HomePublicationsBEng Thesis

Smart Predictive Maintenance using Azure Services
Ref: CISTER-TR-191216       Publication Date: 8, Oct, 2019

Abstract:
With the constant evolution of technology and with home appliances having an increasing influence on people’s lives, it is mandatory to monitor the way these devices are working to make sure they do not show any flaws or risk of being damaged. In order to do this monitorization and avoid this type of situations, CISTER Research Center started focusing on a way to deliver an application that could detect whether a home appliance is functioning properly or not. This way, it could ease the process of fixing the device as the companies would not lose much time in the process. Based on this big necessity, Smart PDM project was born. The original Smart-PDM project lacked maturity as it was only able to acquire the home appliance’s data by using a specific device called Sonoff Pow and save it on a local database. With the Sonoff Pow ready, the focus was then on finding a way to understand how to detect flaws on a home appliance. This new project, which acts as a new iteration of the Smart-PDM project, has the goal of finishing what the previous iteration could not, the data analysis part using algorithms, and in addition doing all this process only supported by Microsoft Azure technologies. The changes on the type of technologies used would force this new iteration to refine what was done on the previous ones. As planned, the new iteration of the Smart PDM project can now detect with a relatively low window of error whether a home appliance is working as it should or not, even though it is not ready to be shown to the public as the UI is not properly updated, the detection needs to be perfected and the process of device configuration shows some flaws.

Tiago Coelho