Water inrush risk assessment during karst tunnel construction based on knowledge decision and data-driven methods
•A set of water inrush risk evaluation indicators during karst tunnel construction were established.•A knowledge decision model of water inrush risk assessment were constructed based on VIKOR model and integrated weights.•A knowledge-data dual-driven DE-GWO-ELM model was proposed to realize interpre...
Uložené v:
| Vydané v: | Tunnelling and underground space technology Ročník 168; s. 107120 |
|---|---|
| Hlavní autori: | , , , , , |
| Médium: | Journal Article |
| Jazyk: | English |
| Vydavateľské údaje: |
Elsevier Ltd
01.02.2026
|
| Predmet: | |
| ISSN: | 0886-7798 |
| On-line prístup: | Získať plný text |
| Tagy: |
Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
|
| Shrnutí: | •A set of water inrush risk evaluation indicators during karst tunnel construction were established.•A knowledge decision model of water inrush risk assessment were constructed based on VIKOR model and integrated weights.•A knowledge-data dual-driven DE-GWO-ELM model was proposed to realize interpreteable of water inrush risk assessment.
Karst tunnels are frequently subject to the combined effects of complex geological conditions, groundwater hydrological characteristics, and construction disturbances, leading to an increased risk of water inrush. To enhance the real-time performance and interpretability of water inrush risk assessment, this study proposes a method based on the integration of knowledge decision and data-driven models. First, a set of water inrush risk evaluation indicators and their benchmark set were established. Then, the analytic hierarchy process and entropy weight method were used to determine the subjective and objective weights, which were fused using game theory to improve the accuracy of the knowledge decision-making model based on the vlsekriterijumska optimizacija i kompromisno resenje (VIKOR) method. Thereafter, the VIKOR results were used as the base data to construct the training sample library for the data-driven model. The differential evolution–gray wolf optimization algorithm was employed to optimize the model hyperparameters, and ultimately, an extreme learning machine water inrush risk assessment model that combined knowledge decision and data-driven approaches was established. By comparing the risk assessment results of both models in typical monitoring sections, the proposed method was verified to effectively and accurately perform water inrush risk assessment with strong real-time performance and interpretability. |
|---|---|
| ISSN: | 0886-7798 |
| DOI: | 10.1016/j.tust.2025.107120 |