The interest in RAMS (Reliability, Availability, Maintainability and Safety) and diagnostics parameters is growing in many different manufacturing fields. These branches of knowledge are nowadays crucial and play a fundamental role in industrial engineering becoming focal part of performance requirements. Modern technologies and business requirements are producing a growth in variety and complexity of manufacturing product and this trend increased number and variety of failures. System downtime and unplanned outages massively affect plant productivity. In many Oil&Gas applications an emergency shutdown produces an interruption of normal running operation, a considerable productivity reduction and a loss of thousands dollars [1-2]. This is the reason why RAMS disciplines together with fault diagnosis and condition monitoring are almost mandatory in Oil&Gas applications where products are forced to endure extreme process and environmental conditions [3]. This thesis is focused on availability improvement and takes into account maintainability and, in particular, reliability roles in order to achieve this kind of target. The goal is to develop a procedure for availability improvement that engineers may used during the early stages of product design. Availability means that a system is “on-line” if it is involved in continuous running condition or “ready to use” in case of on-demand” usage. As said before, in modern systems there are a great variety of factors that can take a system off-line, ranging from scheduled maintenance downtime to catastrophic failures. The goal of improving system availability is to detect incipient failures, minimize downtime and minimize the time needed to restore the system to normal working conditions. Obviously the margin of downtime tolerance is directly associated with the system application and this requirement impose the complexity and the corresponding cost of the solution [4-6]. Reliability prediction is the main focus of this study since it turned out to be best method in RAM (Reliability, Availability and Maintainability) analysis for industrial applications: reliability prediction is very helpful in order to evaluate design feasibility, compare design choices, identify potential failure areas, trade-off system design factors and track reliability improvement. This is the reason why the best solution to improve system availability in the early product design stages turned out to be reliability-oriented since it provides reliability feedback to design engineers in order to reduce re-design costs and time for upgrades. This thesis is organized as follows: Chapter 1 contains a brief description of Life Data Analysis focusing on the comparison of two failure distributions, Exponential and Weibull. The second Chapter shows the best Availability improvement methods starting from standby redundancy and comparing cold and warm standby solutions. Chapter 3 deepens the Reliability Allocation procedures starting from a review of the methods described in literature and showing a new solution to achieve allocation parameters in complex systems; this Chapter contains also the description of a new Reliability Importance procedure (Credible Improvement Potential) and its application on Auxiliary Systems of a gas turbine. Chapter 4 describes the Condition-based Maintenance using Markov models with some applications in case of complex repair solutions and standby spares; Chapter 5 shows the basis of fault detection, isolation, reconfiguration, diagnostics and Condition Monitoring. This Chapter contains both on-board and logic solver diagnostics with a detailed application on a gas turbine safety loop and corresponding Probability of Failure on Demand (PFD) assessment [7-8]. The final Chapter describes the Reliability Assessment Loop with the brand new approach proposed and show the potential of the tool that was developed to achieve a reliability prediction in the early product design stages.

Simulation models and reliability assessment for gas turbine auxiliary systems / Venzi, Matteo. - (2017).

Simulation models and reliability assessment for gas turbine auxiliary systems

VENZI, MATTEO
2017

Abstract

The interest in RAMS (Reliability, Availability, Maintainability and Safety) and diagnostics parameters is growing in many different manufacturing fields. These branches of knowledge are nowadays crucial and play a fundamental role in industrial engineering becoming focal part of performance requirements. Modern technologies and business requirements are producing a growth in variety and complexity of manufacturing product and this trend increased number and variety of failures. System downtime and unplanned outages massively affect plant productivity. In many Oil&Gas applications an emergency shutdown produces an interruption of normal running operation, a considerable productivity reduction and a loss of thousands dollars [1-2]. This is the reason why RAMS disciplines together with fault diagnosis and condition monitoring are almost mandatory in Oil&Gas applications where products are forced to endure extreme process and environmental conditions [3]. This thesis is focused on availability improvement and takes into account maintainability and, in particular, reliability roles in order to achieve this kind of target. The goal is to develop a procedure for availability improvement that engineers may used during the early stages of product design. Availability means that a system is “on-line” if it is involved in continuous running condition or “ready to use” in case of on-demand” usage. As said before, in modern systems there are a great variety of factors that can take a system off-line, ranging from scheduled maintenance downtime to catastrophic failures. The goal of improving system availability is to detect incipient failures, minimize downtime and minimize the time needed to restore the system to normal working conditions. Obviously the margin of downtime tolerance is directly associated with the system application and this requirement impose the complexity and the corresponding cost of the solution [4-6]. Reliability prediction is the main focus of this study since it turned out to be best method in RAM (Reliability, Availability and Maintainability) analysis for industrial applications: reliability prediction is very helpful in order to evaluate design feasibility, compare design choices, identify potential failure areas, trade-off system design factors and track reliability improvement. This is the reason why the best solution to improve system availability in the early product design stages turned out to be reliability-oriented since it provides reliability feedback to design engineers in order to reduce re-design costs and time for upgrades. This thesis is organized as follows: Chapter 1 contains a brief description of Life Data Analysis focusing on the comparison of two failure distributions, Exponential and Weibull. The second Chapter shows the best Availability improvement methods starting from standby redundancy and comparing cold and warm standby solutions. Chapter 3 deepens the Reliability Allocation procedures starting from a review of the methods described in literature and showing a new solution to achieve allocation parameters in complex systems; this Chapter contains also the description of a new Reliability Importance procedure (Credible Improvement Potential) and its application on Auxiliary Systems of a gas turbine. Chapter 4 describes the Condition-based Maintenance using Markov models with some applications in case of complex repair solutions and standby spares; Chapter 5 shows the basis of fault detection, isolation, reconfiguration, diagnostics and Condition Monitoring. This Chapter contains both on-board and logic solver diagnostics with a detailed application on a gas turbine safety loop and corresponding Probability of Failure on Demand (PFD) assessment [7-8]. The final Chapter describes the Reliability Assessment Loop with the brand new approach proposed and show the potential of the tool that was developed to achieve a reliability prediction in the early product design stages.
2017
Marcantonio Catelani
ITALIA
Venzi, Matteo
File in questo prodotto:
File Dimensione Formato  
Tesi_dottorato_Matteo_Venzi.pdf

accesso aperto

Descrizione: Tesi di dottorato
Tipologia: Tesi di dottorato
Licenza: Open Access
Dimensione 19.2 MB
Formato Adobe PDF
19.2 MB Adobe PDF

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

Utilizza questo identificatore per citare o creare un link a questa risorsa: https://hdl.handle.net/2158/1081038
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact