Development Of Machine Learning User Interface For Pump Diagnostics
Loading...
Date
2022-07-01
Authors
Lee, Zhao Yang
Journal Title
Journal ISSN
Volume Title
Publisher
Universiti Sains Malaysia
Abstract
The main objectives for this project are focusing on the development of user
interface that can connect with the machine learning build in Microsoft Azure for pump
diagnostic purpose. Water pump is a very common hydraulic machine which will
convert the rotational mechanical energy become hydraulic energy while the hydraulic
energy is in the form of pressure energy. Even though the water pump is designed to be
low-maintenance, highly efficient, and simple to operate, but we cannot observe the
blockage condition of the water pump from its exterior due to the fully enclosed system.
The blockage of the pump inlet could result in cavitation or mechanical parts breakdown
which would increase the maintenance cost. Machine Learning is one of the ways as a
preventive method by applying the data collected from the clogging experiment in the
vibration lab to build up a machine learning model for classification of flow blockage levels
in the centrifugal pump. The data collected for this machine learning model is using the
statistically significant features from vibration and acoustic analysis. The features extracted
of time domain and frequency domain in vibration and acoustic will use as database of a
Support Vector Machine (SVM) algorithms by using MATLAB R2021a. The result from
the SVM algorithms will be used as database for the machine learning in Microsoft Azure.
Build up a user interface by using Visual Studio Code (VSC) to run the coding of Cascading
Style Sheet (CSS), Hyper Text Markup Language (HTML) and JavaScript (JS) as a
webpage and connect to Azure Machine Learning Model and this will allow the user from
using the model from a webpage when they have active internet with any devices.