Research on intelligent semi-active control algorithms and seismic reliability based on machine learning.

Area of Science:

• Civil Engineering
• Structural Engineering
• Seismic Engineering

Background:

• Existing semi-active control algorithms lack robustness.
• Current seismic reliability evaluation methods have limited generalization.
• Aqueduct structures require robust seismic control strategies.

Purpose of the Study:

• To propose a novel probability density evolution approach for seismic reliability assessment.
• To develop an intelligent semi-active control algorithm using machine learning.
• To evaluate the effectiveness of semi-active control for aqueducts under random seismic events.

Main Methods:

• Probability density evolution for seismic reliability analysis.
• Machine learning for developing an intelligent semi-active control algorithm.
• Engineering case studies to validate control strategies.

Main Results:

• The proposed method enables comprehensive evaluation of control effectiveness.
• The machine learning-based semi-active control algorithm significantly improves seismic reliability.
• Demonstrated higher seismic reliability compared to the uncontrolled state under random seismic effects.

Conclusions:

• The novel probability density evolution approach enhances seismic reliability assessment.
• Machine learning-powered semi-active control offers superior performance for aqueducts.
• Intelligent control strategies are effective for mitigating seismic risks in structures.