{"id":"37e2978f-854c-445b-b203-2f2d428fca8c","arxiv_id":"2504.18905","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A tighter convex inner approximation plus fairness-aware allocation indicates that roughly 5% annual solar curtailment can increase a test feeder's solar hosting capacity by about 50% without line or voltage violations, with positive net profit when carbon revenue is counted.","lead":"This paper improves a method for calculating how much rooftop solar a distribution grid can safely host, using tighter math bounds and fairness-aware allocation. On a simulated 36-bus feeder, it shows that allowing roughly 5% annual solar curtailment can raise the safe solar capacity by about 50%, with carbon savings that offset curtailment costs.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'no negative grid impacts' claim is only as strong as the perfect 5-minute demand forecast used to compute each dynamic hosting capacity bound; Section 5 leaves this as future work, so the abstract's 50%/5% result is not yet robust to realistic forecast error.","rationale":"I read the paper as a methodological extension of [12] plus a case study. The SOC epigraph bound in Eq. (10) is a standard rotated-cone relaxation, and the fairness-aware formulations in Eqs. (14)-(15) are coherent; I do not see an internal proof error that would invalidate the static CIA construction. The main soft spot is the transfer from perfect-foresight 5-minute calculations to the unconditional headline. The reader's weakest assumption identifies the same issue, and the manuscript's Section 5 acknowledges it, which is why the concern is load-bearing but not fatal to the paper's mathematical claims. The quantitative economic claim also deserves a caveat because capital costs are excluded (Section 4.4.1), but I treat that as secondary to the operational guarantee. Since no code or data is shipped, the simulation results cannot be independently checked, reinforcing a CONDITIONAL rather than ACCEPT verdict. My recommendation is to keep the reader's CONDITIONAL verdict; the concern would be fully settled by the forecast-error backtest described above.","tokens_in":20429,"tokens_out":22691,"duration_ms":260105,"concrete_test":"Backtest on the same modified IEEE-37 feeder: for each 5-minute interval, compute p+_g(t) using a one-step-ahead forecast of each node's net demand (e.g., persistent forecast plus zero-mean Gaussian noise with 5% and 10% standard deviation), apply the curtailment rule P_curt = max(0, P_new - p+_forecast) to the actual PV/demand traces, and run an AC power-flow check for voltage and branch-current limits over the full year. If any violations occur in the lower-demand (higher-PV) error cases, the abstract's 'no negative grid impacts' must be qualified as forecast-dependent; if the box remains admissible under the tested errors, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim in the abstract ('increase PV hosting capacity by at least 50% with no negative grid impacts') is built on the CIA guarantee that every injection inside the dynamic hyperrectangle is AC-admissible. That guarantee is conditional on the exact nodal demand profile used when solving (P1) at each 5-minute interval. The curtailment rule in Eq. (24) caps PV output at p+_g,i(t), but if the actual load at that interval is lower than the value used to compute p+ (or the forecast is stale), the true injection vector can sit outside the rectangle for which admissibility was certified. The paper's own Section 5 lists 'uncertainty in net-demand at each node' and 'geographically diverse solar PV generation profiles' as future work, so the manuscript concedes the gap. Yet the headline statement is presented without this caveat. This is not a criticism of the SOC bound or the fairness formulation; rather, it means the demonstrated 50%/5% result is a perfect-foresight planning result, not a robustness guarantee for field deployment. Without a forecast-error study, the phrase 'no negative grid impacts' overstates what the simulations establish.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper extends the convex inner approximation (CIA) method of [12] for radial distribution feeders by replacing the Taylor-based upper bound on branch currents with a second-order-cone (SOC) epigraph constraint that uses four corner combinations of the proxy power-flow variables. It then formulates dynamic hosting capacity (DHC) as a sequence of convex programs and introduces an ε-fairness constraint based on the L1-L2 norm inequality, with two objective variants (linear/logarithmic and demand-weighted). On a modified IEEE-37 feeder with one year of 5-minute demand and solar data, the authors report that a 50% increase in PV capacity is possible with roughly 5% annual solar curtailment, and that a demand-proportional fair allocation (Scenario 1-F2, ε=0.85) yields higher net profit when curtailment cost is balanced against carbon revenue at $100/tCO2. The paper also quantifies temporal and spatial Jain Fairness Index values and compares results across six allocation scenarios, demand levels, and carbon prices.","tokens_in":20630,"tokens_out":13704,"duration_ms":131605,"significance":"If the numerical results are corrected, the paper makes a useful contribution: the SOC upper bound is a valid convex restriction and is shown on test systems to be tighter than the bound from [12]; the ε-fairness constraint provides a tunable and convex way to control spatial equity in host-capacity allocation; and the case study addresses a practically important trade-off between curtailment, hosting capacity, and carbon value. The paper also deserves credit for benchmarking the inner approximation against actual AC power-flow solutions via MatPower sweeps (Figs. 3–4) and for explicitly relating ε-fairness to Jain's Fairness Index in Eq. (16). However, the headline quantitative claims currently rest on a unit inconsistency in the carbon calculation, a numerical claim in the abstract that is not supported by Table IV, and an unqualified 'no negative grid impacts' statement that assumes perfect 5-minute foresight.","major_comments":[{"comment":"The carbon-avoidance equation mixes units. In the text, P_add is in MW, Δt is in hours (5 minutes), mgrid is defined as gCO2/MWh near Eq. (30), and mpv is defined as gCO2/kWh in the same section; these quantities cannot be subtracted as written. If both emission factors are intended in gCO2/kWh, then the product P_add (MW) times Δt (h) must be multiplied by 1000 to convert MWh to kWh before applying the g-to-tonne factor of 10^-6, and that factor of 1000 is missing. If mgrid is truly in gCO2/MWh, then mpv must be converted to the same units. Either way, the carbon revenue and net-profit numbers in Table IV and Figs. 17–19 are not reliable as reported. Please correct the units, re-run the economic calculations, and update the abstract and conclusions accordingly.","section":"Section 4.4, Eq. (30)"},{"comment":"The abstract states that 'with no more than 5% annual solar PV energy curtailed, it is possible to increase solar PV hosting capacity by at least 50%.' Table IV, however, reports the curtailed-energy percentage at the 50% capacity-increase point as 6.5% for Scenario 3, 6.9% for Scenario 4 and Scenario 1-F2, 7.3% for Scenario 1-F1, and up to 10% for Scenario 1. No scenario in the table satisfies the 'no more than 5%' condition at 50% increase. Please reconcile the numerical claim with the reported results, or report the capacity-increase level at which curtailment is actually no more than 5% and adjust the abstract and conclusions accordingly.","section":"Abstract and Table IV"},{"comment":"The 'no negative grid impacts' guarantee is certified only for the exact 5-minute nodal demand profile and the single shared PV profile used to solve (P1). The curtailment rule in Eq. (24) caps PV output at p+_g,i(t), but if the actual demand at an interval is lower than the value used in the optimization, or if solar output differs across nodes, the realized injection vector can lie outside the hyperrectangle for which AC admissibility was certified. Since Section 5 lists uncertainty in net-demand and geographically diverse PV generation profiles as future work, the abstract's unconditional 'no negative grid impacts' statement overstates what the simulations establish. Please add a forecast-error sensitivity study or explicitly qualify the result as a perfect-foresight planning result.","section":"Section 5 and Eq. (24)"},{"comment":"The paper claims that the proposed SOC upper bound satisfies l+_SOC ≤ l+_from[12] for all branch currents and uses this dominance to argue for a larger hyperrectangle in Table I. This dominance is only demonstrated by simulation for one injection pattern on the IEEE-37 network; no proof is given that the SOC bound always dominates the Taylor-based bound from [12]. If the dominance is not guaranteed, then the 'more accurate and larger inner approximation' claim is case-dependent. Please either provide a proof or state the comparison as an empirical observation and soften the corresponding claims in the introduction and Section 2.4.","section":"Section 2.4, Eq. (10) and Fig. 4"}],"minor_comments":[{"comment":"The choice of ε=0.85 is made on the same case study and tuned for capacity increases below 60%. Since the headline result uses this value, please state the selection criterion explicitly and discuss how sensitive the main conclusions are to ε outside the tested range.","section":"Footnote 5, Section 3.2"},{"comment":"The variable mpv is introduced as a constant (40 gCO2/kWh) but appears as mpv(t) in Eq. (30); please clarify whether it is time-varying or constant.","section":"Section 4.4, Eq. (30)"},{"comment":"The inequalities in (12k)–(12l) appear reversed relative to the definitions of P+ and P− in (4), unless overline/underline notation was lost in typesetting. Please check the direction of the bound constraints or clarify the notation.","section":"Section 2.5, Eqs. (12k)–(12l)"},{"comment":"The LaTeX artifacts 'radicaltp/radicalvertex' should be replaced with a properly typeset square root, and the assumption α_i > 0 should be stated in the text before Eq. (15).","section":"Section 3.2, Eq. (15)"},{"comment":"The parenthetical 'where base case corresponds to the minimum base energy across all scenarios' is confusing, since Ebase in Eq. (25) is scenario-specific through Lpv,i in Eq. (19); please clarify the definition.","section":"Section 4.2, Eq. (25)"}],"recommendation":"major_revision","confidential_remarks":"This appears to be a conference paper submitted to a journal venue. The core SOC-bound method is mathematically sound and the fairness framework is useful, but the unit error in Eq. (30), the abstract/Table IV numerical mismatch on the 5% curtailment claim, and the unqualified perfect-foresight 'no negative grid impacts' statement are load-bearing and need to be fixed before publication. These issues are correctable within the scope of the manuscript, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a solid, incremental paper. The genuinely new piece is the SOC upper bound in Eq. (10), which replaces the conservative branch-current envelope from [12] with an epigraph relaxation using four P/Q combinations. It is a standard SOCP trick, but it works: the paper benchmarks it against MatPower AC power flows and shows the tighter envelope enlarges the admissible hyperrectangle. I believe the dominance over [12] is only demonstrated numerically, not proven, but the comparison is honest and the error plot is convincing.\n\nThe fairness part is also a real extension: taking the epsilon-fairness constraint from [19] and applying it to dynamic hosting capacity, with spatial and temporal JFI post-processing, is new. The scenario comparison is clear. The 50%-capacity-increase-for-5%-curtailment headline is a plausible planning result on the modified IEEE-37 feeder, and the carbon-revenue analysis is a reasonable thought experiment, not a field claim.\n\nSoft spots, in order of importance. First, the phrase 'no negative grid impacts' outruns the simulation. The CIA guarantee is per time step, conditional on the exact 5-minute demand and solar profiles used to compute p+_g. The curtailment rule in Eq. (24) caps PV at p+_g,i(t), but if actual load is lower than the forecast embedded in p+, the realized injection can sit outside the certified rectangle. The paper's own Section 5 lists net-demand uncertainty and geographically diverse PV as future work, so this is a conceded gap, not a hidden one. The headline should be read as 'under perfect foresight.' Second, Eq. (30) mixes units: mgrid(t) is labeled gCO2/MWh while mpv is gCO2/kWh, and the 10^-6 factor is only right on one convention. Third, epsilon = 0.85 is tuned on the same case, and the economic 'net positive' excludes capital costs. Those are moderate caveats, not fatal flaws. No code or data is shipped, so I could not reproduce the simulation figures, but the external MatPower check gives some confidence.\n\nI disagree with any reading that calls this circular. The fairness parameter is tuned, not fitted to force the conclusion, and the SOC bound is benchmarked externally. Self-citation to [12] is appropriate given the paper is a direct extension.\n\nBottom line: worth a serious referee. It should go to review, and the revision should fix the unit issue and qualify the headline. If I were working on CIA operating envelopes, I'd cite the SOC bound. I'd bring it to reading group, mostly to discuss whether the perfect-foresight caveat is acceptable for a planning result.","headline":"A modest but real tightening of the CIA branch-current bound, a clean fairness extension, and an interesting 50%-for-5% curtailment result that is a perfect-foresight simulation, not yet a robustness guarantee.","tokens_in":21220,"tokens_out":2478,"would_cite":true,"duration_ms":24233,"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":"Tighter grid bounds enable 50% more solar with 5% curtailment","keywords":["dynamic hosting capacity","convex inner approximation","radial distribution networks","solar PV curtailment","fair resource allocation","second-order cone","carbon emissions","IEEE-37 feeder"],"falsifier":"Repeat the Scenario 1-F2 dynamic hosting-capacity computation on the same feeder, then run an actual AC power-flow simulation at every 5-minute time step using nodal demand and solar measurements with, say, plus or minus 10% forecast error or with measured diverse PV profiles instead of one shared profile; if any point inside the advertised hyperrectangle violates voltage or branch-current limits, the central guarantee fails.","tokens_in":20159,"feed_emoji":"☀️","tokens_out":6118,"duration_ms":59559,"temperature":0.7,"pith_summary":"This paper argues that dynamic hosting capacity, the amount of solar PV a distribution feeder can safely accept at each time step, can expand substantially if curtailment is treated as deliberate, fairness-aware flexibility rather than avoided at all cost. It does this by tightening the convex inner approximation used to certify safe injections: branch currents are bounded with a second-order-cone constraint that lies much closer to the true AC power-flow relation than the previous conservative envelope. On a modified IEEE-37 feeder using 5-minute demand and solar data, the authors show that increasing PV capacity by 50% while curtailing no more than about 5% of annual solar energy causes no negative grid impacts, and yields net economic profit when avoided carbon dioxide is valued at $100/tCO2. A demand-proportional fairness constraint is shown to allocate capacity more equitably across nodes and to produce lower curtailment and higher net profit than unweighted or location-dominated allocations. The practical upshot is that static, conservative hosting limits may be needlessly suppressing solar adoption, and that modest planned curtailment with equitable allocation can unlock more clean energy safely.","feed_headline":"Tighter grid bounds enable 50% more solar with 5% curtailment","feed_subtitle":"Fair PV allocation and tighter convex bounds keep voltages safe; carbon revenue covers curtailment costs on an IEEE-37 feeder.","key_machinery":"The load-bearing object is the convex inner approximation (CIA): a hyperrectangle of nodal power injections inside the AC-admissible set, computed by solving the convex problem (P1) twice, once for upper and once for lower injection limits. The improvement is a new upper bound, Eq. (10), which replaces the conservative branch-current bound from the earlier method with a second-order-cone constraint, $\\|(2P^\\bullet_{ij},\\,2Q^\\bullet_{ij},\\,l^+_{ij}-v^-_i)\\|_2 \\le l^+_{ij}+v^-_i$, derived as the epigraph of $l_{ij} = (P_{ij}^2+Q_{ij}^2)/v_i$. Because the cone bound tracks the true branch-current relation much more closely, the voltage proxy $V^+$ is smaller, the feasible rectangle is larger, and every point inside it remains certified safe. Fairness is then imposed with the L1-L2 norm inequality $(1-\\epsilon+\\epsilon/\\sqrt{N})\\|p\\|_2 \\le \\|p\\|_1$, which is second-order-cone representable and explicitly enforces at least $\\epsilon$-fair allocations; with demand-proportional weights it becomes Eq. (15).","core_discovery":"The central discovery is that the SOC-based upper bound in Eq. (10) is a tighter convex envelope for branch currents than the bound used in the prior convex inner approximation, so every hyperrectangle of independently controllable injections fits inside the true AC-admissible set while having larger volume. This makes the hosting-capacity computation less conservative without sacrificing the guarantee that every point in the computed rectangle is AC-admissible. On the modified IEEE-37 feeder, the method yields a dynamic hosting capacity under Scenario 1-F2, an unweighted linear objective with demand-proportional fairness at epsilon = 0.85, such that a 50% increase over the static base PV capacity requires curtailment of only about 5 to 7% of annual solar energy, and at a $100/tCO2 carbon price the carbon revenue exceeds the curtailment cost, giving positive net profit in the range of roughly 30% capacity increase. Fairness is quantified with Jain's fairness index applied to the ratio of hosting capacity to demand, and the paper finds that fairer allocations reduce both PV curtailment and CO2 emissions compared with unfair allocations.","pith_inferences":["Editorial inference: the no-negative-grid-impacts guarantee depends on exact 5-minute demand and solar forecasts; with forecast error or geographically diverse PV profiles, the advertised safety margin would need a robust or chance-constrained reformulation, which the paper lists as future work.","Editorial inference: temporal fairness is only measured after the fact in this paper, so embedding it directly as a multi-period optimization constraint is a natural next step that would likely reduce the sharp temporal drops in hosting capacity seen for unfair scenarios.","Editorial inference: the same tightened inner approximation could be applied beyond solar hosting to storage, electric vehicles, or demand response, wherever independent DER injections must be certified AC-admissible.","Editorial inference: a utility could tune both the fairness parameter epsilon and the capacity increase against its local carbon price and curtailment tariff, turning the paper's curves into a direct investment rule."],"forward_implications":["The tighter SOC bound directly enlarges the hosting-capacity hyperrectangle, so utilities can allow more distributed solar at the same level of safety assurance.","A 50% increase in installed PV capacity can be absorbed with only about 5% annual energy curtailment, meaning curtailment need not be a barrier to aggressive solar targets.","Fair, demand-proportional allocation reduces both curtailment and CO2 emissions compared with unweighted or demand-weighted linear objectives that let some nodes dominate.","When carbon revenue is counted at $100/tCO2, net profit peaks near a 30% capacity increase and stays positive over a meaningful range, so carbon pricing can justify deliberate curtailment.","Locational grid emissions matter: the same method yields far larger carbon reductions in fossil-heavy regions than in regions with already clean generation."],"supporting_citations":[{"why":"Supplies the convex inner approximation and hyperrectangle method that this paper tightens with the new SOC-based branch-current bound.","marker":"[12]"},{"why":"Provides the DistFlow equations used to model the radial network physics in the hosting-capacity formulation.","marker":"[13]"},{"why":"Contributes the SOC-representable epsilon-fairness constraint that the paper adds to Scenarios 1-F1 and 1-F2.","marker":"[19]"},{"why":"Provides the realistic feeder demand dataset used for the time-varying 5-minute demand profiles in the case study.","marker":"[26]"},{"why":"Gives the utility curtailment usage charge used to compute the economic cost of curtailed PV energy.","marker":"[30]"},{"why":"Supplies the marginal operating emissions rate data used to estimate avoided CO2 emissions for Vermont and Eastern Ohio.","marker":"[34]"},{"why":"Provides the load-flow tool used to compute the nominal operating point around which the convex envelopes are constructed.","marker":"[24]"}],"fun_headline_variants":["50% more solar with 5% curtailment via tighter convex bounds","Tighter grid math unlocks 50% PV boost with minimal curtailment","Fair PV hosting: 50% capacity gain, <5% energy lost","Convex upgrade lets grids host 50% more solar, cut CO2","Solar hosting up 50% with just 5% curtailment, study shows"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire safety guarantee rests on the claim that every power injection inside the computed hyperrectangle is AC-admissible, which is inherited from the convex inner approximation and, in the dynamic case, presumes exact 5-minute demand and solar output at every node with a single shared solar profile and perfect foresight.","fun_headline_variants_meta":{"raw":{"variants":["50% more solar with 5% curtailment via tighter convex bounds","Tighter grid math unlocks 50% PV boost with minimal curtailment","Fair PV hosting: 50% capacity gain, <5% energy lost","Convex upgrade lets grids host 50% more solar, cut CO2","Solar hosting up 50% with just 5% curtailment, study shows"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000738,"raw_usage":{"total_tokens":3308,"prompt_tokens":965,"completion_tokens":2343,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":581,"completion_tokens_details":{"reasoning_tokens":2240}},"tokens_in":581,"tokens_out":2343,"duration_ms":16800,"temperature":1.0,"reasoning_tokens":2240,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T10:08:06.592668+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Repeat the Scenario 1-F2 dynamic hosting-capacity computation on the same feeder, then run an actual AC power-flow simulation at every 5-minute time step using nodal demand and solar measurements with, say, plus or minus 10% forecast error or with measured diverse PV profiles instead of one shared profile; if any point inside the advertised hyperrectangle violates voltage or branch-current limits, the central guarantee fails.","supporting_citations":[{"cited_title":"Grid-aware aggregation a nd real-time disaggregation of distributed energy resources in radial n etworks,","cited_arxiv_id":null,"evidence_quote":"Supplies the convex inner approximation and hyperrectangle method that this paper tightens with the new SOC-based branch-current bound."},{"cited_title":"Optimal sizing of capacitors plac ed on a radial distribution system,","cited_arxiv_id":null,"evidence_quote":"Provides the DistFlow equations used to model the radial network physics in the hosting-capacity formulation."},{"cited_title":"A Parametric, Second-Order Cone Representable Model of Fairness for Decision-Making Problems","cited_arxiv_id":"2412.05143","evidence_quote":"Contributes the SOC-representable epsilon-fairness constraint that the paper adds to Scenarios 1-F1 and 1-F2."},{"cited_title":"Hierarchical, grid-aware, and economically optimal coo rdination of distributed energy resources in realistic distribution systems,","cited_arxiv_id":null,"evidence_quote":"Provides the realistic feeder demand dataset used for the time-varying 5-minute demand profiles in the case study."},{"cited_title":"Rates and Regulations,","cited_arxiv_id":null,"evidence_quote":"Gives the utility curtailment usage charge used to compute the economic cost of curtailed PV energy."},{"cited_title":"WattTime Data Signals API,","cited_arxiv_id":null,"evidence_quote":"Supplies the marginal operating emissions rate data used to estimate avoided CO2 emissions for Vermont and Eastern Ohio."},{"cited_title":"Powe rModels.jl: An Open-Source Framework for Exploring Power Flow Formulat ions,","cited_arxiv_id":null,"evidence_quote":"Provides the load-flow tool used to compute the nominal operating point around which the convex envelopes are constructed."}],"review_version":1}