Related Experiment Video
Updated: Sep 6, 2025
04:54
Author Spotlight: A Stable Phantom Material for Optical and Acoustic Imaging
Published on: June 16, 2023
3.1K
Fiber-Optic Hydrophone Based on Michelson's Interferometer with Active Stabilization for Liquid Volume Measurement
Welton Sthel Duque1, Camilo Arturo Rodríguez Díaz1, Arnaldo Gomes Leal-Junior1
1Telecommunications Laboratory (LABTEL), Graduate Program in Electrical Engineering, Federal University of Espírito Santo (UFES), Vitória 29075-910, ES, Brazil.
Sensors (Basel, Switzerland)
|June 24, 2022
Summary
This study introduces a fiber-optic hydrophone (FOH) system for precise liquid volume measurement in industrial tanks. Utilizing ultrasound waves and machine learning, the FOH sensor achieves 99.4% accuracy in determining liquid volumes.
Area of Science:
- Optical Fiber Sensing
- Acoustic Wave Detection
- Machine Learning Applications
Background:
- Optical fiber sensing has been utilized across various industries since the 1970s.
- Continuous liquid volume measurement in production tanks requires reliable sensing solutions.
- Fiber-optic hydrophone (FOH) sensors offer a potential method for detecting ultrasound acoustic waves.
Purpose of the Study:
- To develop and evaluate a fiber-optic hydrophone (FOH) system for accurate liquid volume measurement.
- To investigate the use of machine learning algorithms for predicting liquid volumes based on acoustic wave patterns.
- To mitigate environmental noise interference in optical fiber sensing systems.
Main Methods:
- A Michelson's interferometer (MI) based FOH system was designed using single mode fiber (SMF) coils.
- An active stabilization mechanism with a piezoelectric actuator (PZT) was employed to reduce noise.
- Ultrasound waves were used to generate acoustic patterns within a test tank, with amplitudes and phases measured.
- Machine learning algorithms, including k-nearest neighbors (k-NN) and Gaussian process regression, were applied for volume prediction.
Main Results:
- The FOH system demonstrated a liquid volume resolution of 1 mL.
- High sensitivity was achieved with 340 mrad/mL and 70 mvolts/mL.
- K-nearest neighbors (k-NN) classification achieved 99.4% accuracy in volume determination.
- Gaussian process regression yielded a root mean squared error (RMSE) of 0.211 mL.
Conclusions:
- The developed FOH system, coupled with machine learning, provides a highly accurate method for liquid volume measurement.
- The system effectively addresses the challenges of nonlinear acoustic wave patterns and environmental noise.
- This technology holds significant potential for continuous monitoring applications in various industrial sectors.
