{"id":"06d8c9b2-0696-4798-8f6d-f72386746bcc","arxiv_id":"2607.18894","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"For a grid-forming converter, the strong-grid sub-synchronous oscillation starts through a supercritical Hopf bifurcation whose limit cycle grows rapidly, and smooth current-limiter approximations can create spurious Hopf bifurcations.","lead":"This paper maps when a grid-forming power converter loses stability to sub-synchronous oscillations and shows the oscillations grow rapidly once they start. It also warns that smoothing a current limiter for numerical analysis can create fake instabilities.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Neglect of PWM/control delay in inner-controller pole-placement tuning could shift the quantitative stability maps and even alter the supercritical classification, making the paper's primary quantitative claims conditional.","rationale":"The paper's central claims are the parametric stability bounds for the inner-controller time constants and the supercritical Hopf with rapid limit-cycle growth. The reader's weakest assumption—neglect of PWM/control delay in the pole-placement tuning—is indeed the most load-bearing because it directly affects these quantitative results. The delay is comparable to the fast inner-loop time constant, and the SSO frequency is in a range where such delay could alter the loop interactions. A concrete sensitivity test with a Padé delay is feasible given the provided Julia code, and if it shows significant shifts, the quantitative stability maps and the supercritical classification would need revision. The spurious-Hopf claim, while not time-domain validated at the weak-grid points, is a secondary contribution and less central to the title's focus on SSO stability. The participation-factor confirmation is also minor. Thus, I agree with the reader's identification, and the verdict remains CONDITIONAL pending this sensitivity check.","tokens_in":11026,"tokens_out":7824,"duration_ms":79066,"concrete_test":"Using the provided Julia code, add a second-order Padé approximation of a 0.5–1 ms control/PWM delay to the inner current loop, recompute the codim-2 continuation of the strong-grid Hopf in the (τ_IVC, τ_ICC) plane at SCR=2.5 (Fig. 8), and recompute the first Lyapunov coefficient at the nominal Hopf (SCR≈4.95). If the Hopf curve shifts by >10% in τ_ICC or the sign of Re(b) flips, the quantitative stability claims and supercriticality results are not robust to delay.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The inner-controller tuning in Section III-C (Eqs. 6–13) explicitly assumes ideal time-scale separation and neglects PWM/control delay. This assumption is load-bearing because the analysed SSO is directly tied to the inner current and voltage loops (Section IV-C), and the nominal current-loop time constant τ_ICC = 1.5 ms is comparable to a realistic 0.5–1 ms control/PWM delay. Such a delay introduces phase lag that modifies the effective inner-loop dynamics near the SSO frequency (~25.6 Hz) and strongly affects the current-loop crossover (~455 Hz). Therefore, the codim-2 Hopf boundaries in Fig. 8 (e.g., the claim that τ_ICC ⪆ 2.5 ms cannot be stabilised at SCR=5) and the strong-grid Hopf location in Fig. 6 could shift substantially, and the computed first Lyapunov coefficient b = −0.00198 − j0.00096 could change sign, reversing the supercritical/subcritical classification. The paper does not justify the delay neglect for this frequency range, nor does it provide a sensitivity analysis, even though the provided Julia code would make such a test straightforward.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a continuation-based bifurcation analysis of a grid-forming (GFM) converter with cascaded inner voltage and current controllers connected to a Thévenin grid. Using numerical continuation and normal-form analysis, it maps stability boundaries with respect to operational parameters (SCR, X/R, active-power set point) and inner-controller time constants τ_IVC and τ_ICC. The authors report that the strong-grid Hopf bifurcation of the inner-controller-related SSO is supercritical, with the emerging limit cycle growing rapidly to unacceptable amplitudes. They further show that smooth approximations of the circular current limiter can introduce spurious Hopf bifurcations in weak grids, which are absent when a hard limiter is used. The paper includes Julia scripts for modelling and continuation, supporting reproducibility.","tokens_in":11323,"tokens_out":10877,"duration_ms":98475,"significance":"If the results stand, the paper provides a useful nonlinear perspective on an important class of GFM stability problems. The codim-2 stability maps for the inner-loop time constants and the warning about smooth limiter approximations are of practical value, and the availability of reproducible code is a significant strength. The key qualitative findings — supercriticality of the strong-grid Hopf and the existence of spurious Hopf points due to smoothing — are internally consistent. However, the quantitative accuracy of the stability boundaries depends on the delay-free, ideal-timescale-separation assumption used to derive the inner-controller tuning, which is not justified for the frequency range of interest. The lack of direct time-domain validation of the limit-cycle and spurious-Hopf claims leaves some uncertainty, but these are addressable with additional simulations.","major_comments":[{"comment":"The inner-loop pole-placement tuning explicitly assumes ideal time-scale separation and neglect of PWM/control delay. This is load-bearing for the quantitative stability maps: with τ_ICC=1.5 ms, the current-loop natural frequency is ≈455 Hz (from Eq. (11)), so a realistic 0.5–1 ms delay introduces ~82°–164° of phase lag at that frequency. The SSO is at 25.6 Hz in Fig. 6, but the Hopf boundaries in Fig. 8 are expressed directly in τ_ICC/τ_IVC space and could shift substantially; the first Lyapunov coefficient b in §IV-B could even change sign. The manuscript states the assumption but provides no justification or sensitivity analysis. Please add a delay model (e.g., Padé approximation) or a sensitivity study quantifying the effect of a realistic delay on the Hopf loci and on b.","section":"§III-C, Eqs. (6)–(13); §IV-C; §IV-B"},{"comment":"The central claim that the strong-grid Hopf is supercritical with rapid onset of unacceptably large oscillations rests entirely on the normal-form coefficient and limit-cycle continuation. No time-domain simulation is shown to confirm that a disturbance near the predicted Hopf point actually produces the predicted limit cycle and amplitude. Given that the manuscript already includes time-domain validation for the smooth limiter (Appendix A), a similar check for the strong-grid limit cycle (e.g., at SCR just above the Hopf) would directly support the main conclusion and rule out numerical artifacts. Please add such a simulation.","section":"§IV-B, Fig. 7"},{"comment":"The spurious-Hopf conclusion is based on the absence of Hopf points in the hard-limiter continuation and their presence in the smooth-approximation continuation. This is internally consistent, but the practical message — that smooth approximations can invalidate bifurcation studies — would be much stronger with a direct time-domain demonstration at a weak-grid operating point (e.g., SCR≈1.1, P*=1 pu) showing that the smooth model exhibits sustained oscillations while the hard-limiter model does not. The current time-domain validation in Appendix A is at SCR=2.5, away from the spurious-Hopf region. This validation is advisable before recommending 'extreme caution' in the conclusions.","section":"§IV-D, Appendix B"}],"minor_comments":[{"comment":"T_s is listed as 47.12 ms in Table I, but the text states T_s = K_i,IVC. Please clarify the units and the exact relationship, as the anti-windup tracking time constant affects the limiter dynamics in Figs. 9–13.","section":"Table I and §III-D"},{"comment":"The phrase 'for relatively small variations of |v_m| (Fig. 4b)' should likely refer to Fig. 4a; Fig. 4b shows δθ.","section":"§IV-A, paragraph around Fig. 4"},{"comment":"Consider replacing 'wide-bandwidth stability issues' with 'wide-band stability issues' to match the terminology of reference [4].","section":"Abstract"},{"comment":"The sequence of SCR values for the spurious Hopf and the return to stability is hard to follow; a small table or a more explicit sentence mapping δ to the two crossing points would improve readability.","section":"§IV-D"},{"comment":"The sentence 'The strong grid Hopf bifurcation remains supercritical only for very small values of δ with a Bautin bifurcation occurring when δ≈0.0011 and SCR≈4.944, after which (i.e., smaller SCR) the bifurcation is instead subcritical' is confusing because the direction of parameter change in δ and SCR is not clearly stated. Please clarify.","section":"Appendix B"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the conference and is generally well-structured with reproducible code. The delay neglect is the key technical issue; if the authors can show that a realistic delay does not change the qualitative conclusions, I would be inclined to accept. The unit inconsistency for T_s should also be fixed. The time-domain validation requests are intended to strengthen the central claims and should be feasible given the provided code. The self-citation [11] is appropriate for prior observation of the SSO."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a competent and genuinely useful bifurcation study of the standard cascaded GFM controller. The genuinely new things are the codim-2 stability-boundary map in (τ_IVC, τ_ICC), the SCR/X-R continuation of the strong-grid Hopf, and the observation that a tanh-smoothed circular current limiter manufactures spurious Hopf bifurcations in weak grids. The supercritical classification is backed by a first Lyapunov coefficient and limit-cycle continuation showing rapid amplitude growth; that's a real result, and the spurious-Hopf story is internally consistent (hard limiter gives no Hopf; smooth did; and the continuation in δ shows the two branches collide). Code and data are provided, so the numerics are checkable.\n\nThe main soft spot is the model's neglect of PWM/control delay. The inner-loop tuning (Eqs. 6–13) explicitly assumes ideal time-scale separation and no delay, but the SSO frequency is ~25.6 Hz and the current-loop crossover is ~455 Hz, so a realistic 0.5–1 ms delay introduces phase lag that could shift the Hopf boundaries in Figs. 6 and 8 and, in principle, flip the sign of the first Lyapunov coefficient. The paper doesn't justify the neglect or run a sensitivity analysis, even though the provided Julia code would make a Padé approximation or delay-state test cheap. This makes the quantitative claims (e.g., the τ_ICC > 2.5 ms limit at SCR=5) conditional, not wrong.\n\nMinor: the participation factor confirmation is mentioned but not shown, and the weak-grid spurious Hopf is not directly time-domain validated at the bifurcation point, though Appendix A's voltage-dip test does show δ=0.05/0.1 trips the spurious instability, which supports the main point. Also, τ_ICC is a 5% settling time, not a time constant; wording is loose but the math is clear.\n\nOverall: the central qualitative claims hold up. The spurious-Hopf artifact is particularly important for anyone doing continuation with smoothed limiters. The reader's conditional verdict is about right. A serious referee should engage, mainly to ask for the delay check and for the participation-factor figure.\n\nWho this is for: power-systems researchers working on GFM converter stability and anyone using continuation with smoothed saturation. I'd take it to a reading group and would cite the spurious-Hopf result. Recommend sending to peer review, with the delay question as the main revision request.","headline":"Solid continuation study of GFM inner-controller SSOs; new codim-2 stability maps and a well-supported spurious-Hopf finding, but delay neglect tempers the quantitative boundaries.","tokens_in":11819,"tokens_out":2025,"would_cite":true,"duration_ms":18462,"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 grid-forming converter's inner controllers can push strong-grid operation into sub-synchronous oscillations that rapidly grow past acceptable levels, via a supercritical Hopf bifurcation.","keywords":["grid-forming converter","sub-synchronous oscillation","Hopf bifurcation","limit cycle continuation","inner voltage controller","inner current controller","current limiter smooth approximation","stability boundary"],"falsifier":"Run the same continuation with a computational delay of one-and-a-half switching periods or a discrete-time current-control model, and check whether the strong-grid Hopf still occurs at the reported SCR values and whether the first Lyapunov coefficient remains negative at the nominal parameters of the paper.","tokens_in":10913,"feed_emoji":"⚡","tokens_out":3385,"duration_ms":34140,"temperature":0.7,"pith_summary":"This paper tries to establish that the strong-grid instability caused by the inner voltage and current controllers of a grid-forming converter is not a marginal linear effect but a nonlinear bifurcation: as grid strength or controller speeds cross a boundary, a stable limit cycle emerges and quickly grows to unacceptable voltage and current swings. Using continuation and normal-form analysis, the authors show the Hopf bifurcation is supercritical and trace the limit cycle's amplitude. They also map two-parameter stability boundaries showing that the inner voltage controller must be slow enough and the inner current controller fast enough. Finally, they show that smooth approximations of the circular current limiter can create spurious weak-grid Hopf bifurcations that disappear with the hard limiter, warning that such approximations can invalidate bifurcation studies.","feed_headline":"Supercritical Hopf triggers fast-growing grid-forming oscillations","feed_subtitle":"Continuation analysis maps when inner controller time constants let strong grids push sub-synchronous swings past acceptable limits.","key_machinery":"The carrying tool is Hopf normal-form theory combined with numerical continuation: the sign of the real part of the first Lyapunov coefficient b in the normal form z_dot = z(jω + a δp + b |z|^2) decides whether the Hopf bifurcation is supercritical (negative) or subcritical (positive), and continuation of the limit cycle with collocation methods tracks how large the oscillation grows. Two-parameter continuation follows the Hopf curve in grids of operational and controller parameters, while a smooth hyperbolic-tangent approximation of the circular current limiter is used to include the non-smooth saturation in the continuation; the spurious Hopf bifurcations arise from this approximation, not","core_discovery":"The paper claims that the strong grid Hopf point of the inner controller-related sub-synchronous oscillation is supercritical: the first Lyapunov coefficient is negative, a stable limit cycle appears around the now-unstable equilibrium, and continuation past the Hopf point shows the oscillation amplitude increasing rapidly with only a small increase in the continuation parameter, quickly reaching unacceptable levels. Codimension-2 continuation of the Hopf point in the inner controller time constants reveals that stability in strong grids requires the inner voltage controller time constant to be sufficiently large and the inner current controller time constant to be sufficiently small, and th","pith_inferences":["A natural testable extension is to repeat the continuation with a computational or pulse-width-modulation delay (or discrete-time sampling) in the inner current loop; because the instability lives in those loops, including delay could shift the Hopf boundaries in the SCR and time-constant maps and could change the supercritical classification.","The spurious-Hopf phenomenon likely generalises beyond this specific limiter: any smooth approximation of a non-smooth saturation element in a converter model may create artificial eigenvalue crossings in continuation analysis, so time-domain validation against the hard nonlinearity should accompany bifurcation results.","The rapid growth of the limit cycle implies that protection systems and the current limiter itself will engage almost immediately after instability onset; a hybrid model combining the hard limiter with limit-cycle continuation could expose interactions between the SSO limit cycle and the saturating current reference.","The analysis treats a single converter against an infinite bus; with multiple grid-forming converters, the effective grid strength seen by each converter changes dynamically, so the depicted Hopf boundary may shift or be replaced by coupled modes in a multi-unit continuation."],"forward_implications":["Crossing the strong-grid stability boundary is not a gentle loss of stability: after the supercritical Hopf point, sustained oscillations quickly become large enough to threaten converter protection and grid voltage quality.","Grid-forming converter designs using the standard cascaded inner controllers must maintain strong time-scale separation in strong grids; if the current loop is slowed by limited switching frequency, the voltage loop must be slowed accordingly, and beyond roughly 2.5 ms current-loop settling time at SCR = 5, no voltage tuning stabilises the system.","The strong-grid instability moves to higher SCR values as the grid X/R ratio decreases, meaning distribution-connected grid-forming converters can tolerate stronger grids than transmission-connected ones, with the SSO frequency falling as low as about 5.7 Hz at X/R = 0.1.","Bifurcation studies that replace hard current-limit saturation with smooth approximations should be validated against the hard-response model, since even ostensibly close approximations can produce false Hopf bifurcations and false criticality conclusions.","The limit-cycle continuation provides a concrete amplitude bound: the region where the equilibrium is locally stable but the limit cycle is already too large is effectively absent, so the local stability boundary can be used as an operational limit."],"fun_headline_variants":["Supercritical Hopf: fast-growing sub-synchronous swings in grid-forming converters","Bifurcation maps show rapid oscillation onset in strong-grid GFM converters","GFM converter inner controllers: Hopf analysis exposes fast oscillation growth","Sub-synchronous oscillations explode past Hopf point in grid-forming converters","Strong-grid GFM instability: supercritical Hopf leads to rapid oscillation spikes"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The inner-controller pole-placement tuning assumes ideal time-scale separation and neglects pulse-width-modulation and control delay; because the sub-synchronous oscillation is created by those same inner loops, adding realistic delay or discrete sampling could move the computed Hopf boundaries and possibly change whether the strong-grid Hopf is supercritical.","fun_headline_variants_meta":{"raw":{"variants":["Supercritical Hopf: fast-growing sub-synchronous swings in grid-forming converters","Bifurcation maps show rapid oscillation onset in strong-grid GFM converters","GFM converter inner controllers: Hopf analysis exposes fast oscillation growth","Sub-synchronous oscillations explode past Hopf point in grid-forming converters","Strong-grid GFM instability: supercritical Hopf leads to rapid oscillation spikes"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000847,"raw_usage":{"total_tokens":3500,"prompt_tokens":702,"completion_tokens":2798,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":446,"completion_tokens_details":{"reasoning_tokens":2701}},"tokens_in":446,"tokens_out":2798,"duration_ms":17286,"temperature":1.0,"reasoning_tokens":2701,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T13:59:46.140597+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same continuation with a computational delay of one-and-a-half switching periods or a discrete-time current-control model, and check whether the strong-grid Hopf still occurs at the reported SCR values and whether the first Lyapunov coefficient remains negative at the nominal parameters of the paper.","supporting_citations":[],"review_version":1}