Shape deformation tracking algorithm for multi-storey building using Fibre Bragg Grating sensor and Neuro-Fuzzy wavelet

Structural Health Monitoring (SHM) is essential for evaluating the safety and durability of multi-storey buildings, especially under progressive loading and deformation. This paper proposes a hybrid neuro-fuzzy wavelet-based system that integrates Fiber Bragg Grating (FBG) sensor networks with machine learning for real-time deformation analysis. The system captures Bragg wavelength shifts due to strain, stress, and angular deformation across multiple floors and extracts features using Continuous Wavelet Transform (CWT) with a Mexican Hat wavelet. These features are interpreted using a Mamdani fuzzy inference system and fed into a Multi-Layer Perceptron (MLP) for structural health prediction. Simulation results on a five-storey building model show high prediction accuracy with an MSE of 0.1594, R² score of 0.931, and MAPE of 5.81%. Cross-validation confirms consistent performance with low variance. The system accurately identifies early-stage collapse risks and localized strain variations with interpretable confidence scores. Compared to existing FBG-based SHM methods, the proposed model offers better physical interpretability, improved numerical accuracy, and reduced computational complexity. This approach demonstrates strong potential for deployment in intelligent SHM systems and early warning frameworks in civil engineering.