3. RELIABILITY | ACCOUNTABILITY
3
3
EI
750 Online BES Generators responded to
the Survey
Summary:
1) Outer Loop Controls preventing or
squelching #1 issue.
2) GO understanding of PFR and GO data
quality
5. RELIABILITY | ACCOUNTABILITY
5
Interconnection: Western
Event ID:
WI_2017-08-08_100816
Event Date & Local Time: 8/8/2017 10:08:16
Report Percent Processed: 0
Report Date: 0
Total Units Online: 1000
Total Unit Submittals: 211
Submittal Rate: 21.1%
Unit Information Completed
Survey
Information
Submittals
Participant
Submittal
Rate
Total Units
Online
Submittal
Rate
EIA 860 Plant Name: 211 100% 21.1%
EIA 860 Plant Code: 205 97% 20.5%
EIA 860 Unit ID: 211 100% 21.1%
GADS ID: 184 87% 18.4%
Planning Case Bus Number: 172 82% 17.2%
Balancing Authority: 0 0% 0.0%
Generator Type: 213 101% 21.3%
Generator Base MVA: 209 99% 20.9%
Inertia Constant (H): 202 96% 20.2%
Droop Setting (%): 201 95% 20.1%
Deadband Setting (Hz): 198 94% 19.8%
Maximum Operating Level (MW): 201 95% 20.1%
Minimum Operating Level (MW): 200 95% 20.0%
Western Interconnection
6. 6 /
GE /
This presentation contains proprietary information of the General Electric Company, USA and is for internal
release only This document shall not be reproduced in whole or in part nor shall its contents be disclosed to
any third party without the written approval of GE Energy
Primary frequency response - GE
Communications
• Gas Turbine PSIB 20150203
• Focuses on PFR at plant level
• Steam turbine TIL-1961-R1
• Deadband checks /
recommendations
Notable issues
• Few recent questions
• PFR should be implemented at highest
plant level closed loop load control and
coordinated as needed inside the outer
loop
• Disabling load control or AGC outside
dead band sometimes done to “free “
governor ….but this disables correction
of frequency / ACE by AGC
• Signal resolution of frequency signal
matters. Turbine speed typically highest
resolution, check frequency meter
resolution if used
• Renewable push demand for speed of
primary response in some global
markets
Example control loop hierarchy
8. 1) Primary Frequency Control – Solution for NERC BAL-001-
TRE (ABB’s pre-approved proposal) (2015 -2018)
2) Generator Governor Frequency Control (Application Guide)
(2015)
3) Strategies that address the challenges of balancing load and
unstable grid frequency (technical paper) – Presented at;
ABB Customer World (3/15/2017) and 2017 ISA POWID
Symposium (6/17/2017)
4) Meeting NERC’s BAL-003 Generator Governor Frequency
Response (technical paper) – Presented at 2015 ISA POWID
Symposium (6/8/2015)
Contact Information:
Daniel Lee (dan.lee@us.abb.com)
Vern Smith (vernon.smith@us.abb.com)
November 29, 2018 Slide 8
ABB process to alert or educate the generator owner/ operators
9. 1) October 30, 2012, NERC published Frequency Response Initiative Report: The Reliability Role of
Frequency Response.
Frequency Response Withdraw is discussed multiple times on multiple pages.
2) Feb 5, 2015, NERC issues an industry advisory specifying the method to improve grid stability by;
Frequency Response Withdraw is discussed in one sentence “Related outer-loop
controls within the DCS, as well as other applicable generating unit or plant controls,
should be set to avoid early withdrawal of primary frequency response.”
3) Dec 15 2015, NERC Operating Committee approved the “Reliability Guideline: Primary
Frequency Control”
Frequency Response Withdraw is only mention in the performance assessment.
Guidelines do not identify withdraw problem or requirements a solution to resolve
frequency response withdraw.
November 29, 2018 Slide 9
NERC emphasis of Primary Frequency Withdraw
10. 1) The survey assesses both the secondary frequency control and primary frequency control
response.
Consider adding Unit Master Setpoint (demand) to data collection. Unit Master setpoint
should be subtracted from the megawatt response in order to evaluate the primary frequency
control. OK to also assess the response of secondary frequency control
2) NERC needs to verify the calculation can correctly compute the front end withdraw behavior.
In my opinion, the current calculation do not.
November 29, 2018 Slide 10
NERC Survey Assessment
11. 1) The Reliability Guidelines for Primary Frequency Control and MOD 27 are related but separate
NERC requirements.
2) NERC intends that the Generator Owners/Operators implement both of these documents.
3) The data collected from the MOD 27 test can be used to calibrate the boiler frequency correction
bias required by the Reliability Guidelines for PFC.
November 29, 2018 Slide 11
NERC PFC Guidelines vs Mod 27
12.
13. Emerson Process Management
Power & Water Solutions
Unit Coordinated Control Updates
Thor Honda
Steam Turbine Business Development Manager
(412) 963-4272
thor.honda@emerson.com
14. Emerson Power & Water Solutions
Unit Coordinated Control (UCC)
• Emerson standard Coordinated
Controller for conventional power
plants (boiler & steam turbine)
• Legacy of Westinghouse Load
Demand Computer (LDC)
• Provides a coordinated “front end”
control strategy that unites boiler
and turbine
• Versions for:
– Drum boilers
– Supercritical “Once Through”
boilers
• Has always had integrated
Frequency Response capability in
all Coordinated modes
• Works with any steam turbine
controls (flyball governor to digital) Emerson Proprietary
15. UCC Periodic Revisions by Subject Matter Expert (SME) Teams
• Emerson has (4) world-wide SME teams
that determine updates to the UCC:
– Boiler Standards SME team
– Boiler Tuning SME team
– Steam Turbine SME team
– Turbine Startup SME team
• Emerson SME teams will meet at least
annually to discuss frequency control and
ensure that our standard is consistent
with NERC’s recommendations
• Updates to the Frequency Control logic
with the UCC were made in 2018 in order
to improve performance and meet
customer preferences.
Emerson Proprietary
16. 2018 Emerson SME Teams’ Decisions Regarding UCC Frequency
Control
• Turbine speed should be used to measure
frequency when available
– Two-out-of-three (2oo3) speed probe
inputs for fault tolerance
– 60-tooth “speed wheel” with Ovation
Speed Detector Module can detect .004
Hz change in 10 msec
• Frequency meters (4-20 mA analog inputs)
should be avoided if possible
• Data link or network inputs for frequency
control should be optimized for:
– Update time
– Resolution
– Fault tolerance
Emerson Proprietary
17. 2018 Emerson SME Teams’ Decisions Regarding UCC Frequency
Control
• Coordinated Boiler Follow is our
preferred operating mode for optimal
Primary Frequency Response
– Turbine valves prioritizing Hz then
MW
– Boiler firing rate prioritizing boiler
pressure control
• The UCC mode should NOT be
automatically changed during a
frequency event
– Some customers had been kicking
their UCC into Boiler Follow during
frequency events
• The turbine must remain tied to the UCC
during a frequency event to allow
secondary control action from AGC or the
operator
Emerson Proprietary
18. 2018 Emerson SME Teams’ Decisions Regarding UCC Frequency
Control
• Maximum and minimum limits can be
used to protect the unit from tripping or
exceeding equipment limits
– Reverse Power protection at low load
– Equipment maximum and minimum
limits
– Upper operating limits during under
frequency events near Pmax
• Rate limiting of the response may be
necessary in some cases in order to
prevent tripping
• Depending on local electrical system
conditions, some extra deadband on
boiler firing rate can prevent unnecessary
responses to small frequency changes
Emerson Proprietary
19. 2018 Emerson SME Teams’ Decisions Regarding UCC Frequency
Control
• Frequency Response Test logic should
be added to UCCs to allow on-line
testing and optimize tuning
• A small software frequency bias that
simulates an actual system frequency
event can be added to the running speed
signals in order to:
– Capture the dynamic unit response
– Provide an opportunity to tune and
optimize unit response
– Validate and improve the governor
model (NERC MODs)
Emerson Proprietary
20. 2018 Emerson SME Teams’ Decisions Regarding UCC Frequency
Control
• Frequency response logic should not be
instantly enabled upon generator breaker
closure
– Not all turbine controls will
automatically pick up enough load
upon synchronization to safely clear
reverse power protection
– Boiler stability will be affected by
synchronization
– Drum level swell with steam flow
increase
– Best to let the unit stabilize and add
enough room to safely respond to a
frequency event
Emerson Proprietary
23. Background
• Poor frequency response is the result of using
megawatt generation control without frequency
error bias.
• The following cases illustrate improvements FPL
made to plant load control logic to improve
frequency response.
24. Cases
Case Plant control Turbine
control
Change
1 Feed forward with
feedback trim and no
freq err bias
Load control
w/no freq err
bias
Added freq bias to
plant control feed
forward and
feedback controller
setpoint
2 Feed forward with
feedback trim and no
freq err bias
Load control
w/freq err bias
Added freq bias to
plant control
feedback controller
setpoint
3 Feedback with no
freq err bias
Speed control
w/freq err bias
Added freq bias to
plant control
feedback controller
setpoint
30. 30
Biography
• Frank C Buttler Jr, P.E.
– Consulting Engineer
– Johnson Services Group
– Southern Company Generation
– Engineering and Construction Services, Technical Services, ERO Support
– Email: x2fcbutt@Southernco.com or fbuttler@bellsouth.net
– Phone 770-401-3944
• BSEE Auburn University, 1978
• Retired from Southern Company after 40 years of service in Power Plant Electrical
and I&C Field Support.
• Started consulting to Southern Company Generation in 2014 for MOD-027-1 Turbine
Frequency Response Testing.
• Over the last four years, performed frequency step testing, modeling simulations,
and frequency response recommendations on 10 Hydro Units, 15 Fossil Steam
Units, and 23 Combustion Turbines for MOD-027-1.
• Performed testing and frequency response recommendations on 9 Fossil Steam
Units and 19 Combine Cycle Units for Outer Loop Controls.
32. 32
Outer Loop Control Philosophy
• Droop control response should be controlled at the lowest level
– As close to the governor controls as possible
• Turbine Controls provides the Droop Response capabilities
• Outer Loop Controls should not affect the Turbine Controls Frequency Megawatt
Response
• Outer Loop Controls if Not Accounted for will NULL out the Frequency Megawatt
Response during a Frequency Event
– Outer Loop Megawatt Controller “Sees’ the Unit Megawatts are off Setpoint Target
and Moves the Megawatt Setpoint to Correct the Error
33. 33
External Outer Loop Controls – Fossil Steam
• GE G3 Tandem Compound Steam Turbine
• Mark VIe Turbine Control System
• 933 MW Output
• Flow Demand/Reference Setpoint set by Emerson Ovation
• 5% Droop
• 0.06% or 0.036 Hz Deadband
• 0.11% or 0.066 Hz or 3.96 RPM Speed Step Change
• Turbine Controls Provides the Outer Loop Controls the Droop Response Flow
Demand Bias Percent or Speed Error after Deadband Percent
• 1.0% Expected Flow Reference Response to the Speed Step Change
• Expected Megawatt Response to the Speed Step Change
– 9.3 MW = 933 MW
0.066 Hz step−0.036 Hz deadband
0.05 droop 60 Hz
– Speed Step was applied on System Frequency so Megawatt Output will vary
Dependent on System Frequency +- Speed Step
– Response will Depend on Linearity of Valve Curves (Flow Demand/Flow
Reference) and Boiler Support/Throttle Pressure
35. 35
External Outer Loop Controls – Fossil Steam
• Fossil Steam Turbine Controls Set the MW Output of the Turbine Generator by
Changing the Flow Demand
• During a Frequency Event, the Fossil Steam Turbine Controls Provide the Outer
Loop Controls with either the Flow Demand Bias or Speed Error after Deadband
• The Flow Demand Bias or Speed Error after Deadband is converted to the Droop
MW Bias based on the Flow Demand Curve
– If Using Flow Demand Bias and Flow Demand is Linear:
‣ Droop MW Bias = Flow Demand Bias (%) * Unit MW Capability
– If Using Speed Error after Deadband and Flow Demand is Linear :
‣ Droop MW Bias = Speed Error/Deadband (%) * Unit MW Capability / Droop
• The Droop MW Bias is subtracted from the Unit MW and Unit Demand Output is
Adjusted as Needed
36. 36
External Outer Loop Controls – Fossil Steam
• During normal unit operation with the frequency within the droop deadband:
– Assume the Unit Setpoint is 500 MW, Unit MW is 500 MW and Droop MW Bias is
0 MW Calculated from the Flow Demand Bias or Speed Error after Deadband.
The Droop MW Bias (0 MW) is subtracted from the Unit MW (500 MW) results in
the Total Corrected MW being 500 MW. The inputs to the PID controller will be
Total Corrected MW of 500 and setpoint of 500. This results in no change of the
Flow Demand setpoint output.
37. 37
External Outer Loop Controls – Fossil Steam
• During normal unit operation with the frequency outside the droop deadband and the
Unit supplying 10 MW response to the frequency deviation over the droop
deadband:
– Assume the Unit Setpoint is 500 MW, Unit MW is 510 MW (increased by 10 MW
due to the Droop Response of the Turbine) and Droop MW Bias is 10 MW
Calculated from the Flow Demand Bias or Speed Error after Deadband. The
Droop MW Bias (10 MW) is subtracted from the Unit MW (510 MW) results in the
Total Corrected MW being 500 MW. The inputs to the PID controller will be Total
Corrected MW of 500 and setpoint of 500. This results in no change of the Flow
Demand setpoint output and thus does not affect the droop response of the
Turbine.
38. 38
External Outer Loop Controls – Fossil Steam
• Starting Megawatt Load at Test – 750 MW
• Negative 3.96 RPM Speed Step Change
• Expected Flow Demand Change
– 1.0%
• Expected Megawatt Response to the Speed Step Change
– 9.3 MW = 933 MW
0.066 Hz step−0.036 Hz deadband
0.05 droop 60 Hz
– Actual Flow Demand Change
‣ Positive/Negative Step:1.0%
• Actual Megawatt Response to the Speed Step Change
– Positive Step: 5 MW
– Negative Step: 28 MW
• Due to non-linear valve curve
46. 46
External Outer Loop Controls – Fossil Steam
• Starting Megawatt Load at Test – 700 MW
• Negative 3.96 RPM Speed Step Change
• Expected Flow Demand Change
– 1.0%
• Expected Megawatt Response to the Speed Step Change
– 9.3 MW = 933 MW
0.066 Hz step−0.036 Hz deadband
0.05 droop 60 Hz
– Actual Flow Demand Change
‣ Positive/Negative Step:1.0%
• Actual Megawatt Response to the Speed Step Change
– Positive Step: 12 MW
– Negative Step:8 MW
• MW Response was greater due to decreasing frequency (speed) during the test
• Outer Loop Controls can help Support Droop Megawatt Response
• Note that Turbine Output will be Dependent on Boiler Operating Conditions at the
Time of the Event
49. 49
Conclusions/Recommendations
• Fossil Steam Operation
– Turbine Controls Provides the Droop Response
– Verify Flow Demand is Linear
– If Flow Demand is not Linear, Work with Turbine Control Vendor to Correct Non-
Linear Characteristics. This will also Provide better MW Control
– If Flow Demand is not Linear and can not be Corrected, Use Outer Loop Control to
Help Augment the Droop Megawatt Response
– May want to disable Droop Response at Unit High Limit