Publication: Real-time diagnostics of centrifugal pump with Raspberry PI
No Thumbnail Available
Date
2024-07-12
Authors
Emil Joseph Donsia
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This research project focuses on the development of real-time diagnostics of centrifugal pumps with Raspberry Pi. The Raspberry Pi is used as the DAQ unit which serves to collect data, preprocess data, handle data, and transmit data. By using the Raspberry Pi to interface the MPU6050 sensor, raw data from the centrifugal pump can be obtained which then undergoes preprocessing to extract the statistical features of both the time domain and frequency domain using FFT. Once the raw data is preprocessed, it can then be handled and transmitted for further analysis, therefore, enabling the Raspberry Pi to be a DAQ unit for this project. By collecting, preprocessing, and handling the raw data, a dataset can be prepared which is then utilized to train the ML model using Python in Jupyter Notebook. The SVM algorithm is utilized train to train the ML model using appropriate techniques to construct a reliable model with great predictive capabilities. Node-RED is utilized to build the IoT application by creating flows that handle incoming live data and employing the trained ML model to make predictions. Additionally, Node-RED is also used to construct the dashboard to visualise real-time monitoring and diagnostics of an operating centrifugal pump. In conclusion, this research project successfully employs the Raspberry Pi as a DAQ unit, developing an ML model with accurate predictive capabilities and an IoT system comprising the pump, DAQ, and Node-RED. By providing engineers with real-time diagnostic information and prognostic insights, the system contributes to the field of pump diagnostics by enhancing the efficiency of maintenance planning and reducing equipment failure rates. The project’s main achievements lie in the utilization of Raspberry Pi as a DAQ unit, the development of an accurate ML model for pump diagnostics, and the integration of the trained model into Node-RED for creating an IoT application for pump diagnostics and monitoring.