{"id":"2070ddac-0eab-4d41-a81e-81b1d86c28b2","arxiv_id":"1908.02176","paper_version":3,"verdict":"UNVERDICTED","confidence":"HIGH","novelty_score":0.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A structured review of the H-formulation finite-element method for computing AC losses in high-temperature superconductors, covering equations, implementations, and applications.","lead":"This paper reviews computer models that use the magnetic field as the main variable to calculate energy losses in high-temperature superconductors. It is a useful summary for engineers who need to estimate AC losses in superconducting cables, coils, and magnets.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 10-50% accuracy claim is too broad: the paper itself documents DC-bias/AC-ripple regimes where the power-law E-J law fails, so the central reliability statement needs qualification.","rationale":"The reader's weakest assumption is the power-law E-J model's validity across all reviewed regimes, and the paper itself flags the DC-bias/AC-ripple cases where that model is questionable. My stress-test confirms that this is the most load-bearing concern about the central accuracy claim: the 10-50% band is stated broadly in Section 4, while Section 3.1 explicitly describes regimes where neither the power-law nor the critical-state model matches experiments. The review also supports its 'de facto standard' claim with citation counts and many application examples, so that part is credible. The typo in Eq. (14) (Hz rather than Hr on the right-hand side) is real and should be corrected, but it is an implementation detail, not a threat to the central claim. The paper remains a competent review rather than a research claim, so the reader's UNVERDICTED verdict is appropriate; the accuracy statement should be qualified, but that is an editorial refinement rather than a change in verdict category.","tokens_in":19863,"tokens_out":6023,"duration_ms":68746,"concrete_test":"Compile every quantitative model-vs-experiment loss ratio reported in the cited papers (including Figures 1 and 3 and refs [22,36,54,74,76,79,81,90,94,96]) into a table, separating pure AC cases from DC-bias/AC-ripple cases. If the DC-bias subset's error distribution falls outside the 10-50% band, or if removing it shifts the band materially, Section 4's uncaveated accuracy claim is not supported by the review's own evidence.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 4 states without qualification that H-formulation AC losses agree with experiment to within 10-50% 'in general.' The load-bearing premise for that band is the power-law E-J law (Eq. 7) with field-dependent Jc and n. Section 3.1 identifies a regime where that premise fails: for DC transport/field with AC ripple (refs [65]-[69]), a finite n lets the current profile relax after transients, so the computed loss depends on where on the slowly descending loss curve it is evaluated, and experiments indicate an E-J relation closer to the critical-state model at low E. Ref [69] concludes that neither the power-law nor the critical-state model captures the observed behavior. Since Section 4's 10-50% figure is presented with no caveat and no citation, and several cited papers achieve agreement only after adding auxiliary losses (e.g. ref [76]) or only in restricted current ranges (refs [58], [94]), the central reliability claim is broader than the reviewed evidence supports. The concern is not that the H formulation is wrong, but that the headline accuracy band is not valid across all claimed operating regimes, and the paper itself supplies the counterexample.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This review article surveys the H-formulation finite-element method for computing AC losses in high-temperature superconductors. It presents the governing equations (Faraday's law with a nonlinear power-law resistivity), the 2D longitudinal and axisymmetric reductions, 3D extensions, and homogenized/multi-scale modeling strategies. It then reviews applications to tapes, coils, cables, magnets, electrical machines, fault current limiters, transformers, and SMES, and concludes with a discussion of popularity, ease of implementation, computational efficiency, and typical accuracy. The central claims are that the H formulation has become the de facto standard for AC-loss simulation and that calculated losses generally agree with experiment to within 10-50%.","tokens_in":20072,"tokens_out":8640,"duration_ms":89474,"significance":"If the central claims are properly qualified, this is a useful reference review for the applied superconductivity community. Its strengths are the broad literature coverage, the clear presentation of the formulation in different geometries, the practical implementation hints (structured meshes, current constraints, air-domain resistivity), and the honest enumeration of drawbacks (large air-domain cost, limited parallelization in COMSOL, black-box code). It also points to publicly available model files, which is valuable for reproducibility. The main weakness is that the headline accuracy statement is asserted rather than derived from a systematic synthesis of the reviewed papers, and the paper itself contains evidence of regimes where that accuracy band does not hold.","major_comments":[{"comment":"The axisymmetric governing equations as printed are not the equations that follow from Eqs. (10)-(12). With Er=Ez=0 and Bθ=0, Faraday's law gives ∂Eθ/∂z = ∂(µHr)/∂t and Eθ/r + ∂Eθ/∂r = -∂(µHz)/∂t. Equation (13) instead has a negative left-hand side matched to a positive right-hand side, and Eq. (14) repeats Hr on the right-hand side instead of Hz. Because this section is the implementation reference for axisymmetric coils and windings, these sign and subscript errors should be corrected or explicitly explained before publication.","section":"Section 2.2, Eqs. (13)-(14)"},{"comment":"The statement that 'the accuracy can be quantified as varying between 10% and 50%' is presented without a methodological basis or qualification. The paper's own Section 3.1 (refs [65], [66], [69]) documents a DC-bias/AC-ripple regime in which the power-law E-J law causes loss values to depend on where on a slowly descending relaxation curve they are evaluated, and ref [69] states that neither the power-law nor the critical-state model captures the observed behavior. Several cited validations are also conditional: Section 3.2.1 (ref [76]) matches experiment only after separately adding copper-lead losses computed with a 3D model, Section 3.3.3 (ref [94]) reports good agreement only above 50% of Ic, and Section 2.4 (ref [46]) reports errors up to 20% for the multi-scale method under a uniform-current starting assumption. The 10-50% range should therefore be presented as regime-dependent, with a short description of how it was compiled and which exceptions apply.","section":"Section 4, accuracy paragraph"}],"minor_comments":[{"comment":"Equation (4) is displayed in a garbled form on the right-hand side; it should express that the divergence of ∂(µH)/∂t is zero once the curl term is removed. Please correct the typesetting so the equation reads ∇·[∂(µH)/∂t] = 0 (up to sign).","section":"Section 2.1, Eq. (4)"},{"comment":"The field dependence notation is inconsistent: the prefactor is written as Ec/Jc while the power-law factor uses Jc(B) and n(B). The prefactor should use Jc(B) as well, or the notation should be defined once to avoid ambiguity in implementation.","section":"Section 2.1, Eq. (7)"},{"comment":"Matlab is not a finite-element software package per se; if the authors mean the PDE Toolbox or a code written in MATLAB, the sentence should say so explicitly for clarity.","section":"Section 4, implementation paragraph"}],"recommendation":"major_revision","confidential_remarks":"This is a competent narrative review by authors who are central to the development of the H formulation, and the heavy citation of their own work is understandable in that context. The main blocking issue is the unqualified 10-50% accuracy claim, which the manuscript's own cited literature contradicts in specific regimes. The equation typos in Section 2.2 should also be fixed because they are the equations an implementer would copy. Once these are addressed, the paper would be a suitable review for the journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a review, not a research paper. It consolidates the H-formulation FEM method for AC losses in HTS. If you are new to the area or need to pick a model, this is a good starting point: the equations are laid out clearly, the variants (2D, axisymmetric, 3D, homogenized, multi-scale) are explained, and the survey of applications (tapes, Roebel cables, CORC, magnets, machines, FCLs) is broad and organized. No new results, but a review doesn't need them.\n\nThe paper does several things well. It explains the divergence-free constraint cleanly. It gives practical implementation tips: element choice, air resistivity, current constraints. It lists pros and cons honestly, including the COMSOL black-box issue and the computational inefficiency. The self-citation pattern is heavy but understandable in a review of a method the authors themselves helped develop, and the text reuse from [77] is disclosed.\n\nThe soft spots. First, Eq. (14) has a typo: the RHS should be ∂(µHz)/∂t, not ∂(µHr)/∂t. Eq. (13) also looks sign-wrong (should be ∂Eθ/∂z = ∂(µHr)/∂t). A new implementer could be misled. Second, and more important, the 10-50% accuracy claim in Section 4 is presented without qualification or citation. The paper's own Section 3.1 describes DC-transport/AC-ripple regimes where the power-law E-J law relaxes over time and neither the power-law nor critical-state model matches experiment [69]. Some cited papers only match after adding auxiliary losses (e.g., copper leads in [76]) or only in restricted current ranges [58,94]. So the general statement is broader than the reviewed evidence supports. This is not a fatal flaw for the review's purpose, but it needs to be fixed: qualify the claim, cite the exceptions, and distinguish the regimes where the method is reliable from those where it is not.\n\nBottom line: this paper is for engineers and graduate students who want to understand and apply the H-formulation. With minor-to-moderate revision (fix the equations, qualify the accuracy claim), it would be a solid reference. A serious referee should see it.","headline":"A useful review of the H-formulation for HTS AC losses; the 10-50% accuracy claim needs qualification and Eq. (14) has a typo, but it deserves peer review.","tokens_in":20630,"tokens_out":5526,"would_cite":true,"duration_ms":49336,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This review establishes that the H-formulation finite-element model, which solves Maxwell's equations using the magnetic field as state variables, has become the de facto standard for calculating AC losses in high-temperature…","keywords":["H-formulation","AC losses","high-temperature superconductors","finite element method","power-law resistivity","homogenization","multi-scale modeling","T-A formulation"],"falsifier":"Measure the electric-field-versus-current-density curve of an HTS coated conductor under DC bias with a small AC ripple, at low electric fields, and compute the cyclic AC loss at two different moments during the slow relaxation that the power-law predicts; if the measured field is systematically below the power-law curve (as the experiments cited in the review suggest) or the loss depends strongly on the evaluation time, the central 10-50% accuracy claim fails.","tokens_in":19619,"feed_emoji":"⚡","tokens_out":6333,"duration_ms":62824,"temperature":0.7,"pith_summary":"This review argues that the H-formulation finite-element model—solving Maxwell's equations with the magnetic field as the state variable—has become the de facto standard for calculating AC losses in high-temperature superconductors. Its central quantitative claim is that computed losses typically agree with measurements to within 10-50%, which the paper presents as satisfactory given uncertainties in tape characterization, assumed uniformity, and measurement noise. The review documents the governing equations, extensions from 2D cross-sections to axisymmetric and 3D geometries, and techniques such as homogenization and multi-scale modeling that make large coils and magnets computable. It also identifies a genuine limitation: for DC-biased conductors with AC ripple, the power-law resistivity model predicts a slow relaxation of current profiles, so loss values depend on when in the cycle they are evaluated. A reader should care because AC losses set the cryogenic burden of every HTS power device, and this method is what practitioners actually use to estimate them before building hardware.","feed_headline":"H-formulation predicts superconductor AC losses to 10-50%","feed_subtitle":"Review finds the H-formulation FEM model now underpins loss estimates for HTS cables, coils and magnets.","key_machinery":"The load-bearing object is the H formulation: Faraday's law rewritten as $\\nabla\\times(\\rho\\nabla\\times\\mathbf{H}) = -\\partial(\\mu\\mathbf{H})/\\partial t$, with the magnetic field components as the state variables. The superconductor enters through a power-law resistivity $\\rho(J) = \\frac{E_c}{J_c(B)}\\left(\\frac{J}{J_c(B)}\\right)^{n(B)-1}$, which captures the nonlinear E-J characteristic, field-dependent critical current density, and flux creep. The divergence-free condition on $\\mathbf{B}$ is enforced by choosing divergence-free initial conditions, and transport currents are imposed via integral constraints on each conductor. This machinery is what lets a single formulation handle tapes, cables, coils, and magnets in 2D, axisymmetric, and 3D geometries, with AC losses computed as the cycle-averaged integral of $\\mathbf{J}\\cdot\\mathbf{E}$ over the superconducting domain.","core_discovery":"The paper's central claim is that the H formulation, implemented in finite elements with a nonlinear power-law resistivity, has become the community's default tool for AC-loss estimation in HTS, and that this status is justified by its accuracy and flexibility. Across the reviewed studies, calculated losses match experiments within 10-50%, and the model has been extended from a single tape to stacks, Roebel and CORC cables, pancake and racetrack coils, fault current limiters, transformers, and multi-thousand-turn magnets. The formulation is invariant to coordinate system, so the same equations serve 2D longitudinal, axisymmetric, and full 3D problems, with integral current constraints imposing transport currents in individual conductors. The review's own evidence, however, shows that the power-law E-J relation produces a slow relaxation of current profiles after transients, and that for DC-biased conductors experiments suggest a critical-state-like behavior at low electric fields, which makes the accuracy claim regime-dependent.","pith_inferences":["If the 10-50% band is taken as a community benchmark, a natural next step is a systematic comparison of the H formulation and the T-A formulation on the same set of coil and cable geometries, reporting both accuracy against experiment and compute time; the review's evidence suggests H will remain more flexible for cases with magnetic materials and 3D effects.","The cited experiments on DC-biased conductors point to a possible refinement: a hybrid E-J law that behaves like the power law during large transients but approaches critical-state-like behavior at low electric fields would likely remove the relaxation ambiguity and extend the formulation's validity.","The review implies that the main source of the 10-50% uncertainty is not the solver but the input data, primarily characterization of Jc(B) anisotropy and tape-to-tape variation; a set of fully characterized reference tapes would let model error be separated from data error.","Because the paper does not quantify how the 10-50% band is distributed across geometries, one could test whether accuracy is systematically worse for 3D twisted cables than for 2D stacks; the review's examples suggest such a gradient but do not state it."],"forward_implications":["Engineers can use the H formulation to estimate AC losses in HTS tapes, cables, coils, and magnets with a stated accuracy of 10-50% against measurements, which is sufficient for many design choices.","Homogenization and multi-scale methods reduce simulation time by factors of 50-60 for large coils while keeping loss differences below about 1% in the cases reviewed.","Full 3D models are needed for twisted conductors, racetrack coils at medium and high currents, and end effects, because 2D planar models can miss these contributions.","For DC-biased conductors with AC ripples, cyclic loss values should be interpreted with caution, because the power-law model predicts a slow relaxation of current profiles and a critical-state-like E-J may be more appropriate at low electric fields.","The H formulation's leading role is not guaranteed: the T-A formulation is faster for thin coated conductors, but H retains an edge in flexibility for other geometries and materials."],"supporting_citations":[{"why":"It is the origin of the method: it first proposed using magnetic-field components as state variables in time-dependent finite-element simulations of superconductors.","marker":"[23]"},{"why":"It is an independent early implementation of the same field-formulation idea in a home-made code, showing the approach was viable before commercial software appeared.","marker":"[24]"},{"why":"It is the first implementation of the H formulation in commercial finite-element software, launching its wide adoption in the applied superconductivity community.","marker":"[25]"},{"why":"It is the independent commercial-software implementation that became a standard reference for the formulation's accuracy and popularity.","marker":"[26]"},{"why":"It provides the weak-form functional-analytic treatment that lets the divergence-free condition on the magnetic flux density be satisfied by construction.","marker":"[28]"},{"why":"It supplies the power-law E-J relation used as the superconducting constitutive model throughout the reviewed studies.","marker":"[31]"},{"why":"It introduces the homogenization technique that makes large coils with many turns tractable by diluting the critical current density over the coil cross-section.","marker":"[45]"},{"why":"It introduces the T-A formulation, the main competitor that the review says may challenge the H formulation for thin coated conductors.","marker":"[104]"}],"fun_headline_variants":["H-formulation FEM: the default for HTS AC loss","Superconductor loss? H-formulation predicts to 10-50%","H-formulation: one model, all HTS geometries","From tape to magnet: H-formulation for AC loss","Why H-formulation rules HTS AC loss modeling"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The accuracy estimate of 10-50% rests on the power-law resistivity model—with field-dependent $J_c(B)$ and $n(B)$—faithfully describing HTS dissipation in the simulated regimes; if that constitutive law is wrong, for example under DC bias with small AC ripple, the stated accuracy band does not hold.","fun_headline_variants_meta":{"raw":{"variants":["H-formulation FEM: the default for HTS AC loss","Superconductor loss? H-formulation predicts to 10-50%","H-formulation: one model, all HTS geometries","From tape to magnet: H-formulation for AC loss","Why H-formulation rules HTS AC loss modeling"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000555,"raw_usage":{"total_tokens":2644,"prompt_tokens":946,"completion_tokens":1698,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":562,"completion_tokens_details":{"reasoning_tokens":1627}},"tokens_in":562,"tokens_out":1698,"duration_ms":14254,"temperature":1.0,"reasoning_tokens":1627,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:51:23.586473+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the electric-field-versus-current-density curve of an HTS coated conductor under DC bias with a small AC ripple, at low electric fields, and compute the cyclic AC loss at two different moments during the slow relaxation that the power-law predicts; if the measured field is systematically below the power-law curve (as the experiments cited in the review suggest) or the loss depends strongly on the evaluation time, the central 10-50% accuracy claim fails.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It is the origin of the method: it first proposed using magnetic-field components as state variables in time-dependent finite-element simulations of superconductors."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It provides the weak-form functional-analytic treatment that lets the divergence-free condition on the magnetic flux density be satisfied by construction."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It introduces the T-A formulation, the main competitor that the review says may challenge the H formulation for thin coated conductors."}],"review_version":1}