Pressure-sensitive platform based on multiplexed in-series macro-bend optical fiber sensors.

Area of Science:

• Optoelectronics
• Sensor Technology
• Data Science

Background:

• Pressure sensing platforms are crucial in various applications.
• Existing methods may face limitations in spatial resolution or complexity.
• Optical fiber sensors offer a promising alternative due to their robustness and sensitivity.

Purpose of the Study:

• To develop and validate a novel pressure-sensitive platform using macro-bend optical fiber sensors.
• To demonstrate the capability of detecting pressure location with fewer sensors than sensing cells.
• To analyze the effectiveness of principal component analysis (PCA) and machine learning algorithms for spectral data interpretation.

Main Methods:

• Instrumentation of a 20x20cm platform with five in-series macro-bend optical fiber sensors, divided into sixteen 5x5cm sensing cells.
• Sensing principle based on wavelength-dependent intensity changes in the visible spectrum of the optical fiber array.
• Data analysis employing principal component analysis (PCA) to reduce spectral data dimensionality, followed by k-nearest neighbors (KNN) classification and support vector regression (SVR).

Main Results:

• PCA successfully reduced spectral data to 12 principal components, explaining 99% of data variance.
• The system demonstrated accurate pressure location prediction with 94% accuracy.
• A mean absolute error of 0.31 kPa was achieved for pressure detection within the 3.74-9.98 kPa range.

Conclusions:

• The developed optical fiber sensor platform effectively detects and localizes pressure.
• The study validates the use of PCA and machine learning for efficient analysis of spectral data from optical sensors.
• This technology shows potential for advanced pressure monitoring applications with reduced sensor count.