Adaptive Compaction Construction Simulation Based on Bayesian Field Theory.

Area of Science:

• Civil Engineering
• Computational Mechanics
• Statistical Modeling

Background:

• Compaction construction is vital for civil engineering projects, requiring accurate duration prediction for effective management.
• Existing simulation models often rely on parametric distributions and require manual hyper-parameter tuning.
• Current methods for determining roller stopping points overlook the randomness of roller movement.

Purpose of the Study:

• To develop a novel adaptive compaction construction simulation method.
• To improve the accuracy and efficiency of predicting compaction duration.
• To enhance construction management and project resource planning.

Main Methods:

• Utilizing Bayesian field theory for adaptive input parameter modeling.
• Implementing a new condition based on roller offset distance to determine stopping moments.
• Conducting simulation experiments on a high earth dam compaction project.

Main Results:

• The Bayesian field theory demonstrates high accuracy and efficiency in compaction simulation.
• The method achieves rapid processing times, with a large dataset processed in 1.54 seconds.
• The enhanced stopping condition significantly reduces the mean absolute error in predicted compaction duration.

Conclusions:

• The proposed adaptive compaction simulation method offers a robust approach for predicting construction duration.
• Bayesian field theory provides an effective framework for adaptive modeling in compaction processes.
• The method contributes to better project resource planning, especially in dynamic construction environments.