capability is needed.
• ? Signal transmission delays 2 msec round trip
  for 150 miles
• but good enough for present closed loop
  control
• disturbances.
• ? Gencos, Transcos, and Discos all trying to
  optimize different
• things although interconnect to each other.

10/14
11
"BACK" to the Future

• Type of Control
• ? Capability to instantly compute optimum
  (defined in some
• way) operating condition infinite bandwidth.
• ? In practice maybe have 100 Hz bandwidth.
• ? Computational speeds a million times greater
  than todays
• high-end computers 1.
• ? Might have to settle for centrally
  coordinated, and
• decentralized control.
• ? Tertiary level now computed in minutes,
  secondary level in
• seconds, and primary level in milli-secs. Thus
  multi-level
• hierarchical control.
• ? Higher level controllers resolve conflict,
  imposes policy.

11/14
12
"BACK" to the Future

• -continued
• ? May be located in satellite, but risky.
• ? Run system closer to safety envelop, produce
  more power
• per dollar invested, but respect security
  /stability.
• ? Leaves us to decide what we wish to optimize.
• ? Speed up computers and controllers. Fast
  control already
• possible by FACTS (micro-secs. to milli-secs.
  response) NOW.
• ? Analytical tools to accomplish this?
• ? Refer to Grand Challenges of Workshop (slides
  pp.13-14)

Ref. 1 B. Fardanesh, System-wide automatic
real time voltage and power flow control for
power systems-coordination of injection (shunt)
controllers with routing (series) Controllers.
Invited paper. IEEE PES Winter Power Meeting, New
York, January 2002.
12/14
13

• Grand Challenges
• ? How to select the type of controllable
  hardware, size it, and
• choose its location in order to optimize some
  criteria.
• FACTS
• Tap changers
• Phase shifters
• Switched capacitors
• ? Integrated network control of items above plus
  generator control
• Dynamic and voltage stability
• Security
• Thermal overloads
• ? Centralized or decentralized control
• Need system wide data
• What should be optimized

13/14
14

• Grand Challenges -continued
• ? Infrastucture
• Communication and data network
• ? Benchmark network
• ? Pilot schemes

14/14