A Review of GFRP Elastic Gridshell Structures and Machine Learning Algorithms
Soheila Kookalani, Hamidreza Alavi, Farzad Pour Rahimian · 2025
Elastic gridshells represent a unique structural system characterized by their ability to achieve complex, curved geometries using initially flat elements that are elastically deformed into shape. These lightweight and efficient structures have gained attention in recent years, particularly with the growing use of innovative materials like glass fiber reinforced polymer (GFRP). This chapter provides a comprehensive review of the definitions and characteristics of elastic gridshells, with an emphasis on GFRP materials, bolted joints, and bracing systems. It discusses various design strategies, form-finding techniques, and construction methods specific to GFRP gridshells, highlighting both the opportunities and challenges presented by this material in structural applications. In addition, the chapter explores a range of machine learning (ML) algorithms, including LR, RR, KNN, DT, RF, AdaBoost, XGBoost, CatBoost, LightGBM, ANN, LSSVM, WLSSVM, PSO-LSSVM, PIN-SVM, ε-TSVM, and WL-ε-TSVM. The chapter further introduces interpretable ML approaches such as PDP, ALE, and SHAP. Several methodologies for developing ML models, including K-fold CV, Taguchi methods, TOPSIS, performance indices, and MOPSO, are also presented to provide a holistic view of ML applications in structural gridshell design and analysis.