An attractive settlement plot is only one part of a foundation decision. For construction teams reviewing geotechnical modelling, the more useful question is whether the performance measure matches the risk being managed. Building movement and overall ground stability deserve separate attention.
Research review: a peer-reviewed Scientific Reports study, published online on 21 October 2026 as the version of record. Civil Insight provides the editorial assessment below.
What was tested
The researchers combined a 1:120 physical model based on Chongqing’s Renhe Tunnel with PLAXIS 2D simulations. Validation used three settlement and six pore-pressure monitoring locations. Separate analyses compared foundations and slope treatments; the latter excluded buildings. Read the full open-access study.
Baseline validation represented 200 mm/day rainfall for 12 hours; the parametric comparisons used 50 mm over 0.1 day. These are different scenarios.
The key comparison
| Foundation | Maximum wet-case settlement (mm) | Safety factor (unitless) |
|---|---|---|
| Shallow | ≈45 | 1.79 |
| Combined | ≈30 | 2.10 |
| Piled raft | ≈20 | 1.86 |
The piled raft minimised predicted settlement; the combined foundation had the highest calculated safety factor. These are model results.
Civil Insight’s assessment of the evidence
The useful contribution is the comparison between two decision criteria. A project team could otherwise favour the smallest displacement while overlooking the wider ground system. The table should prompt a discussion about which response governs, where it is evaluated, and what other consequences remain outside that measure.
Connecting an instrumented experiment to a numerical analysis is stronger than presenting an unexplained contour plot. It gives readers a physical benchmark against which to question the calculations. However, agreement with a benchmark and prediction at a new construction site are different levels of evidence.
Gravity similitude was relaxed. Checking against the physical experiment did not establish independent field validation. The numerical analysis used plane-strain conditions.
For transfer to practice, those boundaries raise specific questions. Does the proposed site have comparable fill, groundwater and geometry? How would plausible changes to permeability or stiffness alter the ranking? Could three-dimensional building or excavation effects change the response? Confidence should depend on those checks, rather than the apparent precision of the reported millimetres.
Why the soil model and its inputs matter
Bentley’s technical explanation of PLAXIS material models describes the Hardening Soil model as using separate loading, unloading and oedometer stiffnesses, with stress-dependent behaviour. Its SoilTest discussion explains how model response can be compared with laboratory data. The manufacturer also cautions that calibration against laboratory conditions can still leave differences in practical applications.
Civil Insight’s interpretation is that a more elaborate soil model creates a greater need to explain the inputs. A reviewer should be able to trace the adopted parameters to suitable evidence and understand the initial groundwater and stress assumptions. Naming the software or constitutive model does not, by itself, establish that the project has been represented adequately.
The rainfall cases also deserve a clear audit trail. A design review should distinguish the total water applied, its duration, initial ground moisture and the treatment of runoff and drainage. Comparing options under a shared scenario can reveal sensitivity; it cannot establish the probability of that scenario at another location.
Australian practice: turn the comparison into a review brief
Queensland Transport and Main Roads’ Geotechnical Design Standard — Minimum Requirements, October 2024, provides relevant local context. For TMR projects, it requires justification of ground profiles, parameters, methods and assumptions. Numerical modelling submissions must document construction stages and boundary conditions, with outputs checked through hand calculations, another numerical method or empirical methods. These are departmental project requirements, rather than a national approval of the reviewed study.
For cuttings, the same standard’s construction guidance requires groundwater and/or movement monitoring, inspection of exposed materials, progressive review of new information and instrument checks during critical work and after significant rainfall. That connects analysis to observation as construction proceeds.
A useful project review brief would therefore set out the decisions the model must inform, the movements and failure mechanisms to assess, the uncertainty to explore, and the observations that could challenge its assumptions. It should also identify who reviews unexpected behaviour and how revised information reaches the design and construction teams. These are Civil Insight’s suggested review questions, not additional findings from the experiment.
Civil Insight assessment: the comparison is valuable for framing foundation decisions because it keeps deformation and stability visible together. The next step for a practitioner is to test that reasoning against the project’s ground model, requirements and monitoring evidence. A universal foundation preference cannot be established from a single case comparison.
Original paper and photo credit
Suliman, L., Zhong, Z., Liu, X. and Zhou, X. (2026). Hydro-mechanical assessment of tunnel–fill slope systems under extreme rainfall: effects of foundations and slope protection. Scientific Reports, 16, 33241. DOI: 10.1038/s41598-026-72810-6.
Publisher record checked 3 November 2026; no correction or retraction notice was identified. Background sources are linked above; they do not validate the study’s numerical results.
Cover: rock-portal excavation at Mitchell Point Tunnel, Oregon, 5 October 2021. Photo: Western Federal Lands / Oregon Department of Transportation, CC BY 2.0. This illustrative photograph depicts a different project and ground type from the Renhe fill-slope study.
