{"id":"bf352e1b-aba0-4a50-ac39-448b7b43ef31","arxiv_id":"1908.06009","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A gradient-based shape optimization using isogeometric analysis and harmonic stator-rotor coupling reduces the EMF total harmonic distortion of a 6-pole permanent magnet machine by more than 75 percent in simulation.","lead":"The paper uses isogeometric analysis to optimize the rotor shape of a 6-pole permanent magnet synchronous machine, cutting total harmonic distortion of the electromotive force by about 75 percent. It couples rotating stator and rotor domains with harmonic basis functions and validates results with a commercial finite-element solver.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The '75% reduction' claim is under-specified: the THD objective (8) uses an index set I from Section II that Section VI never states, so the reported reduction may not apply to total THD.","rationale":"The reader's weakest_assumption correctly identifies the unspecified harmonic index set I as the main point where the central numerical claim is not fully secured. My reading of Section II, Section VI, and Table I confirms that the paper defines THD_I but never states which I was used in the optimization, nor in the independent JMAG validation. The JMAG comparison is a genuine strength: it independently reproduces the IGA trend and shows a substantial reduction even with nonlinear materials. That support, however, is tied to the same objective definition; if the objective excluded certain harmonics, the validation does not establish a 75% reduction of total THD. The methodology itself appears internally coherent: the harmonic stator-rotor coupling, Schur complement reduction, shape derivative, and gradient-based update are all described in a plausible and self-consistent way. The typo in Eq. (19) is a minor issue, not load-bearing. The missing specification of I is the one condition that, if violated, would change the quantitative headline, and it can be settled by a simple reporting or recomputation check. Since the reader already issued CONDITIONAL for this reason, my stress test does not move the verdict; it sharpens the condition under which the paper should be accepted.","tokens_in":12247,"tokens_out":2830,"duration_ms":30547,"concrete_test":"Request the exact index set I used in Eq. (8), either from the authors or from the Zenodo artifact [28], and recompute THD for the original and optimized geometries with I = {2,...,Nα/2-1} (all resolvable harmonics, including even-order terms and harmonics above order 19) using the exported IGES geometry in both GeoPDEs and JMAG with linear material laws. If the reduction remains at or above 75%, the concern is resolved; if the reduction drops materially, the headline claim must be restated as a reduction of the restricted harmonic measure rather than of total THD.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim is that shape optimization reduces total harmonic distortion by more than 75% (Section VI, Table I: IGA linear 0.099268 -> 0.024793; JMAG linear 0.099754 -> 0.02234). The objective actually minimized, however, is THD_I(E) from Eq. (8), with I introduced in Section II as 'an index set of frequencies to consider, e.g., one may disregard frequencies that cannot be diminished by shape optimization.' Nowhere in Section VI or the validation is the particular I used in the optimization stated. If I excluded high-order harmonics (or even harmonics, or harmonics above 19, as Figure 9's axis might suggest), the optimized geometry may still contain large unreduced harmonics, and both the GeoPDEs and JMAG THD entries in Table I would be values of THD_I, not of full THD. The independent JMAG computation validates the same restricted functional, so it does not by itself resolve the ambiguity. This concern is not an accusation of cherry-picking; it is a missing definition that prevents the central 75% reduction from being checked as stated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper develops and demonstrates an isogeometric-analysis-based shape optimization workflow for a 6-pole permanent-magnet synchronous machine. The stator and rotor are discretized as separate multipatch NURBS domains and coupled through harmonic basis functions on the air-gap interface, allowing efficient evaluation of many rotor positions. The optimization objective is the total harmonic distortion (THD) of the electromotive force, computed from the flux linkage over 120 rotor positions; the shape derivative is obtained by an adjoint approach, and descent directions are computed by solving an auxiliary vector problem. The authors report a reduction of THD by more than 75% after 59 iterations, and they validate the optimized geometry with independent JMAG simulations using both linear and nonlinear material models. The central claim is the method's ability to produce a robust, independently confirmed reduction in EMF distortion.","tokens_in":12441,"tokens_out":12777,"duration_ms":141992,"significance":"If the reported results hold, the paper makes a useful contribution: it combines exact IGA geometry, harmonic stator-rotor coupling, and adjoint shape calculus in a way that is substantially faster than solving the full system for each rotor position, and it demonstrates the workflow on a realistic machine with an independent commercial-solver check. The JMAG validation with linear and nonlinear materials is a particular strength, as is the availability of the geometry data via the Zenodo DOI in [28]. The main obstacle to accepting the quantitative claim is the unspecified harmonic index set I in the THD objective; once that is clarified, the contribution should be judged on its engineering value and the quality of the numerical comparison.","major_comments":[{"comment":"The optimization problem (8) minimizes THD_I(E(u(t,Omega))), where I is introduced in Section II as an index set of frequencies to consider, with the remark that one may disregard frequencies that cannot be diminished by shape optimization. Section VI and Table I report the THD values for the original and optimized designs and state that the THD is reduced by more than 75%, but they never specify which I was used in the optimization or in the reported numbers. Figure 9 shows Fourier coefficients only up to harmonic 19, whereas with N_alpha=120 the discrete Fourier representation contains coefficients up to order 60. If the implemented I is a proper subset, the claimed reduction applies to that restricted functional and not necessarily to the total THD, and the JMAG comparison in Fig. 8 validates only the same restricted functional. Please state I explicitly, and in addition report the total THD (with all available harmonics included) for the original and optimized geometries for both IGA and JMAG.","section":"II, Eq. (4); VI, Table I and Fig. 8"}],"minor_comments":[{"comment":"In Eq. (19), the coefficient B_k is written with (M_a)_{k,j}, but from the definition it should be (M_b)_{k,j}; the derivative formula below uses M_b correctly. Please correct the typo.","section":"III-B, Eq. (19)"},{"comment":"Please state the harmonic orders l_k used for the N_Gamma=36 coupling basis. The number 36 alone does not determine the approximation space, and the choice may affect the accuracy of the Fourier coefficients that feed the objective.","section":"IV-B"},{"comment":"The admissible set A and the set of control points that are allowed to move are not specified. For reproducibility, state which control points are design variables and how the descent field W from Eq. (39) is transferred to control-point displacements.","section":"V"},{"comment":"The sentence claiming for the first time shape optimization of a rotating electric machine discretized with IGA is stronger than reference [5] supports; please qualify this claim in light of [5] and related work.","section":"I"},{"comment":"The Fourier-coefficient plots show only odd harmonics up to 19; please indicate the truncation and whether the omitted coefficients are negligible in the reported THD.","section":"VI, Fig. 9"}],"recommendation":"major_revision","confidential_remarks":"For the editor: the stress-test concern is confirmed — the missing specification of the index set I is the main obstacle, and it is fixable. The paper's methodology and independent validation are otherwise sound, and I recommend major revision rather than rejection. I saw no sign of circularity beyond the standard use of the same forward model in optimization and evaluation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a genuinely useful, incremental contribution. The new element is not any single component—both IGA shape optimization for electrical machines and harmonic stator-rotor coupling exist in the authors' prior work—but the combination for a rotating machine with a THD objective, plus a careful independent check. The adjoint/shape-calculus derivation is standard but cleanly executed. The JMAG validation, for linear and nonlinear materials, is real credit: the optimized geometry is tested in a different solver and still shows a large reduction (about 68% under saturation). The Zenodo DOI for geometry/results makes the experiment reproducible in principle. I found no circular reasoning, no fitted constants, no invented entities.\n\nThe load-bearing soft spot is an underspecified objective. The THD in Eq. (8) is THD_I, with an index set I introduced in Section II that the authors explicitly allow to be restricted ('one may disregard frequencies that cannot be diminished by shape optimization'). Section VI never states which I was actually used. Figure 9 plots harmonics only up to 19; Table I labels the quantity THD(E). If the optimization used a restricted I (for example, odd harmonics up to 19, or any subset), the reported 75% reduction is a reduction in that restricted functional, and the JMAG validation—while independent—validates the same restricted functional. The Fourier plots suggest the dominant harmonics lie within the displayed range, so I doubt the effect is cherry-picking, but the claim cannot be checked as stated. That is an addressable revision, not a fatal flaw. The paper also has a small typo in Eq. (19), where Ma appears instead of Mb.\n\nWho this is for: people working on isogeometric analysis for electrical machines, and anyone using adjoint shape optimization on rotating electromechanical devices. It deserves a serious referee. I would send it out and ask for the harmonic set to be stated explicitly, with a statement that full THD behaves comparably, plus the typo fix. My own verdict would be accept after minor revision.","headline":"Solid IGA shape-optimization paper with real validation; the unspecified harmonic index set in the THD objective is the flaw to fix before trusting the 75% claim.","tokens_in":12955,"tokens_out":2543,"would_cite":true,"duration_ms":24024,"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 paper claims that the rotor shape of a 6-pole permanent magnet synchronous machine can be optimized through isogeometric analysis with harmonic stator-rotor coupling, reducing the total harmonic distortion of the electromotive force…","keywords":["electric machines","isogeometric analysis","shape optimization","harmonic stator-rotor coupling","total harmonic distortion","permanent magnet synchronous machine","shape derivative","NURBS"],"falsifier":"Recompute the THD of the original and optimized geometries with the Fourier sum extended to every harmonic resolvable by the 120-step time discretization, using the published machine geometry; if the reduction from 0.099 to 0.025 shrinks below 75 percent, or if an unlisted high-order harmonic dominates the optimized spectrum, the central quantitative claim fails.","tokens_in":12036,"feed_emoji":"⚙️","tokens_out":6062,"duration_ms":49744,"temperature":0.7,"pith_summary":"The paper works out a gradient-based shape optimization pipeline for a rotating electric machine entirely within isogeometric analysis. Because the geometry is represented by NURBS curves, moving control points reshapes the rotor without remeshing, and rotation is handled by coupling the stator and rotor through harmonic basis functions at the air gap. The authors minimize the total harmonic distortion of the electromotive force and report a reduction from 0.099 to 0.025 in 59 iterations, with an independent finite-element solver confirming a close drop. A nonlinear-material check still shows a 68 percent reduction, so the linear-model optimum is not an artifact of ignoring saturation.","feed_headline":"Shape optimization cuts motor voltage distortion by 75%","feed_subtitle":"Tuning the rotor shape directly through its CAD curves cuts distortion, and an independent solver confirms the gain.","key_machinery":"The central machinery is the harmonic stator-rotor coupling: the rotor and stator are separate multipatch IGA domains, and continuity of the magnetic vector potential across the air gap is enforced by Lagrange multipliers expanded in Fourier harmonics $e^{-i\\ell_k\\theta}$. This produces a saddle-point system whose interface part is a small Schur complement, $K_{\\mathrm{int}}(\\alpha) = G_{\\mathrm{rt}}^H K_{\\mathrm{rt}}^{-1} G_{\\mathrm{rt}} + R(\\alpha) G_{\\mathrm{st}}^H K_{\\mathrm{st}}^{-1} G_{\\mathrm{st}} R(\\alpha)$, where the rotation angle $\\alpha$ enters only through the diagonal matrix $R(\\alpha)$. For optimization, the shape derivative (equation 20) converts the THD objective into a descent vector field, and an $H^1$-type Riesz lifting with bilinear form $b(W,Z)=\\int_D (DW:DZ + W\\cdot Z)\\,dx$ extracts the update direction. Only the small interface problem is re-solved for each of the 120 rotor positions, which is why the per-iteration cost stays low.","core_discovery":"The central discovery is that harmonic stator-rotor coupling makes shape optimization of a rotating machine tractable in an isogeometric setting. The simulation splits into an offline phase that factors the stator and rotor stiffness matrices once and an online phase where only a small interface system in the Fourier coefficients of the magnetic field is solved for each rotor angle. The shape derivative of the THD objective is derived through shape calculus with an adjoint equation, and the resulting descent field moves the NURBS control points of the rotor. With 59 iterations the IGA model lowers THD(E) from 0.099268 to 0.024793; an independent finite-element computation gives 0.099754 to 0.02234 with linear materials and 0.10639 to 0.034106 with nonlinear materials. The optimized geometry suppresses the dominant Fourier harmonics of the electromotive force while leaving the permanent magnets unchanged.","pith_inferences":["An implication the authors leave implicit: the same adjoint-based shape derivative can be applied to other design objectives, such as torque ripple or cogging torque, without changing the coupling machinery.","A reader should not treat the 75 percent figure as the total harmonic distortion until the harmonic index set $I$ is specified; re-running with all harmonics included is the natural test.","A testable extension would start the descent from several different initial rotor shapes to see whether the optimized design is a stable local optimum or depends on the starting geometry.","The near-constant online cost in the number of IGA degrees of freedom suggests the method scales to finer discretizations and possibly 3D extensions where classical moving-band methods become expensive."],"forward_implications":["If the central claim is right, gradient-based shape optimization of rotating machines no longer requires remeshing on every geometry update.","The optimized rotor cuts EMF total harmonic distortion by a factor of about four, and the independent validation shows the improvement is not a quirk of the IGA discretization.","With nonlinear magnetic materials the same optimized shape still gives a 68 percent THD reduction, suggesting the linear-model design remains effective under saturation.","Because the optimized design is delivered as a NURBS geometry, it can be exported directly into standard CAD and CAE workflows."],"supporting_citations":[{"why":"Supplies the harmonic stator-rotor coupling formulation and the saddle-point system that the whole rotation model is built on.","marker":"[19]"},{"why":"Provides the shape-calculus and adjoint derivation steps that the paper adapts to the THD objective.","marker":"[21]"},{"why":"Introduces harmonic weighting functions at the sliding interface, the basis for the Fourier expansion of the magnetic field.","marker":"[23]"},{"why":"Establishes isogeometric analysis, the framework that lets control-point movement replace remeshing.","marker":"[1]"},{"why":"Prior work showing IGA shape optimization with applications to electric machines, the methodological starting point.","marker":"[5]"},{"why":"Defines the geometry and material coefficients of the 6-pole test machine used in all experiments.","marker":"[27]"},{"why":"Provides the downloadable machine geometry used to reproduce and validate the optimization results.","marker":"[28]"},{"why":"The isogeometric analysis implementation in which the forward problem and optimization are computed.","marker":"[29]"}],"fun_headline_variants":["Isogeometric rotor shaping cuts motor voltage distortion 75%","Harmonic coupling enables 75% THD reduction in motor design","Shape-optimized rotors shrink voltage distortion by 75%","IGA-based rotor optimization trims motor THD 75%, verified","Rotor geometry tuning via IGA cuts motor distortion 75%"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the index set $I$ of harmonics used in the THD objective really covers the harmonics that matter; the paper defines $I$ in the model section but never states which set the numerical optimization used.","fun_headline_variants_meta":{"raw":{"variants":["Isogeometric rotor shaping cuts motor voltage distortion 75%","Harmonic coupling enables 75% THD reduction in motor design","Shape-optimized rotors shrink voltage distortion by 75%","IGA-based rotor optimization trims motor THD 75%, verified","Rotor geometry tuning via IGA cuts motor distortion 75%"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00092,"raw_usage":{"total_tokens":3876,"prompt_tokens":807,"completion_tokens":3069,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":423,"completion_tokens_details":{"reasoning_tokens":2978}},"tokens_in":423,"tokens_out":3069,"duration_ms":22754,"temperature":1.0,"reasoning_tokens":2978,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:08:20.963013+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the THD of the original and optimized geometries with the Fourier sum extended to every harmonic resolvable by the 120-step time discretization, using the published machine geometry; if the reduction from 0.099 to 0.025 shrinks below 75 percent, or if an unlisted high-order harmonic dominates the optimized spectrum, the central quantitative claim fails.","supporting_citations":[{"cited_title":"Shape optimization of an electric motor subject to nonlinear magnetostatics,","cited_arxiv_id":null,"evidence_quote":"Provides the shape-calculus and adjoint derivation steps that the paper adapts to the THD objective."},{"cited_title":"Harmonic weighting functions at the sliding interface of a finite-element machine model incorporating angular displacement,","cited_arxiv_id":null,"evidence_quote":"Introduces harmonic weighting functions at the sliding interface, the basis for the Fourier expansion of the magnetic field."},{"cited_title":"Isogeo- metric simulation and shape optimization with applications to electrical machines,","cited_arxiv_id":null,"evidence_quote":"Prior work showing IGA shape optimization with applications to electric machines, the methodological starting point."},{"cited_title":"Numerical methods for the estimation of the impact of geometric uncertainties on the performance of electromagnetic devices,","cited_arxiv_id":null,"evidence_quote":"Defines the geometry and material coefficients of the 6-pole test machine used in all experiments."},{"cited_title":"Results of the shape optimization of rotating electric machines using isogeometric analysis,","cited_arxiv_id":null,"evidence_quote":"Provides the downloadable machine geometry used to reproduce and validate the optimization results."},{"cited_title":"GeoPDEs: A research tool for isogeometric analysis of PDEs,","cited_arxiv_id":null,"evidence_quote":"The isogeometric analysis implementation in which the forward problem and optimization are computed."}],"review_version":1}