Zhuojin Pipe Fitting Co., LtdNumerical Simulation and Data-Driven Fusion Prediction of Steel Pipe Column Deformation During Cluster Pit Excavation
Background and Engineering Context
Cluster pit excavation, a common construction method for deep foundation pits in urban environments, induces significant ground displacement that affects the structural response of adjacent steel pipe columns. Steel pipe columns, used extensively in diaphragm wall systems and bracing systems, are subjected to complex loading conditions during excavation, including lateral earth pressure, bending moments, and axial forces. The accurate prediction of steel pipe column deformation is critical for ensuring structural safety, controlling construction risks, and optimizing design parameters.
This study addresses the challenge of predicting steel pipe column deformation during cluster pit excavation by integrating numerical simulation with data-driven fusion prediction methods. The cluster pit configuration introduces additional complexity compared to single-pit excavation, as the interaction between adjacent excavation units creates a coupled deformation field that affects the load distribution on the steel pipe columns.
Numerical Simulation Framework
The numerical simulation employs a three-dimensional finite element model (FEM) to capture the soil-structure interaction during excavation. The model includes the following key components:
| Component | Modeling Approach | Key Parameters |
|---|---|---|
| Soil | Mohr-Coulomb / Cam-Clay | E, ν, φ, c, K0 |
| Steel Pipe Column | Elastic-plastic beam | E = 206 GPa, σy = 345-460 MPa |
| Diaphragm Wall | Shell element | t, E, D |
| Bracing System | Truss element | E, A, initial stress |
| Excavation Sequence | Step-by-step unloading | Stage-by-stage removal |
The soil model is typically calibrated using in-situ testing data, including standard penetration tests (SPT), cone penetration tests (CPT), and pressuremeter tests. The steel pipe column is modeled as a high-strength structural steel (typically Q345 or Q420) with a wall thickness of 12-20 mm and an outer diameter of 600-1200 mm, depending on the excavation depth and load requirements.
The excavation process is simulated in stages, with each stage representing a specific construction sequence: diaphragm wall construction, internal bracing installation, soil excavation to successive levels, and finally, foundation construction. The deformation of the steel pipe columns is tracked at each stage, with particular attention to the maximum lateral displacement, bending moment, and axial force.
Data-Driven Fusion Prediction Method
The data-driven fusion prediction method integrates multiple data sources to improve the accuracy of deformation prediction. The approach combines:
- Numerical Simulation Data: The FEM results provide a baseline prediction of steel pipe column deformation under the assumed boundary conditions and material parameters.
- Monitoring Data: Real-time monitoring data from inclinometers, strain gauges, and displacement sensors installed on the steel pipe columns provide actual deformation measurements during construction.
- Historical Data: Deformation patterns from similar projects in comparable soil conditions and excavation configurations are incorporated as training data for the prediction model.
The fusion prediction model employs a weighted integration approach where the numerical simulation provides the structural response framework, while the monitoring data corrects the model parameters in real time. This adaptive approach significantly improves prediction accuracy compared to using either numerical simulation alone or empirical methods alone.
Prediction Accuracy Assessment
The accuracy of the fusion prediction method is evaluated by comparing predicted and measured deformation values at multiple monitoring points:
| Monitoring Point | Measured Deformation (mm) | FEM Prediction (mm) | Fusion Prediction (mm) | Error Reduction |
|---|---|---|---|---|
| Column Top | 12.5 | 10.8 | 11.9 | 8.8% |
| Column Mid-Height | 8.3 | 7.1 | 7.9 | 10.8% |
| Column Base | 3.2 | 2.8 | 3.0 | 6.3% |
| Maximum Bending Moment | 450 kN·m | 380 kN·m | 430 kN·m | 5.6% |
The fusion prediction method consistently reduces prediction error by 5-15% compared to pure FEM simulation, demonstrating the value of integrating real-time monitoring data into the predictive model.
Key Technical Findings
The study reveals several important technical insights regarding steel pipe column behavior during cluster pit excavation:
- Deformation Pattern: The steel pipe columns exhibit a characteristic bending mode with maximum lateral displacement at approximately 0.4-0.6 times the excavation depth from the ground surface. The deformation pattern is influenced by the bracing configuration, soil stiffness, and excavation sequence.
- Cluster Pit Interaction Effect: Adjacent excavation units induce additional lateral displacement of 10-25% compared to single-pit excavation, due to the cumulative effect of soil unloading from multiple excavation faces.
- Critical Load Condition: The maximum bending moment occurs when the excavation reaches 60-70% of the total depth, corresponding to the stage of maximum soil displacement and minimum passive resistance.
- Axial Force Distribution: The axial force in the steel pipe columns increases with excavation depth, with the maximum value occurring at the column base where the reaction from the foundation slab is transmitted.
Engineering Practice Applications
The findings from this study have direct applications in the design and construction of deep foundation pit systems:
- Design Optimization: The predicted deformation values can be used to optimize the steel pipe column section size, wall thickness, and bracing configuration, reducing material usage while maintaining structural safety.
- Construction Monitoring: The fusion prediction model can serve as a real-time early warning system, alerting construction teams when measured deformations deviate from predicted values by more than the acceptable threshold.
- Risk Management: The identification of critical deformation stages and load conditions enables proactive risk management, including the implementation of additional bracing or dewatering measures before critical thresholds are reached.
- Cost Control: Accurate deformation prediction reduces the need for over-conservative design, resulting in material savings of 10-20% for the steel pipe column system.
Study Insights and Implications
The integration of numerical simulation with data-driven fusion prediction represents a significant advancement in the prediction of steel pipe column deformation during cluster pit excavation. The approach addresses the inherent limitations of pure numerical simulation, which relies on assumed material parameters and boundary conditions that may not accurately reflect the actual construction conditions.
For practicing engineers, the key implication is that deformation prediction should not rely on a single method. The fusion approach, which combines the physical understanding provided by FEM with the empirical correction provided by monitoring data, offers the most reliable prediction capability. This approach is particularly valuable in complex cluster pit configurations where the interaction between multiple excavation units creates deformation patterns that are difficult to predict using traditional methods alone.
The study also highlights the importance of real-time data acquisition and processing in construction monitoring. The effectiveness of the fusion prediction model depends on the timeliness and accuracy of the monitoring data, which requires investment in instrumentation and data processing infrastructure. Future developments should focus on automating the data fusion process to enable real-time prediction and decision support during construction, ultimately improving the safety and efficiency of deep excavation projects in urban environments.