Back analysis of mechanical parameters based on GPSO-BP neural network and its application.
Source check: Source checked. Not independent scientific confirmation.
Finding: For one Zhaotong Tunnel segment, a GA-PSO-optimized back-propagation network inferred rock parameters from displacement inputs; forward calculations matched monitored roof and step displacements with mean relative error 4.34% and outperformed three comparison networks in that case. Generalizability is limited to this field application.
Technology: A genetic-algorithm/particle-swarm-optimized back-propagation network inferred four surrounding-rock parameters from displacement monitoring; the estimates were used in forward tunnel calculations.
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
- AI contribution
- The learned inversion model supported a field-case parameter estimate checked against monitored tunnel displacement.
- Test and result
- In one Zhaotong Tunnel field segment, forward-predicted roof/step displacements had reported relative errors of 1.11%, -6.88%, -0.41% and -8.97% (mean relative error 4.34%), outperforming BP, PSO-BP and GA-BP in that case.
- Limitation / still unknown
- Training used 25 simulation schemes and testing five additional schemes; the field comparison was one tunnel segment. The paper notes limited generalization and weaker cohesion fit.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://www.nature.com/articles/s41598-025-86989-7
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 22 Nov 2026. Review scope: Primary Scientific Reports full text and PMC copy, including field application and parameter-validation Results; publication date checked..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.