Blowout is one of the most dreaded accident for Oil and Gas companies. It is of particular concern during the drilling phase of deep-water oil & gas wells. This is due to the largely uncertain and extremely harsh environmental conditions that affect the design, drilling and operation activities of these wells. Seeking new technological solutions to prevent blowout has led, in the last few decades, to develop Managed Pressure Drilling (MPD) techniques. MPD offers many advantages compared to conventional drilling techniques, such as the capability (1) to detect the gas influx that initiates the kick that might lead to blowout and (2) to optimally control and circulate out this influx, to avoid the blowout. Nevertheless, effects of uncertainties on the MPD functionality is not fully understood and satisfactorily modelled. In this work, we propose a Dynamic Event Tree (DET) modelling framework of the scenario of kick escalation into blowout. The framework allows accounting for the uncertainties that affect not only the kick variables, but also the time and delay of the kick detection. The uncertainties affecting the kick variables are evaluated from kick events records taken from 2000 oil wells drilled in the Niger Delta, whereas the uncertainties affecting the time and delay of kick detection are evaluated by simulating the performance of the tool embedded into the MPD for kick detection (which applies the CUSUM statistical test to differential flow measurements), and assuming a possibilistic distribution for the confirmation time needed by operator to take counteracting measures with respect to the evolving accidental scenario. A Hybrid Monte Carlo and Possibilistic method is utilized to represent and propagate uncertainties associated to the events occurring throughout the DET. Results are compared with those of a Purely Probabilistic method in support of the blowout probability quantification.

Blowout is one of the most dreaded accident for Oil and Gas companies. It is of particular concern during the drilling phase of deep-water oil & gas wells. This is due to the largely uncertain and extremely harsh environmental conditions that affect the design, drilling and operation activities of these wells. Seeking new technological solutions to prevent blowout has led, in the last few decades, to develop Managed Pressure Drilling (MPD) techniques. MPD offers many advantages compared to conventional drilling techniques, such as the capability (1) to detect the gas influx that initiates the kick that might lead to blowout and (2) to optimally control and circulate out this influx, to avoid the blowout. Nevertheless, effects of uncertainties on the MPD functionality is not fully understood and satisfactorily modelled. In this work, we propose a Dynamic Event Tree (DET) modelling framework of the scenario of kick escalation into blowout. The framework allows accounting for the uncertainties that affect not only the kick variables, but also the time and delay of the kick detection. The uncertainties affecting the kick variables are evaluated from kick events records taken from 2000 oil wells drilled in the Niger Delta, whereas the uncertainties affecting the time and delay of kick detection are evaluated by simulating the performance of the tool embedded into the MPD for kick detection (which applies the CUSUM statistical test to differential flow measurements), and assuming a possibilistic distribution for the confirmation time needed by operator to take counteracting measures with respect to the evolving accidental scenario. A Hybrid Monte Carlo and Possibilistic method is utilized to represent and propagate uncertainties associated to the events occurring throughout the DET. Results are compared with those of a Purely Probabilistic method in support of the blowout probability quantification.

Blowout accident probability quantification by a dynamic event tree with hybrid probabilistic and possibilistic uncertainty treatment

ESLAMIAN, ALIREZA
2016/2017

Abstract

Blowout is one of the most dreaded accident for Oil and Gas companies. It is of particular concern during the drilling phase of deep-water oil & gas wells. This is due to the largely uncertain and extremely harsh environmental conditions that affect the design, drilling and operation activities of these wells. Seeking new technological solutions to prevent blowout has led, in the last few decades, to develop Managed Pressure Drilling (MPD) techniques. MPD offers many advantages compared to conventional drilling techniques, such as the capability (1) to detect the gas influx that initiates the kick that might lead to blowout and (2) to optimally control and circulate out this influx, to avoid the blowout. Nevertheless, effects of uncertainties on the MPD functionality is not fully understood and satisfactorily modelled. In this work, we propose a Dynamic Event Tree (DET) modelling framework of the scenario of kick escalation into blowout. The framework allows accounting for the uncertainties that affect not only the kick variables, but also the time and delay of the kick detection. The uncertainties affecting the kick variables are evaluated from kick events records taken from 2000 oil wells drilled in the Niger Delta, whereas the uncertainties affecting the time and delay of kick detection are evaluated by simulating the performance of the tool embedded into the MPD for kick detection (which applies the CUSUM statistical test to differential flow measurements), and assuming a possibilistic distribution for the confirmation time needed by operator to take counteracting measures with respect to the evolving accidental scenario. A Hybrid Monte Carlo and Possibilistic method is utilized to represent and propagate uncertainties associated to the events occurring throughout the DET. Results are compared with those of a Purely Probabilistic method in support of the blowout probability quantification.
BARALDI, PIERO
ZIO, ENRICO
ING - Scuola di Ingegneria Industriale e dell'Informazione
27-lug-2017
2016/2017
Blowout is one of the most dreaded accident for Oil and Gas companies. It is of particular concern during the drilling phase of deep-water oil & gas wells. This is due to the largely uncertain and extremely harsh environmental conditions that affect the design, drilling and operation activities of these wells. Seeking new technological solutions to prevent blowout has led, in the last few decades, to develop Managed Pressure Drilling (MPD) techniques. MPD offers many advantages compared to conventional drilling techniques, such as the capability (1) to detect the gas influx that initiates the kick that might lead to blowout and (2) to optimally control and circulate out this influx, to avoid the blowout. Nevertheless, effects of uncertainties on the MPD functionality is not fully understood and satisfactorily modelled. In this work, we propose a Dynamic Event Tree (DET) modelling framework of the scenario of kick escalation into blowout. The framework allows accounting for the uncertainties that affect not only the kick variables, but also the time and delay of the kick detection. The uncertainties affecting the kick variables are evaluated from kick events records taken from 2000 oil wells drilled in the Niger Delta, whereas the uncertainties affecting the time and delay of kick detection are evaluated by simulating the performance of the tool embedded into the MPD for kick detection (which applies the CUSUM statistical test to differential flow measurements), and assuming a possibilistic distribution for the confirmation time needed by operator to take counteracting measures with respect to the evolving accidental scenario. A Hybrid Monte Carlo and Possibilistic method is utilized to represent and propagate uncertainties associated to the events occurring throughout the DET. Results are compared with those of a Purely Probabilistic method in support of the blowout probability quantification.
Tesi di laurea Magistrale
File allegati
File Dimensione Formato  
thesis.pdf

non accessibile

Descrizione: thesis text
Dimensione 2.21 MB
Formato Adobe PDF
2.21 MB Adobe PDF   Visualizza/Apri
Fig.rar

non accessibile

Descrizione: Figures
Dimensione 7.06 MB
Formato RAR
7.06 MB RAR   Visualizza/Apri

I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10589/135356