Predictive Maintenance in HVAC System utilizing Machine Learning
Publicerad
Författare
Typ
Examensarbete för masterexamen
Modellbyggare
Tidskriftstitel
ISSN
Volymtitel
Utgivare
Sammanfattning
The fourth industrial revolution is present in today’s landscape of industrial engi- neering and
digitalization has emerged to be a vital part of an organization’s product portfolio. Industry
4,0 endorses companies an opportunity to make a superior in- formed fact-based decision.
Digitalization and creating more data-driven decision making is considered to be lucrative
and innovative enough to push organisations a step closer to Industry 4,0.
Swegon aims to investigate if it is feasible to implement a predictive type of main- tenance to
forecast when the wreckage is approaching in the HVAC systems. To guide Swegon AB closer
to the ideal Industry 4,0, a current situation analysis was conducted to examine if the
predictive type of maintenance is viable on Swegon current data by utilizing Machine
Learning.
A collaboration of Cross-Industry Process for Data Mining (CRISP-DM) and Prod- uct Development
methodologies have been utilized to prepare and create an under- standing of the input data,
to build the Machine Learning model based on input data from Swegon and also to measure the
overall potential of the input data. An analysis of the Machine Learning model was conducted and
this resulted in several recommendations for not only Swegon but every company trying to implement
pre- dictive maintenance using machine learning to continue pushing the organisation
closer towards Industry 4.0 and to accomplish a predictive type of maintenance
Beskrivning
Ämne/nyckelord
Predictive Maintenance, HVAC System, machine learning, wreckage