Multi-Source Monitoring Data Fusion Comprehensive Evaluation Method for the Safety Status of Deep Foundation Pit

Bo Wu, Wei Yu, Guowang Meng, Shixiang Xu, Qinshan Wang, Dianbin Cao, Chenxu ZhaoSustainability 2023 · DOI

TL;DR

This paper introduces a novel "Game-Cloud Model" to comprehensively assess the safety of deep foundation pits (DFPs) for subway construction. By fusing data from multiple monitoring sources and intelligently combining subjective expert knowledge with objective data analysis, the method provides a more reliable safety evaluation than traditional single-indicator warnings, leading to better risk management and fewer construction delays.

What Problem Does This Paper Address?

Constructing deep foundation pits for subway stations is inherently risky, with numerous uncertainties. A major challenge is the discrepancy between individual monitoring alerts and the actual overall safety status of the pit. This can lead to "excessive monitoring warnings" where a single sensor triggers an alarm, but the overall structure remains safe, causing unnecessary project delays and increased costs. Existing methods often struggle to integrate diverse monitoring data effectively and account for the complex interdependencies between different indicators. The paper aims to bridge this gap by developing a robust, multi-indicator assessment system that provides a more accurate and holistic view of DFP safety.

Methods

The core of this research lies in a novel "Game-Cloud Model" for safety assessment. Here's a breakdown of the key components:

  • Quantitative Assessment Index System: The authors first establish a comprehensive set of 11 monitoring indicators, carefully selected based on national standards for DFP safety classification (specifically Grade 1, the most critical). This ensures the indicators are relevant and directly reflect the pit's stability.
  • Hybrid Weight Determination (Game Theory): Recognizing that both expert judgment (subjective) and data patterns (objective) are crucial, the paper employs a two-pronged approach to assign weights to the indicators:
  • Subjective Weights: The Grey Decision-Making Trial and Evaluation Laboratory (GDEMATEL) method is used, which leverages expert opinions while accounting for interdependencies between indicators.
  • Objective Weights: The entropy method is applied to objectively determine weights based on the information content and variability within the monitoring data.
  • Game Theory Integration: To reconcile the subjective and objective weights and minimize discrepancies, game theory is introduced. This approach finds an optimal balance, creating a more robust and reliable combined weight for each indicator.
  • Cloud Model (CM) for Uncertainty: The Cloud Model is used to handle the inherent fuzziness and randomness in converting qualitative safety levels into quantitative values. Specifically, a "forward cloud generator" is employed to create membership functions for each indicator at different safety levels. This allows for a probabilistic representation of safety status.
  • Comprehensive Evaluation: Finally, the combined weights and the cloud model's output are integrated to calculate a synthetic safety level for the DFP. This involves determining the "synthetic certainty degree" for each safety level and selecting the level with the highest degree.

Datasets

The study utilizes on-site monitoring data from a real-world subway construction project: the Martyrs Cemetery Station on the M1 line. The data covers a 19-day period, specifically focusing on an "alarm zone" where monitoring values exceeded predefined thresholds. The paper provides specific monitoring values for various indicators (C1-C11) on a particular day (July 4, 2018) and a time series of synthetic safety levels from June 21 to July 9, 2018. No new benchmark datasets were introduced; rather, existing monitoring data from a specific project was used for validation.

Key Insights and Ideas

  • The Power of Fusion: The paper strongly advocates for data fusion, demonstrating that combining multiple monitoring sources provides a far more nuanced and accurate picture of DFP safety than relying on single-indicator alerts.
  • Bridging Subjectivity and Objectivity: The innovative use of game theory to calibrate subjective (expert-based) and objective (data-driven) weights is a significant contribution. This approach mitigates the biases inherent in each method, leading to more trustworthy indicator importance.
  • Handling Uncertainty with Clouds: The Cloud Model proves effective in translating complex, fuzzy safety concepts into quantifiable metrics, which is crucial for DFP safety assessment where precise numerical values can be misleading.
  • Change Rate vs. Cumulative Value: The study highlights that for the same monitoring project, the change rate of an indicator often receives a higher weight than its cumulative value. This is because change rates are more sensitive and timely in detecting emerging risks.
  • Key Indicators for Stability: The analysis identifies lateral displacement of the structure (C1), change rate of horizontal displacement in deep layers (C2), and axial force in support (C9) as the most influential indicators for DFP stability.

Key Results

  • Dynamic Safety Assessment: The proposed Game-CM method was applied to a 19-day dynamic assessment of a DFP.
  • Consistent Safety Levels: Throughout the assessment period, the synthetic safety level of the DFP consistently ranged between Grade I (Negligible) and Grade II (Acceptable), aligning well with on-site observations.
  • Improved Accuracy: The Game-CM method's results were found to be more consistent with actual site conditions and follow-up monitoring compared to single-indicator warnings and a traditional Fuzzy Comprehensive Evaluation (TFCE) method.
  • Reduced Ambiguity: While the TFCE method sometimes exaggerated safety levels (e.g., assigning Grade I when risks were present), the Game-CM method provided a more realistic assessment by considering the uncertainty and interrelationships of indicators.

Limitations and Future Work

  • Ambiguous Grade Boundaries: The authors acknowledge that some evaluation results exhibited "ambiguous grade boundaries," suggesting that further research is needed to refine the model and mitigate this ambiguity.
  • Computational Feasibility: While the paper states the method can be realized through Excel and MATLAB, the complexity of the calculations might still pose a challenge for immediate, real-time implementation in all scenarios.
  • Further Refinement: The paper explicitly calls for "further research is needed to improve the model and mitigate the ambiguity in the evaluation results." This implies ongoing development and potential enhancements to the Game-CM framework.

Summary updated

Abstract

Construction of the deep foundation pit (DFP) in subway stations is fraught with significant uncertainties, which may cause project delays due to discrepancies between single-indicator monitoring warning information and actual conditions at the site. Therefore, this article proposes a safety assessment method for DFP based on the Game-Cloud Model. An entirely quantitative assessment index system is established with on-site monitoring projects according to the design safety classification of DFP. Considering the one-sidedness of using a single method to determine the weights of assessment indices, game theory is introduced to calibrate the subjective and objective weights determined by the grey decision-making trial and evaluation laboratory (GDEMATEL) and the entropy method, respectively. Next, we use the forward cloud generator of the cloud model (CM) to generate the safety level membership function of the evaluation indicators. Finally, we quantitatively calculate the synthetic safety level of DFP using the comprehensive evaluation approach. A 19-day dynamic assessment was conducted on the actual engineering project by the proposed method. The results indicated that the synthetic safety level of the assessed area ranged between grades Ⅰ and Ⅱ, corresponding to Negligible and Acceptable in the acceptance criteria. Compared with the single-indicator monitoring warning results, it was more in line with on-site observation, which verified its reliability and practicality.
BibTeX
@inproceedings{wu2023multi,
  author = {Bo Wu and Wei Yu and Guowang Meng and Shixiang Xu and Qinshan Wang and Dianbin Cao and Chenxu Zhao},
  title = {{Multi-Source Monitoring Data Fusion Comprehensive Evaluation Method for the Safety Status of Deep Foundation Pit}},
  booktitle = {Sustainability},
  year = {2023},
  doi = {10.3390/su151511809},
}
Figure