By leveraging modern technologies such as Artificial Intelligence, Big Data analysis and Internet of Things, Agriculture 4.0 aims to optimise the entire agriculture and food production chain, allowing optimal management of resources, improving product quality and reducing waste and pollution. However, purely mechanical agricultural machinery cannot join the Digital Agriculture Revolution due to the lack of sensors and connectivity. To address this problem, in this thesis a cost-effective real-time condition monitoring system for rotating agricultural machines is presented. The proposed solution is based on vibration analysis and only requires a triaxial accelerometer to be fixed to the body of the implement, allowing non-invasive plug-and-play mounting. The system is able to recognize the operational state of the machinery, to estimate the angular velocity of the power take-off, and to detect anomalies and faults. These results are achieved using computationally efficient methods coming from the fields of Machine Learning, statistics, signal processing and Deep Learning, which are suited for running in real time on devices with limited computational resources. The effectiveness of the devised system is proved by experimental tests in which real-world data are used.
By leveraging modern technologies such as Artificial Intelligence, Big Data analysis and Internet of Things, Agriculture 4.0 aims to optimise the entire agriculture and food production chain, allowing optimal management of resources, improving product quality and reducing waste and pollution. However, purely mechanical agricultural machinery cannot join the Digital Agriculture Revolution due to the lack of sensors and connectivity. To address this problem, in this thesis a cost-effective real-time condition monitoring system for rotating agricultural machines is presented. The proposed solution is based on vibration analysis and only requires a triaxial accelerometer to be fixed to the body of the implement, allowing non-invasive plug-and-play mounting. The system is able to recognize the operational state of the machinery, to estimate the angular velocity of the power take-off, and to detect anomalies and faults. These results are achieved using computationally efficient methods coming from the fields of Machine Learning, statistics, signal processing and Deep Learning, which are suited for running in real time on devices with limited computational resources. The effectiveness of the devised system is proved by experimental tests in which real-world data are used.
Real-time condition monitoring in agricultural machinery via vibration analysis
MARITAN, ALESSIO
2021/2022
Abstract
By leveraging modern technologies such as Artificial Intelligence, Big Data analysis and Internet of Things, Agriculture 4.0 aims to optimise the entire agriculture and food production chain, allowing optimal management of resources, improving product quality and reducing waste and pollution. However, purely mechanical agricultural machinery cannot join the Digital Agriculture Revolution due to the lack of sensors and connectivity. To address this problem, in this thesis a cost-effective real-time condition monitoring system for rotating agricultural machines is presented. The proposed solution is based on vibration analysis and only requires a triaxial accelerometer to be fixed to the body of the implement, allowing non-invasive plug-and-play mounting. The system is able to recognize the operational state of the machinery, to estimate the angular velocity of the power take-off, and to detect anomalies and faults. These results are achieved using computationally efficient methods coming from the fields of Machine Learning, statistics, signal processing and Deep Learning, which are suited for running in real time on devices with limited computational resources. The effectiveness of the devised system is proved by experimental tests in which real-world data are used.File | Dimensione | Formato | |
---|---|---|---|
Maritan_Alessio.pdf
embargo fino al 04/09/2025
Dimensione
23.16 MB
Formato
Adobe PDF
|
23.16 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/33168