{"id":"67254ae7-ba4a-4cc6-ad8e-d9932129cf2b","arxiv_id":"1908.00752","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A power network is stable if each bus is output-feedback passive with an index exceeding the network's passivity shortage, computed from an energy function's Hessian at the operating point.","lead":"Here is a simple per-bus rule for keeping a power grid stable when different buses are controlled by different kinds of devices. The rule is based on a passivity index, and the authors show it works on a small three-bus grid, but many of the mathematical proofs are deferred.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 2's stated Lyapunov argument yields only \\dot W ≤ 0, not negative definiteness; asymptotic stability is not established without an additional invariance or detectability argument, which Condition C1 does not supply.","rationale":"The reader's weakest-assumption pick, Assumption 1 (lossless network), is a real scope limitation, but the theorem explicitly restricts itself to that case, so it is not a correctness risk within the paper's stated domain. The more load-bearing issue is that Theorem 2's asymptotic-stability conclusion is not established even inside the lossless setting: the suggested Lyapunov function has only a nonpositive derivative, and no LaSalle or detectability argument is supplied. The reader did note that the main theorem is stated without proof, so there is partial agreement, but the reader did not identify the specific derivative-semidefiniteness problem. The paper has credible supporting evidence — the passivity framework is standard, and the 3-bus eigenvalue sweeps and transient simulations are consistent with the claimed threshold — so I would not reject it; I would keep it conditional pending a complete proof or a corrected theorem with an additional detectability condition.","tokens_in":9440,"tokens_out":8849,"duration_ms":101000,"concrete_test":"Complete the proof of Theorem 2 by explicitly computing \\dot W from (7) and (10) and characterizing the largest invariant set contained in {\\dot W = 0} under (4) with Condition C1. If this set contains any point with y ≠ y* or with internal states not converging to x*, then Theorem 2 is false as stated; in that case, the theorem should be reissued with an additional strict-dissipativity or detectability assumption, and the three propositions should be rechecked for that assumption. As a numerical companion, run a small perturbation of the Section V lossless 3-bus system at σ = −λ + ε and monitor \\dot W over time: if \\dot W remains at machine precision while the state oscillates without decaying, the Lyapunov argument is insufficient for asymptotic stability.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (Theorem 2) is supported only by the assertion that W(x) = Σ_i S_i(x_i) + S_N(y) + ((σ+λ−ε)/2)||y−y*||² is a Lyapunov function. Combining Definition 4 (inequality (7)) and Lemma 1 (equation (10)) gives, with ΔP_i = P_i−P_i* and ΔQv_i = Q_i/V_i − Q_i*/V_i*, Σ_i \\dot S_i ≤ −[ΔP;ΔQv]ᵀ \\dot y − σ(y−y*)ᵀ \\dot y, and \\dot S_N = [ΔP;ΔQv]ᵀ \\dot y − (λ−ε)(y−y*)ᵀ \\dot y. Adding the derivative of the quadratic term yields only \\dot W ≤ 0. This is negative semidefinite, not negative definite; equality occurs whenever the passivity inequalities in (7) are tight. Proving asymptotic stability therefore requires a LaSalle invariance argument or a zero-state detectability property of the heterogeneous bus dynamics. Condition C1 contains no such property, and the paper gives no characterization of the largest invariant set inside {\\dot W = 0}. The omitted proof is thus not a routine space-saving detail: the supplied argument establishes at best Lyapunov stability, and even that requires S_N to be locally positive definite, which depends on the unproved Lemma 1. Because Theorem 2 is the basis for the entire distributed stability criterion, this gap is load-bearing.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a local passivity-index condition for system-wide stability of lossless power networks with heterogeneous, nonlinear bus dynamics. It introduces a supply rate that depends on the derivative of the output, defines output-feedback passivity with a passivity index relative to an equilibrium, and identifies the network's passivity index λ as the smallest nonzero eigenvalue of the Hessian of a network energy function. The central claim, Theorem 2, is that if every bus is output-feedback passive with index σ > −λ and has a storage function with a strict local minimum at the equilibrium, then the equilibrium is asymptotically stable. The paper gives controller designs for a synchronous generator and for two types of droop-controlled inverters that satisfy the condition, and it reports 3-bus simulations in lossless and lossy cases, including an eigenvalue-based tightness study and critical-clearing-time comparisons. The proof of Theorem 2 and all supporting propositions are omitted due to space limits, and the provided Lyapunov-function sketch appears to yield only a semidefinite derivative, so the main stability claim is not established in the submitted text.","tokens_in":9746,"tokens_out":8373,"duration_ms":89322,"significance":"If the main theorem can be fully proved and made correct, the approach is significant: it would give a scalable, model-agnostic sufficient condition for stability of heterogeneous power systems, with explicit controller-design guidance and a testable threshold σ > −λ. The paper's strengths are its falsifiable small-signal predictions, the parameter-sweep verification around the predicted threshold, and the absence of fitted constants. However, the central proof is incomplete, the network storage lemma appears to be internally inconsistent, and the advertised 'distributed' nature of the condition is weakened by the global dependence of λ. These issues are load-bearing; the contribution is conditional on substantial revision.","major_comments":[{"comment":"The proof sketch is insufficient for the claimed asymptotic stability. Combining Definition 4 (inequality (7)) and Lemma 1 (equation (10)) yields for W(x) = Σ_i S_i(x_i) + S_N(y) + ((σ+λ−ε)/2)||y−y*||² only the inequality Ẇ ≤ 0, not Ẇ < 0. Asymptotic stability therefore requires either a strict-decrease argument or a LaSalle invariance argument together with a detectability or invariant-set condition. Condition C1 contains no detectability assumption, and the paper does not characterize the largest invariant set inside {Ẇ = 0}. Because Theorem 2 is the central claim, this gap is load-bearing: please supply the complete proof or a corrected statement.","section":"Section III-B, Theorem 2"},{"comment":"Lemma 1's assertion that S_N(y) > 0 for all y ≠ y* in a neighborhood is inconsistent with the rotational symmetry of W_N. Along the curve y = y* + α col(1_n, 0_n), one has W_N(y) = W_N(y*) and (y−y*)^T ∇W_N(y*) = 0, so \tilde W_N(y) = 0 and S_N(y) = −((λ−ε)/2) n α² < 0 for any α ≠ 0. Thus S_N is not positive definite in any neighborhood of y* unless the angle reference is fixed or the analysis is performed on the quotient space modulo global phase shifts. The manuscript does not state such a restriction, so Lemma 1 is false as written; this also affects the positive definiteness of the total Lyapunov function.","section":"Section III-A, Lemma 1"},{"comment":"The criterion is not distributed in the advertised sense. The threshold λ is the smallest nonzero eigenvalue of ∇² W_N(y*), which depends on the full network topology, line parameters, and the equilibrium voltage-angle profile. A single bus cannot evaluate σ > −λ from local information unless λ is broadcast as a network-wide parameter or bounded by locally computable quantities. Please either qualify the 'distributed' claim, provide a local bound on λ, or present a distributed estimation procedure.","section":"Section III-A/B and Abstract"},{"comment":"The paper omits the proofs of Lemma 1, Theorem 2, and Propositions 3–5, all of which are central to the claimed results. The phrase 'omitted here due to space limit' is not acceptable for a journal submission; in particular, the storage-function calculations for Propositions 3–5 must be written out to verify that inequality (7) is satisfied with the stated gains, and the derivative computation for Lemma 1 must be shown explicitly.","section":"Sections III and IV"}],"minor_comments":[{"comment":"In the sentence 'for each scale factor s, we set σ = −λ + ρ where ρ rangers from −1 to 1', 'rangers' should be 'ranges'.","section":"Section V-B"},{"comment":"The word 'receptively' in Table III's caption should be 'respectively'.","section":"Section V-C"},{"comment":"The row 'Transmission Lines x r 0.12, 0.01' should be formatted as x = 0.12, r = 0.01 to avoid ambiguity.","section":"Table I"},{"comment":"The formula for the state space, 'X µ×R×R>0', appears corrupted; it should be written as a product of the auxiliary state space, R, and R_{>0}.","section":"Section II-A"},{"comment":"The phrase 'the excess of passivity of each bus dynamics percolates into the network' is informal; consider replacing it with a precise statement about the interconnection of storage functions.","section":"Section III-B, Remark 4"}],"recommendation":"major_revision","confidential_remarks":"The main idea is potentially publishable, but the current manuscript is closer to a detailed extended abstract than a complete journal paper. The most serious issues are the unproved and apparently incomplete Lyapunov argument for Theorem 2 and the false positive-definiteness claim in Lemma 1 caused by rotational symmetry. Both must be addressed with full proofs, and the global nature of λ should be disclosed honestly. If the authors can fix these points, the result would be a worthwhile contribution; in the present form I cannot recommend acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a plausible passivity-index stability certificate for heterogeneous power grids, with a real idea and honest simulations, but the central theorem is not actually proved in the text. The displayed Lyapunov argument only gives \\dot W ≤ 0, not asymptotic stability, and the “distributed” label oversells a condition that needs the global smallest-nonzero-eigenvalue λ of the network energy Hessian.\n\nWhat is genuinely new: the derivative supply rate in (6), the network passivity index λ defined from the Hessian of the network energy, and the sufficient condition σ > −λ per bus. The three controller designs for synchronous generators and two kinds of droop inverters are concrete and checkable. The 3-bus eigenvalue sweep in Fig. 5 is a good piece of evidence: it suggests the condition is almost necessary in the lossless case, and the CCT tables show larger σ improves transient stability. That is real, reproducible-looking support, not curve fitting.\n\nSoft spots, in order of size. First and load-bearing: Theorem 2 is stated with “proof omitted.” Combining Definition 4 with Lemma 1, the candidate W = ΣS_i + S_N + ((σ+λ−ε)/2)||y−y*||² yields \\dot W ≤ 0, not negative definiteness. Equality can occur whenever the passivity inequality is tight. Asymptotic stability needs a LaSalle invariance argument or a zero-state detectability property, and Condition C1 contains neither. The paper asserts asymptotic stability, but the supplied argument reaches at best Lyapunov stability. Lemma 1, which gives the positive definiteness of S_N and the derivative formula, is also unproved. This is not a routine space-saving omission; it is the core gap.\n\nSecond, calling the condition distributed is generous. λ is a global quantity — the smallest nonzero eigenvalue of the network Hessian at the equilibrium — and every bus uses the same global threshold σ > −λ. A local bound on λ or a distributed way to compute it would make the title honest. Third, Assumption 1 (lossless network) is central. The lossy simulations show the boundary shifts, so the theorem does not cover them; the concluding claim that the condition “can still be valid” with resistance is a simulation observation, not a theorem.\n\nThe citation pattern is fine: passivity/dissipativity background is standard, and the self-citation [26] is to related work by the same group. The paper shows clear thinking and the simulations are consistent with the proposed condition, so the gaps look fixable rather than fatal. Who should read it: people working on passivity-based or compositional stability analysis for inverter-rich grids, and anyone who wants a model-agnostic stability certificate. It deserves a serious referee — the idea is worth referee time — but the referee should demand full proofs of Lemma 1 and Theorem 2, and a decision between adding a detectability/invariance argument or revising the stability claim. My recommendation: send it out, expect major revision.","headline":"A plausible passivity-index stability certificate with a load-bearing proof gap: the main theorem's Lyapunov argument yields only semidefinite derivative, and the distributed condition depends on a global eigenvalue.","tokens_in":10243,"tokens_out":3617,"would_cite":false,"duration_ms":32353,"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 per-bus passivity index can certify grid-wide stability.","keywords":["power system stability","passivity index","output feedback passivity","heterogeneous bus dynamics","distributed stability analytics","lossless power network","droop control","synchronous generator"],"falsifier":"On the paper's 3-bus test system at a fixed loading, tune every bus to a passivity index exactly $-\\lambda + \\delta$ with $\\delta > 0$ arbitrarily small, then compute the full Jacobian's eigenvalues; if any eigenvalue crosses into the right half-plane while all buses still satisfy $\\sigma > -\\lambda$, Theorem 2 is false. The same procedure with $\\delta < 0$ should make the system unstable if the condition is genuinely tight.","tokens_in":9212,"feed_emoji":"⚡","tokens_out":5970,"duration_ms":50501,"temperature":0.7,"pith_summary":"This paper tries to turn system-wide stability analysis of power grids from a centralized, model-specific computation into a local check on each connected device. It claims that in a lossless transmission network, an equilibrium is asymptotically stable whenever every bus's dynamics is output feedback passive with a passivity index larger than the network's own passivity shortage, measured by the smallest nonzero eigenvalue of the network energy function's Hessian. Because the condition is stated in terms of input-output behavior rather than detailed equations, it applies to heterogeneous nonlinear bus models—synchronous generators and droop-controlled inverters included—and can be met by tuning existing local controllers. If right, this gives a scalable stability certificate for grids with many diverse power-electronic devices, and the paper's 3-bus eigenvalue study suggests the condition is close to necessary.","feed_headline":"A per-bus passivity index can certify grid-wide stability","feed_subtitle":"In lossless grids, system-wide stability follows when every bus's passivity index exceeds the network's deficit—no global model required.","key_machinery":"The carrying object is the passivity index of a dynamical system with respect to a specialized supply rate that includes derivatives of the output. A bus is OFP($\\sigma$) when there is a storage function $S_i$ satisfying $\\dot{S}_i \\le -(P_i-P_i^*)\\dot{\\theta}_i - (Q_i/V_i - Q_i^*/V_i^*)\\dot{V}_i - \\sigma(y_i-y_i^*)^T\\dot{y}_i$. The network's passivity index is $\\lambda$, the smallest nonzero eigenvalue of the Hessian of the network energy function $W_N(y) = \\sum_i -\\frac{1}{2}B_{ii}V_i^2 - \\sum_{(i,j)} B_{ij}V_iV_j\\cos\\theta_{ij}$ at the equilibrium; the zero eigenvalue corresponds to rotational symmetry of phase angles. Theorem 2 works by combining each bus's storage function with a shifted network storage function into the Lyapunov candidate $W = \\sum_i S_i + S_N + \\frac{\\sigma+\\lambda-\\varepsilon}{2}\\|y-y^*\\|^2$, whose decrease follows once every $\\sigma$ exceeds $-\\lambda$.","core_discovery":"The central claim is Theorem 2: for any equilibrium of the interconnected system (4), if every bus dynamics is output feedback passive in the sense of Definition 4 with passivity index $\\sigma > -\\lambda$, where $\\lambda$ is the smallest nonzero eigenvalue of the Hessian of the network energy function $W_N$ at the equilibrium, and each storage function has a strict local minimum at the equilibrium, then the equilibrium is asymptotically stable. The paper introduces a supply rate with differential at one port, $-(P_i-P_i^*)\\dot{\\theta}_i - (Q_i/V_i - Q_i^*/V_i^*)\\dot{V}_i$, and proves the network itself is dissipative with passivity index $\\lambda$. The stability proof assembles a Lyapunov function from the bus storage functions plus a shifted network storage function. The authors verify on a 3-bus system with three different device types that violating $\\sigma > -\\lambda$ marks the boundary of small-signal stability in the lossless case, and that lossy lines require a slightly larger index.","pith_inferences":["The input-output nature of C1 suggests the same certificate could be applied to devices whose internal models are unknown or proprietary, so long as their passivity index can be measured or estimated from terminal behavior.","The monotone relation between $\\sigma$ and critical clearing time hints that passivity index could serve as a tunable robustness margin for transient stability, not just a yes/no stability certificate.","One could test whether the network's passivity shortage $\\lambda$, computed from the energy Hessian, can be decomposed line by line, which would make the whole stability check fully distributed down to individual transmission lines rather than requiring knowledge of the full network.","Extending the supply-rate framework to resistive networks might be possible by treating conductance as an explicit sink of passivity, yielding a modified threshold that depends on the conductance matrix instead of the lossless $\\lambda$."],"forward_implications":["If Theorem 2 is correct, engineers can certify stability of a heterogeneous grid bus by bus, using only each device's input-output passivity index and the network's smallest nonzero energy-Hessian eigenvalue.","The same condition doubles as a controller-tuning rule: the paper gives explicit PI gains for synchronous generators and droop-gain bounds for two inverter types that make the device satisfy C1.","Because the condition is nearly necessary in the lossless 3-bus study, the gap between sufficient and necessary conditions appears small; stability margins essentially coincide with the passivity index boundary.","A larger passivity index beyond the threshold also gives better transient performance, measured by longer critical clearing times in the simulations.","In lossy networks the theorem's bound no longer holds; the lossy simulations require a larger $\\sigma$, indicating resistance acts as an additional passivity demand."],"supporting_citations":[{"why":"Supplies the definitions of dissipativity, passivity, and output feedback passivity with passivity index $\\sigma$ that the paper's condition C1 builds on.","marker":"[19]"},{"why":"Provides the network energy function and the observation that its Hessian has a zero eigenvalue from phase-angle rotational symmetry, giving the network's passivity index $\\lambda$.","marker":"[24]"},{"why":"Supplies the flux-decay synchronous generator model (11) and the energy-based stability framework whose network-preserving view the paper extends.","marker":"[18]"},{"why":"Provides the conventional droop-controlled inverter model (13), one of the three bus dynamics for which passivity-percolating control is designed.","marker":"[8]"},{"why":"Supplies the quadratic droop controller model (15) and its voltage stabilization setting, used as another example of C1 satisfiable by control.","marker":"[22]"},{"why":"Cites direct energy-function methods for power system stability analysis as the centralized baseline that the distributed condition is meant to replace.","marker":"[4]"},{"why":"Gives the port-Hamiltonian approach to power network modeling that motivates the differential supply rate and network storage function.","marker":"[15]"},{"why":"Relates the differential supply rate (6) to phasor-circuit and Brayton-Moser power formulations, grounding the specialized passivity notion.","marker":"[26]"}],"fun_headline_variants":["Per-bus passivity index certifies grid-wide stability","Local passivity indices guarantee whole-system stability","Heterogeneous buses: local passivity → global stability","No global model needed: per-bus indices ensure stability","One index per bus: system-wide stability in lossless grids"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is Assumption 1, that the transmission network is lossless (zero conductance); if line resistance is appreciable, the theorem's threshold $\\sigma > -\\lambda$ no longer applies, and the paper's own lossy simulations need a larger passivity index for stability.","fun_headline_variants_meta":{"raw":{"variants":["Per-bus passivity index certifies grid-wide stability","Local passivity indices guarantee whole-system stability","Heterogeneous buses: local passivity → global stability","No global model needed: per-bus indices ensure stability","One index per bus: system-wide stability in lossless grids"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000217,"raw_usage":{"total_tokens":1429,"prompt_tokens":930,"completion_tokens":499,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":546,"completion_tokens_details":{"reasoning_tokens":420}},"tokens_in":546,"tokens_out":499,"duration_ms":5066,"temperature":1.0,"reasoning_tokens":420,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:33:22.598131+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On the paper's 3-bus test system at a fixed loading, tune every bus to a passivity index exactly $-\\lambda + \\delta$ with $\\delta > 0$ arbitrarily small, then compute the full Jacobian's eigenvalues; if any eigenvalue crosses into the right half-plane while all buses still satisfy $\\sigma > -\\lambda$, Theorem 2 is false. The same procedure with $\\delta < 0$ should make the system unstable if the condition is genuinely tight.","supporting_citations":[{"cited_title":"Synchronization in complex oscillator networks and smart grids,","cited_arxiv_id":null,"evidence_quote":"Provides the network energy function and the observation that its Hessian has a zero eigenvalue from phase-angle rotational symmetry, giving the network's passivity index $\\lambda$."}],"review_version":1}