Evidence room for a neighborhood emergency planner. Question dated ; this packet is frozen at that date, the release date of the ERCOT assessment, and reports no later events. ST1
Should ERCOT's published winter reserve be represented as a guarantee for an extreme cold event?
ERCOT's winter assessment says it "anticipates" sufficient installed capacity for forecasted peak demand, and its reserve rows measure "Capacity Available for Operating Reserves, Normal Operating Conditions." That is a planning estimate, not an operational guarantee. In its own harshest column (extreme peak load with extreme generation outages) only 1,352 MW remains. ST1
The 2011 FERC/NERC staff report found that ERCOT and others "had adequate reserve margins, based on anticipated generator availability" going into that storm, "But those reserves proved insufficient." ST2
Decision proposed: plan the neighborhood response for possible rotating outages whenever extreme cold is forecast, regardless of the published reserve (details below). No probability is assigned and no motive is alleged. ST1 ST2
Remaining operating reserve in the four published columns
Baseline reconciliation, all four columns
Each column starts from the same reserve: Total Resources 82,513 MW minus Peak Demand 57,699 MW equals 24,814 MW. Separate adjustments are then subtracted. Every column below adds up exactly to the printed totals. ST1
| Row | 1 Forecasted Season Peak Load | 2 Extreme Peak Load / Typical Generation Outages | 3 Forecasted Season Peak Load / Extreme Low Wind Output | 4 Extreme Peak Load / Extreme Generation Outages |
|---|---|---|---|---|
| [a] Total Resources | 82,513 | 82,513 | 82,513 | 82,513 |
| [b] Peak Demand | 57,699 | 57,699 | 57,699 | 57,699 |
| [c] Reserve Capacity before uses (a minus b) | 24,814 | 24,814 | 24,814 | 24,814 |
| Seasonal Load Adjustment | 0 | 9,509 | 0 | 9,509 |
| Typical Maintenance Outages, Thermal | 4,074 | 4,074 | 4,074 | 4,074 |
| Typical Forced Outages, Thermal | 4,542 | 5,339 | 4,542 | 5,339 |
| 95th Percentile Forced Outages, Thermal | 0 | 0 | 0 | 4,540 |
| Low Wind Output Adjustment | 0 | 0 | 5,279 | 0 |
| [d] Total Uses of Reserve Capacity | 8,616 | 18,922 | 13,895 | 23,462 |
| [e] Capacity Available for Operating Reserves, Normal Operating Conditions (c minus d) | 16,198 | 5,892 | 10,919 | 1,352 |
| Derived: thermal forced outages, typical plus 95th percentile | 4,542 | 5,339 | 4,542 | 9,879 |
- The extreme load cases add 9,509 MW to peak demand, matching the stated extreme winter forecast of 67,208 MW, which the SARA says is "Based on the 2011 winter." ST1
- The page 1 "unit outage forecast" of 8,616 MW equals column 1 maintenance plus typical forced outages. ST1
- The low wind column subtracts 5,279 MW, leaving a stated low wind output level of 1,791 MW. ST1
- The printed resource line items sum to 82,514 MW, 1 MW above the printed Total Resources, consistent with rounding. The printed total is used because the printed reserve is built from it. ST1
Total Resources line items as printed (MW) ST1
| Line item | MW |
|---|---|
| Operational Resources (thermal and hydro) | 67,529 |
| Switchable Capacity Total | 3,710 |
| Less Switchable Capacity Unavailable to ERCOT | -568 |
| Available Mothballed Capacity | 0 |
| Capacity from Private Use Networks | 3,631 |
| Coastal Wind, Peak Average Capacity Contribution | 1,480 |
| Panhandle Wind, Peak Average Capacity Contribution | 1,411 |
| Other Wind, Peak Average Capacity Contribution | 3,251 |
| Solar Utility-Scale, Peak Average Capacity Contribution | 269 |
| Storage, Peak Average Capacity Contribution | 0 |
| RMR Capacity to be under Contract | 0 |
| Capacity Pending Retirement | 0 |
| Non-Synchronous Ties, Capacity Contribution | 838 |
| Planned Thermal Resources with Signed IA, Air Permits and Water Rights | 0 |
| Planned Coastal Wind with Signed IA | 371 |
| Planned Panhandle Wind with Signed IA | 0 |
| Planned Other Wind with Signed IA | 557 |
| Planned Solar Utility-Scale | 35 |
| Planned Storage | 0 |
| Sum of line items | 82,514 |
| [a] Total Resources, as printed | 82,513 |
Where column 4 runs out: a sensitivity, not a prediction
Column 4 already counts 9,879 MW of thermal forced outages (5,339 MW typical plus 4,540 MW at the 95th percentile). If every other column 4 assumption stayed fixed, a thermal forced outage total of 11,231 MW, that is 1,352 MW more, would bring capacity available for operating reserves to exactly zero. Including thermal maintenance outages, that is 15,305 MW of thermal capacity out. This is arithmetic on the published table, not a measured outage or a forecast. ST1
Sensitivity control: hypothetical extra thermal forced outages in column 4
Bounded from 0 to 4,000 MW of hypothetical extra outage. The published column stays fixed beside it. Inputs come from page 2, column 4. ST1
Above zero Baseline: with no extra outage, column 4 keeps 1,352 MW. Zero is reached at 1,352 MW of extra thermal forced outages (11,231 MW total). ST1
What the 2011 FERC/NERC report actually warned
Exact words from the staff report on the February 1-5, 2011 Southwest cold weather event, printed pages 7 to 10. ST2
"They all had adequate reserve margins, based on anticipated generator availability. But those reserves proved insufficient for the extraordinary amount of capacity that was lost during the event from trips, derates, and failures to start." ST2
"But over the course of that day and the next, a total of 193 ERCOT generating units failed or were derated, representing a cumulative loss of 29,729 MW." ST2
"Combining forced outages with scheduled outages, approximately one-third of the total ERCOT fleet was unavailable at the lowest point of the event." ST2
"Generators and natural gas producers suffered severe losses of capacity despite having received accurate forecasts of the storm." ST2
"However, the poor performance of many of these generating units and wells suggests that these procedures were either inadequate or were not adequately followed." ST2
"These recommendations were not mandatory, and over the course of time implementation lapsed. Many of the generators that experienced outages in 1989 failed again in 2011." ST2
Cumulative loss is not a simultaneous outage
29,729 MW
The cumulative loss from 193 units that failed or were derated "over the course of that day and the next." Units counted here were not necessarily all out at the same moment, so this figure is not a concurrent outage and cannot be set against a single-hour reserve. ST2
About one-third
"approximately one-third of the total ERCOT fleet," combining forced and scheduled outages, was unavailable at the lowest point. The excerpt gives no MW figure for that moment, so none is supplied here. ST2
On the first day ERCOT had 3100 MW of responsive reserves against a minimum requirement of 2300 MW; reserves "eventually dropped below the level of safe operation." Responsive reserves are a real-time quantity and are not the same row as the SARA planning figure. ST2 ST1
For the Southwest as a whole, 67 percent of generator failures (by MWh) were due directly to weather-related causes, and at least another 12 percent were indirectly attributable to weather through natural gas curtailments and fuel switching difficulties. These are shares of past failures, not a probability of any future event. ST2
Timeline in the packet
- Cold weather caused many generators to trip, derate, or fail to start; PUCT winterization recommendations were not mandatory and implementation lapsed. ST2
- Rolling blackouts; in ERCOT they affected 3.2 million customers. Many generators that failed in 1989 failed again. ST2
- FERC and NERC staffs publish the report quoted above. ST2
- ERCOT releases the final Winter 2020/2021 SARA, covering December 2020 to February 2021. ST1
Normal operating reserve versus an operational guarantee
The SARA row is titled "Capacity Available for Operating Reserves, Normal Operating Conditions." It is a seasonal planning estimate for a peak hour built from historical averages and percentiles. It does not promise that this many megawatts will be deliverable at every instant of a cold event, and the summary itself uses the word "anticipates." ST1
2011 showed the gap between the two: margins that looked adequate beforehand were overwhelmed as capacity was lost during the event. ST2
What the SARA does not prove, and this packet cannot bound
- Outage definitions. Forced outage rows are historical averages (and a 95th percentile) for December to February weekdays, hours ending 7 am to 10 am, over 2017/18, 2018/19 and 2019/20. The packet does not say how trips, derates and failures to start are counted, so they cannot be matched to the 2011 categories. ST1 ST2
- Fuel supply. Fuel limitation derates are included only "at units in north Texas," with the extreme case adding those from the January 17, 2018 cold weather event. Wider gas curtailment, which ST2 links to failures, is not separately bounded. ST1 ST2
- Combined stresses. No column pairs extreme peak load with extreme low wind. Simply adding the column 3 wind adjustment to column 4 would exceed its remaining reserve by 3,927 MW, but the SARA neither publishes nor endorses that combination; it marks an unbounded case, not a result. ST1
- Duration. The table is one peak snapshot. It does not model consecutive cold days or losses that accumulate, which is how ST2 reports 2011. ST1 ST2
- Demand ceiling. The extreme load case uses the 2011 winter as its basis. Demand above 67,208 MW is not covered. ST1
- Common-cause cold failures. Frozen sensing lines, equipment, water lines and valves, blade icing and low temperature limits drove 2011 failures. Whether winterization is now adequate or followed is not quantified in the SARA. ST2 ST1
- No probability. Neither document gives a probability for an extreme winter storm, and none is assigned here. ST1 ST2
Recommended preparedness decision
Decision: adopt a neighborhood cold-weather plan that assumes rotating outages are possible during any extreme cold forecast, whatever the published reserve. Trigger it on the weather forecast, not on grid notices, because in 2011 losses came "despite having received accurate forecasts of the storm." ST2
- Confirm a warming location with heat that does not depend on the grid.
- Refresh a contact list for residents who rely on powered medical equipment or live alone, with a check-in plan that works without power.
- Publish pipe protection and safe outage heating guidance before the cold arrives.
- Plan for losses that build over more than one day, not a single peak hour; the 2011 loss accumulated over two days. ST2
In public materials, describe the SARA figures as planning margins under normal operating conditions, with column 4 at 1,352 MW, never as a guarantee. ST1
Sources and method
- ST1: ERCOT Winter 2020/2021 SARA, released November 5, 2020, pages 1 and 2.
- ST2: FERC/NERC 2011 cold weather report, August 2011, printed pages 7 to 10.
Links carry the original addresses; opening them needs a connection. Figures were parsed from the packet excerpts by model.py and verified independently by check.py. No named-person claims are made. Nothing after November 5, 2020 is reported. ST1