VALLEY FORGE, Pa. — PJM experienced 77 emergency procedures in May, staff told attendees at last week’s Operating Committee meeting.

Calling it a “busy month,” PJM’s Chris Pilong said the emergency procedures included the first time the RTO has had to order load shedding since implementing its Capacity Performance rules in 2015. (See PJM Experiences First Load Shed in the CP Era.)

The events resulted in a portion of the load forecasting error exceeding its 3% target for the first time since July. The on-peak forecasting error was 3.08%, and the off-peak 1.69%, putting the overall error at 2.38%.

While the error increased in some transmission zones, East Kentucky Power Cooperative posted a 3.3% error, the lowest level in the past 10 quarters.

Load Shed Event

Pilong explained that five facilities were involved in the event. Three 138-kV lines in the area were on planned outages that day. A transformer and additional line tripping out of service triggered “multiple” contingency overloads, which potentially could have resulted in a cascading outage if another facility was lost, Pilong said. Based on that analysis, PJM ordered a pre-emptive load shed to reduce the contingency flow on the Edison-Kankakee line. Within 15 minutes of issuing the order, the transformer was restored, and PJM canceled the load shed nine minutes later.

“Given the timeline, we didn’t need to, but we were definitely looking at [dispatching demand response or behind-the-meter generation in area] and considering those as well,” Pilong said.

The load shed triggered performance assessment intervals (PAIs) that lasted about 30 minutes. While PAIs can trigger significant nonperformance penalties or performance bonuses, none resulted from the event, staff said. The incident was isolated to a small area of northwest Indiana that includes fewer than four generation owners, so PJM’s confidentiality rules prevent staff from releasing any additional information without the owners’ agreement. PJM’s Adam Keech said staff are working with owners to see if they can agree on releasing anything else.

Keech said PJM determined which units were involved by looking at any units that could have increased output to help alleviate the constraint for which the load was shed.

Event Analysis to Follow

While he couldn’t provide specifics on why the event yielded no penalties or bonuses, Keech advised stakeholders to “just remember we are in a year where we are not 100% CP,” referring to the interim base capacity designation PJM implemented as it transitions to the CP requirement that a resource always be available. Base capacity doesn’t have that requirement.

GT Power Group’s Dave Pratzon asked that staff analyze why the three 138-kV lines were allowed to be on planned outages simultaneously because it potentially puts “a few unlucky generators” at financial risk for something they can’t control.

“That’s potentially a large dollar impact for something that potentially has nothing to do at all with generator issues,” he said.

That question and the cause of the facilities tripping are “exactly what we’re looking at as part of the follow up,” Pilong said.

Several stakeholders asked PJM to find better ways to communicate the extent of the incident. RTO staff said they can only target messaging to the level of the transmission zone, even though the event affected a much smaller area, causing many stakeholders to wonder whether they were involved or not. Pilong said the conditions would have to be exactly the same for any refinement of the communication to be more selective, and that’s “probably unlikely.”

Besides, response to the event was unexpectedly quiet, despite the potential confusion.

“Oddly, we only got one phone call,” Pilong said. “It was, to be honest, a little bit surprising.”

Beyond that call, “there was no other anomalous behavior that was obvious or impactful,” he said, adding that system operators’ advice was the same as it would have been for any unit: follow the dispatch signal PJM provides.

Related Updates

Later in the meeting, PJM’s Alpa Jani explained that the load-shed directive was posted at 1:34 p.m. and was effective for 1:22 p.m. Any units that receive system notifications for the AEP transmission zone received the message because the area around the Edison substation where the equipment tripped is not defined as a subzone.

In another presentation, PJM’s Pete Langbein discussed how better “coordination” with behind-the-meter generation, also known as non-wholesale distributed energy resources, could help decrease load forecast errors or mitigate load sheds.

PJM is proposing to identify all such non-wholesale DER of greater than 1 MW on an annual basis, primarily through public Energy Information Administration data. Transmission owners would verify the data and include additions as available so they can be modeled in PJM’s planning and operations tools. The TOs would communicate downstream to the resources as necessary during events to avoid load sheds or dumps. Langbein said draft manual and Tariff language is being introduced in the DER Subcommittee and will move through stakeholder endorsement from there.

Security Initiatives

PJM’s Colin Brisson reviewed security initiatives planned for the RTO this year.

“Critical infrastructure in geopolitics is becoming a higher-priority target” and has hit the energy sector, he said. “We’re actually catching up to the curve where many companies are at.”

PJM is implementing geo-IP blocking, which blocks outside computers from interacting with the RTO’s network if its unique digital signature (or IP address) originates from “high-risk countries,” which Brisson didn’t identify. The technology will be rolled out “increasingly” throughout the year, he said.

The RTO is also implementing two-step verification, which means that along with providing the right password, users will have to tie their accounts to their devices using a “token” to log onto PJM’s network. Once a token is verified, users will be able to log on from that device without going through the process again. Training will begin on Aug. 15 and “full production” to members is scheduled for Oct. 10.

DMS

PJM’s Maria Baptiste announced the Data Management Subcommittee has decided to stop scheduling DMS-Joint meetings and instead hold them on an ad hoc basis as needed to address issues because of “very limited participation.” DMS-Confidential meetings will continue on their existing schedule, and several parts of the DMS-Joint will transition to the Confidential group, including reviewing NERC lessons learned.

The subcommittee will still have work to do. PJM’s Shaun Murphy announced that staff plan to ask the DMS to investigate why the quality of phasor measurement unit (PMU) data has been degrading. He presented a graph showing spikes in error percentages in various transmission zones through the RTO since February 2017. The issues include time, synch and drop errors, planned outages, missing samples and issues with engineering limits, such as threshold, noise and topology.

“On average, we’re starting to see they typical error rate starting to climb,” he said.

The DMS will investigate the impact of the data quality on applications that use the PMU data, enhancing the definition of “data quality,” improving real-time data quality monitoring, reviewing data quality requirements in manuals and guidelines for device outages.

30-Minute Reserve Vote Deferred

PJM had hoped to receive OC endorsement for its planned procurement of 3,784 MW for real-time 30-minute operating reserves, but the vote was deferred because the topic wasn’t included as a voting item on the agenda and came near the end of the three-hour meeting. Based on an analysis of potential reserve shortages, PJM estimates it should secure nearly 3,800 MW of a new 30-minute real-time reserve product. (See “30-Minute Reserves Target Set,” PJM Operating Committee Briefs: May 1, 2018.)

Synch Reserve Response

The RTO experienced one synchronized reserve event of more than 10 minutes in the first quarter, PJM’s David Kimmel said. Of the 1,897 MW estimated for the Tier 1 response, 510 MW responded, or 27%. Demand-side response was assigned all of the Tier 2 response. Of the 113 MW assigned, 58 MW responded, or 51%.

There were three events altogether, all of which occurred in January. Overall, 37% of Tier 1 estimates actually responded, or 2,029 MW. All of the 933 MW of generation assigned Tier 2 response responded, while 341 MW responded of the 397 MW of demand-side response assigned to Tier 2, or 86%.

The events resulted in $1.15 million of Tier 1 credits and $6,666 of Tier 2 penalties.

Skepticism of Gen Capability Changes Continues

Stakeholders remain skeptical of PJM’s plan to revise procedures for generators’ capability testing requirements, which has the potential to reduce generators’ capacity injection rights (CIRs). For several months, PJM’s Jerry Bell has been presenting data analyses to justify the changes to using median capacity factors, arguing that the RTO’s current methods using average capacity factors overestimate what units can realistically be expected to provide. But stakeholders have been concerned about losing value they’ve already paid for. (See “CIR Questions,” PJM Operating Committee Briefs: May 1, 2018.)

Generators are concerned that some existing or planned CIRs could be potentially stranded through PJM’s proposal because it would reduce how a plant’s output is measured for the purposes of qualifying for CIRs.

“PJM is being kind of cavalier with other people’s investments. … There are other ways to do this,” Dayton Power and Light’s John Horstmann said. “I don’t think you’ve addressed the transition nor the compensation adequately. … These interconnection investment costs are not linear.”

He reiterated a request for a special session to discuss the implication of the proposed changes, to which PJM staff ultimately agreed.

Bell’s presentation last week focused on the relationship between summer weather and production from hydroelectric dams. Among PJM’s proposed changes is limiting facility testing to July and August and eliminating June from the testing window. Bell’s analysis showed that hydro capability dips in July and August compared to June.

“As river temperature increases, generator capability wanes, but the majority of the capability decrease can be attributed to the cooling towers that are placed in service incrementally as river temperature increases and control of thermal discharge is needed,” Bell said. “These are the kinds of issues I’m having and why I want to see full plant testing.”

He said a “blanket” RTO calculation is infeasible because conditions vary throughout the RTO’s footprint and there will always be a situation where the analysis won’t be applicable, “so I’d rather just have everybody test in July or August.”

Several generation owners expressed concerns with the plan, such as the constraints of being able to test during a more compressed timeline.

“We just don’t know how we would get this done in two months,” Exelon’s Sharon Midgley said.

“PJM is kind of cavalier with other people’s investments. … There are other ways to do this,” Horstmann said. “I don’t think you’ve addressed it adequately. … The investments are not linear.”

“I’m open to suggestions, but … I want to make sure that everybody understands that when you use the average capacity factor, you are overstating your ability to meet load during peaks and we need to rectify that situation,” Bell said.

Some stakeholders suggested tailoring the requirements to specific unit characteristics, though Bell envisioned some concerns with that.

“Then it becomes somewhat discriminatory to some folks … but if we can work that out, I don’t have a problem at all,” he said.

He said units with “questionable test” results would likely be the first asked to retest under the new rules, but “there will probably be some folks that I would never even look at them.” Other units likely to be contacted are those whose ambient conditions change during the season.

John Brodbeck of EDP Renewables said the plan creates CIR issues for generation in the interconnection queue that will fund network upgrades and “it sort of cries out for a problem statement.” PJM staff did not respond to the suggestion.

— Rory D. Sweeney