Moisture content monitoring in industrial-scale composting systems using low-cost sensor-based machine learning...

Area of Science:

• Environmental Science
• Agricultural Engineering
• Sensor Technology

Background:

• Effective composting relies on optimal moisture levels for efficiency and reduced environmental impact.
• Accurate, real-time moisture monitoring in industrial composting remains a significant technological challenge.
• Existing methods for moisture determination can be costly or lack real-time capabilities.

Purpose of the Study:

• To design and develop a hardware and software model for a low-cost capacitive moisture sensor.
• To enable self-adjustment of the sensor using machine learning techniques for improved accuracy.
• To address the need for accessible, real-time moisture monitoring in industrial composting.

Main Methods:

• Utilized organic compost samples with varied waste compositions and composting stages.
• Applied machine learning algorithms for the self-calibration and adjustment of the capacitive sensor.
• Validated sensor predictions against the established gravimetric method in a laboratory setting.

Main Results:

• The developed sensor model demonstrated high efficiency and reliability in measuring compost moisture.
• Achieved a strong correlation coefficient of 0.9939 between gravimetric analysis and sensor predictions.
• The machine learning-based self-adjustment significantly improved sensor performance.

Conclusions:

• The proposed low-cost capacitive moisture sensor with machine learning self-adjustment is an effective and reliable tool.
• This technology can significantly enhance the efficiency and reduce the environmental impact of industrial composting.
• The validated model offers a practical solution for real-time moisture monitoring in composting operations.