Subsurface Safety Valves: Safety Asset or Safety Liability?

Abstract

Summary This paper summarizes the methods used to compare the risk of a blowout for a well completion with a subsurface safety valve (SSSV) to a completion without an SSSV. These methods, which could be applied to any field, include a combination of SSSV reliability and conventional risk analyses. The Kuparuk River Unit Working-Interest Owners recently formed a group to examine the risks associated with installing and maintaining SSSV's in the Kuparuk field. Considering Kuparuk field operating conditions, the group was charged with determining whether SSSV's are a safety asset or whether the numerous operating and maintenance procedures make them a safety liability. The results indicate that, for the Kuparuk River Unit, an SSSV becomes a safety liability when the mean time between SSSV failures is less than 1 year. Because current SSSV mean time to failure (MTTF) at Kuparuk is approximately 1,000 days, they are considered a safety asset Introduction The use of SSSV's in onshore North Slope development wells was adopted as a statutory requirement to prevent oil spills caused by casing collapse in a permafrost environment. The primary reason for installing an SSSV in a well completion is to reduce the risk of a blowout, an uncontrolled flow of well fluid to the environment. However, a number of wireline and workover operations specifically conducted to service SSSV's in an individual well are likely. These operations, in turn, represent additional risk of a blowout. A typical risk analysis was used to compare risk with and without SSSV'S. The steps of the analysis are as follows.Define the equipment system, in this case the two alternative well completions to be studied.Develop the possible failure modes that could affect the reliability of the system.Build a fault tree to describe these failure relations.Develop component failure rates for each branch of the fault tree, including the development of the relationship of SSSV availability to the other components of the fault tree.Calculate the total failure probability as a function of SSSV MTTF. Not all the component failure rates were derived easily because of the nature of the data required. Thus some of the failure rates are based on engineering judgment. This does not affect greatly the comparison of two very similar well completions, because the relative failure rate is more important than the absolute failure rate. When a component failure rate appeared to have significant influence on the results a range of values was used in the calculation. System Definition The standard production well completion for the Kuparuk field (Fig. 1) has a tubing-retrievable subsurface safety valve (TRSSSV) installed at approximately 2,000 ft [610 m] measured depth below the surface. In the rare event of a TRSSSV failure, the first repair method is by wireline manipulation. If this is unsuccessful, the TRSSSV is locked in the open position, and a wireline-retrievable subsurface safety valve (WRSSSV) is used as an insert inside the TRSSSV. If the WRSSSV fails, it can be retrieved and replaced by wireline operations. If the failed WRSSSV cannot be retrieved or repaired, however, a workover is required. A workover is required also in the event of a leak in the subsurface hydraulic control line. These Kuparuk field operational procedures were analyzed and used to develop the well-system model under study. System Failure Modes To analyze the completion reliability, the possible failure mechanisms were identified. These included completion equipment failure, operational failures, and several catastrophic events, such as drilling-rig derrick collapse, plane crash, and vehicle collision into wellheads. Certain plane crash, and vehicle collision into wellheads. Certain risks that could not be quantified-such as acts of war, acts of sabotage, and acts of God (earthquakes, etc.) were omitted from the study. Fault Tree Development The application of fault-tree risk analysis to well completion systems is well documented by Woodward. The reliability of a component or system of components is the probability that it will perform its functions for a specified probability that it will perform its functions for a specified time interval and is predicted generally by a single-parameter exponential distribution: (1) where R(t) is the probability the component will operate without failure for time period t under the stated operating conditions. JPT p. 1813