Investigation Report: 'A Consortium Blockchain-Based Secure Storage and Access Control Scheme for Power Data' — Suspect Mislabeling of Citations and Fabricated Performance Charts

Academic fraud report · Geng Detector

Summary

This report evaluates a 2026 paper by Liu Chun, Zhang Linghao, Jiang Changsong et al. in the Journal of University of Electronic Science and Technology of China (DOI: 10.12178/1001-0548.2025016). The verdict is highly suspicious. Three confirmed issues are identified: (1) References [22] and [25] appear to have been swapped — Section 5.2 cites 'Bestie' as ref. [25], but the bibliography shows ref. [25] is the Bloom filter paper by Tarkoma et al., while ref. [22] is the Bestie paper by Chen et al., indicating copy-paste or templated writing; (2) Figure 7 plots per-user communication overhead as a function of total system user count (1 000–7 000), which is logically incoherent because a single retrieval's ciphertext/key size should not scale with the unrelated total user population; (3) Figure 7's x-axis labels are '1 000, 2 000, 3 000, 4 000, 5 000, 5 000, 7 000', showing a duplicated tick and layout disorder consistent with manually fabricated data. A fourth suspected issue — unusually clean linear scaling across Figures 3, 4, and 6 — remains unverified. Confidence is moderate-to-high for findings 1–3; the overall judgment is preliminary and subject to institutional investigation.

Verdict

Highly suspicious (🟠). Three independent, confirmed anomalies (citation mislabeling, logically incoherent communication-overhead plot, and malformed axis labels) converge with one suspected issue (suspiciously smooth linear scaling across multiple performance charts). The combination is consistent with templated writing and synthetic experimental data, but final determination of academic misconduct requires official institutional investigation.

Key findings

Citation swap between [22] and [25] (Section 5.2, Section 1.3, References). Section 5.2 attributes the 'Bestie' scheme to ref. [25], while Section 1.3 attributes the Bloom filter background to ref. [22]. The reference list, however, lists Chen et al.'s Bestie paper as [22] and Tarkoma et al.'s Bloom filter paper as [25]. The numbering is fully inverted.
Communication overhead plotted against total user count (Section 5.3, Figure 7). The x-axis represents the number of users in the system, yet the y-axis purports to measure per-request communication overhead. A single retrieval's ciphertext/key size is not expected to scale linearly with unrelated system-wide user population, suggesting the underlying experiment or chart was repurposed from a computation-cost experiment.
Malformed x-axis tick labels in Figure 7. The sequence '1 000, 2 000, 3 000, 4 000, 5 000, 5 000, 7 000' contains a duplicated '5 000' and inconsistent spacing, indicating manual fabrication of axis parameters rather than data-driven plotting.
Suspected (unverified) — overly perfect linear scaling in Figures 3, 4, and 6. Real blockchain smart-contract experiments on Ganache typically exhibit noise, GC pauses, and step-wise thresholds; uniformly textbook-straight lines are suspicious but not confirmed.

Evidence highlights

Section 5.2 states: 'In the existing scheme Bestie[25]…' (claimed Bestie reference is [25] in text).
Reference list: [22] = 'CHEN T Y … Bestie: Very practical searchable encryption …'; [25] = 'TARKOMA S … Theory and practice of bloom filters …'. Numbers are swapped.
Figure 7 x-axis tick labels (transcribed verbatim from the report): 1 000 2 000 3 000 4 000 5 000 5 000 7 000.
Figure 7 y-axis variable: communication overhead per data-access request; x-axis variable: total user count.
Implementation platform cited in the paper: Ganache + web3j (Ethereum smart-contract testbed).
DOI: 10.12178/1001-0548.2025016

Notes

Confidence is high for Findings 1–3; the citation swap and axis duplication are observable from the text and figures alone.
Finding 4 is explicitly marked 'insufficient evidence' in the source report and should be treated as a hypothesis requiring raw experimental logs to confirm or refute.
The authors have not been contacted as part of this review; conclusions are preliminary.
Recommended follow-up: request raw Ganache/web3j test logs; query the journal's editorial office regarding peer-review diligence; cross-check the team's other publications for similar patterns.