Machine learning-based fusion of experimental and synthetic data for reliable prediction of steel connection stiffness

The development of robust prediction tools based on Machine Learning (ML) techniques requires the availability of complete, consistent, accurate, and numerous datasets. The application of ML in structural engineering has been limited since, although real size experiments provide complete and accurate data, they are time consuming and expensive. On the other hand, validated Finite Element Models (FEM) provide consistent and numerous results but they usually require large computational time and cost, and could be subjected to convergence issues due to the complexity of the simulated problem. Hybrid approaches to combine experimental and synthetic datasets have emerged as an alternative to improve the reliability of ML model predictions. In this paper, we explore two hybrid methods to propose a robust approach for the prediction of the Extended Hollo-Bolt (EHB) connection stiffness: i) Artificial Neural Networks (ANN) with data fusion, and ii) ML methods with Particle Swarm Optimization (PSO). Based on the analysis of a dataset that combines the experimental results with a synthetic dataset based on FEM, we concluded that ML (using ANN) with PSO is suitable for the prediction of the connection stiffness, given the limited number of experimental data points. However, ANN with data fusion shows a promising method for cases with more availability of experimental data