polito.it
Politecnico di Torino (logo)

Use of Machine Learning to Optimize Drilling Equipment Performance and Life Cycle Focus: Predictive Maintenance

Michel Kattar

Use of Machine Learning to Optimize Drilling Equipment Performance and Life Cycle Focus: Predictive Maintenance.

Rel. Marilena Cardu. Politecnico di Torino, Corso di laurea magistrale in Petroleum And Mining Engineering (Ingegneria Del Petrolio E Mineraria), 2022

[img]
Preview
PDF (Tesi_di_laurea) - Tesi
Licenza: Creative Commons Attribution Non-commercial No Derivatives.

Download (5MB) | Preview
Abstract:

In the oil and gas industry, any unplanned breakdown that stops the operations in progress can cost the companies millions of dollars. Being able to predict machinery failures and perform a scheduled maintenance ahead of time, won’t only increase the life of the equipment but will also reduce the sudden breakdowns saving the industry a lot of money. To add on that, even the maintenance costs will be reduced since unnecessary ones will not take place anymore. This study talks about all the steps done from real data collection, data cleaning and processing as well as the model generation for the failure prediction of some components of the drilling rig. These equipment are the blower and the lube oil of the top drive system, as well as the filter, the heat exchanger and the hydraulic pumps of the hydraulic power unit. The analytics and modelling parts were all performed using ThingWorx analytics. All the steps are explained and discussed on how and why they were performed. The model created for the blower didn’t give a good accuracy and was the only model possible to create from all these components due to some data quality issues. Even though the results weren’t as good as expected, a huge potential for future work is present to achieve the initially set goal of performing a predictive maintenance program. This potential and how can it be achieved is also discussed in this work.

Relatori: Marilena Cardu
Anno accademico: 2022/23
Tipo di pubblicazione: Elettronica
Numero di pagine: 101
Soggetti:
Corso di laurea: Corso di laurea magistrale in Petroleum And Mining Engineering (Ingegneria Del Petrolio E Mineraria)
Classe di laurea: Nuovo ordinamento > Laurea magistrale > LM-35 - INGEGNERIA PER L'AMBIENTE E IL TERRITORIO
Aziende collaboratrici: Drillmec Spa
URI: http://webthesis.biblio.polito.it/id/eprint/24329
Modifica (riservato agli operatori) Modifica (riservato agli operatori)