High-sensitivity ocean temperature sensor using a reflective optical microfiber coupler and machine learning methods

Area of Science:

• Oceanography
• Materials Science
• Sensor Technology

Background:

• Accurate seawater temperature monitoring is crucial for understanding ocean dynamics and climate change.
• Existing sensors face challenges with durability, multi-parameter interference, and environmental adaptability.
• Novel sensor designs are needed to overcome these limitations for reliable oceanographic data collection.

Purpose of the Study:

• To develop and evaluate a novel optical microfiber coupler sensor with a reflective silver mirror (OMCM) for high-precision seawater temperature measurement.
• To enhance sensor performance through encapsulation in polydimethylsiloxane (PDMS) and a pressure-immune metal casing.
• To investigate the effectiveness of machine learning algorithms for demodulating sensor signals and reducing errors.

Main Methods:

• Fabrication of an OMCM sensor encapsulated in PDMS and housed in a specialized metal casing.
• Experimental evaluation of temperature sensitivity and compressive strength.
• Application of particle swarm optimization support vector regression (PSO-SVR) for signal demodulation and comparison with traditional methods.

Main Results:

• The OMCM sensor demonstrated enhanced sensitivity and durability.
• The specialized casing effectively isolated temperature measurements from pressure interference.
• PSO-SVR significantly reduced the mean absolute percentage error (MAPE) from 2.16% to 0.157% compared to traditional methods.

Conclusions:

• The novel OMCM sensor provides a robust and accurate solution for seawater temperature monitoring.
• Machine learning demodulation, particularly PSO-SVR, substantially improves measurement precision.
• This technology offers an effective tool for high-precision ocean environmental temperature monitoring in harsh conditions.