Automatic Time Error Control

1
Automatic Time Error Control
2
What is Auto Time Error Control?

• A revised ACE equation and Inadvertent
  Interchange Payback procedures to cause every
  control area to determine its own contribution
  to TE and then to feedback into ACE a correction
  factor to reduce that contribution.

3
How does it work?

• When a control area experiences meter error, or
  schedule error, or EMS problems, or generation
  response problems, the prevailing system
  frequency changes.
• The other 33 control areas in WECC will respond
  to correct frequency through their individual
  frequency bias term in ACE.
• The control area causing the frequency error is
  said to have created primary time error.
• The 33 control areas responding to correct
  frequency are said to have created secondary
  time error.

4
How does it work?

• Time error is directly related to inadvertent
  interchange.
• All control areas have procedures in place to
  determine their hourly inadvertent interchange.
• Converting hourly inadvertent interchange into
  primary inadvertent interchange a control area
  can observe just that portion of WECC time error
  that they alone caused.

5
How does it work?

• Time error is driven to zero by the combination
  of these actions. Inadvertent interchange and
  time error pass through zero at the same time.

6
Inadvertent Payback

• Is ATE really Inadvertent Payback? No.
• There is relationship between primary time error
  and primary inadvertent. Decomposition of ?n and
  Inn was patented in 1982.
• These decomposed components are directly related
  to each other. Inn -(Bs ?n) ?n / 6
• If you drive one component to zero, the other
  must go to zero.
• We use Primary Inadvertent as the control
  parameter because it is an extension of existing
  Inadvertent accounting methods. We could have
  introduced an entirely new primary time error
  accounting method for ATE parameters in ACE, but
  it would have created unnecessary development
  effort. (See EPRI TR-107813 Appendix E)

7
Inadvertent Payback

• Since you know who did what to whom, you could
• develop a settlement process
• Requires all II accounts balance
• Requires all frequency bias settings are known
• Requires N(N-1) matrix for every hour
• Single Step Corrective Control
• Already have necessary information - ?s, In,
  Bs, n
• Simple hourly accounting process to compute Inn
  for on-peak and off-peak.
• Simple hourly accounting process to track TE
• No need to compute secondary inadvertent
• interchange

8
Frequency Issue

• Does ATE oppose prevailing time error?
• Good control in the first place would negate the
  need for any form of auto time error control
• there are some control areas that do not put
  forth resources to control properly.
• Auto Time Error method is designed to place 100
  of the responsibility to correct primary time
  error on all control areas responsible for
  creating it.
• If all control areas correct their accumulated
  primary time error properly, then WECC time
  error stays well bounded.
• It only takes one control area to shirk their
  responsibility to defeat coordinated time error
  control

9
Frequency Issue

• PWG is simulating two methods of ATE that result
  in an Interconnection-wide ACE0.
• That means the control ACE ? NERC ACE.
• That means all compliance reporting could use one
  ACE
• New methods still subject to control area
  non-participation.
• Any approach is subject to gaming

10
CPS Degradation

• Does ATE degrade CPS performance? No.
• The NERC Resources Subcommittee tracks CPS
  performance. We are successfully balancing CPS
  compliance with the WECC request to improve time
  error performance by reducing manual corrections.
• WECC control areas are exceeding CPS1.
• There is abundant margin for more relaxed
  control.
• Some WECC control areas are marginally meeting
  CPS2.
• Reasons include
• Failure to follow FAQ guidance
• Overt control decisions to just barely meet CPS2
  and nothing more

11
Frequency Performance

• Frequency Error performance steadily decreasing
  toward expected CPS1 profile.
• Short term averaging intervals show more evidence
  of relaxed control.
• Long term averaging intervals show more evidence
  of time error control effects.

12
Other Issues

• Is L10 discretionary? No.
• NERC developed L10 to meet the necessary and
  sufficient conditions to be technically
  defendable.
• As a statistical measure, it is recognized that
  some control action can occur at the tails of a
  central tending distribution.
• NERC gave every control area discretion in
  selecting its frequency bias setting.
• Larger bias larger limits relatively greater
  obligations

13
Time Error Control Survey

• Most CA computing hourly Primary Inadvertent
  correctly
• Factoring in effects of Manual TE corrections was
  a problem. (Fixed)
• Some TE equipment faulty. (Now using ICCP TE
  sources)
• Most CA computing Accumulated Primary Inadvertent
  Correctly.
• Some problems with month-end adjustments
• Some CA apply the ATE correction term improperly.
• Sign backward, too much limitation on magnitude.
  (Not resolved)

14
Action Items

• Revise FAQ
• DCS scenarios
• Difference between control ACE and CPS ACE
• Proceed with CA survey recommendations
• Simulate two alternative ATE methods in ACE
  equation to allow control ACE to be compliant as
  the NERC ACE for CPS DCS
• Method 1 continuous incremental change in Fs
• Method 2 Larger exponentially decaying H rather
  than equally weighted H.
• Continue to assist CA with implementation