The application of OREDA (offshore reliability data) in developing valve criticality analysis criteria

In oil and gas processing plants, extensive networks of hundreds to thousands of valves are vital in controlling the entire process. However, not all valves have the same importance in process control. Therefore, identifying and prioritizing critical valves is essential for efficient and safe operation. Assessing the importance of valves requires a risk-based approach. However, quantifying the associated risks becomes challenging, especially when limited field data makes it difficult to model the probability of valve failure. One possible solution is to use Offshore Reliability Data (OREDA) as a valuable failure probability data source. OREDA provides valuable insight into historical failure and performance data related to oil and gas equipment such as valves. To go one step further, the new proposal presented in this paper includes implementing a modified OREDA approach that considers the cumulative operating time of the valve. By considering cumulative operating hours, a more comprehensive and accurate assessment of failure probability can be achieved, enabling more informed decision-making regarding valve criticality and prioritization of maintenance.


Introduction
An oil and gas processing plant is a complex facility that contains a complex network of pipelines, tanks, compressors, pumps, valves, and other critical components that jointly extract, process, and refine valuable hydrocarbon resources.This equipment should be managed so that the integrity of all equipment can work properly.The integrity of all equipment or assets should be managed through asset integrity management (AIM).
The AIM ensures that plant operations, including process control, remain uninterrupted by maintaining equipment efficiency.AIM can be defined as the management system for ensuring the integrity of assets throughout the life cycle of the assets [1].The AIM involves identifying potential risks, implementing measures to manage and mitigate them, and regularly monitoring and assessing the asset's condition to ensure it performs its intended function safely and effectively.
Of the equipment mentioned above, valves are vital in controlling the entire process.Failure of offshore oil and gas valves can have significant negative consequences and risks.Consequences include loss of property, loss of production due to plant shutdowns, health, safety, and environmental (HSE) issues such as oil and gas spills, air pollution, and human losses [2].However, not all valves have the same importance in process control.Therefore, identifying and prioritizing critical valves is essential

Criticality Analysis
The initial step to implement RBM is identifying the assets register and identifying the critical item.Criticality analysis refers to assigning a criticality rating to an asset based on a set of criteria.In the oil and gas industry, it is commonly used to determine the criticality of equipment before prioritizing maintenance strategies.Identifying critical equipment is known as the equipment criticality analysis (ECA).Once the critical items are identified, the maintenance task of the critical item can be developed based on its dominant failure mode [3].There are several approaches to determining the criticality of equipment.Failure Mode Effects and Criticality Analysis (FMECA) is considered the earliest structured method in criticality analysis and then followed by other methods.

MIL-STD-1629A
The MIL-STD-1629A is one of the earliest official guideline documents for criticality analysis [4].This standard aims to identify potential modes of failure, their associated effects, and the criticality of allocating resources and effort to mitigation and prevention.
The FMECA Table documents the identification and definition of the system or equipment under analysis.The FMECA document includes identifying potential modes of failure for the equipment or systems, their associated causes, and the severity of each mode's impact on the system/equipment.Every failure mode can be assigned by criticality number (C ), as shown in equation (1).

=
(1) Cm is the criticality number for a failure mode,  is the condition probability of mission loss, is the failure mode ratio, p is the part failure rate, and t is the duration of the applicable mission phase.
In the case of equipment or a system having more than one failure mode, the criticality of it can be represented by criticality numbers (C ).The criticality number (C ) is the sum of the failure mode criticality number (C ), as shown in equation ( 2).

= ∑
(2) Cr is a criticality number for equipment or a system under investigation.The number of failure modes in the equipment or a system that fall under a particular criticality classification is represented by n, which can take values from 0, 1, 2, …, j.The last failure mode in the equipment or a system under the criticality classification is represented by j.The criticality number shows the criticality of the equipment or a system relative to others.

Risk Priority Number (RPN)
An alternative metric to measure the criticality is a risk priority number (RPN), as found in [5].This document describes the approaches contained in MIL-STD-1629A.RPN metric is practicably useful when quantitative failure rate data is unavailable.Examples of the use of RPN can be found in [6] [7][8] [9].Equation (3) shows the formula to calculate RPN.
S is the severity of the effect of failure, P is the probability of failure, and D is the ease of detection.
The severity of the equipment failure can also determine the asset's criticality.NORSOK Z-008 helps identify critical equipment for maintenance through its consequence classification process.The process considers the potential consequences of equipment failure, such as safety, environmental, and production impacts.Equipment with higher criticality levels requires more frequent or specialized maintenance tasks [10].

Consequence-Based Criticality Approach
A consequence-based criticality is a systematic approach to assessing the criticality of an asset based on its potential risks.It is a method of evaluating equipment's criticality based on its failure consequences.The approach involves identifying the potential consequences of equipment failure and then assigning a criticality rating based on the severity of those consequences.Equipment with higher criticality levels requires more frequent or specialized maintenance tasks.NORSOK Z-008 is one of the official reference codes for determining the consequence-based criticality of the equipment [10].The use of this standard in determining the critical item and maintenance strategies in subsea production systems can be found in [11], while in a gas central processing plant can be found in [12].

Risk-Based Criticality Approach
Risk is also often used as the criticality metric of equipment.In this approach, the higher the risk, the more critical the equipment is.In terms of AIM, risk can be interpreted as the event of the failure of the equipment.The formal definition of failure is the loss of function of a system, structure, asset, or component to perform its required or intended function(s).In the case of risk-based inspection (RBI), failure represents a loss of containment [13].
Risk is expressed as the combination of the equipment's probability of failure (PoF) and the consequence of failure (CoF).Mathematically, the risk is the multiplication of PoF and CoF, as seen in equation (4).
A risk matrix or iso-risk line is commonly a risk presentation.Examples of risk as criticality criteria in oil and gas industries can be found in [14][15].For the screening purpose, a 3x3 risk matrix will be used to assess the criticality of the valve, as in Figure 1.The corporate policy sets the matrix dimension, probability ratings, consequence ratings, and risk level.The risk level can be determined by combining the probability and consequence rating of the assessed valve.The risk level corresponds to the criticality of the valve.The high, medium, and low risk corresponds to the equipment criticality's C1, C2, and C3, respectively.The probability ratings are determined based on the number of failures per year.Since the probability ratings are hard to determine based on real historical data, the secondary data will refer to the reputable source of failure database.The OREDA (offshore and onshore reliability data) is considered the most reputable reliability and failure database source.The consequence ratings are measured based on failure impact concerning the ultimate impact on plant operation, personnel injury, environmental impact, and financial loss.Table 1

Safety Critical Equipment
Safety critical equipment (SCE) refers to equipment required for the safe operation of a facility.The equipment in the SCE list, if it is damaged, could result in a serious accident or death.In asset integrity management, SCE is identified and managed to ensure that it is fit for purpose and will continue to function safely throughout its life.
SCE regulations in Europe emphasize functionality, availability, and reliability while surviving the fire/explosion incidents they should prevent and mitigate.As a result of this consequence-based approach, SCE is often defined to include most items on the master equipment list.It is not surprising since the origins of the regulations stem from the Piper Alpha disaster [1].
On the other hand, asset integrity management may require a narrower definition of safety-critical equipment based on risk instead of consequence.Equipment criticality aims to identify the subset of equipment critical to managing major incident hazards.Before and during a major incident, this equipment ensures the prevention, control, and mitigation of major hazards.Much of this safety-critical equipment will require preventive maintenance and calibration to ensure its reliability to function on demand.Some equipment may become critical or increase in criticality during its life cycle due to deterioration and increased risk of failures, such as a corroded pressure vessel subject to a fitness-forservice assessment requiring more frequent inspections and a greater monitoring effort.

The object of the study
The object of the study is the central processing plant (CPP) and connecting systems to CPP.The block area layout of the systems is shown in Figure 2. Sour gases from wells groups 1, 2, and 5 are sent to the CPP for further processing via trunkline.Wells group 3 is for future development, while wells group 4 is abandoned.The CPP will deliver sweet gas to customers via a gas metering station through the pipeline.The stabilized condensate is also sent to the tanker ship in the condensate jetty station through the pipeline.The produced water is sent to the produced water tank through the pipeline.The produced water will be reinjected into the wells by the production water injection pump.
The CPP consists of an acid gas removal unit (AGRU), a gas dehydration unit, a condensate stabilization system, a produced water treatment system, and other necessary supporting systems.The gas from the well is normally wet and contains impurities.The gas needs to be purified and dried.The gas being purified is known as sour gas.An AGRU is a unit that removes hydrogen sulfide and carbon dioxide from sour gas.Amine solution is widely used to purify sour gas.The purifying process normally occurs in an amine scrubber.Lean amine is the condition of amine before it absorbs hydrogen sulfide and carbon dioxide in the gas-sweetening process.Rich amine refers to an amine's condition when it is rich in hydrogen sulfide and carbon dioxide.Rich amine is recycled by removing hydrogen sulfide and carbon dioxide from the regenerator unit.
Dehydrating the sour gas is essential after removing hydrogen sulfide and carbon dioxide.Gas dehydration units remove water vapor from natural gas.Triethylene glycol (TEG) dehydration is one of the oil and gas industry's most widely used dehydration systems because of its low operating costs and low CAPEX.
Stabilizing natural gas condensates reduces the vapor pressure to a value that ensures safety during transportation and storage, as well as separating lighter from heavier condensates.Stabilized natural gas condensate removes C1-C4 hydrocarbons or is fractionated in stages sequentially, removing ethane (C2), propane (C3), and butane (C4).Stabilized LPG/gasoline is stripped of lighter ends, making it less flammable.
In oil and gas refineries, it also produces water.Produces water needs to be treated before it is utilized for other purposes.Hydrocyclone unit is often used in produced water treatment since it separates solids from liquids with different densities.Hydrocyclone is also able to separate liquids with different densities.

Valve Criticality Analysis
A risk-based criticality approach will be adopted to assign the criticality of the valves in the gas processing plant.However, quantifying the associated risks becomes challenging, especially when limited field data makes it difficult to model the probability of valve failure.The analysis is started by registering all valves.It is recommended to assess which valves are categorized as safety critical valves.The safety critical valve will be categorized as C1 or the most critical or high-risk.The other valves will be assessed further to determine whether they will be categorized as C1, C2, or C3. Figure 3 shows the proposed flowchart to conduct a risk-based valve criticality analysis.

Valves register
The valve criticality analysis is started by documenting all the valves operating in the CPP.As many as 4080 valves are identified in the CPP.The types of identified valves are ball valves, blowdown valves, shutdown valves, butterfly valves, control valves, gate valves, globe valves, needle valves, and safety valves.Table 3 shows the detailed distribution of valves by area in the CPP.

Safety critical valve
A safety critical valve, as in SCE, is essential for the facility's safe operation.Figure 3 depicts the flowchart to identify the safety-critical valve (SCV).The flowchart is developed based on the one discussed in [1].Once the valve is identified as the SCV, the valve is grouped as the most critical.Table 2 summarizes the safety critical valve in the CPP facility.The SCV is determined by using the flowchart shown in Figure 4.  Of the total valves, 338 (8.28%) are categorized as SCV, while the remaining 3742 (91.72%) are categorized as non-SCV.Table 4 summarizes the distribution of SCV and non-SCV based on the type of valves.

Credible failure modes
A failure can be defined as the termination or degradation of an item to perform its required function.An item's failure can occur due to one or more failure modes.On the other hand, a failure mode may happen due to one or more causes known as failure cause(s).In this case, the item denotes any hardware level in the asset hierarchy.
The OREDA (offshore and onshore reliability data) handbook is one of the most credible reliability databases of the oil and gas industry equipment.The data provided in the database covers both topsite and subsea installation.

The OREDA.
There is a project in the oil and gas industry to collect data on equipment reliability, including failure rates and failure mode distribution.Phase I of the project was initiated in 1981 and called the OREDA (offshore reliability data) project.Phase I of data collection was initiated in 1981.Up to 1998, reliability data has been collected for approximately 24,000 units of offshore equipment, including approximately 33,000 failures.[16].The program finished phase XIII in 2020.
In the beginning, the OREDA project collected only topside equipment of equipment in offshore platforms.The database is now extended to collect the equipment of the subsea as well as onshore installation in the oil and gas industry.The OREDA, now becoming offshore and onshore reliability data, provides valuable insight into historical failure and performance data related to oil and gas equipment such as valves.
The 6 th edition OREDA handbook organizes the data based on the system and equipment class.The equipment class is the subcategory of the system.Table 5 summarizes the systems and equipment class of the OREDA handbook.complete loss of an equipment unit's capability to provide output.Degraded failure is not critical but prevents an equipment unit from providing its output within specifications.Incipient failure is a failure that does not immediately cause loss of a unit's capability of providing its output but which, if not attended to, could result in a critical or degraded failure shortly.An unknown failure is a failure whose severity was not recorded or deduced.
For valve critical analysis, only critical failure will be referred.The following failure modes of valves lead to valve failure: Abnormal instrument reading, Delayed operation, External leakage -Process medium, External leakage -Utility medium, Fail to close on demand, Fail to open on demand, Fail to regulate, High output, Internal leakage, Valve leakage in a closed position, Low output, Other, Unknown.

Assessment of the valve probability of failure
The assessed valve's failure probability will be referred to the OREDA database.Databases provide valve failure data at various levels of detail.At the top level, the database summarizes all collected valve failure populations.The following levels provide more detailed information about the valve's type, application, and dimensions.Every level provides the following information: the number of valve failures, the operational time (both calendar and operation time), the failure rate, and active repair hours.For the assessment of valve PoF, only the detailed failure information matched to the assessed valve will be considered.
Table 6 shows samples of detailed information on valve reliability data for critical failure taken from the OREDA handbook.The mean, upper, and lower limit of the failure rate was estimated using equation ( 5), (6), and (7).The failure rate in OREDA is estimated based on the assumption that the equipment is in a useful life.Therefore, the failure of the equipment is also assumed to be random, with a constant failure rate ().Statistically, the time to failure of the equipment will follow the exponential distribution.The maximum likelihood estimate (MLE) theory can estimate the failure rate of the failure data.In the case homogenous sample, that is, the failure data come from identical items operating under the same operation and environmental condition, then the maximum likelihood estimator of the parameter  is given by equation ( 5) In equation ( 5), n and  represent the number of failures and aggregated time in service, respectively.The uncertainty of the estimate is presented with a 90% confidence interval.Mathematically, it is shown in equation (6).
In equation ( 6), L and U denote the lower and upper bound of the failure rate, respectively.For exponential distribution, the upper and lower bound of equation ( 6) can be rewritten in terms of chisquare distribution as in equation ( 7).
. , ≤ ≤ . , ( In equation ( 7), z0.95,2n and z0.05,2(n+1) symbolize the upper 95% and 5% percentile of the chi-square distribution with 2n and (2n+1) degrees of freedom, respectively.It is suggested that the reliability data of the assessed equipment is also calculated in the same way as in the OREDA handbook.Unfortunately, most field data is not ideal.In this study, the number of failures of each valve cannot be obtained, and only the accumulating operating hours are available.It is found that the accumulative operating hours of the valve can be either more or less than that of the OREDA handbook for a similar valve.It is recommended not to disregard the accumulative operating hours of the existing valve when assessing the PoF of the valve.Therefore, the PoF represented by the valve's failure rate will be calculated by combining the failure mode listed in the OREDA handbook and the accumulative operating hour of the existing valve.The failure rate for the existing valve will be calculated based on equation ( 5), but it is modified to equation ( 8).* = * In equation ( 8), * , n and   represent the calculated failure rate based on the actual accumulative operating hours of the analyzed valve, the number of failures of the selected valve in the OREDA handbook similar to the analyzed valve, and the actual aggregate operating hours of the analyzed valve.The upper and lower bound of the calculated failure rate can be obtained using equation (7).Table 7 shows the samples of calculated failure rates based on cumulative operating hours of the homogenous sample of valves installed on the field.The plant started operating in November 2015 and has operated for 61320 hours.It means if there are 45 homogenous valves, the aggregate operating hours of the homogenous valve are 2.75 x 10 6 hours.The failure rate mean of the existing valves is estimated based on equation (8).The failure rate lower and upper bound of the actual valves is estimated based on equation (7) by substituting the appropriate parameters.
This study sets the probability ratings based on the number of failures per year.Therefore, the estimated valve failure rate unit shall be converted from failure per hour to failure per year.Only the mean failure rate will be referred to determine the probability rating of the valve.The probability rating of the assessed valve is determined based on the mean failure rate as in Table 1.

Assessment of the valve consequence of failure
Four categories of valve failure consequences should be determined.Those four categories include the ultimate impact on the plant operation, the impact on personnel health, the environmental effect, and IOP Publishing doi:10.1088/1755-1315/1298/1/01201510 the financial loss.Each category is governed by specific criteria outlined comprehensively in Table 2.This structured approach helps to measure the severity of the consequences of failure.The final consequence rating is determined by selecting the highest rating from the four categories analyzed.Of those four consequence ratings, the ultimate impact seems to have a more severe impact than others.
The ultimate impact of valve failure can be predicted by studying the plant operation, safety chart, process flow diagrams, piping and instrument diagrams, and other supporting documents.The impact of valve failure on the personnel is determined by examining the number of personnel working around the facilities.The environmental effect is estimated based on the released hydrocarbon when the leak has occurred.The financial loss is estimated based on production loss and repair costs.Table 8 shows the sample of consequence estimation of valve failure.

Assessment of risk level and criticality of the valve
The risk level assessed valve can be determined by combining the probability and consequence ratings in a risk matrix.This study uses a 3x3 matrix, as in Figure 1.The risk level of the valve will determine the criticality of the valve.Table 9 summarizes the risk level evaluation and the criticality of the sample valves discussed in the previous sections.The criticality analysis covers 4080 valves in the plant.The valves are assessed similarly to those discussed in this section.Figure 5 shows the valve criticality distribution grouped by the combination of its probability and consequence ratings.As many as 3237 (79%) valves are categorized as low risk or having criticality C3.The other 493 (12%) valves are categorized as medium risk or criticality C2, and the other 350 (9%) valves are categorized as high risk or criticality C1.

Discussion and Conclusions
Valve criticality analysis can be determined using various methodologies.One of the proposed methodologies can be found in [18].The methodology proposed was intended to identify critical valves in the company's water supply network based on their potential impact on customer service levels.The project uses an all-main hydraulic model of each water pressure zone to assess the criticality of thousands of valves.
Other methodologies may refer to [19].The proposed method classified the valves based on a formally agreed priority ranking criteria described qualitatively.Three valve categories have been proposed: 1, 2, and 3.The most critical valve is designated Category 1, while the least critical valve is designated Category 3.
The first methodology mentioned above is more operational-based and designed to respond to the decreased service level.This methodology considers the impact of valve failure on the overall system and identifies valves critical to maintaining the system's service level.The second methodology is more for quick screening purposes.It is used to identify valves that require further assessment and analysis.This methodology is useful when there are many valves to assess, and a quick screening is required to prioritize the valves for further analysis.
The proposed methodology for valve criticality assessment is a risk-based valve criticality assessment that considers the operation history of the valves in running hours and the valves' operation risk.This methodology aims to identify valves at high risk of failure due to age, usage, and other factors.The results of the proposed methodology will be utilized to determine appropriate maintenance strategies for the valves.
Assessing 4080 valves using the proposed risk-based criticality methodology has resulted in valuable insights into their significance within the operational framework.The categorization of valve criticality into C1, C2, and C3 classifications serves as a comprehensive means of prioritizing maintenance efforts.Among these, C1 valves, representing the highest criticality, are of utmost concern due to their pivotal role in safety.
This study identified a notable distribution of criticality, with 9% categorized as C1, 12% as C2, and 79% as C3 valves.Identifying all safety-critical valves falling under the C1 category underscores the methodology's accuracy in recognizing potential hazards.The recommendation to prioritize highpriority maintenance for C1 valves aligns to ensure safety and operational continuity.Similarly, C2 valves, though less critical, merit significant attention.
The integration of reliability-centered maintenance (RCM) for C1 and C2 valves, considering the nature of their failure modes, is a prudent step toward optimizing maintenance strategies.Lastly, the flexibility in extending maintenance intervals for C3 valves within regulatory limits recognizes their lower impact on safety and operational integrity.This study establishes a robust framework for valuing valve criticality and paves the way for informed decision-making in resource allocation and maintenance planning within the broader operational landscape.
The discussion presented here reinforces the risk-based criticality methodology's significance in effectively managing valve assets.The clear categorization of valves based on their criticality aids in streamlining maintenance efforts and resource allocation.The prominence of C1 valves in the safety context necessitates their prioritized attention.Moreover, applying reliability-centered maintenance (RCM) to C1 and C2 valves offers a structured approach to addressing potential failure modes.While C2 valves are of lower criticality, they remain integral to operational efficiency and warrant diligent upkeep.The measured approach to C3 valves recognizes their lower impact and permits maintenance intervals to align with regulatory requirements.This study underscores the importance of adopting a holistic and systematic methodology for assessing and managing valve criticality, ensuring safe and efficient operations in the oil and gas industry.

References
[1] American Institute of Chemical Engineers.Center for Chemical Process Safety., Guidelines for Asset Integrity Management.American Institute of Chemical Engineers, 2016.

Figure 1 .
Figure 1.Risk matrix and the equipment criticality.

Figure 2 .
Figure 2. Block area layout of CPP.

Figure 3 .
Figure 3. Flowchart to determine valve critical equipment.

Figure 4 .
Figure 4. Flowchart to determine valve critical equipment.

Figure 5 .
Figure 5. Valve criticality distribution based on a combination of probability and consequence ratings.
and Table2show the probability and consequence criteria to determine the assessed valve's risk level and criticality.

Table 3 .
Distribution of type of valves in CPP.

Table 4 .
Distribution of SCV based on the type of valves.
4.3.2.Valve failure modes.A valve failure can be total or partial, known as total failure and partial failure.The OREDA classified the severity of failure within one of the following categories: critical failure, incipient failure, degraded failure, and unknown.Critical failure causes an immediate and

Table 6 .
Samples of detailed information on OREDA valve reliability data.

Table 7 .
Samples of calculated failure rate based on cumulative operating hours.

Table 8 .
Samples of estimated consequences of failure and ratings.

Table 9 .
The risk level and criticality of sample valves.