Enhancing Measurement Precision in FBG Sensors Through Robust Temperature Compensation Mechanisms

With the development of Fiber Bragg Grating (FBG) sensors, applications such as Structural Health Monitoring (SHM) systems, aerospace systems, and medical diagnostics have become highly accurate and more demanding. However, the use of FBG sensors is limited when the temperature varies since it brings about lateral cross-sensitivity errors in performance due to the wavelength shifting with temperature changes. This paper suggests two simple methods of avoiding this extreme that are inexpensive to compute: linear regression and polynomial fitting. The better temperature compensation model is incorporated to check the overall performance. Linear regression is a straight-line fitting method that applies to data that does not exhibit much curvature in the 25% improvement in measurement accuracy. In cases where complex temperature dependencies are present, polynomial fits are offered which improves accuracy by up to 30% by accounting for the non-linear tail showing more complex characteristics. Investigations using laboratory experimental data show the success of these techniques and show that indeed polynomial fitting worked better than linear fitting in measurements that were time-varying. Relatively easy to apply and real-time capable are the unique attributes of the proposed methods in comparison to earlier approaches such as dual FBG and reference stripping techniques. The results appropriately indicate their contribution to improving the dependability of FBG sensors under various conditions. The work will go on to consider how adaptive models and machine learning can help future FBG sensor measurements become more robust.