System operators in the WECC footprint were too quick to declare energy emergency alerts (EEAs) during several recent events, resulting in unnecessary load sheds on two occasions, according to a new NERC “lessons learned” notice.

The notice detailed five incidents in which software glitches, communication lapses and generator malfunctions contributed to unnecessary EEAs, two of which resulted in the shedding of a combined 250 MW of load.

Donnie Bielak, PJM’s manager of reliability engineering, briefed the RTO’s Operating Committee on the lessons Thursday. “Some of these were actually EEA-3s and result in load shed, so some of the mitigating actions were quite severe,” he said.

One of the load-shed incidents occurred in the evening of a hotter-than-forecast day as solar PV resources were declining, when an unidentified reliability coordinator declared an EEA-1 for a balancing authority at the request of the BA, referred to in the document as “BA 1.” (As is customary, although the notice identified WECC as the source of the lessons, it did not identify the entities involved.) About the same time, two generators totaling 530 MW tripped offline, causing the RC to place BA 1 and a second BA within the same reserve sharing group (RSG) in an EEA-3.

Under EEA-1, the BA seeks to use all available generation to meet firm load, firm transactions and reserve commitments; non-firm wholesale energy sales are curtailed unless they are recallable to meet reserve requirements. Under EEA-3, firm load interruptions are imminent or in progress.

The RSG’s computer program for requesting contingency reserve assistance from other RSG members requires the system operator to fill out information in a pop-up display. Because BA 1 submitted its request first, the program required BA 2 to acknowledge BA 1’s request before submitting its own. But the system did not credit BA 2 with the assistance it was providing to BA 1 because BA 2 filled out its request before acknowledging the request from BA 1. As a result, the program required BA 2 to provide about 200 MW more generation than was needed for the RSG to recover its reporting area control error (ACE).

“It appeared that they were short and they actually weren’t,” Bielak explained.

In addition, about 60 MW of contingency reserves were not available because a resource failed to start. The combination of events made it appear that BA 2 could not recover its ACE within 15 minutes, prompting a BA 2 system operator to shed 150 MW of firm load.

“What was not known to the system operator at the time was that the two units tripped one minute and six seconds apart, so the resource loss was outside the one-minute threshold of a reportable balancing contingency event, making the requirement to recover the reporting ACE within 15 minutes not applicable per BAL-0024,” NERC said.

In another incident, operators reduced available transfer capability for a major transmission corridor threatened by a wildfire, reducing an RSG’s ability to deliver contingency reserves to some of its member BAs. One of the BAs asked its RC to put it into an EEA, but because the BA was participating in the RSG, the contingency reserves from the zone were sufficient to recover from the largest, most severe single contingency. “After-the-fact review per the RSG and the BA indicated the BA did not need to request an EEA,” NERC said.

In another case, a BA fell short of its required contingency reserves as load was increasing, causing its RC to declare an EEA-3. It was later discovered that the BA was never short generation but that its tool was not reporting all available generation to system operators.

Misreading BAL-002

The second incident that resulted in load shedding occurred when a BA lost generation and asked for reserves from its RSG but fell short of its required generation when some units it requested to help failed to start. To recover the BA’s reporting ACE within 15 minutes, the system operator shed about 100 MW of load.

It was later determined that, because the BA was in an RSG and the amount of lost generation was less than the threshold for a reportable balancing contingency event, the BA did not have to recover their reporting ACE within 15 minutes. Thus, the system operator was not required to shed load under BAL-002.

Several changes resulted from the incidents, including:

• weekly start-up tests of units relied on for replacement reserves;
• an initiative to develop a new way of determining generation replacement reserves that takes account of factors such as increased outages, market dynamics and variable generation;
• development of a new load-forecasting tool, which is being used on a trial basis; and
• enhanced operator training on topics, including three-part communication, implementing interruptible loads and responding to data integrity issues.

Changes also resulted from a fifth incident, when a 300-MW resource was lost in the evening as solar resources were dropping, causing the BA to fall below its required levels and triggering an EEA-3. The incident resulted in the introduction to the day-ahead load forecast of a “high confidence” band — assuring load will come within the high and low ranges of the forecast. It also caused the entity to increase its coordination with its thermal generators to ensure units are positioned in their fastest ramp rate ranges before the start of the evening solar ramp.

NERC said the incidents indicated that BAs participating in an RSG need to understand when they are acting as a member of the RSG or as an independent BA. “BAs in [two cases] dropped load per their individual limits but not per their RSG obligations,” NERC said. “They were focusing on recovering their individual ACE.”

It said BAs participating in an RSG should provide periodic refresher training to their system operators on the applicability of BAL-002 and procedures to determine when they are and are not considered an active member of the RSG.

It also said RSGs should validate their programs for multiple contingency reserve activations to ensure their computer programs do not miss prior reserve activations. “BAs should have a process to validate that all available reserves are accounted for and properly displayed for the BA system operators to be aware of in case they need to be called upon,” NERC said.

Wind Generation Cutoffs During Cold Weather

The Midwest Reliability Organization was the source of a second lessons learned report that NERC said highlighted the need to ensure grid operators know whether wind turbines in their footprints can operate in extreme cold. [Editor’s Note: An earlier version of this article mistakenly attributed the lesson to WECC.]

The report stemmed from extreme cold weather on Jan. 29-31, 2019, when unplanned wind generation outages triggered a maximum generation event, resulting in a registered entity calling on demand response, behind-the-meter generation and voluntary reductions to avoid emergency power purchases.

Actual overnight temperatures for the period were a few degrees lower than forecasted, and when temperatures fell below -21 degrees Fahrenheit, some wind farms shut down their turbines to avoid damage to their gear boxes.

On Jan. 30, the temperature was 4 degrees lower than expected and wind output was 6 GW below the day-ahead forecast, a 50% shortfall. Fears of insufficient generation to meet the morning peak resulted in the maximum generation event.

The entity had expected 8.5 GW of wind generation in its unit commitment for the Jan. 30 peak, which reflected a 1-GW reduction because of unit cutoffs from cold temperatures. But 98 of 216 wind generators had to be derated or shut down, leaving the entity with only about 4 GW of wind for the peak load hour. The units that continued to operate had heating sources in the gearbox to prevent the oil from freezing, allowing them to operate at -40 F.

The entity’s deployment of load management resources, along with school and business closings, reduced demand by at least 3 GW, allowing it to avoid emergency power purchases.

“Wind unit owners should prepare for extreme cold-weather performance and promptly communicate anticipated operating parameters and data to their BA, RC and [transmission operator] to ensure readiness and provide situational awareness in both operations and planning,” NERC said.