Yihong Li ,Zhiqiang Hu

School of Engineering,Newcastle University,Newcastle upon Tyne NE1 7RU,United Kingdom

Keywords:Offshore decommissioning Multi-attribute Decision-making Cost assessment MCDA

ABSTRACT With the development of the offshore fossil energy industry,the designing life of many offshore oil and gas facilities will end.The decommissioning of these facilities has become an urgent task due to unpredictable costs,high risks,and environmental protection issues of public concern.Decision-making,as the core in the pre-decommissioning stage,plays a decisive role in the cost,risk,and impact of the entire decommissioning.Therefore,the multi-attribute decision-making model has attracted much attention from industry and academia.An efficient,accurate,and simply using multi-attribute decision-making model can enable governments,energy companies,other marine users,and environmental protection organizations to reasonably fulfill their concerns.It is of great significance to all parties.This review mainly studies the multi-attribute decision-making models that have been used in the decommissioning of offshore oil and gas facilities,and conducts a more detailed interpretation of them,including the relevant regulations,frameworks,methodology,preferences and advantages and disadvantages of different models.In addition,a more comprehensive review of the cost assessment model,an important part of the decision-making model,is carried out,including the general framework and methodology of the cost assessment model,and the accuracy of the models is explored.And then the current evaluation method of accuracy of the cost assessment model raises the author’s personal doubts.At the end of the article,this paper names two core problems of the current decision-making model,that is,the lack of basic data and the incomplete MCDA (Multi-criteria Decision Analysis) method.This review can provide a comprehensive reference and feasible research directions for future scholars who aims to study the decommissioning of offshore oil and gas facilities especially in the North Sea in the UK and point out the direction for the industry to improve its current multi-attribute decision-making models.

1.Introduction

1.1.Background

The development of offshore oil and gas facilities has been more than a hundred years since 1897 [1].Oil and gas exploration,collection and storage technologies are growing rapidly,more safety equipment also been implemented to decrease any chance of accident.Although accidents that are not friendly to the environment and human lives occur from time to time,there have been more prevention and treatment methods makes industry do not worry that too much [2].There are about 7500 offshore oil and gas facilities and supporting platforms in the world until the beginning of 21st centuries,which include but are not limited to caissons,compliant towers,fixed platforms,Floating Production,Storage,and Offloading (FPSO) vessels,Mobile Offshore Production Units (MOPU),Tension Leg Platforms (TLP),Semi-submersible platforms (SEMI),Spars,Well Protectors (WP),Subsea Templates(SSTMP) [3].These facilities mainly located in the North Sea,the Gulf of Mexico and the offshore areas near California,and the offshore areas of Southeast Asia [4].About 85% of them need to be decommissioned in the next few decades [5],and according to the international and regional conventions that countries have participated in,most of the facilities need to be completely removed onshore for dismantling and recycling in North Sea,which means high decommissioning costs,a huge risk to the marine environment and the safety issues of human lives.A standard fixed platform has a design life ranging from 20 to 30 years depending on the structural design and the situation of sea area [6].However,due to some reason,some of infrastructure even processed more than 50 years [7].

Although international conventions [8) and regional [9]and national regulations [10]related to the decommissioning of offshore oil and gas facilities have been established since 1958,and the relevant decision-making theoretical frameworks are flourishing [11,12],the engineering process of facility decommissioning is also very clear.For most facilities,it is basically the reverse installation process from well closure to seabed cleaning.The actual situation,such as reducing construction risks,reducing the impact on the environment,or dealing with larger modules,has some modifications in the process [13].The specific decommissioning process is shown in Fig.1 below.It can be said that all the theories and theories required for the decommissioning of offshore oil and gas facilities the technology and equipment required for practical operation have constituted the basic system.

However,the lack of efficient,accurate,and easy-to-use multiattribute decision-making models is a major problem currently faced by decision makers like government and energy companies.Being able to consider multiple concerns and comprehensively evaluate the cost,risk and environment of decommissioning projects is very important for decision-making and management.The multi-attribute decision-making model can make scientific decisions to balance costs,environmental and societal impacts,and risks before decommissioning begins;in the process of the project,the operation of the target can be monitored in real time;after the project is completed,it is necessary to continuously monitor the remaining project to ensure that it will not produce unexpected environmental impacts and pose no threat to other marine users.

1.2.Decommissioning procedure

Before starting to study the multi-attribute decision-making model,it is important to understand the decommissioning procedure.The engineering procedure for the decommissioning of offshore oil and gas facilities are generally adopted by the industry,that is,they are generally the inverse process of the installation process [14],but appropriate modifications will be made based on decision-making,equipment,technology,safety,and environmental considerations [5,15].The entire process from the initial planning to the continuous monitoring after the completion of the project has been widely recognized by industry and academia.In addition to the different laws and regulations applied by the governments of various countries,there are differences in the details.Take the differences between the North Sea surrounding countries and major oil producing countries in Southeast Asia as an example.Due to the OSPAR Convention [16]and the laws and regulations of the United Kingdom [17],Norway [18]and other countries,most offshore oil and gas facilities must be completely removed and dismantled onshore [19].Oversized facilities can be decommissioned in situ,but rig-to-reef project shall not be allowed [20],but in the Southeast Asian waters,the rig-to-reef project has already been carried out [21].This difference is the difference in the final engineering steps caused by legal restrictions,but other aspects are basically the same.Fig.1 shows generic engineering flowchart,which roughly describes the steps necessary for the entire decommissioning engineering stage.

The entire decommissioning process includes three stages:pre-decommissioning,decommissioning execution,postdecommissioning and project management throughout,a total of four parts.Lloyd’s Register and some research [25]associations also take into account the late production stage,but this overlaps more with the content of the pre-decommissioning part,so how to divide is often defined according to the service provided by the country or service provider.

The first is the pre-decommissioning stage.The objective of this stage is to make decisions and plans for the implementation of decommissioning projects,prepare materials and personnel,and obtain government permits [26].There are ten main steps,including responsibility division,platform engineering information collection,cost,risk and environmental impact assessment,decision-making,engineering simulation,resource mobilization,and government document acquisition.It usually takes about two to three years to complete all tasks [14].During this period,third parties,such as classification societies or professional engineering companies,may also join at this time to provide professional consulting services for energy companies and the government [27].Although this is not shown in the figure,it is mainly because of the larger of energy companies do not make this choice for ordinary decommissioning projects due to business classified and other considerations,but only consider them when facing more complex projects.This conclusion can be drawn from most of the United Kingdom Continental Shelf (UKCS) decommissioning reports currently public [28,29].

Fig.1.General decommissioning procedures for offshore oil and gas facilities,obtained by the research integration of multiple scholars.The dotted line is based on target facilities and decommissioning decisions to determine whether to implement.For example,conductor removal will not receive attention in subsea Well Plug &Abandonment(Well P&A),and some platforms do not have underwater facilities,so there is no need to consider the decommissioning of underwater facilities [3,14,22-24].

This stage is also a stage where academic research is currently widely concerned.The main research contributions of the academic community at this stage are decision-making frameworks[12,30) and evaluation tools [12,31,32].The purpose of the research is to enable decision-makers to make decisions more intuitively,simply,and with justification.However,different decision-making frameworks are focused,some focus on environmental protection [33],some focus on cost reduction [34],some focus on low risk [35],and of course,some are more balanced [5].However,there are currently few evaluation tools developed correspondingly,which cannot fully meet the focus and needs of the decisionmaking framework.

The second part is the decommissioning execution stage,the main content is the implementation of the decommissioning project.The steps in this part depend on the type of platform:fixed,floating,semi-submersible;wellhead type: platform well,underwater well;decommissioning decision: complete removal,partial removal,and reuse [15,26,36-41].There are different processes,the main difference is whether there is a conductor and whether to move the substructure and whether remove pipelines.In many decision models,huge decision branches are provided,as shown in Fig.2 [42,43].But in general,the fundamental difference lies in whether to move the substructure,because the removal of topside is almost always necessary in various conventions and regulations and not clearly demand for pipeline so most of decision maker would flush pipelines then leave them situ.The main difference in cost and risk assessment of the substructure is whether to move.

At this stage,the focus of engineering and academic institute is mainly on risk management,construction efficiency,and marine environmental protection.Specific examples include the development of new equipment [47]or management software [44],structural collision research [48],structural wind and wave response research [49],hydrocarbon leakage prevention and control [50],and the application of dynamic risk control that integrates weather,hydrology,engineering and other factors [51].

The third stage is of relatively small importance and is used as a summary and clean-up after the completion of the decommissioning project.The main content is validation,monitoring and liability release,and provide the government with a close-out report,stating that the platform facilities have been decommissioned in accordance with the requirements [52].The research at this stage is mainly focused on the marine environment,studying whether the hydrocarbons,radioactive materials,heavy metals and other contents contained in the leftover drill cuttings,subsea infrastructure and other items [50]and structures will affect the life of marine organisms,especially marine mammals [53].For projects that apply rig-to-reef,there are research on the release of hydrocarbons,heavy metals,radioactive materials,structural stability and the development of marine biological communities in artificial reefs[54-57].In terms of engineering,due to third-party validation,this part is not valued by the offshore oil industry.

The project management part runs through the entire offshore oil and gas facility decommissioning project,and the main purpose is to plan the decommissioning project in the early stage.Maintain the supply of materials,personnel,funds and equipment in the mid-term,ensure the project nodes,try to maintain the risk level within an acceptable range,ensure the marine environment of the project area,coordinate the implementation of offshore and onshore projects,and handle any immediate incidents to ensure project execution the reputation of institutions and related companies,etc.Write and submit the report later,release the responsibility,and ensure that the monitoring obligation continues.

Fig.2.Decommissioning options for fixed platforms and subsea facilities (including pipelines).The research content is integrated by multiple scholars [5,42-46].

Although this part is listed separately for display,it is not integrated into the three stages because this part is very important.Although it continues to run through the three stages,it is difficult to define which part of its activities belong,but it also needs to be taken seriously.

This paper mainly focusses on reviewing the two tool models in the pre-decommissioning stage: the decision tool model and the cost evaluation mathematical model,as well as the knowledge related to these two models,such as the theoretical framework and the boundary conditions formed by relevant laws and regulations.,The universal method used,and the form and effect of the actual model.

2.Methodology

2.1.Quantitative method

For the decommissioning of offshore oil and gas facilities,quantitative methods are widely used in cost assessment,energy use and gas emissions,risk assessment,and material statistics.These methods include but are not limited to theoretical method,equivalent cost method,regression analysis,material and energy flow method.Table 1 describes the use of these methods in the decision-making model for the decommissioning of offshore oil and gas facilities.In many cases,these methods are just a submethod among the main methods of evaluating models and will be used in combination with qualitative methods.When used alone,it is often used to calculate the cost and the probability of loss of life.

Table 1 Methods and their description.Description.

Theoretical method refers to the method of calculation using the theoretical formula that already exists in the industry.The industry has already derived calculation methods such as gas emissions,energy use,and hydrocarbon permeability based on the theoretical basis of physics or chemistry and can further calculate some costs based on market details.Although the results of this method will have small deviations,the theoretical framework is reliable,and the results used for evaluation are also easy to convince the public.However,this method cannot calculate items that have not been clearly related,so it has great limitations.This type of method will be widely used in the estimation of gas emissions,energy use,structural strength and other related cost estimation and risk assessment parts in decision-making models [35].

Equivalent cost method is a method that connects multiple physical quantities and costs.For example,the water depth,weight,number of structures,etc.are linked to the cost currency.This method is very similar to regression analysis,and the method actually used is also a regression method,so it will be explained in detail in the regression analysis [46].

Regression analysis is widely used,and its specific theory will not be repeated in this article.For details,please refer to Claudia’s related papers [64].This article mainly discusses the application of regression analysis in the decommissioning of offshore oil and gas facilities.Regression analysis is mainly used to find the relationship between variables,especially the relationship between variables that have not been determined by theory.In decommissioning,regression analysis is often used after logical relationship analysis,such as: looking for the relationship between decommissioning costs and engineering time,water depth,module weight,etc.[61].There must be a logical relationship between these variables,and similar conclusions can be drawn from the analysis of the data correlation.However,due to many unknown variables,traditional theoretical analysis cannot completely determine what the relationship between these variables is.At this time,regression analysis can be based on historical data,through mathematical calculations and modern computer technology,to give the most likely relationship curve between the two and the expression of the curve.The application of regression analysis in decommissioning is not only to estimate costs,but also to estimate construction time,or to predict sea conditions to reduce construction risks [62].However,the basis of reliable regression analysis is based on sufficient historical data.In other words,if there is not enough historical data,the use of regression analysis is meaningless and inaccurate.This is also the reason why the evaluation accuracy of the Platform Abandonment Estimating System (PAES) mentioned below was low in the early stage of establishment [65].

Fig.3.Material and energy flow analysis.Adapted from Kirchain and Ekins content [67,69].Where TMR means Total Material Requirement;DMI means Direct Material Input,TFO means Total Field Output.

In addition,for the regression analysis of the decommissioning of offshore oil and gas facilities,there is a common problem among scholars who currently use this method,that is,the historical data of all studies is insufficient in quantity,even for studies with a large amount of data,no more than 100 group reliable data [32].And the regression model established by using these historical data,the verification process often one or two groups of the same historical data which establishes a major potential problem with the prior models is that the original regression analysis is likely to be over-fitting [66].Although the argument of overfitting is usually used in the field of machine learning rather than decommissioning,in the author’s actual research,it is found that this phenomenon often occurs when the sample data of the research decommissioning cost evaluation is small.Under such potential problems,several indicators used to test the results of regression analysis,such as the coefficient of determination R2,no longer have reference value,because even if R2performs well,like reaching 0.8 or 0.9,the results of the regression analysis may be excessive fitted,so it cannot be used.The expression of R2is as follows:

Whereyiis the actual value;

is the mean of the actual value;

is the estimated value obtained by linear regression.

The basic principle of Material and Energy Flow Analysis(MEFA) is that energy and matter do not live or die and will only be transferred in other forms.The characteristic of this method is to establish a boundary,and then calculate all the energy and matter entering and leaving the boundary to obtain the required information,such as emissions,heat radiation,workmanship,pollution,etc.[67].Fig.3 shows a schematic diagram of MEFA.This method is more used in the research of urban systems rather than in the industry [68].Ekins et al.[42,63] innovatively applied this method to the decommissioning of offshore oil and gas facilities.This method can estimate many values related to decommissioning,and make it link with costs,gas emissions,energy use,material statistics,etc.The difficulty is that it requires a high degree of information control.Generally,only the institution that implements decommissioning can grasp the first-hand and most detailed material flow data,and third-party researchers will not be able to obtain these data.This is also the biggest limitation of the method.

The above are the quantitative methods that have been used in the decision-making model for the decommissioning of offshore oil and gas facilities.These methods are mainly used in cost assessment and risk assessment,as well as in the calculation of energy use and gas emissions.It is not difficult to see that since other aspects cannot be quantified well,qualitative methods will be used for evaluation.

2.2.Qualitative method

Since many aspects cannot be quantitatively analyzed to obtain results,the use of qualitative methods is very common when making option decisions.These qualitative methods are the same as quantitative methods.They will only be used alone in certain assessments.In most cases,they are combined with quantitative methods as part of comprehensive analysis.

There are many types of qualitative methods used in decommissioning,including but not limited to: expert scoring method,comparative evaluation method,case study method,risk matrix method,etc.Even many evaluation methods do not have specific names but are collectively referred to as qualitative evaluation methods.These methods generally use three qualitative evaluation systems,the first is weighting,the second is scoring matrix,and the third is comparative appraisal [70-73].Many qualitative methods are based on these three systems,adding the latest technology and integrating the needs of related fields to form new methods.It needs to be particularly pointed out that the weighting method can be either a quantitative method or a qualitative method according to the way it is used.The main difference lies in whether the weighting selection is subjectively determined by the decision maker.If the weight is set according to the proportion of each attribute value in the total,there is no subjective intervention,which is a quantitative method;if the weight value is determined subjectively by the decision maker,a qualitative method is preferred.Therefore,the detailed description of this method is placed in the explanation of Simple Multi-Attribute Rating Technique with Swing weight (SMARTS) below.This part will mainly focus on the scoring matrix system and the method of comparative appraisal system.

The first is the scoring matrix system.The characteristic of this type of qualitative analysis method is to establish a matrix that includes all the options that need to be evaluated,the assessment criteria applicable to each option,and a set of unified degree evaluation lists.This list can be in the form of a score from 0 to 100,or it can be in the form of a grade from the worst to the best.The following of this paper will mention MCDA as an example,for five criteria and their sub-criteria,and all parts of decommissioning options,combined with the evaluation indicators,show as Table 2.Decision makers only need to fill in the identified scores in the blanks according to their own needs,then the results are weighted and added,and other data processing methods,to the final total score,to decide which option to choose.The method of this system is often used in combination with the weighting method and the expert evaluation method.The advantage is that it helps to play the role of experts,can well reflect the preferences of decision makers,and can prevent improper behavior.However,the shortcomings are also obvious.Experts are not necessarily familiar with the required scoring field and the scoring is not effective.The introduction of weights also substitutes the shortcomings of the weight method that the weight cannot be set well,and it is difficult to eliminate subjective and emotional evaluation negative effects of results [74,75].

Table 2 Scoring matrix example [75].

Table 3 The five criteria and some of their contents were obtained by the author based on the literature of multiple scholars [5,99,100].

The Comparative appraisal system is widely used.It can be said that any similar part of any decommissioned project can be evaluated using comparative appraisal methods.Such methods include,but are not limited to,the ranking method,the mandatory classification method,the key point allocation method,the pairwise comparison method,the critical event comparison method,the target management method and the comprehensive method [76].The core idea of comparative appraisal is similar to control variables and proportional scaling.In decommissioning,for example,when the same type of platform is like the water depth,the energy company will base on the weight ratio of the platform,the offshore distance,the number of wellheads,the length of pipelines and other different quantities,to evaluate the decommissioning cost range of un-decommissioned oil and gas facilities.In terms of risk assessment,these methods can assess the likelihood and consequences of similar accidents in different scenarios based on the scenes and consequences of accidents that have occurred.The use of such methods is flexible and diverse,and often does not require the cooperation of many experts.However,because the executors of the comparative appraisal system may not have suffi-cient knowledge of the project,they may be underestimated or over-evaluated,and the results obtained are often inaccurate and,in many cases,not strong convincing.In addition,this method requires a large amount of historical sample data,especially for comprehensive and complex projects such as the decommissioning of offshore oil and gas facilities.There are often only one or two comparable data in the hands of energy companies,and the evaluation results can be imagined.

The above is a brief summary of the qualitative analysis methods used in the decommissioning of offshore oil and gas facilities.It is undeniable that the use of qualitative methods is inevitable and necessary in decommissioning.This kind of method has strong applicability and easy to use.So far,many scientific methods have been incorporated to avoid and reduce the influence of excessive subjective consciousness on the evaluation process.Although there are still widespread hidden dangers of excessive subjectivity,and sometimes it is necessary to set basic thresholds for the quality of evaluators,so that the efficiency of the implementation of some methods is still not high but returning to the decommissioning decision-making itself is also based on the subjective concerns of all parties.It has formed five criteria generally applicable in the field of decommissioning.Therefore,it can be said that the use of qualitative methods is necessary in the field of decommissioning decision-making and evaluation.Related decision-makers and decision-making tool developers need to pay attention to that such methods should be gradually reduced after the gradual increase in historical decommissioning data in the future.Try to use quantitative methods to measure objective variables,such as cost,risk and environmental pollution assessment.

2.3.Comprehensive method

In the decision-making model,it is rare to use a simple quantitative and qualitative method alone.In many cases,the method used in the model is a combination of quantitative and qualitative methods.This type of method pioneered the combination of quantitative and qualitative methods to obtain evaluation results.

Such methods used in multi-attribute decision making models include but are not limited to: Decision Tree Method,Goal Programming [42],Semi-quantitative and Qualitative Methodologies(SQ) [77],Analytic Hierarchy Process (AHP) [72],Elimination and Choice Expressing the Reality (ELECTRE) [78],Multi-Attribute Utility Theory (MAUT) [79],Mixed Integer Programming (MIP) [80],Net Environmental Benefit Analysis (NEBA) [81,82],Oracle Multicriterial General Assessment of Decommissioning (OMEGA) [83],Preference Ranking Organization Method (PROMETHEE) [84],Simple Additive Weighting (SAW) [5],Strengths Weakness Opportunities and Threats (SWOT) [85],Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) [86],Simple Multi-Attribute Rating Technique with Swing weight (SMARTS) [87],Comparative Assessment (CA) [69],Multi-Criteria Decision Analysis (MCDA)[45]and many other methods.The application of these methods in the industrial field includes but is not limited to the oil and gas field,wind power generation field,mining,nuclear power,road and bridge,transportation,etc.The decision-making models used in the decommissioning of offshore oil and gas facilities are mainly AHP,SAW,SMARTS,MCDA and CA.Although these methods have different names,they overlap to some extent in the field of offshore oil and gas facility decommissioning.For example,MCDA is currently the core method of CA,and AHP and SMARTS are the most used methods of MCDA.Therefore,AHP,SAW,SMARTS and MAUT will be briefly introduced next.As for the application of other methodologies,readers are requested to understand by themselves based on references.

2.3.1.AHP

Analytical Hierarchy Process (AHP) is a structured technology used to analyze complex decisions.It was developed by Thomas L.Saaty in the 1970s [88]and has been extensively researched and expanded to this day.The content of the method is to classify and layer the targets,assign weights to them according to the importance of the targets and the preferences of decision makers,then analyze the weights based on pairwise comparisons between the targets,and finally use linear algebra to calculate the results for each target calculate the score.Fig.4 shows how to use AHP [89].This method is intuitive and simple.After a long period of development,in the report [90],a quantitative scale was used instead of a qualitative scale.It is used to make decommissioning decisions for subsea structures.

2.3.2.SAW

Simple Additive Weighting (SAW) is also called scoring method or weighted linear combination method.The simple explanation is to find the weighted sum of the performance levels of each alternative on all attributes.Its mathematical expression is as follows[91):

WhereRij: Qualified performance level

Maxxij: The maximum value of each row and column

Minxij: The minimum value of each row and column.

xij: Rows and columns of the matrix

WithRijis the normalized performance rating of Ai alternatives.

Vi: The final value of the alternative

Wj: The Specified weight

A larger value ofViindicates that Ai’s alternatives are preferred.

The simplicity of SAW can greatly improve the efficiency of decision-making,and this method has a significant advantage,that is,due to preset preferences and weights,the SAW method can perform more accurate judgments.In the decommissioning of oil and gas,this method does not appear to be accepted by all decision-making participants and stakeholders,because the preferences of all parties are different.Therefore,Fowler et al.[5]proposed a need for flexibility that can cover all decommissioning options and backup options the importance of method.

2.3.3.SMARTS

Simple Multi-Attribute Rating Technique with Swing weight can be seen from the name that SMART is improved by adding the swing weight method.The SMART method was first proposed by Edwards in the 1970s and is a simple and intuitive decision-making method.After improvement,there were SMARTS and SMART Exploiting Ranks (SMARTER).At present,SMART has been eliminated and no longer used,and SMARTS and SMARTER methods are more widely used,because the latter two can better improve the phenomenon of excessive subjective decision-making in the SMART method [87,92].

Fig.4.Steps to use AHP,PASC is Priorities of the Alternatives in each Sub-criterion,PSC is the Priorities of the Sub-criteria for each Criterion,PAC is the Priorities of the Alternatives in each Criterion,PC is the Priorities of each Criterion,PA is the final Priorities of the Alternatives [89].

The use of SMARTS is mainly divided into nine steps,except that the swing weighting method is applied in the seventh and eighth steps,the other steps are the same as SMART.The SMARTER is similar,except that the weight is directly calculated using formula [4) in the eighth step.

Where K is the number of attributes.

The SAMRTS method was first used by Bernstein in the decision-making model for the decommissioning of offshore oil and gas facilities,mainly in the PLATFORM mentioned above.The use of this method makes the weight distribution in decision making more flexible,and its subjective influence is reduced [46,93].

2.3.4.MAUT

The Multi-Attribute Utility Theory method was first established in the 1970s [94].It follows the same logic as the traditional utility theory.The criteria are normalized within the range of [0,1]according to their purpose,and then a weighted score for each available alternative can be obtained [95].The feature of this method is additive independence,which means that a person’s preferences have no interaction between attributes,and the value of one attribute is not affected by other attribute levels.In this way,the utility function of each attribute can be evaluated separately and evaluated according to the weight of combining them into a multiattribute utility function.The specific method is as follows [96]:

1 Identification and organizational attributes.

2 Define a scale for each attribute.

3 Define a single attribute utility function,and divide the possible ratings of each attribute from the worst 0 to the best 100%

4 Choose swing weight or equivalent cost to construct the stakeholder’s relative value or cost preference model for each attribute.SMARTS is used here to get the weighting.

5 Integrate swing weights and attribute scores into the overall multi-attribute utility of each decision option.

6 The specific mathematical expression is shown in formula [5].

The MAUT method has been used in the decommissioning of nuclear [79)and offshore oil and gas platform decommissioning[46,93].Some scholars have compared AHP and MAUT in the field of nuclear energy and found that the potential disadvantage of the latter is that it is difficult to generate a utility function for each criterion,while AHP requires a method of generating a comparison matrix for each criterion and sub-criterion is very troublesome[97].

3.Multi-attribute decision-making models

3.1.Multi-attribute decision-making models’ description

The decision-making model for the decommissioning of offshore oil and gas facilities is an important part in offshore decommissioning project,this framework can give decision makers and relevant information demanders an intuitive decision basis.It is generally believed that this type of model has five major categories,namely environmental,economic,social,health and safety(risk),and others.Although according to the requirements of the British government,it is necessary to include technical feasibility in the model,in fact,the purpose of technical feasibility analysis mainly serves the selection of equipment,so that the result can be aggregated into the cost and risk part.In this paper,there is no discussion,readers can refer to relevant British government documents according to their needs.The contents of these categories are shown in Table 3 [5].The ultimate goal of these decisionmaking models is to use qualitative or quantitative methods,or MCDA (multi-criteria decision analysis) method [36,98],as the main method,this comparative support tool mixed by methods such as weighting or expert review,to perform calculations based on the focus of the model,so as to arrive at the most appropriate decommissioning plan under the model.However,it should be noted that the results obtained by the MCDA method are not the so-called optimal solution but one of an acceptable solution [36],which will be expanded in Section 2.2.

The existing decision-making model has been developed specifically for the decommissioning of offshore oil and gas facilities[101],but more of it is borrowed and modified from other field like onshore oil and gas industry [102],and marine environmental industry [11,33],but these models generally take the decommissioning option as the basic logical route which mentioned in Fig.3.According to the selection of different options,the corresponding evaluation results are obtained.Finally,according to the different weight arrangements for each focus,the evaluation results are summed to obtain the overall performance evaluation is to compare the pros and cons of the results and arrive at a suitable oil and gas facility decommissioning decision.Table 4 shows the names,established year,main contents and characteristics of some decision-making models.Some details about these models shall be given after Table 4.

The PLATFORM is a decision-making model dedicated to the decommissioning of offshore oil and gas facilities developed jointly by Bernstein and MMS.This model is currently for comprehensive evaluation.The method used is the MCDA method mentioned above,which combines the weight method,quantitative method and qualitative method.The developed tool uses historical data from decommissioned platforms in the Gulf of Mexico in the United States to obtain parameters and weight values suitable for offshore installations in the Gulf of Mexico.The model has been developed in multiple versions,from being only used for fixed shallow water platforms to deep water floating platforms and CGBS[32,93,105,106].And this model is currently known to the author,and the only model for interactive software exists.However,the model did not consider engineering risks,marine wind and wave risks,etc.,when making decision-making assessments,and provided more room for cost surplus.Therefore,its decision-making ability needs to be improved atwwww present.

DAPSI (W)R (M) represents the basic concerns of the framework,namely: Drivers,Activities,Pressures,State changes,Impacts(on Welfare),and Responses (as Measures) [13].The framework was enhanced by DPSIR,Drivers-Pressures-State-Impact-Response,and the purpose of the initial development was to establish a causal model to describe the relationship between society and nature[30,107].Just like the original DPSIR,DAPSI (W)R (M) is also a framework that focuses on causality and response.It was used to manage environmental systems.Later,through the Cooper [108],UK National Ecosystem Assessment Follow-On (UKNEAFO) project[109,110],Elliot [111],Smyth et al.[85]and others further developed [11,112-114].The purpose of this framework is to answer eight questions.Correspondingly,there are 7 corresponding resources to solve these problems.See Table 5.For the decommissioning of oil and gas platforms,Table 6 can give the decommissioning content corresponding to the relevant elements of the framework.

Table 4 Models’ information.

Table 5 DAPSI(W)R(M) core questions and resources [13].

Table 6 Elements of the DAPSI(W)R(M) framework of relevance to decommissioning [13].

The framework was established to deal with the decommissioning in the Marine Protect Area (MPA),so it pays more attention to environmental and social response.However,newer data cannot be applied,and security,social impact,technical and cost issues cannot be combined.This framework provides decision results and play a role of consultation,but it cannot make decisions by its own.It is worth mentioning that the creator of the DAPSI(W)R (M) framework Burdon et al.[13]mentioned in their article a number of tools and frameworks related to the decommissioning of offshore oil and gas facilities for MPAs.However,according to the author’s own inspection,development of these frameworks or the original purpose of the tool was only for the protection of marine life,and its function overlapped with the decommissioning of offshore facilities but could not be completely relied upon.But the related research results about Net Gain mentioned by it are no longer available.

BPEO,Best Practicable Environmental Option,together with ALARP (As Low As Reasonably Practicable) [115],is the basic decision-making framework and risk control framework currently adopted for the decommissioning of offshore oil and gas facilities in the UK.BPEO was applied by the UK RCEP (Royal Commission on Environmental Pollution) in 1995 to manage industrial waste[115].It is defined as "the outcome of a systematic and consultative decision-making procedure which emphasizes the protection and conservation of the environment across land,air and waste" [116].In terms of actual decommissioning issues,EIA (Environmental Impact Assessment),EA (Economic Appraisal),SIA (Social Impact Assessment) [103,117].will be combined for decision-making,and the whole is called Integrated Assessment.The decision-making process of BPEO is mainly divided into 5 steps,corresponding to 7 methodologies,to integrate the environmental and social assessment results.The process and corresponding methodology are shown in Fig.5.The methodology will be explained in detail in the later section.Based on the BPEO framework,the Environmental Impact Assessment (EIA) method currently adopted by the British government and industry,as well as the comparative assessment method,has been very mature [42,103,117].According to the requirements of BEIS (Department for Business,Energy and Industrial Strategy) [75],these two methods integrate five aspects of cost,environment,risk,social impact,and technical feasibility.They are currently the mainstream methods.Although the five aspects are treated equally in the evaluation method (equal weights are given),in actual use,their subjective components still have an impact on the decommissioning decision.Secondly,these two methods do not consider the impact of risks on costs,environment,and society.In actual operations,industrial methods are still used to calculate the risks separately.In my opinion,this aspect needs to be improved.

Fig.5.BPEO procedure [103].

Comparative Assessment is an assessment model widely used in many fields.In the United Kingdom,the CA model used in the evaluation of the decommissioning of offshore oil and gas facilities is a model developed after integrating the frameworks of BPEO,ALARP,and EIA.This is the first model that uses five criteria as the evaluation criteria at the structural level.It is also the decisionmaking model used for the decommissioning of offshore oil and gas facilities,which is more mature using MCDA so far.The main focus of CA is the five criteria,and its evaluation process is shown in Fig.6 [104].Many methods are used to evaluate the five criteria,including quantitative methods such as theoretical formulas,qualitative methods based on comparison,and expert evaluation methods.The more advanced point of CA is the addition of sensitivity analysis to weaken the subjective bias caused using more qualitative methods and expert assessments.This is not available in other decision-making models,so this model can be considered more advanced.

However,since the formulation of criteria and sub-criteria is artificial,there may be overlap or high correlation between the criteria,which leads to multiple evaluations of a certain criterion.Secondly,there are many qualitative methods and expert evaluation methods used in the evaluation process.The use of quantitative methods is also relatively traditional and cannot reflect the uncertainty and randomness of actual projects [36].

According to the description of each frame in Table 4,combined with the five criteria classification of the decommissioned frame in Table 3,the radar chart in Fig.7 is obtained to visually show the performance of each decision frame in the five criteria.This part is more subjective and designed by the author based on the understanding of the three models.

It is true that this type of model is scientific and intuitive,but based on actual conditions,it is not widely accepted.Different from the decommissioning of onshore facilities,it is not only the traditional risks and costs,but the decommissioning of offshore oil and gas facilities will have an impact in many areas.The marine environment and the impact on other users are also valued by the public.Brent Spar,which was decommissioned by Shell in 1995,is a typical example [19,26].Although this example has been widely mentioned by scholars in the field of offshore oil and offshore platforms,it also has a strong position in the history of marine environmental protection.It must be said that the initial decommissioning decision of Brent Spar is scientific,low-risk and low-cost of.However,the public’s concern is that the decision poses a potential threat to the marine environment and may harms the benefit of other water users.As a result,after a long negotiation,Shell compromised with environmental protection organizations and dragged the submerged body of brent spar onshore for disposal at any cost.

3.2.Cost assessment model

3.2.1.Costassessmentmodelframeworks

As an important part of the decision-making model,the cost evaluation model is also an important part of the decommissioning decision-makers.The accuracy of the assessment results not only affects the duration of the project,but also affects the environment,risks,government approvals and many other aspects.Therefore,academia and industry have always paid great attention to more accurate and versatile cost evaluation models.However,the cost assessment of the decommissioning of offshore oil and gas facilities is different from the decommissioning of onshore oil and gas facilities.Its data sufficiency [118,119],uncertainty,and insuffi-cient historical experience make it difficult to develop accurate cost assessment tools and cannot guarantee stable accuracy.It is known that among the currently published models,the higher accuracy can only be controlled at an average of about 35% (the weather impact is directly assessed as 20% of the total cost,while providing an error space of 15%) [70].Especially with increasing emphasis on marine environmental protection,an accurate cost assessment model means that the further optimization of project schedule,environmental impact,and risk control is very important and valuable.

Fig.6.The CA phases [42].

Fig.7.Models’ features comparison.

At present,the U.S.and U.K.offshore industries have each developed a set of cost assessment models based on their own technology and conditions.The U.S.model is the PAES tool which is mainly developed by ProServ Offshore [65].In the United Kingdom,a model form combining CA and EIA is adopted [103,104],that is,based on the historical data of decommissioning of similar facilities,evaluation is performed to estimate the cost of decommissioning and the environmental impact.

The models used by the two countries are like some extent,because according to PAES development records,the model also uses comparative assessment methods in some parts of the evaluation,such as gas emissions,energy use,waste management and other aspects [120].The main difference between these two models is the assessment of detailed costs.The United States uses a compound regression method to establish an empirical formula to estimate the cost.There is no empirical formula in the method presented in the UK.The evaluation method is still developed by various energy companies and oil service companies based on the data they have,and the framework is the same.

Back to the beginning,the academia has already given a more authoritative explanation on how to establish a cost evaluation model.Two mainstream frameworks show as Fig.8,which are topdown and bottom-up frameworks [119].The top-down framework mainly uses historical data,starting from a larger level of data,ignoring part of the cost accounting that is too detailed,after data processing-such as standardization-and then regression,there are various regression methods,and the following methods are mainly the method used in data analysis.For example: get the topside structure decommissioning cost,total emissions,construction duration and other data of some decommissioned platforms,as well as the basic parameters of the topside structure’s weight,scale,number of modules,etc.,through the aggregation of the same type of data on these platforms,data processing,data analysis,get its functional relationship,and then use it to evaluate the cost of topside retirement of other structures.This method requires a large amount of accurate historical data and is a framework based on data analysis methods.(Fig.9)

Fig.8.Two cost model frameworks [119].

Fig.9.Asset retirement obligation balance Equation [119].

And bottom-up focuses on the detail level,that is,starting with each task of the decommissioning project,cost estimation is carried out separately,and then summarized.From a theoretical perspective,a theoretical model is established for each step of the decommissioning of oil and gas facilities,combined with engineering guidelines and engineering data,and from the perspective of man-hours or unit volume,the relationship with the cost is established to conduct cost evaluation.

These two types of frameworks currently have their own advantages,but it is difficult to achieve a better unification in form.From a theoretical point of view,the model built by the bottomup framework is more pertinent.For a certain type of oil and gas facility in a certain area of the sea,the cost can be estimated in detail and accurately without the need for similar historical data of decommissioning projects.However,the establishment process requires the collaboration of multiple experts to evaluate engineering standards,parameter settings,etc.The process is cumbersome and complex and does not have wide applicability.The top-down framework is simpler,which is conducive to the industry’s own development,and the versatility of the resulting model will be more advantageous.But first,the framework requires more historical data of decommissioned projects as a database,and the degree of data sufficiency will directly affect the accuracy of the evaluation results.However,the sharing of such information is taboo in the industry because it involves issues such as antitrust laws and trade secrets.At the national level,such as the British BEIS,although such data can be obtained,it cannot be granted the right to share such data [121].Secondly,the model evaluation results under this framework are often not as accurate as the model evaluation results under the bottom-up framework.Although in the future,ample data may improve the accuracy,it is still difficult to match the accuracy of the bottom-up framework.The model is comparable.

In recent years,some scholars have used economics or other industry methods to innovate the decommissioning cost assessment framework from different perspectives [119,122],mainly based on the decommissioning methods of nuclear industry facilities and chemical industry facilities.Cost assessment for the decommissioning of offshore facilities.But its model framework does not deviate from the two modes mentioned above.We look forward to the emergence of more novel and effective models to solve this urgent problem.

3.2.2.Severalcostassessmentmodels

Platformabandonmentestimatingsystem(PAES): Based on the decommissioning data of three platforms in the Gulf of Mexico,ProServ Offshore developed a simple algorithm for the decommissioning of oil and gas platforms in the GOM region in 2000 [123].The main content of this algorithm is to fix some costs,such as HLV (Heavy Lift Vessel) mobilization/demobilization costs,and daily costs.And part of the cost,such as the cost of the mobile platform is returned.The sum of all values is the method of decommissioning cost of oil and gas facilities to estimate the cost of platform decommissioning.At the beginning of the model establishment,due to the small amount of data,the accuracy of this algorithm is not good enough,and the estimated value is often more than 25% lower than the actual cost.

Later,as the sample data gradually increased,the PAES system was also improved.Especially with the addition of Mark J.Kaiser and Brock B Bernstein,as well as organizations such as ICF (International Coaching Federation) and BSEE (Bureau of Safety and Environmental Enforcement) [3,31,124],the model now has even better performance.The model determines that the management cost of platform decommissioning accounts for 8% of the total cost,the weather factor is preset to 20%,and the cost flexible interval is 15%.This is a cost evaluation model that is very suitable for rough estimation by industry and government agencies.Although it has been improved so far,the definition of its accuracy is still large fluctuation,between 6% and 258% [3],but it is still the best in the cost evaluation mathematical model.This mathematical model is also used in the decision model PLATFORM mentioned above.

MarkJ.Kaiser’smodels: Mark J.Kaiser has made many optimizations based on the PAES model and established a variety of cost evaluation mathematical models from different angles.It has a very high reference value and use value.Although the scope of application of the model is partially restricted,it is not applicable to all offshore oil and gas facilities at present,and the actual parameters have regional differences.For different countries and regions,it needs to be re-studied.But the template he provided is worthy of in-depth study by related scholars.

In Kaiser’s 2003 model [61],he innovatively classified the platforms according to four legs less than or equal to four legs and greater than four legs,and concluded that in each of the six main decommissioning activities,four leg platforms and eight legs The average cost of the leg platform and the standardized data.And get the key parameters of the six main activities of the four-leg and eight-leg platform,such as: water depth,number of wells,deck weight,maximum module weight,integrity,etc.,and get regression equations.But the disadvantage is that the coefficient determination R2value of some of the regression equations is too low,many of which are even lower than 0.5,which is not unavailable for regression analysis.Perhaps because of this,the accuracy of the model is not specified in the original text.

Later,in his 2006 article [125],a multi-parameter model was developed.This model provides even as many as seven parameters for the cost of each process in the decommissioning process,including but not limited to: water depth,number of wells,work category,well development degree,workboat deployment,season,waiting season,number of construction period days,etc.,Some of the parameters are two-dimensional,that is,only 0 and 1 are used to distinguish whether to use the parameter.This model is a standard Top-Down model,which requires a lot of data and is very detailed,and at least three models have been developed for each activity to choose from.Although the accuracy of this model is not clearly stated in the original text,it is known that it must be very high.However,the data required for this model is too detailed,which is unrealistic for researchers or model users without data support.

In 2014,Professor Kaiser developed a dedicated decommissioning cost assessment model for the Gulf of Mexico deep water fixed platform and the compliance tower platform [65].Deep water means a platform with a depth of more than 400 ft.The model uses data from 53 deep-water fixed platforms and compliance towers.The decommissioning costs of such platforms are often very high,so the impact is huge.In this model,more detailed platform decommissioning division and decommissioning options are applied.In this model,regression analysis is further used.Except for wet tree retirement costs and riser removal costs,which are determined according to unit prices,regression methods are used for other parts.The difference is that in the regression analysis of each part,they are classified according to the number of wells,conductors’ number,water depth,etc.,to obtain a more accurate regression equation.It is worth mentioning that not only linear regression analysis is used,but nonlinear regression analysis is also used more frequently.

ComparativeAssessment(CA): Although CA is a decision-making model,it also includes the cost assessment part.As the most intuitive,simple,no software calculation,and can consider multiple evaluation methods,the cost assessment part of CA can be considered as the most versatile and reliable model in the current mature technology.

In the CA,a strict assessment process is the key to ensuring the reliability of the result.Fig.8 shows the CA process and its explanation.In the entire process,the setting of criteria and sub-criteria is the most important part [104].Strictly speaking,the cost assessment part of the CA cannot be called a model,but is only a part of the comprehensive assessment,which requires relevant experts to evaluate according to Asset Retirement Obligations (ARO),and the result is often reported as a percentage of the lowest cost to the highest cost of options or report the difference in orders of magnitude.The accuracy of this assessment method is compared with the Decommissioning Programme and Close Report reports of the decommissioned offshore oil and gas facilities in the UKCS area.It can be concluded that the accuracy fluctuates greatly.The better evaluation results only differ by several million pounds [126],and there are individual actual costs that exceed the expected results by 100% [127,128].

LogarithmTransformationModel(LTM): The Logarithm Transformation Model,also known as LTM [60],was developed by Malaysian scholars in 2015 to evaluate the cost of a full-water depth platform.The model largely draws on the results of Professor Kaiser and applies data from the decommissioning of offshore oil and gas facilities in Malaysia,making the model localizable.In this model,regression analysis is widely used,but unlike other models that mostly use linear regression analysis,this model uses a lot of nonlinear regression analysis.When substituting data,calculations are made based on Malaysia’s own working hours and efficiency.The entire model also deals with the decommissioning project step by step.Unlike the model of PAES and Professor Kaiser,this model also makes an empirical equation for the project management part of the cost to make a more detailed cost estimate,which is undoubtedly conceptually An improvement to improve accuracy,but for the actual effect,although R2and correlation analysis data are given,because the amount of data cannot be determined,and the verification part of the entire model only uses one example,whether the regression equation is over Saturation is not known.Because it is well known that the Top-Down mode uses historical data for fitting.It is a posterior model,so the over-fitting situation cannot be ignored.Even if R2performs well,it is very likely to have occurred.Fitting situation,so only a good prediction for one case,and cannot be applied to the prediction of other cases [129].

SettledLiabilityModel(SLM): Settled Liability Data Model is the latest research result of Professor Kaiser in 2015 [119].He broke away from the original Top-Down and Bottom-Up frameworks,but used economic statistics,which is refreshing.The reason for the development of this cost assessment model is that more professional and detailed oil and gas facility decommissioning data cannot be easily obtained.These sensitive data are related to the commercial secrets of energy companies.Often these data are only available to energy companies,oil and gas service companies hired,and relevant government departments.Such information opacity is important to researchers or policy makers in other related fields.Extremely unfriendly.Therefore,in this model,the physical characteristics of oil and gas facilities are no longer important,while Settled Liabilities,Working Interest,the number of decommissioned wells,and the number of various types of facilities of related companies are more important and easier to obtain from the public.Obtained from the data.After Eq.(6) to Eq.(7) are calculated,the average annual decommissioning cost can be obtained,in units of millions of dollars per structure.The model is loaded with data every year,and the coefficients are different every year.It is not difficult to know the development intention of this model.It is not to accurately provide energy companies or oil service companies with a more accurate decommissioning cost assessment,but to the government.Policy and strategy makers such as institutions and market analysts provide reference.This is a very novel model that combines the knowledge of statistical economics and offshore oil engineering.

Where,AROtmeans Asset Retirement Obligations at the beginning of year t;

LItmeans Liabilities Incurred in year t;

LStmeans Liabilities Settled in year t;

AEtmeans Accretion Expense in year t;

REVtmeans Revisions in Estimated Liabilities in year t;

DECt(Ci)means Decommissioning Cost of companyCiin year t;

αiis coefficients;

ACTIVITYj,t(Ci)means Ownership position of companyCiperforming decommissioning ACTIVITY j in year t.

4.Conclusion

The development of the multi-attribute decision-making model for the decommissioning of offshore oil and gas facilities has so far possessed certain functionality.The industries of the United Kingdom and the United States have also cooperated with the academic community to develop a multi-attribute decision-making model suitable for their respective countries.The performance of these models has been stable so far.Although they can basically meet the concerns of all parties for retirement,there are still shortcomings.These challenges are mainly concentrated in the following two aspects.

(1) The first challenge for the development of offshore oil facility decommissioning model is the abundance and detail of the available data [118].As can be seen from the above,in terms of algorithms and methodology,regression analysis and MCDA have become or will soon become mainstream,and are also recognized as the most reliable and efficient algorithms and methodology.Although MCDA can get rid of the support for data to some extent by using experts’ opinions,the use of too many qualitative methods and expertdefined criteria makes this method not objective enough and inefficient in actual use.For regression analysis,the abundance and detail of the data greatly affect the performance of the regression equation.

(2) Secondly,for MCDA,this methodology will be the mainstream of future offshore oil and gas facility decommissioning assessments.The methodology is intuitive,easy to use,can integrate concerns in multiple fields,and does not require complex software for calculation advantages.It is reported that it has been used in the decommissioning of multiple industrial facilities.However,an important step of the methodology,namely the formulation of criteria and the formulation of sub-criteria,requires experts to discuss.The formulation of these criteria is likely to have repetition and high correlation,and it is likely that the same type of criteria has been evaluated multiple times in the actual evaluation.As far as sub-criteria are concerned,incorporating uncertainty and randomness into qualitative and quantitative evaluation is a problem that needs to be solved in the future.This means that the entire MCDA needs to be complicated and integrated into the most advanced technology to increase its ability to adapt to random problems [36].

The decommissioning of offshore oil facilities is complicated and tedious.Although the existing decision-making methods have multiple attributes,they still need to be improved in terms of attribute types,ease of use and efficiency.This improvement requires not only the development of algorithms and theories,but also the richness and accuracy of data.In terms of methodology,drawing on and modifying the methodology of other industries,such as the use of methodology in the field of economics,is considered feasible and novel.This may be a very good progress in the direction of macro retirement.It is foreseeable that MCDA will become the mainstream methodology for future offshore oil facility decommissioning projects.Combined with more advanced algorithms and data,this method will shine in many industrial fields.Secondly,as far as data is concerned,much progress has not been made.Energy companies and governments mainly control data.Due to the sensitivity of trade secrets and the existence of antitrust laws,it is temporarily not feasible to share these data at will.

This article groundbreakingly summarizes most of the multiattribute decision-making models currently applied to the decommissioning of offshore oil and gas facilities.And analyzes their pros and cons and the methodologies used in detail,including those that are rarely used in the engineering world,statistical economy the methodology used in the academic field.The cost evaluation model is also meticulously summarized and analyzed,because this part is often considered extremely important in the multi-attribute decision-making model.In addition,in the summary of the cost evaluation model,the author groundbreakingly questioned the verification method of the cost evaluation mathematical model.The author believes that when using retrospective analysis,the use of traditional R2to verify the pros and cons of the model has been challenged.The reason is the insufficient historical data mentioned above may lead to overfitting of the regression equation.

However,this article is only limited to the content related to the decommissioning of offshore oil and gas facilities.In fact,the MCDA method is involved in nuclear energy,chemical industry,mining and other fields.The relevant literature can be obtained in the MCDA method description chapter,and readers can also refer to it.The literature of Martin et al.[36]has a very comprehensive review.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.