Sensitivity of DC Network Representation for GIC Analysis

Adam Mate; Aniruddh Mishra; Arthur K. Barnes; Jose E. Tabarez

REVIEW 4 major objections 5 minor 19 references

Sensitivity of DC Network Representation for GIC Analysis

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This paper shows that, in a uniform electric field, 100% GIC blocker placement always drives transformer reactive power loss to zero, while in non-uniform fields only series blocking devices guarantee zero loss.

desk verdict The uniform-field zero-loss theorem is clean and worth knowing, but the paper's secondary claim about substation vs neutral blockers is a single small case study with unquantified confounds. read the letter →

arxiv 2505.23016 v1 pith:QZEFPRSA submitted 2025-05-29 eess.SY cs.SY

Aniruddh Mishra , Arthur K. Barnes , Jose E. Tabarez , Adam Mate This is my paper

classification eess.SYcs.SY

keywords geomagneticallyinducedcurrentneutralblockerssubstationtransformerwindingconfigurationsdcnetworkrepresentationuniformelectricfieldreactivepowerlossseriescapacitorblocking

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Geomagnetic disturbances push quasi-dc currents through transformer neutrals, and planners decide where to place blocking devices using a dc model of the transmission network. This paper asks how much the modeling choices, especially the representation of the blocker itself, change the predicted reactive-power loss. It argues that in a uniform electric field, every kind of 100% blocker placement drives the loss to zero because a uniform field cannot drive current around closed loops. In a non-uniform field that safety net disappears: neutral and substation blockers leave losses, and only series-capacitor blocking on every line guarantees zero loss. The practical point is that the common uniform-field assumption can conceal vulnerabilities that a realistic non-uniform storm would expose.

What carries the argument

The central object is the dc equivalent network built from transformer winding configurations and line resistances, with GMD branches representing every path to ground. The argument's load-bearing mathematical fact is that a uniform electric field is conservative: $\oint \vec E \cdot d\ell = \sum V_k = 0$ around any closed loop, so mesh-current analysis gives $\vec I = 0$ in a fully blocked network. That identity turns the numerical blocker comparison into a structural statement about the graph of the dc network.

What would settle it

Re-run the non-uniform-field experiment with identical solver settings and generator step-up transformer treatment in both implementations, and with the artificial 25 kOhm grounding branches removed; if the gap between substation-blocking and neutral-blocking $Q_{\mathrm{loss}}$ shrinks to zero, the 'minor error' conclusion is an artifact of the implementation differences rather than of blocker representation.

Watch

Extended reading notes

Core claim

The paper's central claim is that the sensitivity of GIC power loss to blocker representation is controlled by whether the induced electric field can drive circulating current around closed loops of the dc network. In a uniform, conservative field, the line integral around any loop is zero, so a fully blocked network has no current and therefore $Q_{\mathrm{loss}}=0$ for every transformer, independent of blocker type. In a non-uniform field, loops such as autotransformer series windings acquire a net induced voltage, so neutral-point blocking still permits current through the series branch, and substation-ground blocking, which replaces several neutral blockers with one ground-point blocker, leaves additional meshes that carry current and overstate the loss. The paper concludes that the substation-ground approximation of a neutral blocker is reasonable, introducing only minor error in non-uniform fields, and that series blocking is the only modeled option guaranteed to drive power loss to zero once the field is non-uniform.

Load-bearing premise

The numerical comparison that backs the 'substation blocking is a reasonable approximation' conclusion assumes the two software setups differ only in blocker placement; in fact they also differ in solver implementation, in artificial 25 kOhm grounding branches, and in how generator step-up transformers are modeled, and the paper does not measure how much those differences contribute.

Editorial extensions

If this is right

With a uniform-field assumption, comparisons of blocker types are insensitive: any 100% placement strategy yields the same zero-loss result, so uniform-field studies cannot rank mitigation options.
In non-uniform fields, substation-level blocking will consistently report higher $Q_{\mathrm{loss}}$ than neutral-point blocking, because the substation approximation leaves extra meshes in the dc graph.
Series-capacitor blocking on every line is the only modeled scheme that guarantees zero reactive-power loss under non-uniform fields, at the cost of added ac reactance and phase shifts.
Using uniform electric fields in planning studies can hide vulnerabilities that non-uniform fields reveal, so conclusions about system safety should be checked against non-uniform field cases.
The finding supports using a substation-ground blocker as a computationally cheaper surrogate for neutral blocking in optimization, with the correction that it overestimates loss.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A natural follow-up is to solve the blocker-placement optimization with explicit neutral blockers and compare total blocker counts against the substation-blocking surrogate; if the surrogate consistently overestimates loss, it may demand more blockers than necessary.
Because the paper's comparison also embeds differences in solver, artificial grounding branches, and generator step-up transformer treatment, a cleaner experiment would isolate the blocker-representation effect by holding those details fixed.
The conservative-field argument is graph-theoretic and does not depend on planarity, so the uniform-field zero-loss conclusion should extend to larger realistic grids; a testable extension is to check on a large synthetic network whether the substation-vs-neutral gap scales with the number of autotransformer meshes.
The paper's suggestion to couple the dc GIC model back to ac power flow could be tested against historical storm events where harmonic-induced control trips occurred, using the modeled $Q_{\mathrm{loss}}$ as the driver.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: this paper has one solid, clean result—under a uniform electric field, 100% blocker coverage forces Qloss to zero regardless of blocker representation, because a conservative field induces no net loop voltage. That derivation is simple but correct, and it matters because a lot of GIC screening still assumes uniform fields. The paper is the first to my knowledge to compare blocker representation choices directly, so this fills a small gap.

Where it's good: the circuit-theoretic argument in Sec. V-C and Eq. (6) is sound and graph-independent. The authors are also appropriately transparent about what they did: PMsGMD doesn't model neutral blockers exactly, so they used PowerWorld for that case; they tell you the 25 kOhm branches and the bypass resistances are there to keep the matrix solver well-conditioned. That honesty is a real credit.

Soft spots: the 'substation blocking is a reasonable approximation with minor error' conclusion in Sec. VIII rests entirely on Fig. 7, a single synthetic grid with two different software implementations. The comparison is confounded—neutral blockers were solved in PowerWorld, substation blockers in PMsGMD; the PowerWorld model adds 25 kOhm grounding branches between every bus and substation ground; and implicit GSU treatment may differ. Any of those could shift the Qloss curves by more than the observed gap. The paper does not ablate these factors or provide an error metric. So the qualitative direction—substation blocking overestimates loss because it leaves meshes closed—is plausible, but the 'minor error' claim is not established by the evidence shown. The uniform-field result stands on its own; the non-uniform part should be treated as a motivating case study, not a verified approximation.

Citation pattern: the literature is properly cited, including [11] on sensitivity to physical parameters and the PowerModelsGMD and PowerWorld docs. No red flags.

Bottom line: a modest but honest paper. It deserves a serious referee; the uniform-field result is durable, and the empirical comparison could be strengthened with ablations or by repeating the neutral-blocker case in PMsGMD once modeled. A revision with those additions would make it solid. I'd cite it if I were working on GIC blocker placement.

Recommendation: send to peer review with a request to address the confound in the non-uniform comparison.

Referee Report

4 major / 5 minor

Summary. The manuscript examines how different dc-network modeling choices for geomagnetically induced current (GIC) studies affect the computed reactive power loss (Qloss). It derives a matrix-based argument (Eqs. 6–7) that, under a uniform electric field, a fully blocked network has zero circulating current and therefore zero Qloss, and it supports this with a small synthetic network. For non-uniform fields, the paper compares four blocker configurations—no blockers, substation blocking, neutral blocking, and series-capacitor blocking—using PMsGMD and PowerWorld. It concludes that in non-uniform fields the substation-ground approximation of a neutral blocker introduces only minor error, that substation blocking overestimates Qloss because of additional meshes, and that series capacitor blocking is needed to drive Qloss to zero. The uniform-field theorem is cleanly derived; the non-uniform-field conclusions rest on a single empirical comparison that involves several implementation differences beyond the blocker representation.

Significance. The uniform-field no-circulation theorem is a clean, reusable analytical result: if the electric field is uniform, every loop in the dc network has zero induced voltage, so after all neutral/substation ground paths are blocked, no circulating current remains regardless of blocker type. This distinction between conservative uniform-field assumptions and non-uniform field conditions is practically important for blocker-placement studies, and the paper is upfront about the limitations of the uniform-field approximation. The case study is openly available, which supports reproducibility. The empirical claim that the substation approximation is reasonable and introduces only minor error, however, is not yet established because the comparison is confounded by solver choice, artificial grounding branches, and implicit-GSU treatment; the paper would require a controlled ablation to support that applied conclusion. With such an ablation, the manuscript would be a useful methodology contribution to GIC modeling and mitigation.

major comments (4)

[VIII and Fig. 7] The conclusion that 'the substation approximation of a transformer neutral blocker is ... reasonable, introducing only minor error' rests entirely on Fig. 7, which compares PMsGMD with substation blocking against PowerWorld with neutral blocking. These two implementations differ not only in blocker placement but also in the solver, in the artificial 25 kOhm grounding branches (Section IV-B), and in the treatment of implicit generator step-up transformers (Section IV-A). No ablation or quantitative error metric is provided to separate these factors, so the gap in Fig. 7 cannot be attributed to the blocker representation alone. Provide the same comparison with a single solver and with the 25 kOhm branches and implicit GSUs toggled on and off, and report a quantitative error metric such as maximum or mean absolute Qloss difference.
[VI-B and Eq. (2)] For the non-uniform-field experiment, PowerWorld generated the line voltages that were then input to PMsGMD, but the neutral-blocker Qloss values are computed by PowerWorld. The manuscript does not state whether both software packages use the same K-factors, base currents, and transformer tables in Eq. (2). If the Qloss formulas or the K-factor assignments differ, the difference between the curves in Fig. 7 is at least partly a post-processing artifact, not a blocker-modeling effect. Specify the exact Qloss computation used by each implementation and, if possible, compute both blocker types with the same Qloss post-processing routine.
[IV-B] The manuscript states that implicit 25 kOhm grounding branches are placed between every bus and its substation and says 'PowerWorld Simulator likely adds this branch to help their matrix solver.' It is not clear whether PMsGMD also includes these branches for the substation-blocking and series-capacitor runs. If these leakage branches are present in one network but not the other, they can alter the current distribution and hence Qloss even when neutral blockers are open. Clarify which implementation includes them, whether the same branches are present in both networks used for Fig. 7, and report the sensitivity of the result to the branch value.
[VIII] The concluding statement that 'in non-uniform electric fields ... series blocking devices are needed to drive power loss to zero' is presented as a general conclusion but is demonstrated only on one synthetic network with one non-uniform-field pattern. There may be non-uniform fields for which 100% neutral blocking also yields zero Qloss (for example, if the field pattern produces zero net voltage across autotransformer loops), or for which a different blocker configuration outperforms series blocking. Qualify this claim to the case study or provide a more general argument beyond the single tested field.

minor comments (5)

[III-B, Eq. (3a)] The formula for the latitude midpoint is written as 'phi = y1 + y2/2'; this should be (y1 + y2)/2 to avoid ambiguity, and the units of the coefficients in Eqs. (4a)-(4b) should be stated explicitly.
[IV-A] The threshold for adding an implicit generator step-up transformer is 'a bus that has a nominal voltage >= 30 kV,' but the justification for this threshold and for the resistance values in Table II is not given beyond the citation; add a sentence explaining the basis for the 30 kV value.
[VI-A] The case study description does not report the total number of buses, lines, and transformers in the synthetic network; adding these counts would make the experiment easier to reproduce and to compare with other studies.
[II-B3, Eq. (2)] The K-factor is cited but not defined; a sentence defining K and its units, and explaining how the effective GIC current is computed from dc branch currents, would make the Qloss calculation self-contained.
[Figs. 6 and 7] The plots show the no-blocker case alongside blocker cases, but the text only describes the blocker cases in the uniform-field result; clarify whether no-blocker Qloss is nonzero and, if so, how that is consistent with the uniform-field theorem.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the uniform-field zero-loss result follows from Eq. (6), and the blocker comparisons are external/empirical.

full rationale

This paper derives no result by assuming its own conclusion. The uniform-field zero-loss claim is obtained in Eqs. (6)-(7): a conservative field has zero line integral around every loop, so mesh currents vanish, and with 100% blocker coverage all ground paths are opened; this is a direct matrix consequence, not a restatement of the blocker model. The Qloss and dc-network equations (1)-(5) are standard and cited from prior literature, and they are not fitted to the reported results. The special resistance values (25 kOhm implicit branches, 5 mOhm bypass, 1 uOhm/1 MOhm bridge resistors) are numerical regularizations or conveniences chosen to stabilize the matrix or enforce the intended equivalent circuits; none is defined in terms of the target Qloss values, so no fitted-input-called-prediction pattern appears. Self-citations (e.g., [10], [12], [3]) point to the software and review articles used as tools or background; the central comparison in Fig. 7 is run against PowerWorld, an independent external implementation, so the 'reasonable approximation' conclusion is supported by external evidence rather than by self-citation. Potential confounds between the two implementations (solver, 25 kOhm branches, implicit GSU treatment) concern internal validity of the empirical comparison, not circularity of the derivation. Accordingly, no circular step is identified.

Assumptions & free parameters 4 free parameters · 6 assumptions · 2 invented entities

The central claims do not require new physics, but they rely on several assumed modeling artifacts. The most consequential is the artificial 25 kOhm grounding branch, which is acknowledged but uncalibrated, and the implicit GSU resistances, which are taken from a synthetic-grid study. The Qloss model and dc-network construction are standard from the cited GIC literature.

free parameters (4)

Implicit substation grounding branch resistance = 25 kOhm
Added between every bus and its substation ground to avoid conductance-matrix singularity; the paper acknowledges it is an artificial branch and does not test sensitivity to its value.
Bypassed series capacitor resistance = 5 mOhm
Used to represent a bypassed series capacitor without an infinite conductance entry; a hand-chosen numerical floor that affects the dc network.
Three-winding autotransformer bridge resistances = 1 uOhm (series), 1 MOhm (common)
Chosen to force the low-side bus voltage onto the star bus and to make the common-winding resistance dominate; arbitrary values not derived from measurements.
Implicit GSU winding resistance = Table II, e.g., 3.623 uOhm at 230 kV
Assumed from a prior synthetic grid study for every generator bus at or above 30 kV, even when the ac model has no such transformer; affects the distribution of losses.

assumptions (6)

standard math Line integral of a uniform electric field around any closed loop is zero (Eq. 6).
The paper proves zero circulating current by applying this conservative-field property; it is a standard vector-calculus result.
standard math In a fully determined linear resistive network, zero voltage sources on all loop equations imply zero loop currents (Eq. 7).
Used to conclude I=0 for uniform fields; assumes the conductance matrix is non-singular, which the paper tries to ensure with artificial grounding branches.
domain assumption GICs can be represented by a dc network formed from positive-sequence ac data, with transformer windings to ground modeled as resistances equal to one-third of winding impedance.
This is the standard GIC modeling assumption cited from [1], [11]; the paper's entire analysis rests on it.
domain assumption Transformer reactive power loss is Qloss = K * I_gic * |V_eta| * I_base (Eq. 2).
Taken from literature [8], [10], [13]; all blocker comparisons are made in terms of this quantity.
ad hoc to paper Artificial 25 kOhm grounding branches do not materially change the current solution.
The paper adds these branches to prevent singularity but never validates that they leave Qloss unchanged; this is an unexamined modeling assumption.
domain assumption Implicit GSU transformers with Table II resistances are valid for generator buses not present in the ac model.
Borrowed from [16]; the paper uses these assumed transformers to create dc ground paths and they affect other transformers' Qloss.

invented entities (2)

Implicit substation grounding branch
purpose: Provides a fictitious path to ground for every bus to avoid a singular conductance matrix.
The paper notes PowerWorld likely adds this branch and does not provide any physical justification or independent validation; it sits in series with blocker placements and could bias the non-uniform-field comparison.
Implicit generator step-up transformer
purpose: Creates a dc ground path for generator buses whose ac models contain no explicit GSU.
An assumed component from prior literature (Table II), not measured; the paper states it is only added to the dc network and is not in the Qloss table, so its effect on other transformers is model-dependent.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Sensitivity of DC Network Representation for GIC Analysis." pith.science (2026). https://pith.science/paper/QZEFPRSA

@misc{pith2026250523016,
  author       = {Pith},
  title        = {Pith review of: Sensitivity of DC Network Representation for GIC Analysis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QZEFPRSA}},
  note         = {Machine review of arXiv:2505.23016}
}

read the original abstract

Geomagnetic disturbances are a threat to the reliability and security of our national critical energy infrastructures. These events specifically result in geomagnetically induced currents, which can cause damage to transformers due to magnetic saturation. In order to mitigate these effects, blocker devices must be placed in optimal locations. Finding this placement requires a dc representation of the ac transmission lines, which this paper discusses. Different decisions in this process, including the method of representing the blocking devices, result in significant variations to the power loss calculations. To analyze these effects, we conclude the paper by comparing the losses on a sample network with different modeling implementations.

Figures

Figures reproduced from arXiv: 2505.23016 by the authors.

**Figure 1.** DC equivalent circuit of a Yg-Yg transformer [12] with the ∆ windings. However, the grounded-wye (Yg) side is represented the same way as the previous model. This is similar to [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗

**Figure 2.** DC equivalent circuit of a Yg-Yg autotransformer [12] Three-Winding Transformers: The primary difference between a three-winding transformer and the previously discussed transformers is the addition of a tertiary bus. In order to represent the three-winding transformer, the ac model contains three separate two winding transformers. The connecting node of these transformers is referred to as the “star” bus. The win… view at source ↗

**Figure 3.** DC equivalent circuit of a three-winding transformer [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

Figures from the paper (4 more)

**Figure 4.** Figure 4: DC equivalent circuit of a Yg-a-∆ transformer The value of aSt(k) is 1 µΩ in order to pull the voltage value of l(k) onto σ(k). In addition, the value of aCf(k) is set to 1 MΩ so that the equivalent resistance from the σ(k) to g(k) is only that of the common winding of…

**Figure 5.** Figure 5: GIC blocker modeling test case This circuit was specifically designed to demonstrate the differences between the different blocker placements. There are two main meshes here, which are affected differently by where the blocking device is located. The loop with only aut…

**Figure 6.** Figure 6: Comparison of qloss with different blocker placements in a uniform electric field 3https://github.com/bluejuniper/gmd-tools [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]

**Figure 7.** Figure 7: Comparison of qloss with different blocker placements in a nonuniform electric field [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

19 extracted references · 19 canonical work pages

[11]

Zheng et al

K. Zheng et al. Effects of system characteristics on geomagnetically induced currents. IEEE Trans. Power Del. , 29(2):890–898, Apr. 2014

work page 2014
[1]

D. H. Boteler et al. Modeling geomagnetically induced currents. Space Weather, 15(1):258–276, Dec. 2016

work page 2016
[2]

T. R. Hutchins et al. The effect of geomagnetic disturbances on the electric grid and appropriate mitigation strategies. In IEEE/PES North American Power Symposium (NAPS) , pages 1–5, August 2011

work page 2011
[3]

Mate et al

A. Mate et al. Relaxation based modeling of GMD induced cascading failures in PowerModelsGMD.jl. In IEEE/PES North American Power Symposium (NAPS), pages 1–7, Nov. 2021

work page 2021
[4]

Haddadi, R

A. Haddadi, R. Hassani, J. Mahseredjian, G. Luc, and A. Rezaei- Zare. Evaluation of simulation methods for analysis of geomagnetic disturbance system impacts. IEEE Transactions on Power Delivery , pages 1–1, 2020

work page 2020
[5]

D. H. Boteler, R. Pirjola, C. Blais, and A. Foss. Development of a GIC simulator. In 2014 IEEE PES General Meeting | Conference & Exposition, pages 1–5. IEEE, 2014

work page 2014
[6]

Zhu et al

H. Zhu et al. Blocking device placement for mitigating the effects of geomagnetically induced currents. IEEE Transactions on Power Systems, 30(4):2081–2089, July 2015

work page 2015
[7]

Characteristics of transformer exciting- current during geomagnetic disturbances

RA Walling and AN Khan. Characteristics of transformer exciting- current during geomagnetic disturbances. IEEE Transactions on Power Delivery, 6(4):1707–1714, 1991

work page 1991

Show all 19 references

[8]

Integration of geomagnetic disturbance modeling into the power flow: A methodology for large-scale system studies

Thomas J Overbye, Trevor R Hutchins, Komal Shetye, Jamie Weber, and Scott Dahman. Integration of geomagnetic disturbance modeling into the power flow: A methodology for large-scale system studies. In IEEE/PES North American Power Symposium (NAPS) , pages 1–7. IEEE, 2012

2012
[9]

J. G. Kappenman et al. GIC mitigation: A neutral blocking/bypass device to prevent the flow of GIC in power Ssystems. IEEE Transactions on Power Delivery, 6(3):1271–1281, July 1991

1991
[10]

A. K. Barnes et al. A review of the GIC blocker placement problem. In IEEE/IAS Industrial and Commercial Power Systems Technical Con- ference, pages 1–6, May 2024

2024
[12]

Mate et al

A. Mate et al. Analyzing and mitigating the impacts of GMD and EMP events on the electrical grid with PowerModelsGMD.jl, Jan. 2021

2021
[13]

T. R. Hutchins et al. Power System Dynamic Performance During the Late-Time (E3) High-Altitude Electromagnetic Pulse. In 2016 Power Systems Computation Conference (PSCC) , pages 1–6, June 2016

2016
[14]

Modeling GMD in PowerWorld Simulator , 2020

PowerWorld Corporation. Modeling GMD in PowerWorld Simulator , 2020

2020
[15]

Computing geomagnetically-induced current in the bulk-power system

Application Guide. Computing geomagnetically-induced current in the bulk-power system. NERC North American Electric Reliability Corporation, 2013

2013
[16]

Statistical considerations in the creation of realistic synthetic power grids for geomagnetic disturbance studies

Adam B Birchfield, Kathleen M Gegner, Ti Xu, Komal S Shetye, and Thomas J Overbye. Statistical considerations in the creation of realistic synthetic power grids for geomagnetic disturbance studies. IEEE Transactions on Power Systems , 32(2):1502–1510, 2016

2016
[17]

Lorrain et al

P. Lorrain et al. Electromagnetic Fields and Waves. W. H. Freeman and Company, San Francisco, 1970

1970
[18]

Currents produced in earthed conductor networks by geomagnetically-induced electric fields

M Lehtinen. Currents produced in earthed conductor networks by geomagnetically-induced electric fields. Ann. Geophys. , 3(4):479–484, 1985

1985
[19]

A 21st century view of the march 1989 magnetic storm

David H Boteler. A 21st century view of the march 1989 magnetic storm. Space Weather, 17(10):1427–1441, 2019

1989

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [11]

Zheng et al

K. Zheng et al. Effects of system characteristics on geomagnetically induced currents. IEEE Trans. Power Del. , 29(2):890–898, Apr. 2014

work page 2014

[2] [1]

D. H. Boteler et al. Modeling geomagnetically induced currents. Space Weather, 15(1):258–276, Dec. 2016

work page 2016

[3] [2]

T. R. Hutchins et al. The effect of geomagnetic disturbances on the electric grid and appropriate mitigation strategies. In IEEE/PES North American Power Symposium (NAPS) , pages 1–5, August 2011

work page 2011

[4] [3]

Mate et al

A. Mate et al. Relaxation based modeling of GMD induced cascading failures in PowerModelsGMD.jl. In IEEE/PES North American Power Symposium (NAPS), pages 1–7, Nov. 2021

work page 2021

[5] [4]

Haddadi, R

A. Haddadi, R. Hassani, J. Mahseredjian, G. Luc, and A. Rezaei- Zare. Evaluation of simulation methods for analysis of geomagnetic disturbance system impacts. IEEE Transactions on Power Delivery , pages 1–1, 2020

work page 2020

[6] [5]

D. H. Boteler, R. Pirjola, C. Blais, and A. Foss. Development of a GIC simulator. In 2014 IEEE PES General Meeting | Conference & Exposition, pages 1–5. IEEE, 2014

work page 2014

[7] [6]

Zhu et al

H. Zhu et al. Blocking device placement for mitigating the effects of geomagnetically induced currents. IEEE Transactions on Power Systems, 30(4):2081–2089, July 2015

work page 2015

[8] [7]

Characteristics of transformer exciting- current during geomagnetic disturbances

RA Walling and AN Khan. Characteristics of transformer exciting- current during geomagnetic disturbances. IEEE Transactions on Power Delivery, 6(4):1707–1714, 1991

work page 1991