{"id":"469483df-25bd-4592-9e01-21f597d20d0d","arxiv_id":"2506.12102","paper_version":1,"verdict":"UNVERDICTED","confidence":"MODERATE","novelty_score":2.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"PVsyst performs as a digital shadow, while SAM can be extended toward a digital twin, based on a case-study comparison of simulated and real photovoltaic production.","lead":"This report compares the solar simulation programs PVsyst and SAM and places them on a scale from static model to digital twin. Using a real rooftop plant in Turkey for reference, it concludes that SAM is the better candidate for digital twin integration, though neither tool is a full digital twin on its own.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The recommendation that SAM is the better digital-twin base is unsupported: the paper never demonstrates bidirectional feedback for SAM, and its own Table 7 marks physical feedback as absent for both tools.","rationale":"The reader's weakest_assumption identifies the same soft spot: the maturity classification is driven by a qualitative feature checklist, not by demonstrated bidirectional feedback. My review confirms and sharpens that concern. The paper's own Table 7 shows that neither simulator possesses native physical feedback, so the distinction between SAM and PVsyst reduces entirely to an asserted extensibility advantage. The quantitative results in Section 4 do not test real-time adaptation and, if anything, show PVsyst with a smaller annual error, so they do not rescue the claim. This is a genuine load-bearing concern because the paper's central conclusion is a recommendation to use SAM as the motor of a solar digital twin. A conditional verdict is appropriate: the recommendation should hold only if the claimed closed-loop capability is actually demonstrated, ideally through a prototype integration of the kind described in the concrete test. This is not an ad hominem critique; the paper is transparent about its qualitative basis, but the central claim nevertheless depends on an unverified architectural assumption. The proposed test is concrete and would settle the matter directly.","tokens_in":9376,"tokens_out":4376,"duration_ms":46871,"concrete_test":"Implement a minimal closed-loop Digital Twin prototype using PySAM (SAM's Python API) that ingests a live or replayed time series of measured irradiance and temperature, automatically updates model parameters (e.g., derate factors or inverter efficiency), and sends a corrective action to a simulated plant control interface, with no manual steps. Repeat the same exercise using PVsyst's scripting/API capabilities. If the SAM loop requires an external orchestrator for data ingestion and feedback, or if PVsyst can be scripted to achieve the same loop with comparable effort, then the claimed architectural advantage of SAM over PVsyst for DT integration is not substantiated, and the recommendation should be downgraded.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim (Section 8) is that SAM's scripting flexibility and API access make it structurally closer to a Digital Twin than PVsyst, which is placed at best as a Digital Shadow. The load-bearing assumption is that the feature checklist in Table 5/Table 7 correctly proxies DT readiness. But the paper's own comparison (Table 7) assigns '×' to both PVsyst and SAM for 'Retroalimentación al sistema físico', the defining characteristic of a DT. SAM is credited with only 'Parcial (vía scripting)' for dynamic updating and 'Parcial (API externa)' for IoT integration, yet the conclusion says SAM 'puede extenderse hacia una configuración DT'. This extension is asserted, not demonstrated. The quantitative study in Section 4 compares static energy estimates (Table 6) and actually favors PVsyst in annual accuracy (177,523 kWh simulated vs 175,515 kWh real, versus SAM's 163,559 kWh); it provides no evidence that SAM's scripting enables closed-loop adaptation. Therefore the central recommendation rests on an unvalidated assumption that external integration with monitoring platforms and AI is both feasible and meaningfully easier with SAM than with PVsyst. If that assumption fails, the paper's main conclusion loses its foundation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reviews the digital-twin concept for photovoltaic systems, adopts the Digital Model/Digital Shadow/Digital Twin (DM/DS/DT) taxonomy, and applies it to two widely used simulators, PVsyst and SAM. Using a reference rooftop installation in Elmali, Turkey, the authors replicate the system in both tools with a common NSRDB weather input, compare monthly simulated energy with reported real data in Table 6, and compute RMSE/MAE-type errors. On the basis of a feature checklist (Tables 5 and 7), the paper concludes that PVsyst operates at best as a Digital Shadow, while SAM has a more flexible architecture that can be extended toward a Digital Twin when combined with monitoring platforms, APIs, and artificial-intelligence modules (Section 8).","tokens_in":9609,"tokens_out":6428,"duration_ms":78102,"significance":"If the central qualitative claim is accepted, the paper provides a useful, practical screening of two commercial simulation tools for digital-twin projects and a concrete case-study benchmark. Its strengths include a real installation with comparatively detailed panel and inverter parameter tables, the use of the same NSRDB weather data in both simulators, a transparent DM/DS/DT checklist, and an explicit admission that neither tool is, by itself, a complete digital twin. The contribution is primarily an expert-assessment/maturity-classification exercise; the quantitative energy comparison is ancillary to the digital-twin conclusion. The result is plausible but currently rests on a small number of subjective checklist entries rather than on a demonstrated closed-loop workflow.","major_comments":[{"comment":"The central conclusion that SAM has greater structural maturity for a Digital Twin is based on the entries 'Actualización dinámica con datos reales: Parcial (vía scripting)' and 'Soporte para integración IoT: Parcial (API externa)' in Table 7, but no evidence or reference is provided showing that SAM's Python/MATLAB scripting can ingest live operational data and update the model accordingly, and Table 7 itself marks 'Retroalimentación al sistema físico' as absent for both tools. As written, the conclusion in Section 8 goes beyond the documented capabilities. The authors should either soften the claim to 'SAM is a more promising candidate for future integration, provided an external control layer is built,' or demonstrate a minimal closed-loop example, such as using PySAM to read a time-series file and re-optimize a parameter, and compare that with what PVsyst can or cannot do.","section":"§5, Table 7; §8"},{"comment":"The maturity rubric is not operationalized. Table 5 uses binary checkmarks for DM/DS/DT, but Table 7 introduces unquantified gradations such as 'Parcial (vía scripting)', 'Parcial (API externa)', 'Limitada', and 'Medio–Alto' without defining thresholds, weights, or an aggregation rule. Consequently, the central ranking of SAM as closer to DT than PVsyst is not reproducible from the stated criteria. The authors should define a scoring protocol, specify what 'Parcial' requires (e.g., a demonstrable data ingestion path, but no automatic feedback), justify each Table 7 entry with a reference or a test, and explain how the entries map onto the DM/DS/DT levels in Table 5.","section":"§3.4, Tables 5 and 7"},{"comment":"Section 8 states that 'El análisis comparativo también evidenció ... limitaciones importantes en su capacidad de adaptación dinámica,' but the comparisons in Section 4 and Table 6 are static monthly energy estimates; they contain no test of dynamic updating, real-time data ingestion, or bidirectional feedback. The digital-maturity ranking therefore derives entirely from the feature checklist in Section 3.4/Table 7, not from the simulation results, and the current wording conflates the two evidence bases. The paper should state this explicitly and either remove the implication that the quantitative study supports the maturity assessment or add a genuinely dynamic experiment.","section":"§4 and §8"}],"minor_comments":[{"comment":"The accuracy comparison should include units in the table header, a description of how the 'Real' energy values were obtained (metering, inverter logs, or the source paper [15]), and an estimate of measurement uncertainty. The monthly residuals in Table 6 show substantial seasonal bias (e.g., SAM under-predicts January by about 25% and over-predicts June by about 22%), which is not discussed; an analysis of this bias and its possible causes (degradation, soiling, inverter clipping, weather-data mismatch) would strengthen the claim that both simulators give 'razonable' annual estimates.","section":"§4, Table 6"},{"comment":"There are numerous typographical and formatting issues: the byline contains 'YGEMELOSDIGITALES' as a single word, 'Pv-syst' and 'PVsyst' are used inconsistently, Table 4 has 'MPTT' instead of 'MPPT' and 'V oltaje' instead of 'Voltaje', Section 6 bullets begin with '.)', and 'imágen' appears instead of 'imagen'. A careful proofread is needed.","section":"Throughout"},{"comment":"Several figures (Figures 5–10, 11–17) are not explicitly referenced in the text with the expected 'Figura N' callouts; add in-text references so the reader can connect the screenshots and error plots to the corresponding simulation steps and metrics.","section":"§3.3 and §4 figures"},{"comment":"References [18] and [19] appear to be web documentation rather than peer-reviewed sources; if used to support the claim that SAM supports scripting and MATLAB linkage, cite the official PySAM documentation with version and access date. Reference [15] is the likely source of the real plant data, but this should be stated explicitly in Section 3.2 rather than only appearing in the bibliography.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper is a competent exploratory comparison with a useful real-case dataset, but it reads more like a technical report than a research article. The central recommendation is plausible but currently supported mainly by an unvalidated checklist; the requested revisions (evidence for SAM's partial DT capabilities, an operationalized rubric, and a clearer separation between the static simulation results and the maturity assessment) are necessary before publication. The manuscript may also need a tighter novelty statement relative to the existing literature on PVsyst/SAM comparisons."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, what this is: a solid engineering report comparing two PV simulation tools on a real plant, with a digital-twin maturity classification. The case study is honest and clear. The authors use NSRDB data for both tools, present monthly and annual comparisons, and explicitly say neither tool alone is a complete DT. The conclusion that SAM is a better base for a future DT is conditional on external sensors and AI modules, so the stress-test note that this is unsupported is a bit too harsh. The paper never claims SAM is a DT; it claims SAM is structurally closer.\n\nThat said, the structural-closeness argument rests entirely on a feature checklist. The quantitative results actually favor PVsyst on annual energy accuracy (177,523 kWh vs. 175,515 kWh real, against SAM's 163,559 kWh), and the paper doesn't reconcile why flexibility should outweigh accuracy in a DT context. If the authors want to convince a skeptical engineer, they need at least a minimal demonstration: feed live data into SAM via PySAM, push a set-point back, and show the loop. Without that, the recommendation is an interpretation, not a verified property.\n\nThe paper's real value is as a decision-support or teaching document for Spanish-speaking engineers. It's well-structured, limitations are stated, and the references look appropriate. No machine-checked proofs or shipped code, so reproducibility is limited. RMSE/MAE figures lack uncertainty bounds, and the input files aren't given.\n\nMy recommendation: if this crosses your desk, don't desk-reject it if the venue publishes applied case studies. Send it to review with a request to either add a proof-of-concept integration or soften the conclusion to a clearly labeled 'potential' claim. For a research journal, novelty is too low; for a practice-oriented journal, it's a solid submission.","headline":"A competent practical comparison of PVsyst and SAM, but the claim that SAM is the better digital-twin base is a plausible opinion backed by a feature checklist, not by a demonstrated closed-loop integration.","tokens_in":10109,"tokens_out":3894,"would_cite":false,"duration_ms":45094,"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":"The paper argues that among photovoltaic simulators, SAM is the realistic base for a solar digital twin, while PVsyst operates at best as a digital shadow.","keywords":["digital twin","photovoltaic systems","PVsyst","SAM","digital shadow","solar energy simulation","digital maturity","case study"],"falsifier":"Connect a live data feed and a control channel to both SAM and PVsyst on the same plant; if SAM cannot update its model from streaming sensor data and send recommendations back in real time, or if a PVsyst-based setup with external scripting achieves the same behavior, the paper's central ranking is falsified.","tokens_in":9218,"feed_emoji":"☀️","tokens_out":5198,"duration_ms":52864,"temperature":0.7,"pith_summary":"This report asks whether widely used photovoltaic simulators can serve as digital twins, and on what evidence. It compares PVsyst and SAM on a real rooftop plant in Turkey, using the same NSRDB weather data, and measures simulated against actual monthly energy production. Its central claim is that PVsyst, however accurate as a design tool, operates at best as a digital shadow—it cannot ingest live operating data or feed actions back to the plant—while SAM's scripting and data-access flexibility make it the realistic foundation for a digital twin once external sensors, APIs, and AI modules are added. If the claim holds, the practical takeaway is that simulator choice for operational use should be driven by integration capacity, not by simulation precision alone.","feed_headline":"SAM outranks PVsyst as a solar digital-twin base","feed_subtitle":"A real-plant comparison shows PVsyst stays a digital shadow, while SAM's scripting can close the feedback loop.","key_machinery":"The argument runs on two coupled devices. First is the three-level maturity taxonomy—digital model (no live data), digital shadow (one-way data inflow), digital twin (bidirectional data and control flow)—which the paper adopts from its references. Second is a feature checklist in Table 5 that assigns checkmarks to criteria such as import of real data, dynamic update, feedback to the physical system, IoT support, machine learning use, scripting/interoperability, and result export; that checklist is what separates PVsyst (shadow) from SAM (between shadow and twin). The quantitative RMSE/MAE comparison in Section 4 supports the accuracy part of the story, while the checklist carries the maturity verdict.","core_discovery":"On the paper's own terms, the discovery is a maturity ranking grounded in the digital-model/digital-shadow/digital-twin taxonomy: PVsyst has high physical fidelity but no native dynamic data loop, so it belongs at the digital-model level or, at best, the digital-shadow level; SAM sits between digital shadow and digital twin, with Python/Matlab scripting, NSRDB data access, and result export that make it extendable to a true digital twin when paired with monitoring platforms, APIs, and artificial intelligence. The paper also reports that both tools give reasonable annual energy estimates for the Elmah case, with monthly deviations captured by RMSE and MAE, but neither tool in its base configuration is a complete digital twin.","pith_inferences":["The paper's error analysis measures energy-estimation accuracy but never exercises the feedback loop, so the maturity conclusion rests on the qualitative feature checklist rather than on the simulation numbers.","A direct test of the paper's recommendation would be to build a SAM-based twin with live sensor data and compare its nowcast accuracy and response time against a static PVsyst model, measuring how much the feedback loop actually improves operations.","The same maturity checklist could be applied to open-source simulation stacks, which already support one-way data ingestion and scripting, and might therefore outrank commercial tools on the path to a digital twin.","The paper places SAM 'between' discrete maturity levels, which suggests that a graded or continuous maturity metric would be more informative than the three-level label."],"forward_implications":["PVsyst remains the stronger tool for design-stage optimization, where historical weather and detailed shading and loss analysis matter more than live data.","SAM can serve as the predictive engine of a solar digital twin, provided a data-acquisition layer with IoT sensors and AI modules is attached.","Choosing a simulator for operational use should weight integration and scripting ability at least as heavily as simulation accuracy.","A full photovoltaic digital twin is not available in either commercial tool out of the box; it has to be assembled from simulator, monitoring, and feedback components."],"supporting_citations":[{"why":"Introduces the digital twin concept and its lifecycle-management origin, which the paper adopts as the starting definition.","marker":"[7]"},{"why":"Supplies the digital model / digital shadow / digital twin taxonomy that structures the paper's maturity ranking.","marker":"[9]"},{"why":"Baseline comparison of PVsyst, SAM, and PVLib that the paper uses to position each simulator on the maturity scale.","marker":"[12]"},{"why":"Provides the Elmah rooftop plant case study and the real production data against which both simulators are measured.","marker":"[15]"},{"why":"Systematic review used as the source of the digital-maturity criteria that appear in the feature checklist.","marker":"[16]"},{"why":"Reports PVsyst's limitations in dynamic data use, supporting the classification of PVsyst as a digital shadow.","marker":"[17]"},{"why":"Documents SAM's Python scripting interface, the key capability that makes SAM extendable toward a digital twin.","marker":"[18]"},{"why":"Shows how to link SAM to MATLAB/Simulink, evidence of the interoperability the paper highlights.","marker":"[19]"}],"fun_headline_variants":["SAM edges PVsyst as solar digital-twin base","SAM, not PVsyst, nears true solar twin","PVsyst shadow, SAM twin-capable: solar tools ranked","Digital twin test: SAM beats PVsyst for real-time loop","Neither tool is a full twin, but SAM gets closer"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The ranking assumes that the paper's feature checklist—real-data import, scripting, dynamic update, feedback—correctly captures what makes a simulator a digital twin, and that SAM's scripting flexibility will actually translate into a working bidirectional loop once external components are added.","fun_headline_variants_meta":{"raw":{"variants":["SAM edges PVsyst as solar digital-twin base","SAM, not PVsyst, nears true solar twin","PVsyst shadow, SAM twin-capable: solar tools ranked","Digital twin test: SAM beats PVsyst for real-time loop","Neither tool is a full twin, but SAM gets closer"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000196,"raw_usage":{"total_tokens":1261,"prompt_tokens":745,"completion_tokens":516,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":361,"completion_tokens_details":{"reasoning_tokens":430}},"tokens_in":361,"tokens_out":516,"duration_ms":4998,"temperature":1.0,"reasoning_tokens":430,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T04:08:40.493780+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Connect a live data feed and a control channel to both SAM and PVsyst on the same plant; if SAM cannot update its model from streaming sensor data and send recommendations back in real time, or if a PVsyst-based setup with external scripting achieves the same behavior, the paper's central ranking is falsified.","supporting_citations":[{"cited_title":"Esto implica que: .)Puede usarse como motor predictivo dentro de un gemelo digital solar","cited_arxiv_id":null,"evidence_quote":"Introduces the digital twin concept and its lifecycle-management origin, which the paper adopts as the starting definition."},{"cited_title":"Producción energética argentina en la transición mundial hacia las energías limpias – Junio 2024","cited_arxiv_id":null,"evidence_quote":"Supplies the digital model / digital shadow / digital twin taxonomy that structures the paper's maturity ranking."},{"cited_title":"ECLIPSE : Envisioning CLoud Induced Perturbations in Solar Energy","cited_arxiv_id":"2104.12419","evidence_quote":"Baseline comparison of PVsyst, SAM, and PVLib that the paper uses to position each simulator on the maturity scale."},{"cited_title":"Intelligent digital twins and the development and management of complex systems.Digital Twin, 2:8, May 2022","cited_arxiv_id":null,"evidence_quote":"Provides the Elmah rooftop plant case study and the real production data against which both simulators are measured."},{"cited_title":"A review on Digital Twins and its Application in the Modeling of Photovoltaic Installations, January 2024","cited_arxiv_id":null,"evidence_quote":"Systematic review used as the source of the digital-maturity criteria that appear in the feature checklist."},{"cited_title":"Enhancing photovoltaic system efficiency through a digital twin framework: A comprehensive modeling approach.International Journal of Thermofluids, 26:101078, March 2025","cited_arxiv_id":null,"evidence_quote":"Reports PVsyst's limitations in dynamic data use, supporting the classification of PVsyst as a digital shadow."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Documents SAM's Python scripting interface, the key capability that makes SAM extendable toward a digital twin."},{"cited_title":"Gemelo digital de parque solar utilizando ROS2","cited_arxiv_id":null,"evidence_quote":"Shows how to link SAM to MATLAB/Simulink, evidence of the interoperability the paper highlights."}],"review_version":1}