AI-assisted optimization design of seismic performance parameters for timber structures

Area of Science:

• Structural Engineering
• Artificial Intelligence
• Sustainable Architecture

Background:

• Timber multi-story buildings offer sustainability and seismic resilience advantages.
• Optimizing seismic parameters like inter-story drift and roof displacement in timber structures using AI remains a challenge.
• Existing frameworks lack integrated AI solutions for comprehensive seismic performance optimization.

Purpose of the Study:

• To introduce an AI-based framework, the Gradient Boosted Random Forest Machine with Scalable Cheetah Optimizer (GBRF-SCO), for optimizing seismic parameters in timber buildings.
• To improve prediction accuracy for seismic response metrics and guide engineers in selecting robust structural designs.
• To enhance the seismic resilience and sustainability of multi-story timber constructions.

Main Methods:

• Utilized a dataset of 4,000 timber building samples from the Timber Seismic Performance Dataset.
• Applied data pre-processing techniques including Robust Scaling and Isolation Forest for normalization and outlier detection.
• Employed t-Distributed Stochastic Neighbor Embedding (t-SNE) for exploratory data analysis and visualization of feature relationships.
• Implemented the GBRF model for seismic response prediction and the SCO for hyperparameter optimization.
• Used Multiple Linear Regression (MLR) to analyze the influence of structural and seismic elements on roof displacement.

Main Results:

• The GBRF-SCO framework achieved a high prediction accuracy of 0.949 for classifying roof displacement levels (low, medium, high) under seismic conditions.
• Demonstrated superior performance compared to conventional regression and ensemble methods.
• Provided insights into key structural and seismic factors influencing timber building seismic performance.
• Successfully optimized seismic performance characteristics, enabling informed design decisions.

Conclusions:

• The GBRF-SCO framework offers a robust and intelligent method for optimizing the seismic performance of timber structures.
• This AI-driven approach significantly improves the design of sustainable and earthquake-resilient multi-story timber buildings.
• The study highlights the potential of integrated AI frameworks in advancing structural engineering for enhanced safety and environmental benefits.