GMD Planning Application Guide Comments 11/12/2013Stakeholder Section/Page number

General

General

Pg 4, line 30

Pg 5, line 10

Pg 10, line 16

General

James McIver, Siemens

Paul Rocha, Centerpoint

Energy

page 13, beginning on line 5

Paul Rocha, Centerpoint

Energy

page 13, beginning on line 5

General Comment

General Comment

page 9, Line 2-4

Paul Rocha, Centerpoint

Energy

Yan Du, MidAmerican

Page 3, Lines 12 and 13

Page 4, Line 20, (and elsewhere in the report)

Pg.4 line 6

Pg.4 line 23

Pg.4 line 31

Pg.6 line 1

Pg.9 line 2

Pg.10 line 11

Pg.10 line 26

Reigh Walling, WES Consulting

John Bee, Exelon

P.10 line 35

General Comment on Tools

General Comment

Equipment Resiliency

Next

John Bee, Exelon

Alberto Ramirez Orquin, UPRM

General Comment - Design Basis Event

Chapter 4

General Comment - "Storm Scaling"

Thomas Popik, Foundation for

Resilient Societies

Transformer Testing (Transformer Modeling Guide)

Preliminary Technical Comments

Thomas Popik, Foundation for

Resilient Societies

Preliminary Technical Comments

Thomas Popik, Foundation for

Resilient Societies

Preliminary Technical Comments

Conclusions

General

Thomas Popik, Foundation for

Resilient Societies

Sergio Garza, PC Reviewer

Page 4 "Initial Screening"

page 8 "Reliability Criteria"

General

Sergio Garza, PC Reviewer

GMD Planning Application Guide Comments 11/12/2013CommentThe document is well-written, with clear organization and easily understood statements of fact. I commend the task force member who have drafted this document -- good job!

My only specific comments relate to this document's linkage with other GMD documents. There appear to be at least three documents linked together (and even as a taskforce member, I become confused as data is linked between them.)

It is very confusing about which type of information is contained in each of these documents. For example, it is not clear to the 1st time reader that the Planning Application Guide is the overview document, with details contained in the other documents. I would also suggest at least one of the documents needs a name change > "GIC Application Guide" sounds like an earlier version of the "Planning Application Guide" and should be labeled differently. It is also likely that there will be confusion about location of system earth modeling details (in the GIC Application Guide, I assume?) and xfmr modeling details (which will be contained in the 2nd modeling document.)

Will Ref [2] be modified to include Pulkkinen's 1 in 100 year storm scenario? (I don't think that info is contained in Version 1, dated April 2013.) There was a very nice storm-scaling approach which was discussed under the heading "Team 3 Application Guide" at Vancouver meeting. Which document will contain that modeling approach ?

Which document will contain necessary modeling information to address the reactive power absorption needs of transformers? I assume that info will be in Ref [3], Xfmr Modeling Guide but I don't think we have seen that document yet?

I think an editorial addition is necessary for the statement "....manufacturers maintain that such capability curves have to be developed for every transformer design and vintage...." Perhaps modification to a more nuanced phrase would be helpful to the industry > "GIC thermal duty is not yet addressed in xfmr standards and only a handful of units have received detailed thermal assessments. Since only a portion of construction configurations and vintages have been addressed, manufacturers maintain that in the near-term such capability curves have to be developed for any transformer judged vulnerable to significant GIC flow...."

CenterPoint Energy appreciates the diligent work of the GMDTF Team charged with drafting the Planning Application Guide (“the Guide”). CenterPoint Energy finds the draft document to generally be a useful guide.

The Guide describes initial screening analysis using “contingencies considered relevant by the planning authority” and further suggests that a 3% voltage fluctuation and operational limits be used as a screening criteria. CenterPoint Energy requests that the GMDTF consider the following red-line changes to this verbiage:

3. Perform power flow analysis using stressed system conditions and generic models for reactive power absorption. A conservative approach is to assume that all transformers are single-phase; however, in some cases this approach will be overly conservative. Loss of reactive power sources such as shunt capacitor banks and SVCs (on protection) should be considered as valid contingencies associated with the GMD event. If the steady state model does not indicate evidence of voltage collapse, then more detailed load flow studies are probably not necessary.

CenterPoint Energy is concerned that using “contingencies considered relevant by the planning authority” is vague and will likely result in inconsistent application across the various NERC regions, and also does not provide useful guidance to planning authorities regarding what contingencies should be considered for this type of analysis. CenterPoint Energy’s proposed change uses the guidance provided on page 9, lines 33-34 regarding contingency selection.

CenterPoint Energy is likewise concerned that referring planners to the planning authority for load levels to be used would likewise be problematic, and suggests a change to indicate stressed system conditions should be used for the analysis.

Finally, a rare (ostensibly one in 100 years, as discussed on Page 8, lines 18-19 of the Guide) event such as the event contemplated in the Guide is much, much less likely than NERC TPL Category B or C events, and therefore should be considered an extreme, NERC TPL Category D type of event. If a one in 100 year GMD event is used for the design basis, the frequency of that event occurring at the most severe field orientation is even less frequent than one in 100 years. Furthermore, that GMD event occurring at the worst field angle coincident with stressed system conditions is even far less likely. Given all these considerations, CenterPoint Energy believes that the performance criteria should be based upon performance criteria for a NERC TPL Category D event; i.e., that voltage collapse does not occur. As it is, the proposed 3% voltage fluctuation criterion is inappropriate because it is far more stringent than the performance criteria for much more common NERC TPL Category B events. For example, it is not uncommon for systems to normally operate in a range of 0.95 to 1.05 per unit voltage, and to allow bus voltages to drop to 0.95 per unit voltage, or lower, for NERC Category B contingencies. Thus, the outage of a single element, without any GIC, could cause voltage fluctuations of 10% or more. Stated otherwise, if a planner considers and models a single capacitor bank outage as a relevant contingency for the GMD initial screening analysis, the voltage may fluctuate more than 3% without any consideration of reactive losses caused by GIC, and that system performance would be quite adequate. If two or more simultaneous (or near simultaneous) capacitor bank outages are modeled as part of the GMD screening analysis, it is more likely that voltage fluctuations greater than 3% could occur without any consideration of reactive losses caused by GIC.

Moreover, using operating limits as a performance criterion is inappropriate due to the short duration of the peak GIC waveform contemplated by the assessment. Most operating limits are based upon “short” term emergency ratings typically in the range of two to four hours, much longer than the time frame that the peak GIC waveform would be present. Given all these considerations, CenterPoint Energy asks that the GMDTF consider the changes proposed by CenterPoint Energy to this portion of the Guide

There is a section of the Guide relating to Generator, Capacitor Bank, SVC, and Protective Relaying Impacts, presumably as these impacts should be considered for screening or system impact assessments. That section begins with a statement that system harmonic analyses are necessary to investigate harmonic-related system impacts of GMD, and then continues with a discussion of how harmonics impacts various pieces of equipment.

There is a separate section on Harmonic Impact Studies on page 15 of the Guide. That section of the Guide indicates that the industry has limited availability of appropriate software tools to perform harmonic analyses and further describes the difficulty of building network models that provide reasonable representation of harmonic characteristics. Later in the Harmonics Impact Section (page 15, lines 26-32), the Guide indicates that even if the modeling obstacles were overcome, the harmonic tolerance level for most equipment is poorly defined except for capacitor banks and, furthermore, that the sensitivity of capacitor bank protection systems to harmonics is not well defined.

In summary, the Harmonics Impact Studies section on page 15 indicates that harmonic impacts are difficult to model and analyze and, even if it could be achieved with considerable effort, it would be problematic to definitively quantify the results. Given this situation, CenterPoint Energy suggests the following changes to the Generator, Capacitor Bank, SVC, and Protective Relaying Impacts section on page 13, to better align that discussion with the Harmonics Impact Studies section on page 15 of the Guide and provide more useful guidance to planners:

"Harmonics caused by half-cycle saturation of transformers during a severe GMD event could cause various pieces of electrical equipment to fail. Theoretically, harmonic impacts could be analyzed as indicated in the Harmonic Impact Studies section of this Guide. However, as indicated in that section, the reality of the situation is that harmonic impacts are difficult to model and analyze and, even if such modeling and analysis could be achieved with considerable effort, it would be problematic to definitively quantify the results because harmonic tolerance of most pieces of equipment is poorly defined.

Given these practical considerations, planners should use judgment when considering harmonic effects on equipment in GMD planning studies. Low-order harmonic current injections can travel considerable distances through the transmission system. Thus, at any location in the system the harmonic currents and voltages may represent the aggregated contribution of multiple GIC-saturated transformers. As such, harmonic withstand capabilities of any given system component are not solely the result of the harmonic injections of a particular transformer (e.g. the harmonic currents flowing into a generator are not solely due to saturation of the GSU transformer). That said, harmonic impacts to equipment should generally be most severe to equipment located in close electrical proximity to the source of the harmonics. Furthermore, a planner can also reasonably assume that harmonic impacts due to GIC current are more severe for higher levels of GIC. Therefore, for planning study purposes, a planner can reasonably model the contingency loss of reactive resources in close proximity to ground-connected transformers that have high level of GIC in planning assessments. Generators can also be impacted by harmonics but, as a practical matter, instances of generator damage or tripping due to harmonics caused by GIC have been exceedingly rare. Therefore, most planners would not be expected to model generator tripping as a contingency in GMD planning analyzes except perhaps in situations where generators or generator step-up transformers at certain specific locations are known to be particularly vulnerable to high levels of GIC based upon historical experience."

In reviewing the Draft Application Guide, the reference [3]-NERC Transformer Modeling Guide has been referred multiple times and seems to be a key document, but it is not readily available as of the comment deadline. We would like to be notified when it becomes available.

It is unlikely to be able to model foreign BES elements for two or more buses in neighboring network as explicitly as modeling elements that we own. The Planning Application Guide should suggest ways on how to make reasonable approximations or assumptions for modeling foreign neighborhood elements.

the Planning Application Guide should provide a clear definition of the term “surge-arrester grounded transformer”, it is also unclear as to whether the “surge-arrester grounded transformers” should be modeled as high resistance branch or not, if yes, what the resistance should be; if no, can those be excluded in the model? In addition, It is also unclear about the meaning of “numerical instability”.

Initial screening should be performed for fields in at least a few different directions.

refers to taking into consideration the “waveshape”, what is the time and magnitude of the waveshape to be used for the study?

This bullet implies that the consequence to generators from excessive GMD is tripping on protection. This is not accurate in practice. As adequately explained elsewhere in this document, the risk to generators is material damage due to rotor heating from harmonics. Also explained elsewhere in the report is that most, if not all, generation protection systems in current use tend to ignore harmonic currents and do not adequately consider their thermal impact on generators. Thus, generators are largely unprotected today from excess harmonics during GMD, and the risk is not merely tripping, but rather serious of not destructive damage.

The report concentrates attention on SVC tripping. This is understandable, given the fact that SVC tripping was the final straw that caused the 1989 GMD-related blackout in Quebec. However, focusing so much on just SVCs takes attention away from the fact that all complex transmission system equipment, including the range of FACTS devices and HVDC, are vulnerable to unanticipated behavior and possible tripping during GMD. The inherent complexity of these systems, and the susceptibility to complex system interactions makes them prone to such behavior. Proposed is to say “FACTS devices, including SVCs, and HVDC…”.

Recommend modifying statement “The system must perform within specified limits…” to read “Evaluate performance of the system with respect to applicable limits…”. This will depend on the severity of the event being studied and required performance during a GMD event, which has not necessarily been established.

Can dissolved gases be predicted from the screening study described? GIC flow through the transformer may be a more efficient screening parameter in an initial analysis.

Recommended change: “voltage collapse can occur when the system does not have enough var reserves to support current operating conditions or recover from a valid contingency”.

Recommend changing to read “The dc model does not need to include ungrounded transformers”. The stations may need to be modeled based on other lines, etc. at the location.

Regarding the statement: ” A simple threshold on the basis of GIC current alone … would be difficult to justify as a screening threshold”. If the transformer has been designed or tested to a specified GIC current threshold, couldn’t this be a valid screening threshold? Some sort of simple screening threshold would be of great value since it may be impractical to perform detailed analysis of every transformer.

Referring to transformer manufacture GIC studies. The cost to have the manufacture perform these studies had been between $20 and $30K three years ago. Is there an expectation to have manufacture GIC studies run for all transformers that are applicable to the GMD standard? Additionally the absorption of reactive power will require manufacturer input, this will be a bigger problem to simulate without manufacturer data. You need to know how much GIC will saturate the core at full load. Transformer factory testing or modeling will be required here as well. This is all part of the $20 or $30 per design.

The guide references PSS/E as a tool with GMD analysis capability, so I looked into that to see what capabilities the program might have. I did not find anything specific to GMD, but the guide may have only been referring to PSS/E as possessing the network modeling portion of the GMD analysis.

Based on the information available to us it must be contended that NERC is unjustifiably too optimistic regarding stress/vulnerability assumptions. Moreover the critical issue of price/performance associated to operational procedures OP remains a big question mark which already sentences a questionable methodological weakness. After all we learned from the power system reliability pioneers like Prof. Roy Billington, Prof. Ronald Allan and others that typically you need a lot of sound probability/statistics and key indices, like proness to failure, failure rate, force outage rate, etc; plus methods like LOLP, frequency and duration, Montecarlo, etc, to understand composite reliability/security and what it takes to achieve it; particularly difficult to emulate for HILF (high-impact low-frequency) events as the one at hand. Anyway, in our case the metrics are simply not there and we have not been able to stochastically assess the postulated OP performance; or attendant price for that matter; consequently I believe there is not enough analytical substance/rigor/experience to layout plausible Reliability Standards at this time; these, by definition, spirit and charter, should instead have reflected sound well-conceived/pondered cost-effective utility system and practices on GMD security.

As a footnote on consistency, and worth mentioning, it can be pointed out the insistence from some key Guide-contributing manufacturers, advocates on the prevailing notion of equipment GIC inherent (yet unquantifiable) robustness; nonetheless at the same time such corporate concerns align behind hefty projects on modular portability (mitigating equipment-damage impact) or even making capacitive blocking devices. One is duty-bound, besides being quite relevant to the process, to wonder why they would do that if the apparatus robust-hardware assumption were so totally reliable.

It would seem as though the next reliability standard stage could include challenging somehow the adequacy of mitigation devices. In that regard let me offer some insights into this matter; obviously OP cannot and will not cope with a GMD of any significance; hence for this case cascading tripouts of transformers will be inevitably leading to unit damage and/or blackouts. But the process still allows for more than one subtle consideration: say each autotransformer is, in circuital terms, basically minimal high-to-low impedance (and star to ground resistance); yet upon tripping, those conditions change from their near zero value to infinity (open circuit). In practice it is like a deployment of a blocking device of sorts at those terminals; it will be a kind of capacitor insertion; except for the transformer outage and for such tripping will force, besides the GIC grid redistribution, a very undesired AC-flow one; the latter arguably leading to cascading outages.

In fact, each GIC-caused transformer tripout, as a consequence of OP limitations and protective insufficiency, will practically behave like a very costly/risky blocking device. Conversely, a well-conceived neutral GIC blocking-device strategy could avoid both outages and AC-flow shifts while presenting a similar DC detour of minimal consequence in comparison to OP outages. The bottom line seems to dismiss the concept of OP being a fine device-avoiding alternative but a much inferior performing blocking device instead.

The Application Guide makes numerous references to a “design basis” event, i.e. the maximum expected threat level or solar storm severity—including maximum geoelectric field—that might be expected during a 1-in-100 year storm. For example, on page 14, line 27 the Application Guide states, “Carry out system impact studies assuming the maximum design-basis geoelectric field.” However, the proposed design basis event is not posted in the revision of the Application Guide available on the NERC website as of August 9, 2013, the deadline for comments. Instead, management of the GMD Task Force chose to partially disclose the design basis event in a PowerPoint presentation at the July 25-26, 2013 GMD Task Force meeting in Vancouver, British Columbia, with key elements of the proposed design basis event disclosed only verbally. As a result, the public has been deprived of key details of the design basis event and cannot adequately comment.

The July 2013 Horton presentation at the GMD Task Force meeting disclosed the above slide [Carrington-type simulation example geoelectric fields], which postulates a maximum geoelectric field of approximately 5 volts per kilometer for a “Carrington type” event at observatories within the United States. The simulation and resulting data plots were developed at University of Michigan by undisclosed researchers using an undisclosed methodology.

A critical element of any GMD planning study would be examination of power transformer vulnerability to damage from Geomagneticially Induced Currents (GIC). In fact, Chapter 4 of the Application Guide is titled “Equipment Impact Assessment” and has the subsection “Transformer Impact Screening Process.” Chapter 4 recommends that electric utilities make use of the “NERC Transformer Modeling Guide” to assess impact of GIC on power transformers. Additionally, the “NERC Transformer Modeling Guide” is listed as Reference No. 3 at the end of the Application Guide. Despite its obvious importance, and despite being included as a key reference for the Application Guide, the “NERC Transformer Modeling Guide” is not posted on the GMD Task Force page of the NERC website as of August 9, 2013, the deadline for comments. Instead, management of the GMD Task Force chose to partially disclose elements of proposed transformer modeling in a PowerPoint presentation at the July 25-26, 2013 GMD Task Force meeting in Vancouver, with key elements of transformer modeling and testing in support of models disclosed only verbally. As a result, the public has been deprived of the right to review and comment on this key document and NERC has been denied benefits of independent assessment of its proposed modeling guidelines.

During the presentation, Dr. Horton verbally disclosed that the above storm scaling data in the Pulkkinen analysis was taken over a thirty-year period not in the United States and Canada, but in Scandinavia. There were no severe solar storms during the thirty year observation period. NERC proposes to use this Scandinavian storm scaling data to underpin a design basis event with geoelectric field of approximately 5 volts/kilometer for latitudes within the United States.

The storm scaling data developed by Dr. Pulkkinen and presented by Dr. Horton at the July 25-26, 2013 GMD Task Force meeting in Vancouver is in marked contrast to preliminary storm scaling data presented by Dr. Pulkkinen at the February 25-27, 2013 meeting of the GMD Task Force in Atlanta, as shown in the above slide.

At the February 2013 GMD Task Force meeting, the maximum postulated geoelectric field was 30-40 volts/kilometer. At the July GMD Task Force meeting, the maximum postulated geoelectric field was 20 volts/kilometer.

Dr. Pulkkinen stated at the February 2013 GMD Task Force meeting that the maximum postulated geoelectric field of 30-40 volts/kilometer was "preliminary" and subject to future change. Nonetheless, his downward revision in maximum geoelectric field is a good example of the substantial uncertainty in modeling geoelectric fields. Moreover, a reduction in the maximum postulated geoelectric field from 40 volts/kilometer to 20 volts/kilometer, if not consistent with geomagnetic risks in the United States, could result in reliance on ineffective operating procedures when only hardware protection would protect public safety.

The NERC transformer tests specified above were, or would be, limited to 17-30 amps of injected direct current. Notably, GIC observed at power transformers during past solar storms has been regularly in excess of 30 amps; while electric utilities and the SUNBURST data sharing consortium at EPRI have generally refused to release GIC data to the public, a graph disclosed by EPRI to the NERC GMD/EMP High Impact Low Frequency Report Working Group on March 21, 2010 shows 20 observations over 30 amps and 6 observations over 100 amps from 1990 to 2010. During a severe solar storm, GIC would be expected to be over 1,000 amps, according to Metatech Report R-319 sponsored by the Federal Energy Regulatory Commission.

During his presentation, Mr. Marti stated that technical limitations preclude transformer tests at above 30 amps injected direct current. Mr. Marti also revealed that all of the above specified transformer tests were, or would be, under “no-load” conditions. In contrast, during an actual solar storm, power transformers would be under load.

On one hand, electric utilities and transformer manufacturers claim that power transformers can withstand dozens or even hundreds of amps of GIC. On the other hand, electric utilities refuse to realistically test power transformers operating under load in the commercial electric grid by injecting direct current over 30 amps. If electric utilities and transformer manufacturers are so confident in the withstand capability of transformers to GIC, why do they refuse to engage in realistic testing using load from commercial customers? Could it be that injecting even moderate levels of simulated GIC into power transformers risks transformer failure and cascading blackout for utility customers?

Measured geoelectric fields that have actually been observed in the United States during moderate storms are significantly higher than the proposed NERC design basis event. For example, AT&T measured a geoelectric field of 8 volts/kilometer on August 4, 1972 between Iowa and Illinois when recorded dB/dt was approximately 800 nanoTesla/minute. During a severe solar storm, such as the 1921 Railroad Storm analyzed in the U.S. Government sponsored Metatech R-319 report, dB/dt of 4,800 nanoTesla/minute could be reasonably expected in this region, which would imply a geoelectric field of approximately 50 volts/kilometer for a design basis event—ten times larger than the proposed NERC design basis event of approximately 5 volts/kilometer within the United States.

Electric utilities have generally refused to release GIC data, but the limited data released can be utilized to provide reasonable estimates of GIC levels and by extension geoelectric field intensity during a severe solar storm. For example, the observed GIC of approximately 60 amps during the November 6, 2001 storm at Hurley Ave in New York caused by a 180 nanoTelsa/minute disturbance would imply a geoelectric field of approximately 1.5 volts/kilometer in that region. During a severe solar storm of 4,800 nanoTesla/minute, simple extrapolation would imply a geoelectric field of approximately 40 volts/kilometer and GIC of 1,600 amps.

The proposed NERC design basis event uses measured geomagnetic fields from Scandinavia adjusted to Quebec, Canada grounding conditions, with observed data over a 30 year period. There were no severe solar storms during the 30 year period of observed data. The soil geology of the United States is different from Scandinavia and Canada. The USGS survey of the United States soil geology is still not complete nor has it been validated using published GIC data and so any model for the United States would have substantial uncertainty.

The NERC Application Guide does not propose safety factors, or other safety allowances for modeling uncertainty. Prudent engineering practice would utilize safety factors of at least two, and as much as four, for events with catastrophic consequences. If a safety factor of four were to be applied to a design basis event based on real world measurements—not unproven models—then the solar storm design basis event should have a geoelectric field of 200 volts/kilometer for locations within the United States.

The proposed NERC design basis event with geoelectric field of only 5 volts/kilometer within the United States could be used to claim that power transformers would be unlikely to fail during even severe solar storms. But real world experience shows that transformers do fail during solar storms with relatively small geoelectric fields. For example, a Generator Step Up (GSU) transformer failed at the Salem nuclear plant during the March 1989 solar storm that caused the Hydro-Quebec blackout. This GSU transformer failure occurred shortly after a geoelectric field of 1.7 volts/kilometer at the nearby Fredericksburg Observatory, according to an analysis by the United States Geological Service (USGS). Other GSU transformer failures during even smaller solar storms include failures at the Seabrook plant on November 8-11, 1998; the Braidwood 1 plant on April 5, 1994; and the Maine Yankee plant on April 28, 1991.

The proposed NERC design basis event does not account for sudden commencement solar storms at low latitudes, despite these conditions having been observed in the real world and producing surprisingly large GIC levels at low latitude locations. For published work on this topic, see “Storm sudden commencement events and the associated geomagnetically induced current risks to ground-based systems at low-latitude and mid-latitude locations,” John Kappenman, SPACE WEATHER, VOL. 1, NO. 3, 1016, 2003.

The proposed NERC design basis event relies on the scientifically unsound and outdated assumption that geoelectric field varies with a power curve of the geographic latitude. In fact, the geoelectric field varies with the magnetic latitude, which can be significantly different than the geographic latitude. In fact, soil geology is a significant determinate of geoelectric field, and differing soil geology is not accounted for in the proposed NERC design basis event. In fact, proximity to water bodies is a significant determinate of geoelectric field, and water proximity is not accounted for in the proposed NERC design basis event. In fact, there are multiple peer-reviewed and published studies showing that low latitudes can have significantly higher geoelectric fields during severe solar storms than higher latitudes.

The proposed NERC design basis event is inconsistent with other published models, such as the model in the U.S. Government-sponsored Metatech R-319 study. The Metatech R-319 study physically modeled the United States bulk power grid and established benchmarking and validation of the grid for a number of storms where GIC data was publicly available. To date no such NERC physical model has been demonstrated nor have any efforts been undertaken to validate a model to verify accuracy of proposed storm levels and associated geoelectric fields; and to submit these data to independent and public review before propounding a model upon which to propose reliability standards.

Within the United States and Canada, high capacity High Voltage Direct Current (HVDC) ties run from north to south for hundreds of miles. Examples include the 2,000 MW Phase II tie running from Quebec to Sandy Pond, Massachusetts and the 3,000 MW Pacific Intertie running from Celilo, Oregon to Sylmar, California. Power generated at high latitudes and then exported to lower latitudes could be interrupted by a solar storm. In fact, had the Phase II tie been in operation at the time of the Hydro-Quebec storm of March 1989, up to 2,000 MW of power import would have been interrupted. In fact, the Phase II tie has already been tripped by a small solar storm. It is fallacious to assume a low-latitude terminus of a HVDC tie would not be affected by a more severe solar storm at higher latitude. The Application Guide should specifically prescribe modeling for north-south HVDC ties, and utilize geoelectric fields calibrated to northern latitudes to gauge risk of power disruption at more southerly terminals of HVDC ties.

The Application Guide proposes complicated, iterative, and subjective procedures for electric utilities to establish a geoelectric field “threshold” at which negative equipment and system impacts might occur. (The term “threshold” is per the specific language of the Application Guide.) There is no sound scientific basis for pretending that a “threshold” geoelectric field can be determined with precision. Moreover, there is no sound scientific basis for pretending that impacts of GIC can be modeled with precision. By progressively altering modeling assumptions until the estimated geoelectric field is below the “threshold,” electric utilities might erroneously conclude no GMD protection is necessary and thereby transfer substantial uncompensated risk to utility customers and to the general public with catastrophic consequences for public safety and the economy. The impact study procedures proposed in the Application Guide are susceptible to gaming and are therefore unsuitable for inclusion in a standard-based regulatory process. In fact, the Application Guide reads as an instruction manual for electric utilities to game equipment and system impact studies; page 14 of the Application Guide in the subsection titled “Integration of Equipment Impact and System Impact Studies” (lines 35-37) reads, “If equipment considerations require mitigating measures, reduce the magnitude of the geoelectric field to the point where there are no equipment issues.”

The assessment procedures proposed in the Application Guide are biased toward paper studies and operating procedures that would not require hardware protection against GMD. Moreover, these operating procedures would not exclude GICs from entering high voltage transmission networks and placing other critical grid infrastructure at risk. Alternatively, if electric utilities were to install neutral ground blocking devices that would block all geomagnetically induced currents—rather than relying upon uncertain protection against the GIC magnitude induced by a postulated but unsubstantiated “threshold” geoelectric field—a wide range of threats could be protected against and inherent uncertainty in impact studies would no longer be a concern. Neutral ground blocking devices would protect against solar storms larger than a “Carrington-type” event and even protect against the approximately 40 volt/kilometer geoelectric field produced by a nuclear EMP attack.

The rosy scenario proposed in the NERC Application Guide and its proposed design basis event does not take into account the enormous economic, legal, safety, and strategic consequences of potentially erroneous technical assumptions. If NERC is incorrect in its modeling guidelines—and the weight of both real world observations and published technical studies indicate that NERC is not only wrong, but grievously wrong—then the deaths of millions of Americans could result. With this potential outcome, the NERC Application Guide and reference documents should be revised substantially, or at least subjected to the most strenuous independent scientific review.

Under the proposed NERC design basis event, the magnitude of geoelectric field during a severe solar storm would be much smaller than both previous real-world observations and alternative models would suggest. Moreover, under the still undisclosed NERC Transformer Modeling Guide, the withstand capability of power transformers to GIC could be much higher than observed transformer failures during solar storms would indicate. Unrealistically low assumptions for geoelectric field combined with unrealistically high assumptions of GIC withstand for power transformers could easily result in studies concluding that electric utilities need take no GMD mitigation measures other than operating procedures. In this way, the explicit mandate of FERC under Order 779 for hardware-based GMD protection could be defied and defeated. Consequently, this prospective use of scientifically-invalid NERC modeling guidelines would foreseeably result in indefinite postponement of hardware protection, negation of ratepayer benefits of excluding GICs from the bulk power system, and preclusion of effective protection against both natural and man-made threats to the electric grid.

If it is intended to be a guiding document, a full numerical example of an assessment of a simple system would be very helpful to include in the guide. If it is intended to be an “awareness” document, then it is OK as is.

Section 6 – Other Studies, this section provides little value as a guide.

As described in page 4 of the report, the reasoning of why the assessment of facilities operated under 200-kV is not critical, is not very clear to me. I believe this is an important factor for developing a future NERC reliability standard in terms of systems that might be exempt (e.g., <200-kV) . It would be especially important to those owners and planners for systems that are exclusively operated below 200-kV.

The Reliability (“withstand”) Criteria, as discussed in page 10, does not appear to support an adoption by NERC of a specific key assumptions (or conditions / scenarios) for conducting the assessment. As I understand it, it describes three possible scenarios or as the report states - schools of thought”. I believe these conditions will be key factors in ensuring assessment consistency. Also, it is not clear if the results are expected to vary significantly between the approaches described.

The task force considered several very good papers as listed in the References section. There is a paper I found that provides some guidance on this matter and it is titled “Integration of Geomagnetic Disturbance Modeling into the Power Flow: A methodology for Large-Scale System Studies” that the task force may wish to consider as an additional reference.

GMD Planning Application Guide Comments 11/12/2013Comment Type

Opinion

Comment

Comment

Comment

Editorial

Opinion

Clarification

Clarification

Response Type

Technical

Technical

Technical

Comment

Comment

Technical

Comment

Opinion

Clarification

Comment

Clarification

Clarification

Comment

Comment

Clarification

Clarification

Comment

Clarification

Comment

Clarification

Comment

Comment

Outside Scope

Outside Scope

Outside Scope

Outside Scope

Outside Scope

Outside Scope

Outside Scope

Outside Scope

Comment

Outside Scope

Outside Scope

Outside Scope

Outside Scope

Outside Scope

Comment

Outside Scope

Outside Scope

Outside Scope

Outside Scope

Comment

Comment

Comment

Opinion

Comment

Comment

Clarification

comment

comment

comment

GMD Planning Application Guide Comments 11/12/2013TF Response

Change incorporated to clarify various guides

Commenter is is correct-GIC App guide

Commenter is correct-Transformer guide (forthcoming)

The text has been changed to address the intent of the comment.

Text has been mofied to include the intent of the comment "While they are simple to use, manufacturers maintain that in the near term, such capability curves have to be developed for every transformer design and vintage in the absence of transformer standards defining thermal duty due to GIC".

Using voltage collapse as the threshold for not carrying out detailed system studies can be misleading considering the inherent innacuracies of generic data. For instance, building a dc network model on the basis of load flow data can cause errors in the order of 35% or higher in the transformer models.

The term "planning authority" probably has unintended consequences. Text has been modified as follows: "Perform power flow analysis using system load levels and stressed system conditions using generic models for transformer reactive power absorption. Loss of reactive power sources such as shunt capacitor banks and SVCs (on protection) should be considered as valid contingencies associated with the GMD event"

The selection of the design-basis event for GMD studies will come from the regulatory front. All four documents state potential criteria but do not recommend what the design basis ought to be. The 3% fluctuation due to transformer var loss alone is an indication of system susceptibility (or lack thereof) with sufficient margin to account for data innacuracies of the less detailed screening study and a trigger to sharpen the pencils.

Load flow studies that take into consideration GIC are time independent. There is no technical contradiction in the use of emergency limits to determine system susceptibility to the peak of a storm

An event that only causes a voltage dip of 3% is unlikely to cause issues due to harmonics.

The application guide addresses this issue.

At this point in time it is preferable to highlight that HV harmonics can be an issue and allow the planners the discretion of how to deal with them.

A surge arrester grounded transformer is a well-defined configuration in the industry. Reference [2] provides guidance as what the numeric value of the resistance is as well as the issues regarding numerical stability in the solution of a dc netwrk that contains very high and very low resistive values.

modified as proposed

Paragraph modified accordingly

This is described in Chapter 3 under Sytem Imact Assesment studies.

This paragraph is meant to highlight "telltale" signs of vulnerability. There are no reports of generator damage directly linked to GIC.

Dissolved gasses cannot be predicted in a screening study at this point in time. GIC flow means different things in different transformer types. In a steady-state analysis the time factor of the GIC wavesape is not taken into consideraton

Changed to "Under such operating conditions voltage collapse can occur when the system does not have enough var resources to support current operating conditions or to recover from a valid contingency (e.g., a fault)."

The paragraph addresses the modelling of transformers within a station. Added a reference to the Application Guide, where what needs to be modelled is discussed in more detail.

A single GIC thershold would be desirable but not possible since it would not take into consideration duration or waveshape

Definition of the peak value of the reference waveshape has been intentionally been left open since it wil be defined by the "benchmark" event of the new NERC GMD standard. The actual waveshape can be obtained by scaling a GMD event such as the March 1989 storm or a synthetic one such as the one proposed by Pulkkinen, or a pre-defined one that could come from a future transformer GIC withstand standard. It would not be prudent at this point in time to make a specific recommendation.

PSS/E offers a GIC calculation module

Not practical to include a hand calculated numerical example. Guide is a reference document describing types of studies and method for assessing GMD vulnerability

A supporting reference is provided on page 8.

The task force is aware.

The guide as written provides a useful explanation of approaches that are available to the planner. Guidance to assess to a specific design basis is more limited; if a Reliability Standard prescribes such a design basis the guide will still be relevant in aiding the planner in conducting the required study.

The intent of the section is to provide an overview of more advanced analysis. Tools and methods are not mature enough to provide more explicit guidance