Physics-Based Calibration of the Fiber Bragg Grating Temperature Response via Particle Swarm Optimization

Fiber Bragg grating (FBG) sensors are widely used for temperature monitoring because of their compact size, wavelength-encoded response, corrosion resistance, multiplexing capability, and immunity to electromagnetic interference. However, the accuracy of physics-based FBG models depends strongly on the thermal and optical parameters adopted in the simulation. This study developed a MATLAB-based physics-driven framework for calibrating the temperature response of a uniform silica FBG using particle swarm optimization (PSO). The model was formulated from the Bragg condition and incorporated temperature-dependent variations in effective refractive index and grating period through the thermo-optic and thermal-expansion coefficients. A published 17-point temperature–wavelength dataset spanning approximately 9.49–90.34 °C was digitized for model calibration and comparison. The extracted experimental sensitivity was 9.9450 pm/°C, compared with 10.13 pm/°C reported in the source study. The baseline model predicted a sensitivity of 14.1799 pm/°C, with RMSE, MAE, MAPE, and R² values of 0.20353 nm, 0.17446 nm, 43.90%, and 0.3151, respectively. Following PSO calibration, the predicted sensitivity decreased to 9.8888 pm/°C, differing by only 0.57% from the extracted experimental value. The calibrated model achieved an RMSE of 0.00521 nm, MAE of 0.00405 nm, MAPE of 1.38%, and R² of 0.99955. These results demonstrate that bounded PSO calibration can substantially improve agreement between a transparent physics-based FBG model and experimental data. Because the same dataset was used for parameter estimation and performance assessment, the results represent in-sample calibration performance rather than independent predictive validation.