{"id":"3ecedfca-3cea-4900-a707-4929def0987a","arxiv_id":"2607.05599","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":2,"one_line_summary":"A Jordan-mode split of a reduced linear power-system model yields certified multi-second reachable tubes for mixed SG/IBR grids in about 0.1 s on a 39-bus case.","lead":"The paper gives a fast way to bound how power-grid frequencies and powers can evolve after uncertain load or generation shocks when the grid mixes synchronous machines with grid-forming and grid-following inverters. Operators could use the resulting tubes to check worst-case frequency and RoCoF without running thousands of detailed simulations.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified beyond the modeling-fidelity caveat already flagged by the reader.","rationale":"The reader correctly separates the certified claim (over-approximations of the reduced linear ODE) from the modeling premise that makes those tubes useful for the EMT plant. The math of the Jordan split, interval embedding (Theorem 1), contraction balls via the logarithmic norm of the 2-by-2 blocks (Theorem 2), and the steady-state intersection (Theorem 4 / Algorithm 1) is standard and does not appear to contain an internal contradiction. The brief GFL frequency exit is explicitly attributed to the quasi-steady frequency-divider approximation, so it does not falsify the certificates. Because that modeling caveat is already the weakest assumption in the reader’s report, and because no deeper technical flaw surfaces on a second pass, the CONDITIONAL verdict and its rationale stand. The suggested concrete test simply quantifies the already-identified gap rather than overturning the claim.","tokens_in":26366,"tokens_out":496,"duration_ms":4900,"concrete_test":"Recompute the reachable tubes of Algorithm 1 on the same IEEE-39 scenarios while replacing the algebraic GFL frequency-divider (8) by a first-order PLL lag whose time constant matches the KauraPLL parameters of Appendix C; if the post-event GFL frequency exit disappears and the remaining EMT trajectories stay inside the new tubes, the modeling-gap concern is quantitatively bounded rather than merely acknowledged.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper’s central claim is carefully scoped: the hybrid interval-plus-contraction construction after real Jordan splitting yields certified over-approximations for the reduced linear ODE (12), not for the stiff EMT plant. Theorems 1–4 and the embedding/contraction arguments are standard and appear internally consistent under the stated multiplicity assumption (Lemma 1). The only soft spot is the one the reader already isolates—the practical usefulness of those tubes for the EMT plant rests on the DC/frequency-divider/reduced-order modeling chain of Section III and Table I. That premise is acknowledged (brief GFL exit in Scenario 1) and is not a hidden inconsistency in the certificates themselves. No stronger load-bearing flaw (e.g., incorrect Schur reduction, invalid logarithmic-norm bound, or unstated singularity of A) is evident from the written argument.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The paper proposes a computationally efficient reachability framework for transmission systems with synchronous generators, grid-forming and grid-following inverters, and bounded power disturbances. Starting from reduced-order device models and a frequency-divider network representation, the authors eliminate algebraic variables by Schur complement to obtain a linear ODE (12). Reachable tubes for this ODE are constructed via a real Jordan decomposition (Lemma 1): non-oscillatory modes are over-approximated by an interval embedding system (Theorem 1), oscillatory modes by logarithmic-norm ball bounds (Theorem 2), with an optional steady-state-informed intersection for piecewise-constant post-event inputs (Theorems 3–4, Algorithm 1). On a modified IEEE 39-bus system the method produces multi-second tubes in ~0.1 s that largely contain high-fidelity dq0 EMT trajectories for two load-step scenarios, with a brief, expected GFL frequency mismatch attributed to the frequency-divider approximation.","tokens_in":26543,"tokens_out":1211,"duration_ms":22672,"significance":"If the reduced-model certificates and the reported runtimes hold, the work offers a practical path to near-real-time set-based assessment of frequency and power envelopes under bounded load/generation uncertainty in mixed SG–IBR grids—something small-signal and single-trajectory transient tools do not provide. Strengths include an explicit, invertible reduction from DAE to ODE, proofs of the interval and contraction bounds (Section VII), a mode-decoupled construction that exploits the real Jordan structure, and a clear scoping that certificates apply to the analytical model while EMT is used only for validation. The sub-second multi-second tubes on a 39-bus system are a concrete computational contribution relative to stiff EMT roll-outs.","major_comments":[{"comment":"Section V and Table III support the efficiency claim only against reduced-model trajectory roll-outs and dq0 EMT, not against established set-propagation tools (zonotopes, support functions, differential inclusions) applied to the same linear ODE (12). The introduction argues that those methods are prohibitive for large networks, but without a head-to-head comparison on (12) the claim that the Jordan-based hybrid method is ‘several orders of magnitude faster’ than existing reachability methods remains incompletely substantiated. A comparison on the IEEE 39-bus reduced model (even for a short horizon) would make the computational contribution load-bearing rather than suggestive.","section":null},{"comment":"The central practical claim is that tubes certified for the reduced ODE remain useful envelopes for the stiff EMT plant (Abstract; contribution c3; Section V). Scenario 1 already shows a brief post-event GFL frequency exit of the tube, correctly attributed to the frequency-divider quasi-steady approximation (Section III-B, Eq. (8)). The manuscript does not systematically quantify when or how far EMT trajectories can leave the tube (e.g., across larger uncertainty sets, consecutive events, or different GFL PLL bandwidths). Without such quantification or an explicit modeling-error discussion, the ‘validation against EMT’ claim risks overstating the operational usefulness of certificates that rigorously hold only for (12).","section":null},{"comment":"Lemma 1 assumes that every eigenvalue of A with nonzero imaginary part has equal algebraic and geometric multiplicity so that oscillatory blocks are pure 2×2 rotation-scaling matrices. The paper states that this is measure-zero and ‘verified numerically via case studies in Section V,’ but Section V does not report the spectrum of A, multiplicities, or conditioning of V. Because Theorems 2–3 and the ball radii depend on this block structure, an explicit numerical check (eigenvalues and Jordan-block sizes for the IEEE 39-bus A) should be included so that the assumption is not left as an unreported claim.","section":null}],"minor_comments":[{"comment":"Figures 2–3 and Appendix D figures would benefit from consistent axis units (p.u. frequency deviation vs. Hz), explicit identification of which bus each panel corresponds to in the caption, and a legend distinguishing EMT, analytical trajectories, and tube bounds.","section":null},{"comment":"Notation for the input set dimension m = |I|+|G|+|F|+N (Section IV) is easy to misread as including all buses rather than load components; a short clarification that p_ℓ is the full N-vector of bus loads would help.","section":null},{"comment":"Table I and Appendix B are valuable but dense; a one-sentence pointer in Section III to which analytical parameters are matched to EMT (bold entries in Appendix C) would help readers who skip the appendices.","section":null},{"comment":"Algorithm 1 uses a fixed step h but does not discuss how h is chosen relative to the fastest retained mode; a brief remark on step-size selection or that the continuous-time bounds are evaluated at discrete samples would reduce ambiguity.","section":null},{"comment":"Minor typographical issues: ‘areachable tube’ (Introduction), inconsistent spacing around citations, and duplicate panel labels ‘(f)’ in Fig. 3.","section":null}],"recommendation":"minor_revision","confidential_remarks":"The mathematical core (reduction, Jordan split, embedding and contraction bounds) appears sound and carefully scoped; the main risk for a top systems journal is overselling EMT ‘validation’ and efficiency versus the reachability literature without direct comparisons. If the authors add a zonotope/CORA-style baseline on (12) and a short modeling-error paragraph, the paper would be a solid contribution. Fit for eess.SY / power-systems dynamics venues is good."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The one thing worth knowing is that this is a clean, usable pipeline rather than a conceptual breakthrough. They take a heterogeneous SG/GFM/GFL DAE, reduce it via frequency-divider and Schur complement to a linear ODE, split modes with a real Jordan form, bound the real modes by an embedding system and the oscillatory modes by logarithmic-norm balls, then optionally intersect with a steady-state-shifted tube. On a modified 39-bus system the multi-second tubes come out in about 0.1 s and mostly contain the PSID dq0-EMT trajectories for the two load-step scenarios they show.\n\nWhat is actually new is the particular combination: frequency-divider reduction tailored to mixed resources, the real-Jordan separation that lets them treat non-oscillatory and oscillatory blocks differently, and the steady-state intersection that trims conservatism after a piecewise-constant event. The math is standard linear-systems material (nonsingularity of A22, inclusion functions, mu_2 bounds) and the proofs in Section VII look correct under the equal algebraic/geometric multiplicity assumption they state. The EMT comparison is honest; they even flag the brief GFL frequency exit right after the event as an expected modeling artifact of the algebraic frequency-divider approximation.\n\nThe soft spot is exactly the one the reader isolates and the stress-test confirms: certificates are for the reduced linear model, not the stiff EMT plant. Usefulness therefore rides on the DC/fixed-voltage/reduced-order modeling chain. That is a modeling premise, not a hidden flaw in the set-propagation theorems, and they do not over-claim. No code is shipped, so reproducibility is mid-tier; the free parameters (uncertainty interval endpoints, step h) are ordinary and not load-bearing. Citations to prior power-system reachability (Althoff, Chen/Domínguez-García, Choi, etc.) are present and fair.\n\nThis is for people who need fast set-based envelopes for mixed-resource frequency dynamics under bounded load/generation uncertainty. A serious editor should send it to referees; the contribution is accept-shaped once the modeling gap is quantified a bit more carefully and ideally code is released. I would read it, cite the pipeline if I am doing set-based power-system work, and bring it to reading group if we are talking about practical reachability for inverter-heavy grids.","headline":"Solid, incremental methods paper: real-Jordan mode split plus interval/contraction bounds gives certified tubes for a reduced linear power-system ODE, with ~0.1 s multi-second runs that largely envelope EMT trajectories.","tokens_in":27219,"tokens_out":589,"would_cite":true,"duration_ms":6298,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"A real Jordan split turns mixed generator-and-inverter grid dynamics into fast certified reachable tubes under bounded load and generation swings.","keywords":["reachability analysis","power systems","Jordan transformation","interval methods","contraction theory","grid-forming inverters","grid-following inverters","frequency divider"],"falsifier":"On the same modified IEEE 39-bus system, introduce a load or generation step whose EMT trajectories systematically and persistently leave the computed reachable tube after the brief PLL transient, or show that the Jordan-based tube computation exceeds the time of a comparable set of reduced-model trajectory roll-outs.","tokens_in":27195,"feed_emoji":"⚡","tokens_out":843,"duration_ms":7208,"temperature":0.7,"pith_summary":"Grid operators increasingly need worst-case envelopes of frequency, power and RoCoF after uncertain load or generation events, not single simulated trajectories. Exact reachability for full electromagnetic-transient models is intractable, so the paper first reduces heterogeneous devices (synchronous machines, grid-forming and grid-following inverters) plus a DC network to a linear ODE via a frequency-divider relation and Schur complement. It then applies a real Jordan change of coordinates that cleanly separates non-oscillatory modes from oscillatory pairs. Interval embedding bounds the former while contraction balls bound the latter; for post-event constant inputs the two tubes are further intersected with a steady-state-shifted homogeneous tube to cut conservatism. On a modified IEEE 39-bus system the resulting multi-second tubes are computed in roughly a tenth of a second and largely contain high-fidelity EMT trajectories, giving operators a practical, certified envelope orders of magnitude faster than repeated EMT runs.","feed_headline":"Jordan split yields multi-second grid tubes in 0.1 s","feed_subtitle":"Certified envelopes for mixed generator-inverter grids under load swings, far faster than EMT.","key_machinery":"Real Jordan transformation of the reduced system matrix: it block-diagonalizes the dynamics into a real-eigenvalue block (handled by interval embedding) and 2-by-2 rotation-scaling blocks (handled by logarithmic-norm contraction balls), so that the full reachable set is recovered as the image of a Cartesian product of an interval and Euclidean balls.","core_discovery":"For the reduced linear ODE obtained from frequency-divider and Schur reduction of a heterogeneous transmission grid, a real Jordan decomposition lets non-oscillatory modes be over-approximated by a single embedding trajectory and oscillatory modes by explicit contraction-ball radii; the product of those sets, transformed back and optionally intersected with a steady-state-informed tube, yields a certified reachable tube under bounded power injections that can be evaluated in sub-second time for multi-second horizons.","pith_inferences":["The same Jordan-plus-contraction pattern could be reused for distribution networks once a suitable reduced linear model is derived.","Tightening the frequency-divider approximation (or restoring a low-order PLL state) would close the short post-event GFL gap without sacrificing the sub-second runtime.","If the method is extended to time-varying or set-valued voltage magnitudes, it could bound voltage-security metrics as well as frequency ones."],"forward_implications":["Operators can obtain multi-second certified frequency and power envelopes in ~0.1 s, fast enough for near-real-time contingency screening.","Sequential multi-event scenarios can be tracked by successive tube intersections without combinatorial explosion of EMT simulations.","Algebraic states (non-swing angles, GFL frequencies) are recovered a posteriori from the dynamic tube via the same linear map used in the reduction.","The same modal split applies to any linear system whose complex eigenvalues have equal algebraic and geometric multiplicity."],"fun_headline_variants":["Jordan split certifies multi-second grid tubes under 0.1 s","Real Jordan modes yield fast reach tubes for mixed grids","Sub-second tubes via Jordan embedding and contraction balls","Certified multi-second reach sets for heterogeneous power grids","Jordan transform: multi-second power tubes computed in sub-seconds"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"The reduced linear model—with fixed voltages, DC power flow, algebraic grid-following frequencies via the frequency divider, and simplified swing and governor dynamics—must stay close enough to the real stiff plant that tubes certified only for the linear ODE remain useful envelopes for high-fidelity electromagnetic-transient trajectories.","fun_headline_variants_meta":{"raw":{"variants":["Jordan split certifies multi-second grid tubes under 0.1 s","Real Jordan modes yield fast reach tubes for mixed grids","Sub-second tubes via Jordan embedding and contraction balls","Certified multi-second reach sets for heterogeneous power grids","Jordan transform: multi-second power tubes computed in sub-seconds"]},"model":"grok-4.5","effort":"low","cost_usd":0.003358,"raw_usage":{"total_tokens":1077,"prompt_tokens":728,"num_sources_used":0,"completion_tokens":86,"cost_in_usd_ticks":33580000,"prompt_tokens_details":{"text_tokens":728,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":263,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":728,"tokens_out":86,"duration_ms":3076,"temperature":1.0,"reasoning_tokens":263,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-11T05:10:58.005614+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"On the same modified IEEE 39-bus system, introduce a load or generation step whose EMT trajectories systematically and persistently leave the computed reachable tube after the brief PLL transient, or show that the Jordan-based tube computation exceeds the time of a comparable set of reduced-model trajectory roll-outs.","supporting_citations":[],"review_version":1}