Safety Science 49 (2011) 852–860
Contents lists available at ScienceDirect
Safety Science journal homepage: www.elsevier.com/locate/ssci
An evaluation of maintenance strategy using risk based inspection Tan Zhaoyang a,d, Li Jianfeng a, Wu Zongzhi a, Zheng Jianhu b,⇑, He Weifeng c a
Center for Urban Public Safety Research, Nankai University, Tianjin, PR China Department of Automobile Engineering, Minjiang University, Fuzhou, Fujian, PR China c Department of Service and Logistics, Military School of Economics, Wuhan, Hubei, PR China d School of Chemical Engineering, Hebei University of Technology, Tianjin, PR China b
a r t i c l e
i n f o
Article history: Received 9 September 2010 Received in revised form 22 October 2010 Accepted 31 January 2011 Available online 26 February 2011 Keywords: Risk based inspection Oil refinery industry Maintenance strategy Analytical hierarchy process
a b s t r a c t Risk based inspection (RBI) methodology was proposed to evaluate the maintenance strategy in industrial process which was constructed in one of the units of Fujian Oil Refinery ISOMAX unit. Using classic definition of risk, both the probability and consequence of accident or failure were investigated respectively under the support of risk-specific code. All equipment in this unit were evaluated and categorized into five risk zone based on the RBI result which covered five levels. In addition, an application of the analytical hierarchy process (AHP) to select the most practicable maintenance strategy for equipment which was located in each risk rating scale was described. To arrange the hierarchic structure and evaluation, four main criteria were defined for pairwise judgments. Finally, four possible alternative strategies were proposed for administrators on the site. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction Risk based inspection (RBI) is a risk based approach to prioritizing and planning inspection, predominantly in the oil and gas industries. This type of inspection planning analyses the likelihood of failure and the consequences of the same in order to develop and inspection plan (Bertolinia et al., 2009). RBI will assist a company to select cost effective and appropriate maintenance and inspection tasks and techniques, to optimize such efforts and cost, to shift from a reactive to a proactive maintenance regime, to produce an auditable system, to give an agreed ‘‘operating window’’, to promotes team work and to implement a risk management tool. The purposes of RBI include: (1) To move away from time based inspection often governed by minimum compliance with rules, regulations and standards for inspection. (2) To apply a strategy of doing what is needed for safeguarding integrity and improving reliability and availability of the asset by planning and executing those inspections that are needed. (3) To provide economic benefits such as fewer inspections, fewer or shorter shutdowns and longer run length. ⇑ Corresponding author. Address: 1 # Chengwenxian Road, Department of Automobile Engineering, Minjiang University, Fuzhou, Fujian 350108, PR China. Tel./fax: +86 591 83761028. E-mail address:
[email protected] (J. Zheng). 0925-7535/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.ssci.2011.01.015
(4) To safeguard integrity. (5) To reduce the risk of failure. Risk based inspection and maintenance methodologies have been mixed with each other gradually during 90s decade, and had been become common by the end of year 2000. The new generation of maintenance strategies has been appeared with risk based management (RBM), reliability centered maintenance (RCM) and condition based maintenance (CBM) characteristics. Since the year 2000, maintenance and safety were separate issues and had independent activities. Some researches described the relation between maintenance activities and main event industries (Khan and Haddara, 2003). Object of maintenance engineers is carrying out maintenance policies to maximize availability and efficiency of equipment, controlling failure and deterioration, guarantee a safe and correct operation and minimizing the costs. These goals will reached by performing new methodologies in maintenance activities. RBM and RBI methodologies were developed by American Society of Mechanical Engineers (ASME) since 1941. Aller et al. (1995) conducted RBI policies for shell facilities. Dey (2004, 2002), Dey et al. (1998) developed a simple risk-based maintenance model for pipeline. Bevilacqua and Braglia (2000) developed an AHP model to select a maintenance strategy. Riskbased strategies in medical equipment were evaluated by some researcher. Ridgway (2001) proposed a new methodology for establishing a logical, fact-based framework for determining which devices should be included in critical devices category. It is based in part on a new FDA-sanctioned definition of what an appropriate
853
Z. Tan et al. / Safety Science 49 (2011) 852–860
regimen of planned maintenance activities for a medical device should include. Dey (2002) has developed a RBI and maintenance model for pipeline in. In this model an AHP based model had been developed by considering main criteria and alternatives. Dey (2004) has developed a risk-based maintenance model for gas and oil pipeline with a case study. Khan et al. (2004) has applied Multi Attributes Decision Making techniques to analyze risk and has proposed a maintenance model considering RBI. Fujiyamaa et al. (2004) has proposed a risk-based maintenance model for steam turbines in power station. The risk-based maintenance (RBM) system has been developed for steam turbine plants coupled with the quick inspection systems. The proposed RBM system utilized the field failure and inspection database accumulated over 30 years. The failure modes were determined for each component of steam turbines and the failure scenarios were described as event trees. The probability of failure was expressed in the form of unreliability functions of operation hours or start-up cycles through the cumulative hazard function method. Bertolini and Bevilacqua (2006) developed a decision support system (DSS) for the inspection staff of oil pipelines based on the decision tree analysis outcome. They proposed a method to predict ‘‘the class’’ of each spillage, with respect to some relevant variables such as, mechanical failure or system malfunction. Reviewing recent papers and researches, no research has been done to use directly RBI results in oil and gas industries in maintenance strategy selection. In this study risk base inspection methodology is accomplished in ISOMAX unit of Fujian Oil Refinery. Fujian Oil Refinery (Fujian Refining and Chemical Company) is located in Quangang district, Quanzhou city, Fujian Province, was founded in January 1989 and fully put into operation in September 1993. May 1997, its crude oil processing capacity of 250 tons/year increased to 400 tons/year. In the end of August 2009, Fujian Refining and Ethylene Project went into operation. The Fujian Oil Refinery has 1200 tons/year refinery and 80 million tons/year of ethylene, 8,00,000 tons/year of polyethylene, 4,00,000 tons/year of polypropylene and 7,00,000 tons/year aromatics production capability. The Bird’s eye view can be found in Fig. 1. The output of implementing RBI which is a risk matrix was drawn. Risk rating scale and equipment in each scale risk rating scale were defined. An AHP decision model was developed for each area by using a Multi Attribute Decision Making method. For proposing the decision model, several criteria were considered and screened into four main criteria. In the next level, main applicable maintenance activities in oil refinery were specified. Pairwise comparison matrixes were developed by Saaty method and finally the most suitable policy for each risk rating scale was proposed. The research methodology and steps are shown in Fig. 1.
1. Fulfilling risk based inspection program by RISKWISE code in ISOMAX unit of Fujian Oil Refinery and extracting risk matrix to identify equipment criticality. 2. Developing AHP decision model to assign proper maintenance policy to each equipment group. 2.1. RBI methodology Inspection activities objective in process industries is distinction and evaluation of equipment damage rate which are working in process continually. Inspection programs are extended through a spectrum based on their effectiveness. In one side there are programs like ‘‘don’t fix it unless it’s broken’’ and on the other side there are comprehensive inspection programs which inspects whole equipment fully and repeatedly in detail. Adjusting inspection intervals has been developed within resent years. Inspection programs were calendar base in previous decade. By developing inspection methods and identifying failure mechanism/rate inspection periods were developed base on equipment status. Standards such as API 510, 570, API STO 653 developed inspection philosophy into on-stream inspection instead of internal inspection. Risk based inspection (RBI) presents recent generation of inspection by adjusting inspection intervals during last decade. RBI improves process equipment safety level and reliability by concentration on equipment which is critical and has more damage mechanism. This method implements inspection produce in a way that effectively can manage and reduces equipment risk level. The risk could be highlighted from the viewpoint of safety, health, and environment and economic by RBI. It proposes risk reduction methods which are cost benefit and effective and also identifies equipment which is in a reasonable risk level and do not need special inspections program (API 580, 2002 and API 581, 2000). 2.2. Risk definition In general, the ‘‘risk based’’ approaches define the risk as a combination of the consequences derived from the range of possible accidents, and the likelihood of these accidents which can be described as follow (Magnusson et al., 1995):
RLTOT ¼
f ¼F X s¼S X t¼T X
RLðs; f ; tÞ
ð1Þ
s¼1 f ¼1 t¼1
where S is the number of source locations; F is number of accident scenarios; T is number of target locations; and RL(s, f, t) is the risk to life for a given source, scenario and target location. As per definition in API 580(2002) and API 581(2000), risk in defined as following:
2. Methodology descriptions
Risk ¼ probability of failure consequence of failure
From Fig. 2, it could be seen that, proposed model in this paper is performed in two main steps as following:
Concept of ‘‘consequence’’ in RBI expresses the effect and results of equipment failure. For example, consider leakage in a pressure vessel which is working in a process plant. Following consequence could be occurred:
Fig. 1. Fujian Oil Refinery overview.
explosive could formation, releasing toxic liquid/gas, harmful material leakage, shutdown (production loss), environment/safety/health consequence.
Risk could be calculated by combing failure risk and consequences. According to RBI methodology to draw a risk matrix and determining each type of equipment risk level, failure probability and consequence must be calculated separately.
854
Z. Tan et al. / Safety Science 49 (2011) 852–860
Fig. 2. Research flowchart.
Based on API 580, RBI can be performed in qualitative, quantitative and semi-quantitative. The results of each method are almost the same, but by qualitative method a unit can be evaluated quickly. Quantitative method involves us in more detail and calculation, but with more accuracy. Semi-quantitative method uses qualitative speed and quantitative accuracy. By the way in each method, two items must be defined to draw risk matrix: likelihood category, consequence category. Each category is consisted of several factors. By computing each factor and summing them likelihood category is calculated from 1 to 5 which represents likelihood variation 102–105/year, in addition consequence category will be calculated from A to E which A has lowest consequence and E represents highest consequence. After computing categories for different types of equipment, a risk will be assigned to the equipment based on original risk matrix. Fig. 3 shows the blank risk matrix.
2.3. RBI construction for ISOMAX ISOMAX unit of Fujian Oil Refinery has been constructed since 1993 and refines 1,50,000 bpd crude oil. In this unit Heavy Hydrocarbons which have not processed in vacuum distillation unit and atmospheric distillation unit, convert to light Hydrocarbons like LPG, Kerosene, Diesel, Lightnafta, and Heavynafta by hydro treating. Huge amount of Hydrogen with Heavy Hydrocarbons enter into four reactors in controlled temperature and pressure and through a Hydro cracking reaction Heavy materials convert to lighter Hydrocarbons as mentioned before. Operational temperature and pressure in this unit is mostly 825 °F and 2500 psi so high temperature and pressure with amount of toxic gas (H2S) has made this unit more important for our study. Overhaul maintenance is done in ISOMAX unit every 3 years based on manufacture time period suggestion. There was not a specific strategy for implementing maintenance activities and
Fig. 3. Risk matrix.
changing the intervals. The aim of this paper is implementing RBI methodology in Fujian Oil Refinery (ISOMAX unit) by using RISKWISE software which is based on API 580 and 581 standards and designing a maintenance decision model according to RBI result by means of AHP technique. RISKWISE is risk-based maintenance software. This software, which has already been adopted by companies worldwide, including, USA, Europe and Japan, is aimed at increasing competitiveness by minimizing downtime and outages without incurring unnecessary maintenance costs. RISKWISE is fully compliant with API and ASME guidelines and is designed for use by the plant engineers and managers. Its user-friendliness is a major attribute as is its automatic maintenance planning and remaining life output. RISKWISE is intuitively designed so that it can be easily learned, and the risk model allows qualitative (Level 1) as well as quantitative (Level 3) input. It includes the full API RP 571 technical modules for all relevant in-service damage mechanisms (description, affected materials and equipment, critical factors, appearance,
Z. Tan et al. / Safety Science 49 (2011) 852–860
prevention, mitigation, inspection and monitoring). Users can undertake a ‘what if’ likelihood and consequence appraisal of each component, to determine the minimum level of inspection and maintenance, to mitigate the risk of failure and optimize the current inspection program. It provides an indication of remaining life for all damage mechanisms, based on an implicit time dimension of risk (risk of failure generally increases over time) which is the only rational basis for setting safe operating periods between inspections. Different types of equipment are located in this unit. Analyzed facilities and equipment in this study were as follow:
process piping, reactors, pressure vessels (tower, drum. . .), heat exchanger, air cooler, furnace, pumps and compressors (casing), and safety valves. Number of types of equipment in this unit is listed in Table 1.
855
(7) Polythionic Acid stress corrosion cracking. (8) Wet H2S damages. (9) Temper Embrittlement. 2.6. Defining failure consequence In this stage, considering the collected data, effective factor on failure consequence were entered to the software. Failure consequence for each type of equipment and pipeline was calculated by software. Software outputs were evaluated by experts and operational personnel and confirmed by them. The calculated consequences were used to define equipment overall risk to be located in risk matrix. Effective factor on failure consequence are as:
Product loss. Pressure factor. Explosion or fire damage potential. Potential toxicity of release. Effect of item failure on production. Location of component in plant Threat to personnel/environment.
2.4. Data gathering
2.7. Defining probability of failure
Required data for entering to software is consisted based on equipment data as following:
In this stage, damage mechanism rate and inspection plan effectiveness in failure identification for each mechanism were defined and entered to software. Each mechanism activation rate was defined by considering operational condition, failure record and inspection result of equipment and pipeline. Mechanism failure probability were defined qualitative for a 36 month period (period of risk analyzing) and entered to software.
design data, operational data, process fluid data and its characteristics, and failure data and previous inspection data.
After collecting data, they were screened by experts to be modified and corrected and then were entered to the RISKWISE software. 2.5. Defining active damage mechanism In this stage, damage mechanism were recognized considering operational condition, process fluid, equipment material, equipment failure record and active damage mechanism. The most important damage factors were as following: (1) (2) (3) (4) (5) (6)
High temperature Hydrogen attack. Low temperature Hydrogen Embrittlement. High temperature H2/H2S corrosion. Naphtenic Acid corrosion. Ammonium Bisulfide corrosion. Chloride stress corrosion cracking.
2.8. Risk analysis Equipment and pipeline risk analysis result of ISOMAX unit is showed in Fig. 4. According to the obtained results, it could be seen that two types of equipment are located in high risk rating scale (5D) which are Heat Exchanger tube 2E-413A, 2E-413B. 81 types of equipment were located in unsatisfactory area and the rest of equipment (150 items) is located in tolerable area. Twelve types of equipment were in acceptable and 95 types of equipment in favorable area. 2.9. Remaining life indicator (RLI) Considering damage progress during time, failure probability will be increased; So (LOF) in time period could be an indicator of remaining life.
RLI ¼ Table 1 Evaluated equipment in ISOMAX. Equipment type
Number
Reactor Vertical vessel Horizontal vessel Shell and tube heat exchanger Air cooler Heater Pump Relief valve Compressor Piping
4 11 23 40 7 5 39 51 4 206
Total
390
Lmax Lcurrent Rl FOS
ð2Þ
That Lm is the maximum failure probability; Lc is current failure; FOS is safety multiplier; and Rl is the failure probability variation rate between first risk assessment period and last risk assessment period. As per:
Rl ¼
LOF1AP LOF3AP DTA
ð3Þ
DTA is time period between first risk assessment period and third risk assessment period and it is double of equipment risk assessment period. Fig. 5 shows RLI of evaluated unit. Results indicate that RLI of 11.76% equipment is less than 36 month and RLI of most equipment is between 36 and 48 months. It means that considering inspection period (36 month). In this case, failure probability of
856
Z. Tan et al. / Safety Science 49 (2011) 852–860 Table 2 Methodology for judgment in AHP.
1. 2. 3. 4. 5.
Fig. 4. Results of risk matrix of ISOMAX unit.
Judgment
Equally
Moderately
Strongly
Very strongly
Extremely
Score
1
2,3
4,5
6,7
8,9
favorable, acceptable, tolerable, unsatisfactory, and critical.
Performable maintenance strategies for favorable and acceptable area are the same, so suitable maintenance policy will be chosen for four areas by AHP method. The Analytic Hierarchy Process (AHP) is a structured technique for dealing with complex decisions. Rather than prescribing a ‘‘correct’’ decision, the AHP helps the decision makers find the one that best suits their needs and their understanding of the problem. Based on mathematics and psychology, it was developed by Thomas L. Saaty in the 1970s and has been extensively studied and refined since then. The AHP provides a comprehensive and rational framework for structuring a decision problem, for representing and quantifying its elements, for relating those elements to overall goals, and for evaluating alternative solutions. It is used around the world in a wide variety of decision situations, in fields such as government, business, industry, healthcare, and education. AHP has three main levels. First level is goal (in this case selecting the best maintenance policy for each mentioned risk rating scale). Second level is consisted of criteria and third level is alternatives (in this case applicable maintenance policies in refinery). Each criteria and alternative will be compared two by two in a pair wise comparison matrix. Numerical values are assigned for comparison based on Table 2 which was proposed by Saaty. The pairwise matrix will be normalized by Saaty method to define weight or priority for each alternative and criteria. Finally, the weight for each branch will be defined each branch that has the greater weight will be chosen. 3.1. Principle criteria selection
Fig. 5. RLI results of ISOMAX equipment.
11.76% equipment will not be zero. It is noticeable that not zeroing the failure probability does not result in equipment failure in this period certainly but failure probability could be less enough that would not result in equipment shutdown.
3. Theory and calculation Risk matrix is used to capture identified risks, estimate their probability of occurrence and impact, and rank the risks based on this information. Risk matrix also provides a capability for documenting how these risks will be handled (action plans) and tracking the effect of this action on associated risks. As discussed before, risk matrix illustrates five risk rating scale.
As mentioned before, second level of an AHP model is included of criteria. Criteria developers have to face a justifiable degree of skepticism from statisticians, economists and other groups of users. This skepticism is partially due to the lack of transparency of some existing indicators, especially as far as methodologies and basic data are concerned. To avoid these risks, the paper puts special emphasis on documentation and metadata. Also, rigorous procedure was adhering to form criteria. Table 3 provides a stylised ‘checklist’ to be followed in the construction of criteria. By reviewing documentation, metadata, maintenance managers and personnel opinions, following criteria were extracted. Safety: safety for personnel, equipment, facilities, environment. Cost: cost can include crew cost and spare part cost. Accessibility: equipment is located in which situation (height, dangerous area, etc.). Outage time: including failure frequency that relates to MTBF1 and outage time that is related to MTTR.2 1 2
Mean time between failure. Mean time to repair.
Z. Tan et al. / Safety Science 49 (2011) 852–860
857
Table 3 Commonly used procedure for building a criterion. Step
Action
1. Theoretical framework Provides the basis for the selection and combination of variables into a meaningful composite indicator under a fitness-for-purpose principle (involvement of experts and stakeholders is envisaged at this step).
To get a clear understanding and definition of the multidimensional phenomenon to be measured To structure the various sub-groups of the phenomenon (if needed) To compile a list of selection criteria for the underlying variables, e.g., input, output, process
2. Data selection Should be based on the analytical soundness, measurability, country coverage, and relevance of the indicators to the phenomenon being measured and relationship to each other. The use of proxy variables should be considered when data are scarce (involvement of experts and stakeholders is envisaged at this step).
To check the quality of the available indicators
To discuss the strengths and weaknesses of each selected indicator To create a summary table on data characteristics, e.g., availability (across country, time), source, type (hard, soft or input, output, process) 3. Imputation of missing data Is needed in order to provide a complete dataset (e.g., by means of single or multiple imputation).
To estimate missing values To provide a measure of the reliability of each imputed value, so as to assess the impact of the imputation on the composite indicator results To discuss the presence of outliers in the dataset
4. Multivariate analysis Should be used to study the overall structure of the dataset, assess its suitability, and guide subsequent methodological choices (e.g., weighting, aggregation).
5. Normalization Should be carried out to render the variables comparable.
6. Weighting and aggregation Should be done along the lines of the underlying theoretical framework.
7. Uncertainty and sensitivity analysis Should be undertaken to assess the robustness of the composite indicator in terms of e.g., the mechanism for including or excluding an indicator, the normalization scheme, the imputation of missing data, the choice of weights, and the aggregation method.
To check the underlying structure of the data along the two main dimensions, namely individual indicators and countries (by means of suitable multivariate methods, e.g., principal components analysis, cluster analysis) To identify groups of indicators or groups of countries that is statistically ‘‘similar’’ and provides an interpretation of the results To compare the statistically determined structure of the data set to the theoretical framework and discuss possible differences To select suitable normalization procedure(s) that respects both the theoretical framework and the data properties To discuss the presence of outliers in the dataset as they may become unintended benchmarks To make scale adjustments, if necessary To transform highly skewed indicators, if necessary To select appropriate weighting and aggregation procedure(s) that respects both the theoretical framework and the data properties. To discuss whether correlation issues among indicators should be accounted for To discuss whether compensability among indicators should be allowed To consider a multi-modeling approach to build the composite indicator, and if available, alternative conceptual scenarios for the selection of the underlying indicators To identify all possible sources of uncertainty in the development of the composite indicator and accompany the composite scores and ranks with uncertainty bounds To conduct sensitivity analysis of the inference (assumptions) and determine what sources of uncertainty are more influential in the scores and/or ranks
8. Back to the data Is needed to reveal the main drivers for an overall good or bad performance. Transparency is primordial to good analysis and policymaking.
9. Links to other indicators Should be made to correlate the composite indicator (or its dimensions) with existing (simple or composite) indicators as well as to identify linkages through regressions
To profile country performance at the indicator level so as to reveal what is driving the composite indicator results. To check for correlation and causality (if possible) To identify if the composite indicator results are overly dominated by few indicators and to explain the relative importance of the sub-components of the composite indicator To correlate the composite indicator with other relevant measures, taking into consideration the results of sensitivity analysis To develop data-driven narratives based on the results
10. Visualization of the results Should receive proper attention, given that the visualization can influence (or help to enhance) interpretability
To identify a coherent set of presentational tools for the targeted audience To select the visualization technique which communicates the most information? To present the composite indicator results in a clear and accurate manner
858
Z. Tan et al. / Safety Science 49 (2011) 852–860
Fig. 6. The proposed AHP model.
Operational condition: erosive situation is more important. Added value: loss production arising from failure affect added value. Feasibility: each maintenance policy must be feasible to implement. Toxic emission effects: including toxic emission arised from a failure. There were many criteria more than above, but most of them overlapped on each other. In addition numerous criteria need more calculation and time consuming. In the other hand large number of criteria does not guarantee the model accuracy. So the criteria must be selected in a way that has less effect on each other. By organizing brain storming meetings, maintenance expert and involved personnel with maintenance in refinery opinion, four below criteria were selected including safety, cost, added value and feasibility.
Table 12 illustrates the final ranking for each branch in each risk rating scale. As is seen, the maintenance strategy for each risk rating scale is highlight in Table 12 proposed maintenance policies for unsatisfactory area is RCM, for critical area PM and for tolerable area and acceptable/favorable area is CM.
Table 4 Weight calculation in criteria level-unsatisfactory area.
Safety Cost Added value Feasibility
1. 2. 3. 4.
Preventive maintenance. Condition based maintenance. Corrective maintenance. Reliable centered maintenance.
Now all levels of AHP model are defined. Fig. 6 illustrates the final scheme of the AHP model.
Cost
Added value
Feasibility
Local weight
0.5607 0.1121 0.1869 0.1401
0.3846 0.0769 0.3076 0.2307
0.6545 0.0545 0.2181 0.0726
0.4800 0.0399 0.3600 0.1200
0.5199 0.0708 0.2681 0.1408
Table 5 Weight calculation in criteria level-critical area.
3.2. Maintenance strategy selection Third level of AHP model is determining the alternative (in this case maintenance policies). A refinery is consisting of different complex equipment which is working in different operational condition. Selecting an appropriate maintenance policy involve with technical requirement and deals whit every unit’s characteristics. A wrong maintenance strategy may affect operational condition, safety and impose excessive cost to management system. The goal of maintenance model design is presenting a model based on risk matrix output and to be applicable in oil refinery plant. Most common maintenance policies which are implemented in oil refineries are as following:
Safety
Safety Cost Added value Feasibility
Safety
Cost
Added value
Feasibility
Local weight
0.6278 0.0896 0.1255 0.1569
0.4666 0.0666 0.2666 0.2000
0.7594 0.0379 0.1518 0.0506
0.4800 0.0399 0.3600 0.1200
0.5834 0.0585 0.2259 0.1318
Table 6 Weight calculation in criteria level-tolerable area.
Safety Cost Added value Feasibility
Safety
Cost
Added value
Feasibility
Local weight
0.4800 0.1200 0.2400 0.1598
0.4000 0.1000 0.3000 0.2000
0.5217 0.0868 0.2608 0.1304
0.4615 0.0769 0.3076 0.1538
0.4658 0.0959 0.2771 0.1610
Table 7 Weight calculation in criteria level-favorable/acceptable area.
Safety Cost Added value Feasibility
Safety
Cost
Added value
Feasibility
Local weight
0.3529 0.1175 0.3529 0.1764
0.4285 0.1428 0.2857 0.1428
0.2857 0.1428 0.2857 0.2857
0.4000 0.2000 0.2000 0.2000
0.3667 0.1507 0.2810 0.2012
3.3. Pair wise comparison matrixes calculation As mentioned before, risk matrix output consists of four main risk rating scales. Suitable maintenance policy must be assign to each area by calculating each policy priority by means of pair wise comparison matrixes. The most important point is that the ranking of criteria and alternatives are different for each risk rating scale, so pair wise comparison matrix must be calculated in each risk rating scale. Ranking are assign to each criteria and alternative based on Saaty ranking table. Results are shown in Tables 4–11.
Table 8 Weight calculation in alternative level based on safety.
PM CBM CM RCM
PM
CBM
CM
RCM
Local weight
0.4545 0.2272 0.0909 0.2272
0.2500 0.1250 0.3750 0.2500
0.4838 0.0322 0.0967 0.3870
0.5333 0.1333 0.0666 0.2666
0.4304 0.1294 0.1570 0.2827
859
Z. Tan et al. / Safety Science 49 (2011) 852–860 Table 9 Weight calculation in alternative level based on cost.
PM CBM CM RCM
PM
CBM
CM
RCM
Local weight
0.0909 0.1818 0.3636 0.3636
0.0588 0.1176 0.5882 0.2352
0.0724 0.0579 0.2898 0.5797
0.1111 0.2222 0.2222 0.4444
0.0833 0.1448 0.3659 0.4057
Table 10 Weight calculation in alternative level based on added value.
PM CBM CM RCM
PM
CBM
CM
RCM
Local weight
0.0833 0.1666 0.2500 0.5000
0.0769 0.1538 0.3076 0.4615
0.1428 0.2142 0.4285 0.2142
0.0477 0.0955 0.5730 0.2865
0.0876 0.1575 0.3897 0.3655
study; the second are the many sub-networks of influences among the elements and clusters of the problem, one for each control criterion. In the AHP, each element in the hierarchy is considered to be independent of all the others—the decision criteria are considered to be independent of one another and the alternatives are considered to be independent of the decision criteria and of each other. But in many real-world cases, there is interdependence among the items and the alternatives. ANP does not require independence among elements, so it can be used as an effective tool in these cases. In the next step, the decision problems in this paper are best studied through the ANP. We will compare the results obtained with it to those obtained using the AHP or any other decision approach with respect to the time it took to obtain the results, the effort involved in making the judgments, and the relevance and accuracy of the results. 4.2. Super Decisions software for decision-making
Table 11 Weight calculation in alternative level based on feasibility.
PM CBM CM RCM
PM
CBM
CM
RCM
Local weight
0.0714 0.1428 0.5000 0.2857
0.0833 0.1666 0.6666 0.0833
0.0755 0.1321 0.5285 0.2642
0.0476 0.3809 0.3809 0.1904
0.0694 0.2056 0.5191 0.2059
4. Discussions
The Super Decisions software implements the Analytic Network Process. It is used for decision-making with dependence and feedback (it implements the Analytic Network Process, ANP, with many additions). Super Decisions extends the Analytic Hierarchy Process (AHP) that uses the same fundamental prioritization process based on deriving priorities through judgments on pairs of elements or from direct measurements. The Super Decisions software will substitute RISKWISE to conduct the calculation of ANP. Also the accuracy and efficiency will be improved.
4.1. Analytic network process 5. Conclusions The analytic network process (ANP) is a more general form of the Analytic Hierarchy Process (AHP) used in multi-criteria decision analysis. AHP structures a decision problem into a hierarchy with a goal, decision criteria, and alternatives, while the ANP structures it as a network. Both then use a system of pairwise comparisons to measure the weights of the components of the structure, and finally to rank the alternatives in the decision. ANP models have two parts: the first is a control hierarchy or network of objectives and criteria that control the interactions in the system under
Risk based inspection methodology was implemented in one of the units of Fujian Oil Refinery. All goals based on Fig. 2 were reached. The results show that by using this method and assisting suitable computational software, a unit with 390 units could be evaluated quickly. In this method as was seen, equipment categorized into five main risk levels. It can be seen that just two units were located in critical risk rating scale which needs special inspection or maintenance program.
Table 12 Final ranking of maintenance strategy. Maintenance policy
Risk area Unsatisfactory
Critical
Tolerable
Acceptable favorable
PM CBM CM RCM
0.2678 0.1546 0.2891 0.2932
0.2905 0.1531 0.2740 0.2877
0.2483 0.1561 0.3032 0.2965
0.2124 0.1590 0.3294 0.3023
Maintenance selection
RCM
PM
CM
CM
860
Z. Tan et al. / Safety Science 49 (2011) 852–860
To assign capable maintenance program, AHP method was applied. AHP technique has proved to be a valid support for selecting maintenance strategy. The hierarchical structure of the proposed AHP combines many features which are important for the maintenance policy: safety, cost, value added and feasibility. Also accepting the fact that there are some differences, the AHP results are not completely different from those directly proposed by maintenance personnel. This is not a surprise because the same experts performed two analyses. But the AHP approach is characterized by some important properties which are considered highly by the maintenance staff. AHP technique makes it possible to approach the decision making problem in a more complete and thorough way, taking several factors into account. This capacity is more difficult to obtain when using conventional methodologies. It must also be considered that the AHP is able to manage a large number of possible alternatives in an efficient way. AHP can integrate both qualitative and quantitative information. With this technique a direct quantitative judgment of the relevant maintenance factors is not necessary required by the maintenance manager. The pairwise comparisons are preferred by the manager when several intangible criteria have to be treated as in the case with maintenance selection. The results and the satisfaction of maintenance management derived by using the proposed methodology confirm how AHP can enhance and improve the understanding of the dynamic of similar complex problem and represents an effective approach to arrive at decision. Acknowledgements The authors appreciate the support of the Young Talents Project of Fujian Province (No. 2007F3078) and the National Science Foundation for Post-doctoral Scientists of China.
References Aller, J.E., Horowitz, N.C., Reynolds, J.T., Weber, B.J., 1995. Risk-based inspection for the petrochemical industry. In: Risk and Safety Assessments, Where is the Balance? Proc of ASME Pressure Vessels and Piping Division Conference, vol. 296, New York. American Petroleum Institute, 2000. Risk-based Inspection Base Resource Document, first ed. API Publication 581. American Petroleum Institute, 2002. Risk-based Inspection, first ed. API Recommended Practice 580. Massimo Bertolini, Maurizio Bevilacqua, 2006. Oil Pipeline Spill . . . A Decision Support System (DSS) for the Inspection Staff of Oil Pipelines. Bertolinia, M., Bevilacquab, M., Ciarapicab, F.E., Giacchettab, G., 2009. Development of risk-based inspection and maintenance procedures for an oil refinery. Journal of Loss Prevention in the Process Industries 22 (2), 244–253. Bevilacqua, M., Braglia, M., 2000. The analytical hierarchy process applied to maintenance strategy selection. Reliability Engineering & System Safety 70, 71– 83. Dey, P.K., 2002. A risk-based maintenance model for inspection and maintenance of cross-country petroleum pipeline. Journal of Quality in Maintenance Engineering 7 (1), 25–41. Dey, P.K., 2004. Decision support system for inspection and maintenance of crosscountry petroleum pipeline. IEEE Transactions on Engineering Management 51, 47–56. Dey, P.K., Ogunlana, S.O., Gupta, S.S., Tabucanon, M.T., 1998. Risk-based maintenance model for cross-country pipelines. Cost Engineering 40 (4), 24–31. Fujiyamaa, Kazunari, Nagaia, Satoshi, Akikunib, Yasunari, et al., 2004. Risk-based inspection and maintenance systems for steam turbines. International Journal of Pressure Vessels and Piping 81, 825–835. Khan, Faisal I., Haddara, Mahmoud M., 2003. Risk-based maintenance (RBM): a quantitative approach for maintenance/inspection scheduling and planning. Journal of Loss Prevention in the Process Industries 16 (6), 561–573. Khan, F.I., Sadiq, R., Haddara, M.M., 2004. Risk-based inspection and maintenance (RBIM): multi-attribute decision-making with aggregative risk analysis. Process Safety and Environmental Protection 82 (6), 398–411. Magnusson Sven Erik, Frantzich Håkan, Harada Kazunori, 1995. Fire Safety Design Based on Calculations Uncertainty Analysis and Safety Verification. Department of Fire Safety Engineering, Lund Institute of Technology, Lund University. Ridgway, M., 2001. Classifying medical devices according to their maintenance sensitivity: a practical risk-based approach to PM program management. Biomedical Instrumentation & Technology 35 (3), 167–176.
