HomePublicationsBEng Thesis

Smart Maintenance of Home Appliances
Ref: CISTER-TR-191214       Publication Date: 8, Oct, 2019

Abstract:
In several countries around the world, home appliances have a minimal warranty of two years since the date they are bought. Due to their electromechanical nature, during its lifecycle several failures can occur, some are related to poor usage, others are related to the malfunctioning of components or even due to the natural wear and tear of certain parts. The project “Smart Maintenance in Home Appliances” (Smart-PDM) aims to prevent and detect home appliance failures by acquiring its energy consumption pattern and detecting variations or abnormal patterns. Any variation that is abnormal is a good indicator that something is not working properly, requiring external intervention in order to identify the failure and possibly repair it. The full project implementation is described in the report including the study done to learn the areas of Internet of Things and Machine Learning. An Android Application was developed to install the sensors on the personal Wi-Fi network and to view the energy consumption of home appliances. A Web Application was implemented with the main functions of storing the consumption information of the home appliances and invoking a Data Analysis component. This component, in addition to analyzing and processing the data, involved the implementation of Machine Learning techniques to predict or classify the health status of home appliances. Finally, the report presents the conclusions drawn essentially from the Machine Learning area, such as what is the best machine learning technique (namely, classification algorithms and pattern matching) for each situation.

Rita Sousa

Document:

View
PDF Download
PDF