1. Construction and simulation of robust electric power futures
4/1/2015
Construction and simulation of robust electric power futures
David Schmalzer, Argonne National Laboratory
Donald Hanson, Argonne National Laboratory
Peter Balash, National Energy Technology Laboratory
Christopher Nichols, National Energy Technology Laboratory
Presented at the 33rd USAEE/IAEE North American Conference
Pittsburgh, Pennsylvania
October, 2015
11/17/2018
Argonne National Laboratory

2. 4/1/2015
acknowledements
Don Hanson and Dave Schmalzer want to thank NETL for support.
Disclaimer:
The work shown does not necessarily reflect the views of Argonne National Laboratory, the University of Chicago, the National Energy Technology Laboratory, or the U.S. Government
11/17/2018
Argonne National Laboratory

3. overview Explain evidence for and theory of cycling
4/1/2015
overview
Explain evidence for and theory of cycling
Show what happens when steep near-term CO2 reductions are required (e.g. EMF-31 Natural Gas study runs under a severe CO2 intensity standard)
Show importance of unit level dispatch
Highlight importance of maintaining fuel and technology diversity
Conclusion: Robust future power scenarios should be designed to avoid severe cycling
11/17/2018
Argonne National Laboratory

4. Cycling Background Aging, cycling, and load following are not new
Extensive literature back at least into the 1990’s
Pervasiveness of cycling and load following has grown
Renewable energy mandates
Low natural gas prices
CFPP fleet has aged
Capital costs of new capacity
Grandfathering of plants in CAA
EPA NSPS disincentives
CO2 regulation uncertainties
Cycling demonstrably reduces unit remaining life
11/17/2018

5. Terminology Btu/ kWh, operating, not test/design
Heat Rate
Btu/ kWh, operating, not test/design
Load following
Operating between design (100%) and ~50% design
Cycling
Unit output to grid goes to zero
Creep
Time dependent deformation below tensile yield
Fatigue
Defect growth from cyclic changes in stress
11/17/2018

6. Major Economic Items Heat Rate Dominates dispatch within CFPP
Replacement power cost
Driven by forced outages, usually equipment failure
11/17/2018

7. Known Physical Processes that Increase Heat Rate and Cause Forced Outages
Wear of seals and turbine blades
Fouling and deposition on heat transfer surfaces and steam turbine blades
Aging of refractories and structural shells, particularly boilers
Component failure from corrosion, fatigue, and creep
Interaction of fatigue and creep under cycling
11/17/2018

8. Component Failure Impacts
Failure of a boiler or condenser tube has very different impact than failure of a turbine or generator
Source: Kumar, N., et al, Power Plant Cycling Costs, NREL/SR
11/17/2018

9. Fleet Heat Rate Trends
(Data source: EIA Table A6. Approximate Heat Rates for Electricity, and Heat Content of Electricity, Published October 28, 2014)
11/17/2018

10. Interaction of Creep and Fatigue
Effect of temperature ramp rate
Source: Fleming and Foster, Aging of Coal Fired Power Plants
EPRI 2001 Report: “Where operational cycling is introduced on a former baseload unit, the residual life can be greatly reduced to between 40% and 60% of the original design life because of the combined effects of creep and fatigue.”
11/17/2018

11. Boiler Material Failures
Waterwall web cracking
Boiler tube corrosion
Source: Lefton, Power Plant Asset Management
Superheater tube attachment fatigue cracking
11/17/2018

12. Steam Turbine Failures
Source: Lefton, S., Power Plant Asset Management, Intertek APTECH
11/17/2018

13. Development of cycling Scenarios
Heat rate guidance from EIA-923 (2014) and EIA-860
Filtered data ranked into quintiles
Midpoints of first, third, and fifth quintile taken to represent heat rates
Operating periods based on literature and judgement
Residual plant life in various operating modes taken to be reasonable
Little hard data on this
Known to be impacted by plant maintenance and investment
Extra-plant factors impact retirement decisions
Should not be taken to represent any specific plant
11/17/2018

14. Operating Modes in scenario Modeling
Runs 24 x 5 but does some load following, 85% effective capacity while running. One warm shutdown and startup per week. Annualized CF 51.6%, residual life 10 years
Mode 2
Runs 16 hours per weekday, averaging 85% effective capacity. Six warm shutdowns and startups per week. Annualized CF 34.4%, residual life 5 years
Mode 3
Runs 12 hours per weekday in two six hour periods, averaging % effective capacity. Eleven warm shutdowns and startup per week. Annualized CF 28.8%, residual life 3 years
All cases assume 85% availability and 30+ years of service at outset of cycling
11/17/2018

15. Major elements in dispatch modeling
Heat rate
Fuel cost
Carbon fees
Capital costs of capacity
Penalty function leading to unit retirement based on (Operating Mode x Duration)
Low CFPP capacity factors force heavy cycling of CFPP, leading to retirements and loss of base load capacity
11/17/2018

16. EMF-31 Specification for a CO2 intensity standard
4/1/2015
EMF-31 Specification for a CO2 intensity standard
EMF-31 was not modeling the Clean Power Plan, which was not finalized at the time.
The allowed amount of CO2 per unit of generation averaged over was specified.
Also the year 2030 intensity was specified.
It turned out for our modeling that the target was very stringent requiring a $50 per ton shadow price on CO2 emissions in the first year of 2020.
This severe effective carbon price resulting in dispatching the NGCC units before the CFPP (i.e., high gas CFs and low coal CFs)
Severe cycling resulted and most of the CFPP existing fleet phased into retirement over the first few years of the program.
11/17/2018
Argonne National Laboratory

17. For example, we show a specific coal-fired unit built in 1964
4/1/2015
For example, we show a specific coal-fired unit built in 1964
Effects on unit Heat Rate and availability due to exposure to cycling, as represented by ESIM model
Heat rate increases and availability decreases will decrease the unit’s capacity factor
11/17/2018
Argonne National Laboratory

18. $50 dollar CO2 price shock in 2020
4/1/2015
$50 dollar CO2 price shock in 2020
A typical CFPP, with capacity factor and penalty score shown below, retired in 2024 due to cycling damages
11/17/2018
Argonne National Laboratory

19. 4/1/2015
Total variable costs of units dispatched and their resulting capacity factors
As shown in the figure, the increasing variable cost curve is flatter than the decreasing CF curve. The latter is constrained by the slope of the load curve.
Under a high carbon tax, the gas units have lower variable costs and are dispatched before the coal units with higher variable costs
11/17/2018
Argonne National Laboratory

20. 4/1/2015
discussion
The relatively flat variable cost curve can also contribute to gas market price volatility, i.e., small changes in gas price can swing large changes in gas demand as NGCC utilization rates change significantly as their loading order shifts.
The companion presentation discussed in more detail the modeling structure and findings
11/17/2018
Argonne National Laboratory