Geng academic fraud review: Phase change doped sand storage for isothermal thermoelectric heat delivery (Energy Materials, DOI: 10.20517/energymater.2026.73)

Academic fraud report · Geng Detector

Summary

Verdict: questionable (yellow). This is a simulation-only study on phase-change thermal energy storage coupled with thermoelectric generation. Re-verification confirms two substantiated concerns: (1) the authors explicitly computed only 10 charge-discharge cycles in their finite-element model yet extrapolated fatigue life to N = 7,300 cycles (20 years at one cycle/day) by fitting accumulated plastic strain to a power-law ratcheting model, without physical validation; and (2) the work contains no experimental data, relying entirely on COMSOL Multiphysics 6.4 and ANSYS Workbench 2024 R1 with idealized assumptions such as a fixed contact resistance of 5 × 10⁻⁵ m² K/W. Two additional allegations—mismatched references and a Figure 9/Figure 10 numbering inconsistency—could not be independently confirmed because the truncated source PDF did not include the REFERENCES section or the relevant figure captions referenced in the original report. Overall, the core methodological concerns regarding extreme extrapolation and absence of experimental corroboration are well supported, while the alleged citation and figure-numbering issues remain unverified.

Verdict

Questionable (🟡). The paper presents serious methodological concerns that undermine the engineering validity of its conclusions, even though some flagged allegations could not be independently re-verified from the truncated source.

Key findings

Unsupported life extrapolation (confirmed): Authors computed only 10 charge-discharge cycles but claimed assessment of structural integrity over 7,300 cycles (20 years at one cycle/day) using a power-law ratcheting fit.
No experimental validation (confirmed): The entire study is simulation-based (COMSOL Multiphysics 6.4, ANSYS Workbench 2024 R1) with idealized assumptions; no physical experiments are reported.
Mismatched citations (unverified): Original report alleged references [30], [31], and [33] are on unrelated topics (Li-ion battery cathode degradation, photocatalytic hydrogen, Li-ion electrolyte); the truncated PDF did not include the REFERENCES section, so this could not be checked.
Figure numbering contradiction (unverified): Original report alleged text refers to "Figure 10" while a caption labels the same content as "Figure 9"; the truncated PDF contains references only to Figures 1–9 and no conflicting captions.
Idealized contact resistance assumption: Stacking contact resistances were arbitrarily fixed at 5 × 10⁻⁵ m² K W⁻¹ without sensitivity analysis or experimental grounding.

Evidence highlights

Source PDF lines 695–697: "Ten complete charge-discharge cycles were computed explicitly."
Source PDF lines 701–706: "Extrapolation to the target design life of N = 7,300 cycles (20 years at one cycle per day) was performed by fitting the accumulated plastic strain data from the first ten cycles to a power-law ratcheting model."
Source PDF line 318: implementation in COMSOL Multiphysics 6.4.
Source PDF line 590: fatigue simulation performed in ANSYS Workbench 2024 R1.
Source PDF lines 530–531: idealized contact resistance of 5 × 10⁻⁵ m² K W⁻¹ per interface.
DOI: 10.20517/energymater.2026.73 (preserved exactly).

Notes

Two of the four flagged issues (Findings 1 and 2) are marked "依据不足 / insufficient evidence" because the truncated source PDF omitted the REFERENCES list and the disputed figure captions. They should not be treated as confirmed until the full manuscript is inspected.
The two confirmed issues (extrapolation of 10 cycles to 7,300 cycles with no physical validation, and a wholly simulation-based methodology with idealized inputs) are themselves sufficient to raise substantial questions about the paper's engineering claims, irrespective of the unverified citation and numbering allegations.
Recommended follow-up: request raw simulation files (.mph) and MATLAB extrapolation code from the authors; post a PubPeer comment focused on the 10→7,300 cycle extrapolation; flag the lack of experimental validation to the journal editor.
This report is AI-assisted and intended for academic discussion only; final determination of misconduct requires an institutional investigation.