Geng Integrity Report: CFD Simulation of a bubble column evaporator (Cappelli, Glennon, Donnellan 2022)

Academic fraud report · Geng Detector

Summary

This review examines the computational fluid dynamics (CFD) paper 'CFD Simulation of a bubble column evaporator' by Cappelli, Glennon and Donnellan, published in the International Journal of Heat and Mass Transfer (188, 2022, 122296, DOI 10.1016/j.ijheatmasstransfer.2021.122296). No experimental images are present, so image-based checks were not applicable. The paper was nevertheless flagged as highly suspicious because the reported numerical evidence exhibits multiple internal contradictions that cannot be reconciled by simple typographic error. Key issues include: (1) the Conclusion reports a control-case evaporation rate of 2.48×10⁻⁵ kg/s that corresponds to a theoretical thermodynamic maximum, whereas the actual CFD control value in Tables 5 and 6 is 2.68×10⁻⁵ kg/s (gas flowrate) and 2.34×10⁻⁵ kg/s (gas temperature); (2) the percent increases for 100–600 °C gas temperatures only match the data when the theoretical maximum is used as the baseline, not the CFD control; (3) Section 4.3 cites gas flowrates of 0.5 and 1.5 L/min while Table 5 lists 2 L/min; (4) Section 5.1 reports bulk liquid temperatures of 221–222 K against an initial 323.13 K, a drop violating energy conservation; and (5) the authors claim a laminar assumption yet specify turbulent wall functions in Table 1. Confidence is high in the arithmetic mismatches, moderate in inferring intent, and limited by absence of raw simulation files.

Verdict

🟠 Highly suspicious. The paper contains multiple severe, mutually reinforcing numerical and methodological inconsistencies. While the absence of experimental images precludes image-manipulation analysis, the textual and tabular evidence strongly suggests either careless manuscript assembly, undisclosed changes to the underlying CFD model, or fabricated/post-hoc adjusted numbers.

Key findings

Conclusion vs. tables mismatch: The Conclusion reports a 1 L/min evaporation rate of 2.48×10⁻⁵ kg/s; Table 5 lists the CFD control value as 2.68×10⁻⁵ kg/s. The 2.48×10⁻⁵ kg/s figure matches the theoretical thermodynamic maximum in Table 3.
Same swap occurs for the 50 °C gas-temperature case: Conclusion gives 2.48×10⁻⁵ kg/s while Table 6 lists 2.34×10⁻⁵ kg/s for the CFD control.
Percentage increases for gas temperatures (100, 200, 300, 600 °C) only reconcile when the theoretical maximum 2.48×10⁻⁵ kg/s is used as the baseline, yielding 2.8%, 10.08%, 13.3%, and 62.5%—matching the paper's stated 3%, 10%, 13%, 62%. Using the actual CFD control (2.34×10⁻⁵) gives 8.9%, 16.6%, 20.1%, 72.2%.
Section 4.3 text references gas flowrates of 0.5 and 1.5 L/min, but Table 5 lists 2 L/min. The reported 24% decrease at 0.5 L/min is arithmetically −26%.
Section 5.1 reports bulk liquid temperatures of 221 K, 222 K, and 321.4 K for flowrates of 3, 6, and 15 g/s, despite Table 2's initial condition of 323.13 K. A 70+ K drop within ~5 s for minor flowrate changes is thermodynamically implausible.
Section 2.2 assumes laminar flow, yet Table 1 specifies epsilonWallFunction and kqRWallFunction boundary conditions consistent with turbulent modeling.

Evidence highlights

DOI: 10.1016/j.ijheatmasstransfer.2021.122296
Table 5 CFD control (1 L/min) evaporation rate: 2.68×10⁻⁵ kg/s
Table 3 theoretical maximum evaporation rate: 2.48×10⁻⁵ kg/s
Table 6 CFD control (50 °C) evaporation rate: 2.34×10⁻⁵ kg/s
Table 6 elevated-temperature CFD values: 100 °C = 2.55×10⁻⁵, 200 °C = 2.73×10⁻⁵, 300 °C = 2.81×10⁻⁵, 600 °C = 4.03×10⁻⁵ kg/s
Table 2 initial bulk liquid temperature: 323.13 K
Section 5.1 reported bulk temperatures: 221 K (3 g/s), 222 K (6 g/s), 321.4 K (15 g/s)
Table 1 boundary conditions: epsilonWallFunction, kqRWallFunction

Notes

Pure CFD study; no experimental micrographs were available, so pixel-level duplication analysis (e.g., PS or AI-image forensics) was not performed.
The arithmetic contradictions are reproducible from the published tables and text alone; raw OpenFOAM case files, solver logs, and post-processing scripts were not supplied and would be required to determine whether the underlying simulation is sound or merely the reporting is erroneous.
The temperature anomaly (221–222 K vs 323.13 K initial) may plausibly be a dropped hundreds digit (e.g., 321–322 K), but this remains unconfirmed and does not excuse the systematic baseline swap in the percentage analysis.
The laminar/turbulent wall-function conflict may indicate the authors actually ran a turbulent simulation while documenting a laminar assumption to simplify methodology, a recurring pattern in low-quality CFD papers.
This report flags integrity concerns; formal determination of misconduct requires institutional investigation and inspection of source data.