FBG-Based Temperature Sensors for Liquid Identification and Liquid Level Estimation via Random Forest

Area of Science:

• Sensor Technology
• Optical Sensing
• Machine Learning Applications

Background:

• Accurate liquid level measurement and fluid classification are crucial in various industrial processes.
• Traditional methods can be limited by environmental factors like temperature variations.
• Fiber Bragg Gratings (FBGs) offer a robust sensing platform for such applications.

Purpose of the Study:

• To develop and validate a liquid level measurement and classification system.
• To assess the performance of a fiber Bragg grating (FBG) temperature sensor array combined with the random forest (RF) algorithm.
• To demonstrate the system's capability in distinguishing between oil and non-oil fluids under controlled temperature conditions.

Main Methods:

• Utilized a fiber Bragg grating (FBG) temperature sensor array for sensing.
• Employed the random forest (RF) algorithm for both fluid classification (oil vs. non-oil) and level measurement.
• Heated fluids using a Peltier device to maintain a constant experimental temperature (318.15 K).

Main Results:

• Achieved 100% accuracy in fluid identification using the RF algorithm.
• Obtained an average root mean squared error (RMSE) of 0.2603 cm for fluid level measurement, with a maximum RMSE below 0.4 cm.
• Successfully identified water, mineral oil, and silicone oil using three distinct FBGs.

Conclusions:

• The proposed FBG-based system with RF algorithm is a feasible tool for fluid identification and level estimation.
• The method demonstrates robustness under temperature variation conditions.
• The system offers practical benefits due to its easy assembly and straightforward operation.