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

• Published in: Bioresource technology Vol. 359; p. 127456
• Main Authors: Moncks, P.C.S., Corrêa, É.K., L. C. Guidoni, L., Moncks, R.B., Corrêa, L.B., Lucia Jr, T., Araujo, R.M., Yamin, A.C., Marques, F.S.
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 01-09-2022
• Subjects: Artificial intelligence; Compost management; IBK; Linear regression; MLP; Remote sensing; MLP; Compost management; IBK; Linear regression; Artificial intelligence; Remote sensing; linear regression; remote sensing; artificial intelligence; compost management
• ISSN: 0960-8524; 1873-2976
• DOI: 10.1016/j.biortech.2022.127456

[Display omitted]
•Low-cost real-time monitoring for industrial composting is a challenge.•Moisture measurement in composting through a capacitive sensor was validated.•Machine learning enables self-adjustment for different composts.•Portable sensors eliminate the need to send samples to laboratories.•The IBK algorithm allows sensor self-adjustment.
Moisture is a key aspect for proper composting, allowing greater efficiency and lower environmental impact. Low-cost real-time moisture determination methods are still a challenge in industrial composting processes. The aim of this study was to design a model of hardware and software that would allow self-adjustment of a low-cost capacitive moisture sensor. Samples of organic composts with distinct waste composition and from different composting stages were used. Machine learning techniques were applied for self-adjustment of the sensor. To validate the model, results obtained in a laboratory by the gravimetric method were used. The proposed model proved to be efficient and reliable in measuring moisture in compost, reaching a correlation coefficient of 0.9939 between the moisture content verified by gravimetric analysis and the prediction obtained by the Sensor Node.