Prediction and Regulation of SCC's Shrinkage Using the PSO-BPNN Model.

Area of Science:

• Civil Engineering
• Materials Science
• Computational Intelligence

Background:

• Shrinkage deformation poses a significant risk to self-compacting concrete (SCC)-filled steel tube structures.
• Understanding and regulating autogenous shrinkage strain is essential for mitigating these risks and ensuring structural integrity.

Purpose of the Study:

• To develop and validate a predictive model for autogenous shrinkage strain in SCC before regulation.
• To investigate experimental methods for compensating shrinkage using expansion and combined expansion-contraction strategies.
• To apply predictive modeling to guide practical shrinkage regulation in an actual bridge project.

Main Methods:

• Construction of a Particle Swarm Optimization-Back Propagation Neural Network (PSO-BPNN) model.
• Prediction of autogenous shrinkage strain using the developed PSO-BPNN model.
• Conducting experimental investigations on shrinkage compensation techniques.
• Validation of the PSO-BPNN model against experimental measurements.

Main Results:

• The PSO-BPNN model demonstrated high accuracy in predicting concrete autogenous shrinkage strain, with a prediction error less than 10% for 28-day self-shrinkage.
• A strong consistency was observed between the predicted and measured values from the PSO-BPNN model.
• The predictive model effectively reduced experimental workload and guided the determination of expansion and shrinkage compensation agent dosages.

Conclusions:

• The PSO-BPNN model is a highly accurate and convenient tool for predicting SCC autogenous shrinkage strain prior to regulation.
• The model provides valuable guidance for implementing effective shrinkage compensation measures in practical engineering applications.
• Accurate shrinkage prediction facilitates optimized material dosages, leading to improved structural performance and reduced experimental effort.