My Final Year Project

The process of maintenance becomes more and more important nowadays when the number of installed industrial robots or machines rise exponentially. The traditional manual maintenance is insufficient to support the advancement of robotic technology and the latest machinery complexity design. Therefore, this project is carried out to develop a self-supervised maintenance system for fault detection and identification by using IOT solution with machine learning model. The overall target of the project is to design a product solution to increase the efficiency of maintenance system.

A Raspberry Pi model 3 is used to construct the Internet of Things (IoT) device to collect the machinery data were and send to the SAS server in every second.

 
Figure 1.0: Technology outlook of the on-cloud analytical system.

Data collected undergone data analysis using Statistical Analysis Software (SAS) and all the procedures are written in SAS language.

 
Figure 1.1: Log file of REST API based on HTTP.

A trained machine learning model was deployed in the system to find out the pattern in different condition to identify the machine fault condition. This system was tested on a simulated model of an industrial robot, Automated Guided Vehicle (AGV) to identify the performance of the device designed and check the functionality of the system.

 
Figure 1.2: Zalpha Series AGV and Simulated Robot Model.

The real-time data and the analysis outcome were visualized in web-based form and APP form on any mobile device. There are some alert notification technique also involved to let the data analysis result to interface with industry personel.

Figure 1.3: Real-time data visualization in web-based form.

 
Figure 1.4: Visualization of machine learning model output in APP form.

This project also was used to participated SAS track Innovate Malaysia Design Challenge and was awarded.

Figure 1.5: Scene during competition.

 
Figure 1.6: The winner of SAS track IMDC.