{"id":"975d59cf-f970-4fd5-aeb1-f462792bb371","arxiv_id":"1908.06330","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A homogenized T-A simulation model reproduces screening-current field effects and field drift in an HTS coil and proposes a 20% current overshoot to suppress drift.","lead":"This paper shows that a fast numerical model, the homogenized T-A formulation, can simulate screening currents and magnetic-field drift in high-temperature superconducting coils in about an hour. A magnet designer could use it to explore how superconductor quality, tape striation, and current overshoot affect field stability.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Homogenized T-A model is unbenchmarked for this 800-turn coil; the predicted SCIF magnitude and the 20% overshoot claim rest on an approximation that may alter current-front penetration.","rationale":"The reader's verdict is CONDITIONAL, and my stress-test supports that conditionality. The central scientific claims — the magnitude of SCIF, the drift dependence on Jc and n, and the specific 20% overshoot prescription — all depend on the homogeneous T-A model faithfully representing the current-front dynamics of an 800-turn insulated coil. The homogenization in Section 2 is a deliberate approximation, but its quantitative impact is not demonstrated in the paper. In an insulated pancake coil, the per-turn transport current is fixed by series connection; the homogeneous bulk enforces only the total current, potentially allowing unphysical current sharing. Additionally, neglecting the parallel-field term in Eq. (6) removes part of the field-angle dependence that controls the local critical current. Without a resolved-tape or experimental benchmark, the numbers reported in Sections 4 and 5, and the overshoot recommendation in Section 6, are not established beyond the level of a plausible demonstration. This is exactly the kind of missing verification that a conditional verdict should require. I do not see an internal inconsistency or a flaw that would force a rejection; the method is coherent and the performance claim is reasonable, but the quantitative predictions need external support. The reader identified the homogenization and the lack of benchmark as the weakness; I agree partially, sharpening the concern to the per-turn current constraint and the need for a reference simulation.","tokens_in":8609,"tokens_out":15461,"duration_ms":162695,"concrete_test":"Run a reference simulation of the same 10-pancake coil with all 800 turns explicitly resolved (individual T-A lines or H-formulation) for at least two cases: the base ramp and the 20% overshoot ramp. Compare the central-field time traces B_sim(t) over the full 30 h plateau and the SCIF hysteresis loop over one 6 h cycle. If the homogeneous-model curves deviate from the resolved-model by more than about 5% of B_n, or if the drift after the 20% overshoot reappears within 30 h in the resolved model, then the paper's quantitative conclusions are not validated. Also check time-step convergence in the homogeneous model during the plateau to confirm the 20% overshoot drift elimination is physical rather than numerical.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing point for the paper's central claim is the validity of the homogenization in Section 2. The model replaces each 80-turn pancake with a continuous bulk (spaces between layers disappear) and drops the parallel-field term in the T equation (Eq. 6). In an insulated coil, the 80 turns are electrically independent and must each carry the same transport current; the homogeneous bulk enforces only total current and so permits unphysical radial redistribution of transport current. It also averages the tape-level field-angle dependence that controls Jc. No comparison against a resolved-tape model (all 800 turns) or against measured SCIF/drift is provided for this coil. Consequently, the quantitative outputs — the 4.1% SCIF in Section 4, the drift rates in Section 5, and the 20% overshoot recommendation in Section 6 — are unverified. If the homogenization changes current-front penetration timing, the SCIF magnitude and the drift-versus-overshoot relationship would change.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript presents a homogeneous T-A formulation for modeling HTS solenoid coils, in which stacks of HTS tapes are replaced by continuous bulk regions. The model is applied to a 10-pancake coil with 80 turns per pancake to study the screening current-induced field (SCIF) and central field drift as functions of the critical current density and the power-law index n. The paper also numerically tests two mitigation strategies: tape striation and current sweep reversal with overshoot. The main reported results include a 4.1% SCIF at peak current for the reference parameters, a maximum drift of 2.2%, and the claim that a 20% current overshoot eliminates the field drift. The authors further report that simulations take about one hour on a desktop computer and claim that this enables real-time simulation of slow ramping cycles of large-scale systems.","tokens_in":8826,"tokens_out":4801,"duration_ms":49619,"significance":"If the homogenized T-A model is quantitatively accurate, its computational speed advantage would be significant and would make systematic studies of large-scale HTS magnets feasible. The parametric investigation of Jc and n on SCIF and drift, and the numerical exploration of striation and overshoot, address questions of practical interest to magnet designers. However, the central quantitative claims rest on an unvalidated homogenization approximation and on simulation-only predictions; no comparison against a resolved-tape model, an H-formulation benchmark, or experimental data is provided for this coil. The significance is therefore conditional on an additional validation step.","major_comments":[{"comment":"The homogeneous-bulk approximation is presented without a benchmark against a resolved-turn model or against experimental data for this 800-turn coil. Replacing each 80-turn pancake by a continuous bulk removes the constraint that every insulated turn must carry the same transport current; only the total pancake current is imposed through Eq. (7). This can alter the radial distribution of transport current and the penetration of current fronts, which directly affects all quantitative outputs: the 4.1% SCIF in Section 4, the drift rates in Section 5, and the 20% overshoot recommendation in Section 6. The paper should add at least one validation case, for example comparing a single pancake against a fully resolved T-A or H-formulation model, or comparing the predicted SCIF against measured data for a similar coil.","section":"Section 2"},{"comment":"The T equation in Eq. (6) uses only the radial field component, ∂B_r/∂t, while the critical current density in Eq. (9) depends on both B_r and B_z. The manuscript acknowledges that the parallel-field contribution is neglected, but it does not quantify the resulting error for the analyzed geometry. In a solenoid, the field angle varies significantly with position, especially near the inner and outer radii, and the omitted term can change the local resistivity and, in turn, the dynamics of current front penetration. Please provide a quantitative estimate of the error (for example, by including the B_z term in a test case) or a physical justification for why this term is negligible here.","section":"Section 2, Eq. (6)"},{"comment":"The field-drift and overshoot results are simulation-only; no comparison to experimental measurements or to an alternative numerical formulation is provided. The conclusion that a 20% overshoot 'eliminates' the field drift is based on the homogenized model within a 30-hour window, and the reappearance of drift after 5.5 hours for the 10% overshoot case indicates that the outcome is sensitive to the time horizon and to the model assumptions. The statement in Section 7 that H-formulation simulations 'would have required computation times in the order of months' is also unsupported by any runtime data for this problem. The authors should either provide a benchmark comparison or substantially weaken the claims, for example by clearly labeling the results as model predictions for the specific simulated conditions.","section":"Sections 5 and 6"}],"minor_comments":[{"comment":"The spelling 'homogenous' is used in the title while the text uses 'homogeneous'; please unify to a single spelling.","section":"Title and throughout"},{"comment":"The sentence 'These ramps are shown in figure' lacks a figure number; it should refer to Figure 10.","section":"Section 6"},{"comment":"The notation Jc0n is used without an explicit definition; please define it as the modified reference critical current density used in place of Jc0 in Eq. (9).","section":"Section 4"},{"comment":"The relationship between the physical tape thickness and the thickness of the homogeneous bulk is not stated; please explain how the bulk dimensions and the HTS layer thickness δ (Table 2, Eq. (7)) are related.","section":"Tables 1 and 2"},{"comment":"The 'real-time' claim in the abstract and conclusions would benefit from clarification: a 1-hour computation time for a 6-hour simulated cycle is faster than the cycle duration, but no extrapolation to 'large-scale systems' is demonstrated; please state the precise meaning of real-time used here.","section":"Section 7"},{"comment":"The sentence describing the case {Jc0, n=10} says it exhibits 'the larger drift' (2.2%) and then mentions two cases with 'the same 1.5% drift'; the wording is confusing and should be revised for clarity.","section":"Section 5"}],"recommendation":"major_revision","confidential_remarks":"The paper is likely within the journal's scope, but the absence of any validation against a resolved-turn model, an H-formulation benchmark, or experimental data is a serious concern for the quantitative claims. The homogenization is based on the authors' earlier work, and the present manuscript does not add a new independent verification. I would recommend that the editor require a concrete validation case before acceptance. The parametric study itself is useful and the computational speed is attractive, so a major revision with a benchmark and softened claims could make the paper acceptable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Ed, quick take.\n\nThis paper applies the authors' own homogeneous T-A method to a 10-pancake, 800-turn solenoidal coil and runs a surprising number of parametric sweeps in about an hour of desktop compute per case. That is the real thing. The novelty is not the method—it is their earlier work, and the Jc dependence was already known (they cite it)—but the demonstration that this coil can be simulated dozens of times in a practical time, plus the systematic look at n-index, striation, and sweep reversal. The finding that the overshoot needed to kill drift is coil-specific, rather than the 1% from earlier references, is a useful practical warning.\n\nWhat is well done: they report mesh and parameters, use a standard E-J power law and anisotropic Jc(B), and are honest that the study is qualitative. The striation results follow the expected trend. The real-time claim is reasonable in the narrow sense: compute time (1h) is less than simulated time (6h or 30h), so a slow ramp could be followed in real time on a desktop. That is not a demonstration of a real-time control loop, but it's a fair statement.\n\nThe soft spot is the one the stress-test flags, and it is real. Section 2 drops the parallel-field term in the T equation and erases the insulation between turns. In a series-wound insulated coil, every turn must carry exactly the same transport current. In the homogeneous bulk, the model only constrains the total current through the stack, so current is allowed to redistribute radially across the turns. Whether this changes current-front penetration and hence SCIF/drift timing is an empirical question, and the paper answers it with no benchmark: no resolved-tape model for this coil, no comparison with an H-formulation on a smaller stack, no measured B(t). So the absolute numbers—4.1% SCIF, 2.2% drift, 20% overshoot—are predictions, not validated results. I would not call them unreliable, just unverified. The authors could address this in a revision by adding one comparison case (e.g., one pancake modeled with all 80 tapes resolved).\n\nThe citation pattern is fine; self-citation is appropriate because the method is theirs. No code or data is shipped, but the parameters given are enough to reproduce the study with reasonable effort.\n\nWho is this for: applied superconductivity people who build HTS magnets and want to know whether the homogeneous T-A model is worth adopting. They will get a good, practical yes-with-caveats. For readers looking for new physics, nothing here.\n\nRecommendation: yes, send it to peer review. It deserves referee time. If I were refereeing, I'd ask for the benchmark as a major revision, but the core contribution—speed, parametric sweep, coil-specific overshoot—is solid enough.","headline":"A useful case study of the homogeneous T-A model on an 800-turn coil, with credible speedups and a useful design hint about overshoot; the main gap is the missing benchmark, so the quantitative predictions should be treated with caution.","tokens_in":9362,"tokens_out":7209,"would_cite":true,"duration_ms":76587,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A homogenized T-A finite-element model makes real-time simulation of slow-ramping large-scale HTS coils feasible and predicts that a 20% transport-current overshoot eliminates central-field drift in an 800-turn coil.","keywords":["T-A formulation","homogeneous model","screening current-induced field","field drift","current sweep reversal","striation","HTS coil","real-time simulation"],"falsifier":"Place Hall sensors at the center of the same 800-turn coil and record the central field through the triangular 26.5 A cycle and a 30 h plateau; if measured SCIF at peak departs from the predicted near-4.1% beyond sensor uncertainty, or if measurable drift reappears after the 20% overshoot, the homogeneous-model prediction fails for this coil.","tokens_in":1756,"feed_emoji":"🧲","tokens_out":2688,"duration_ms":96196,"temperature":0.7,"pith_summary":"This paper claims that a homogenized T-A finite-element model can simulate screening-current-induced field (SCIF) and field drift in an 800-turn HTS solenoidal coil in about one hour per run on a desktop computer, making real-time simulation of slow ramping cycles feasible. It uses that speed to vary the critical current density $J_c$ and the $n$ power-law index, and shows that both SCIF and drift grow with $J_c$ and $n$, while lower values reduce them at the cost of tape performance. It also tests two remedies: striating tapes into narrow strips reduces SCIF, and a 20% transport-current overshoot eliminates central-field drift for this coil, where 5% and 10% overshoots only postpone it. The practical consequence is that drift mitigation must be evaluated coil by coil rather than assumed from literature values.","feed_headline":"20% current overshoot wipes out superconducting coil drift","feed_subtitle":"A one-hour simulation of an 800-turn magnet shows how to suppress the field drift that plagues NMR and MRI.","key_machinery":"The machinery is the T-A homogeneous formulation. The current vector potential $\\mathbf{T}$ is defined only along the 1D HTS layers, or inside the homogeneous bulks obtained by replacing tape stacks, while the magnetic vector potential $\\mathbf{A}$ is defined over the whole domain. Transport current enters through a Dirichlet boundary condition $I = (T_1 - T_2)\\delta$, imposing a sheet current. Conduction follows the E-J power law $\\rho = E_c/J_c(\\mathbf{B})\\,|\\mathbf{J}/J_c(\\mathbf{B})|^{n-1}$, with an anisotropic field-dependent $J_c$, so the tape-parallel field component that is neglected in the $\\mathbf{T}$ equation still affects the local critical current density. Homogenization is the limit where inter-layer gaps vanish and the surrounding medium is insulating, collapsing the number of degrees of freedom so that 6-h and 30-h cycles simulate in about one hour.","core_discovery":"On the paper's own terms, the discovery is that replacing each stack of HTS tapes in a solenoid by one homogenized bulk, with the current vector potential $\\mathbf{T}$ defined only in the bulk and the magnetic vector potential $\\mathbf{A}$ everywhere, brings computational cost low enough for real-time simulation of slow ramps while retaining the essential screening-current physics. For the 800-turn test coil (10 pancakes, 80 turns each) at 26.5 A, the model gives a central field of $B_{\\mathrm{sim}} = 0.345$ T against a nominal uniform-current value of $B_n = 0.36$ T, a SCIF of about 4.1%. Systematically varying $J_{c0}$ and $n$ shows that lower $J_c$ and lower $n$ shrink SCIF and drift, that smaller $n$ rounds the hysteresis loop, and that the largest drift (2.2% over 30 h) appears at $n=10$. Two remedies are tested: striation of tapes into up to eight uncoupled strips reduces SCIF, and a 20% current overshoot eliminates drift for this coil while 5% and 10% overshoots only delay its reappearance.","pith_inferences":["An extension left implicit is that the speed of the homogeneous model could support a real-time control loop that adjusts the overshoot level during magnet operation, not just a pre-computed ramp.","The 20% overshoot threshold is likely sensitive to coil geometry, ramp rate, $J_c(\\mathbf{B})$ anisotropy, and operating current; a useful extension would map the threshold across these parameters.","A direct experimental check would compare the predicted near-4.1% SCIF and log-linear drift against Hall-probe measurements on the same coil; the paper does not report such a comparison.","Because the simulated drift grows linearly in log time, longer plateaus could be extrapolated from the 30 h data, but that extrapolation is not made in the paper."],"forward_implications":["For the studied 800-turn coil, a 20% transport-current overshoot removes central-field drift over the simulated 30 h plateau, while 5% and 10% overshoots only delay it.","The overshoot level needed to stabilize the field must be chosen per coil and per operating cycle; the paper explicitly notes that a 1% overshoot reported earlier is insufficient for this coil.","Striating tapes into 2, 4, or 8 uncoupled strips reduces SCIF because each strip develops its own current fronts.","Lowering $J_c$ and $n$ reduces SCIF and drift, but this trades away the high critical current and sharp transition that make 2G HTS tapes attractive for high-field magnets.","Because a 30-hour drift run computes in about one hour, the same approach can be used to screen operating cycles before energizing real magnets."],"supporting_citations":[{"why":"Defines the T-A homogeneous model and shows that it permits real-time simulation of large-scale HTS systems, the computational claim this paper relies on.","marker":"[18]"},{"why":"Introduces the T-A formulation that this paper extends by homogenization.","marker":"[16]"},{"why":"Extends the T-A formulation to coils and stacks with many turns, providing the baseline for large-turn models.","marker":"[17]"},{"why":"Supplies the homogenization technique for converting tape stacks and coils into bulk regions.","marker":"[13]"},{"why":"Provides the multi-scale method and the specific 10-pancake coil case study used here.","marker":"[15]"},{"why":"Defines screening current-induced field and field drift and reviews current sweep reversal as a remedy.","marker":"[2]"},{"why":"Reports measurements of screening-current field in YBCO solenoids that motivate the drift and overshoot study.","marker":"[22]"},{"why":"Shows that striation reduces the magnetic field generated by shielding currents, the remedy tested in Section 6.","marker":"[3]"},{"why":"Establishes the H formulation whose computational cost the T-A homogeneous model is meant to overcome.","marker":"[12]"}],"fun_headline_variants":["20% overshoot erases field drift in HTS magnet","T-A homogenized model runs real-time HTS coil simulation","How to kill screening current drift in HTS magnets","Modeling HTS coil drift without the computational wall","Striation and overshoot tame HTS coil field drift"],"cache_read_input_tokens":11520,"weakest_assumption_plain":"The load-bearing premise is that replacing the many thin tapes in each pancake by one solid block, with no current transfer between turns, still represents how screening currents penetrate; if homogenization distorts the current-front pattern, the predicted SCIF and drift timing would shift.","fun_headline_variants_meta":{"raw":{"variants":["20% overshoot erases field drift in HTS magnet","T-A homogenized model runs real-time HTS coil simulation","How to kill screening current drift in HTS magnets","Modeling HTS coil drift without the computational wall","Striation and overshoot tame HTS coil field drift"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00084,"raw_usage":{"total_tokens":3685,"prompt_tokens":997,"completion_tokens":2688,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":613,"completion_tokens_details":{"reasoning_tokens":2616}},"tokens_in":613,"tokens_out":2688,"duration_ms":19042,"temperature":1.0,"reasoning_tokens":2616,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:48:25.200200+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Place Hall sensors at the center of the same 800-turn coil and record the central field through the triangular 26.5 A cycle and a 30 h plateau; if measured SCIF at peak departs from the predicted near-4.1% beyond sensor uncertainty, or if measurable drift reappears after the 20% overshoot, the homogeneous-model prediction fails for this coil.","supporting_citations":[{"cited_title":"and Grilli, F","cited_arxiv_id":null,"evidence_quote":"Defines the T-A homogeneous model and shows that it permits real-time simulation of large-scale HTS systems, the computational claim this paper relies on."},{"cited_title":"and Yuan, W","cited_arxiv_id":null,"evidence_quote":"Introduces the T-A formulation that this paper extends by homogenization."},{"cited_title":"and Yuan, W","cited_arxiv_id":null,"evidence_quote":"Extends the T-A formulation to coils and stacks with many turns, providing the baseline for large-turn models."},{"cited_title":"and Sørensen, M","cited_arxiv_id":null,"evidence_quote":"Supplies the homogenization technique for converting tape stacks and coils into bulk regions."},{"cited_title":"and Grilli, F","cited_arxiv_id":null,"evidence_quote":"Provides the multi-scale method and the specific 10-pancake coil case study used here."},{"cited_title":"and Yanagisawa, Y","cited_arxiv_id":null,"evidence_quote":"Defines screening current-induced field and field drift and reviews current sweep reversal as a remedy."},{"cited_title":"and Kiyoshi, T","cited_arxiv_id":null,"evidence_quote":"Reports measurements of screening-current field in YBCO solenoids that motivate the drift and overshoot study."},{"cited_title":"and Akachi, K","cited_arxiv_id":null,"evidence_quote":"Shows that striation reduces the magnetic field generated by shielding currents, the remedy tested in Section 6."},{"cited_title":"and Martini, L","cited_arxiv_id":null,"evidence_quote":"Establishes the H formulation whose computational cost the T-A homogeneous model is meant to overcome."}],"review_version":1}