Valve Workshop

Assessment of Valve Failures in the Offshore Oil & Gas Sector

John Peters, TUV NEL

Ravi Sharma, HSE, Offshore Division

28'^ October 2003

Valve Workshop 28"^ October 2003

ASSESSMENT OF VALVE FAILURES IN THE OFFSHORE OIL & GAS SECTOR

Mr John Peters, TUV-NEL Ltd Mr Ravi Shanna, HSE, Offshore Division

1 INTRODUCTION

This paper summarises the findings of a recent Health & Safety Executive (HSE) study [1] into valve failures in the offshore industry by OSD (Offshore Safety Division), undertaken by TUV-NEL Ltd.

The study looked at identifying and assessing the potential underlying causes of valve failures and valve problems associated with offshore valve installations. The study is part of a wider HSE initiative to reduce hydrocarbon releases.

HSE recognise that there are a very large number of valves used by the industry however there are frequent reports of valve problems or failures. There are four basic generic valve types used - isolation, control, non-return and relief, but many different models of each type. They are used in a wide range of sizes, eg from 12 -1200 mm nb, with line pressures ranging from a fraction of a bar up to 100+ bar and some up to 500 bar or higher.

1.1 Study Objectives

As there are ft"equent reports of valve problems and failures, the objective of the study was to assess valve failure data and identify if there were any underlying reasons and causes for the vatve problems. Also to fdentrty any generic trends associated with a valve type, operating sector or fluid. The study findings vi/ill provide valve trending information and identify areas where potential improvements or preventative measures might be implemented, since potential valve problems/ failures are often safety related.

1.2 Data Sets

For tfie study, potential existing electronic data sets were identified which contained suitable valve related failure or problem data, which could be anonymised and used for the study.

One of the data sets originated from the OSD's RIDDOR Hydrocarbons Releases Database which catalogues all hydrocarbon release incidents. By separating out the valve related incidents, this formed a basic data set (DS1). The other two datasets (DS2 & DS3) were from Duty Holders (operators)-

Table l - List of Data Sets

Data Set

DS1

DS2

DS3

Details

Valve related entries from the HSE, OSD RIDDOR database- reportable hydrocarbon releases on offshore installations-Valve related data from an Operators database; this tracks any problems which cause production to be reduced or stopped in any of their fields and production systems-Data collated by an Operator - to record valves being removed & the reason for removal - from all their fields and production systems.

No of entries

250

41

1900

Period covered

Dec'93 to March '99

Jan'OO to Aug 2001

Mar '94 to Nov '00

Valve Workshop 28"" October 2003

As each data set was originally set up for other purposes, the format and detail of valve related information recorded which is of specific interest to this study varies significantly and has limited standardised information for direct comparison However all the data sets were populated in a reasonably consistent way. often using pre-defined critena for data input and classification.

As a result each data set was first revised and the data assessed separately. Then some data was re-classified so that common definitions could be used for subsequent inter-comparison across the data sets,

2 DATA CLASSIFICATION

In this report the word valve is used in two ways, the sense being determined by the context in which it is used - either a valve on its own, or in the wider/ collective sense, meaning the 'valve, actuator (manual or powered) and associated controls'

Although the phrase 'valve failure' is used in the report title, in the temis of this study the meaning is somewhat wider. It may be that the valve itself has physically failed - e.g. valve seized, valve body or seal leaking. Alternatively in the wider context, it could be that something has happened or not functioned which has caused the valve to have effectively failed; for example a hydraulic power supply has failed, or the valve is not receiving a control signal (controls failure). As a result there are many causes of valve failure; some being directly attributable to a valve, whilst others are indirectly related to a valves operation.

2.1 Valve Type

There are six valve types and they are classified In Table 2.

Table 2 - List of Valve Types

VALVE TYPE COMMENT Bleed Bleed, vent, used for instrumentation lines-

smatl sizes DN S to 25 mm nb ^flanged ends) Block Any type of block or isolation valve- eg ball, butterfly, gate, plug.

diaphragm Check Non-return valves- all types Choke Choke valve- specialised high pressure drop/ flow control valves Control Pressure/ flow control valves, modulating operation Relief Pressure relief, safety valves ESDV A specialist function valve - where identified within a Data-set;

usually a type of block valve ESDV" = Emergency Shutdown Valve

2.2 Operating Sector

This is the operating system where the valve is installed, identified as follows: Ancillaries - eg, methanol, nitrogen, flare gas Fiowlines & Manifolds" Gas compression, gas* Imports/ exports" Metering" Processing" Separation" Utilities - e.g. diesel fuel, gas fijel, compressed air, Water Systems ~ e.g. produced water, fire-water main, cooling, water Injecfion, sea-

water # = oii. gas. possibly some condensate + = gas, possibly some condensate

Valve Workshop 28"^ October 2003

2.3 Valve Size

This is identified by three size ranges in temis of nominal pipe diameter (D):

D 275 (mm's)

2.4 Valve Problem, Classification Level

The valve problem (failure mode) or initial reason for valve removal is split into three classification levels as detailed in Table 3,

Table 3 - Valve Problem Classification Levels

Level

1

2

3

Classification

Initial valve problem

Primary problem

Underlying cause

Description This is the initial classification assigned to the valve problem, or initial reason for valve removal; based upon first impressions or first symptoms This is the classification assigned after the problem has been investigated further or after the valve has been removed/stripped or examined more closely This is the final assessment after taking further account of all data, information and circumstances leading up to the valve problem.

Each data-set had a variation of this type of information, so all the data was re-configured into the classification levels and sub categories, to enable the data-sets to be compared and assessed with each other. The sub categories of the three classification levels are given in Table 4.

Table 4 - Sub-categories of the Classification Levels

LEVEL 1 INITIAL VALVE PROBLEM Failed to operate (open,

close) Through valve leakage Extemal leakage Difficult operation External corrosion Valve not operating

properly Other reasons (e.g,

redundant, specification change)

Reason not specified

LEVEL 2 PRIMARY PROBLEM

Valve seized Stem, seal problem ActuaUon problem (eg. electrics, hydraulics.

pneumatics) Control system problem

(e.g. communications faulty, software problem)

Human error Seat, seal problem Body/ bonnet, flange.

tnjnnion problem Erosion Design Defect Materials defect Corrosion Valve not stripped Not known

LEVEL 3 UNDERLYING CAUSE

Inadequate maintenance Design inadequate, materials

deficient Lack of training, inexperienced

staff Corrosion Sand erosion System software, control system.

signal data communications Human error Incon-ectly specified Quality Assurance (QA) issue-

Procedures of manufacturer/ supplier deficient, (eg incorrect materials fitted to valve)

Poor manufacture- e.g. poor welding

Quality Assurance issue -Procedures of operator deficient (e.g. dismantling procedure incorrect; process operating procedure defective)

Not commissioned properiy Undefined, not known

Valve Workshop 2 8 * October 2003

C O N C L U S I O N S

3.1 Genera!

The type of valve information, detail and quality held In the data sets is very variable because the three data-sets were originally set up for a different purpose, not specifically for this valve failures study. Only some of the data fields are common throughout, so Inter-comparison is often limited to two sets

The lengtti of time that a valve has been In service is an essential factor that should be considered when looking at valve failure and reliability, however this was not available in any of the data-sets. If a valve has only been in use for a relatively short period of time -several weeks or months, prior to failing, then this would very likely cause questions to be raised. However if the valve had been in use for a number of years, then it may well have failed due to 'fair wear and tear,

3.2 Valve Types

The distribution of the valve types across ^ e three data sets Is shown in Fig 1.

Fig 1 -Disblbution of Valve Types tn Data Sets

BLOCK CHECK BLEED CHOKE

VALVE TYPES,

RELIEF CONTROL

3.2.1 Block Valves

The most dominant valve type is the block valve- ranging from 51% to 85% across the data sets, with seat leakage and difficult to operate as major problems. The high figure of 85% in DS3 is perhaps exaggerated by the crude valve classification met