Models for the Multicriteria Selection of Options for Decommissioning Projects for Offshore Oil and Gas Structures

Abstract

Companies are striving to optimize the decommissioning process of oil and gas facilities in order to reduce overall costs. Various options and criteria for making decisions on the liquidation of objects are considered, which are based on a multicriteria approach that allows optimizing this complex process. The most important characteristics of a reliable model for optimizing options for decommissioning oil and gas facilities are defined as follows: the option should take into account qualitative and quantitative criteria. To minimize the subjectivity of human judgment, a systematic poll of skilled performers should be conducted using a well-structured technique. It is shown that multicriteria decision analysis can be considered as an adequate model for choosing options for decommissioning oil and gas assets. Based on a test case that reflects the practice of decommissioning offshore platforms in the state of California (USA), the logic and algorithm of calculations are reproduced when choosing the best options for decommissioning facilities. The presented calculation scheme is quite universal and can be used at other facilities, including Russian ones. In order to determine the optimal choice, three options for the project model for the decommissioning of oil and gas facilities of offshore structures (platforms) were considered: complete liquidation, partial liquidation, and leaving in place.

1. Introduction

Most companies are postponing their decommissioning activities due to cash flow restrictions; others are taking advantage of low oil prices to speed up decommissioning activities as some of the related oilfield service costs are also reduced.

However, decommissioning is an emerging and growing market. Global decommissioning projects are expected to increase from about USD 2.4 billion per year in 2015 to USD 13 billion per year by 2040, with approximately 2000 offshore structures to be decommissioned between 2021 and 2040. Companies are looking for ways to streamline the decommissioning process to reduce overall costs [1].

In this article, we consider various decommissioning options and decision criteria, in particular a multicriteria decision analysis that makes it possible to optimize the decommissioning process based on previous studies [1,2,3]. When an offshore platform is considered as a whole, the options for decommissioning include leaving it in place, partial decommissioning, and full decommissioning. However, in practice, different decommissioning options may apply to individual platform components, depending on the results of its evaluation. Key components expected to be removed during decommissioning include wells, topsides, the substructure, and the subsea infrastructure.

There is a marked tendency towards the optimization of a decommissioning project when a platform is considered on a component-by-component basis rather than as a whole because the preferred decommissioning option for one component may not be appropriate for another. Most topsides weigh between 1000 and 30,000 tons, and their dismantling (small-piece dismantling or single-lift method) is an expensive operation requiring the use of large sea crafts; however, it is a mandatory requirement in international decommissioning laws such as the United Nations Convention on the Law of the Sea.

However, as shown in Figure 1, the largest of these components are the topsides and substructure.

On the other side, there is a number of decommissioning options for steel shells, such as onshore disposition, capsizing on site, leaving in place, etc. The suitability of a decommissioning option for any given structure depends largely on its configuration, weight, material, equipment, and location, so no decommissioning option is universal for all possible scenarios. Therefore, decommissioning requires a highly individualized approach, where the best option is selected based on existing or generally accepted requirements for the decommissioning of a structure.

It is also important to mention the importance of production techniques in offshore platforms such as low-water-salinity projects (Experimental study of the low salinity water injection process in the presence of scale inhibitor and various nanoparticles, 2022).

2. Existing Approaches to Optimizing the Choice of Decommissioning Options

Compared to other aspects of the life cycle of an offshore platform, the application of modeling techniques in decommissioning has not been well studied.

However, thanks to the global increase in the number of platforms requiring decommissioning and their availability, significant progress has been made in this area over the past decade as practical experience is gained in implementing such necessary, and costly, projects.

A number of independent publications describing offshore decommissioning operations in various regions of the world that deserve attention can be found in the public domain [2,3]. Governments of oil-producing countries are working to reduce costs by encouraging decommissioning knowledge sharing between operators. For example, Decom North Sea launched an online tool called the Late-Life Planning Portal in 2017 to promote collaboration between regulators, operators, and the supply chain to improve decommissioning practices (Decom North Sea 2017).

Similarly, it is proposed to use the Contract Verification Process (CVP) as an initial assessment prior to the detailed engineering design and execution of the decommissioning project. These developments have led to a rapid growth in decommissioning knowledge, which has stimulated research in this area and increased efforts to streamline and speed up the decision-making process in selecting the appropriate decommissioning option from a range of alternatives. Decision models that have been proposed for evaluating decommissioning options, depending on the nature of the criteria they work with, can be broadly classified into quantitative and combined approaches.

2.1. Quantitative Approaches

These approaches include decision models that refer only to measurable information about the decommissioning project in order to realistically minimize subjectivity and bias in the results obtained. Quantitative approaches are preferable when measurable data are available and easily obtainable. They produce results that can be easily shared, verified, and replicated.

However, the application of quantitative approaches to the evaluation of decommissioning options is limited due to the lack of data for the evaluation of several criteria used in the selection of decommissioning options. In particular, quantitative indicators are not fully developed for criteria such as technical feasibility and public opinion.

It is concluded that most studies using these approaches tend to focus on only a few decision criteria. In addition, obtaining the necessary data for these models is expected to be costly and unpredictably difficult.

2.1.1. Methodology Based on Ecosystem Assessment

The ecosystem approach is used in management to assess options for decommissioning offshore platforms by analyzing all interactions within the ecosystem where the platform is located. Recently, a Remotely Operated Vehicle (ROV) survey was conducted in the Gulf of Mexico on a deep-sea platform with a fixed steel jacket to obtain data that were then used to rank three decommissioning options based on their ecosystem services. The results obtained, based on facts and empirical data, showed that it is better to leave the jacket in the sea environment. However, applying this method for another platform requires extensive research, and the model has a limited scope because it primarily considers environmental decommissioning criteria. Thus, it can only serve as supporting evidence for a more thorough assessment of options in decision making.

2.1.2. Methodology Based on 4D and 5D Information Model of a Structure

This methodology is based on Building Information Modeling (BIM) technology, which can be used to evaluate decommissioning options by visualizing the schedules, costs, and resources involved in their implementation. This approach is useful for project management when considering decommissioning and can help reduce non-productive time, control costs, and monitor the rational use of resources. However, it does not indicate which option is best for a given offshore platform; in addition, environmental safety and other aspects of the project are not sufficiently taken into account. Obviously, this oversimplification of decommissioning projects makes this method unattractive.

Conclusion: The use of quantitative approaches is favorable for progress in ROVs and other technologies that can be used to measure the structural and ecological health of offshore platforms. However, at present, their exclusive use for the full evaluation of decommissioning options is not entirely adequate because other criteria, which are difficult to quantify, are also of equal or even greater importance to stakeholders.

The theoretical method is based on the use of theoretical formulas. However, using this method, it is impossible to estimate parameters that do not have clear relationships, which is a considerable limitation. The equivalent cost method combines several physical parameters with a cost value. For example, water depth, weight of structures, number of structures, etc., are associated with monetary units. This method is similar to regression analysis and is also a widely used regression method. Such an analysis is mainly used to determine the relationship between variables that have not been theoretically determined [5].

In decommissioning tasks, regression analysis is often used after analyzing logical relationships, such as finding the relationship between decommissioning costs and engineering works, water depth, module weight, etc. [5]. Regression analysis is based on historical data and gives a curve of the relationship between the parameters.

The aim of material and energy flow analysis (MEFA) is to obtain a variety of information such as on emissions, thermal radiation, workmanship, pollution, etc. This method was applied to the decommissioning of offshore oil and gas facilities [5]. The challenge is that the analysis requires a high degree of information control. As a general rule, only the decommissioning organization can obtain the most detailed data first hand. This is the biggest limitation of the method.

Table 1. Quantitative methods and their description.

Method	Description	Application
Theoretical method	▪ Theoretical formula for data substitution; ▪ High reliability; ▪ Lack of flexibility; ▪ Universal.	▪ Cost estimation. Risk assessment; ▪ Energy consumption and gas emissions; ▪ Commercial fishery; ▪ Employment status.
Equivalent cost method	▪ Regression analysis of relationships.	▪ Cost estimation.
Regression analysis	▪ Connection; ▪ Universally applicable; ▪ High demand for data; ▪ Simple process.	▪ Cost estimation. Risk assessment; ▪ Sea conditions modeling.
Material and energy flow analysis	▪ Boundary conditions; ▪ High reliability; ▪ High demand for data; ▪ Cross-domain application.	▪ Cost estimation; ▪ Estimation of energy consumption and gas emissions; ▪ Environmental assessment.

shows the use of quantitative methods in decision-making models for the decommissioning of offshore oil and gas facilities. In many cases, these methods are only auxiliary to the main methods of model estimation and are used in combination with qualitative methods [5].

2.2. Qualitative Methods

Since many technical and economic aspects cannot be analyzed quantitatively, qualitative methods are widely used in making decisions about the choice of decommissioning options.

There are many types of qualitative methods used in decommissioning, including (the list is not complete) the expert evaluation method, comparative evaluation method, risk matrix method, etc. [5]. Three systems of qualitative evaluation are commonly used: weighting, evaluation matrix, and comparative evaluation.

Many qualitative methods are based on these three systems, to which the latest technologies are added. The weighting method can be either quantitative or qualitative, depending on how it is used. The main difference is whether the choice of weighting is subjectively determined by the decision maker (DM) [5].

If the decision maker determines the weight value subjectively, a qualitative method is preferable. Its detailed description is given in the explanation.

This section focuses on the scoring matrix system and comparative evaluation system. A feature of scoring matrices as a qualitative analysis is the creation of a matrix that includes all options necessary for evaluation, evaluation criteria applicable to each option, and a set of unified significance estimates [5]. This list can be presented as scores ranging from 0 to 100, or as ratings from worst to best.

Decommissioning options, combined with estimated figures, are presented in

Table 2. Scoring matrix example.

Evaluation Criteria	Subcriteria	Platform Decommissioning Options
		Full Decommissioning	Partial Decommissioning	Leaving in Place	Disposal at Sea
Safety	Personnel risk	1	2	2	1
	Risk to other sea users	1	2	3	1
	Risk to those on land	2	1	1	1
Environmental	Impact on the marine environment	1	2	3	1
	Energy/resource consumption	3	2	1	1
	Other environmental consequences	2	3	3	2
Technical	Risk of major project failure	2	2	2	3
Social	Impact on fisheries	1	2	3	2
	Services	3	2	1	3
	Communities	3	3	3	3
Economic		3	2	1	1
Sum		24	26	26	21
Rating		2	3	3	1
Scores		High = 3; Medium = 2; Low = 1

. Decision makers only need to assign certain scores according to their judgment; then, to select an option, the results are weighted and added to the total score.

The system of comparative evaluations is widely used in practice. Such methods include: the ranking method, mandatory classification method, point distribution model, pairwise comparison method, critical event comparison method, target management method, and complex method.

The basic idea behind comparative evaluation is similar to control variables and proportional scaling. For example, in a decommissioning operation where the same type of platform is used with approximately the same water depth, the energy provider will be based on the platform weight ratio, distance from shore, number of wellheads, length of pipelines, and other parameters to estimate the range of decommissioning costs [5].

In addition, this method requires a large amount of historical data, especially for complex projects such as the decommissioning of offshore oil and gas facilities [5]. However, qualitative analysis methods are easy to use and have wide applicability.

2.3. Combined Approaches

These approaches include decision models that are designed to evaluate decommissioning options in terms of both qualitative and quantitative criteria. They are developed using information obtained from platform data analysis and discussions with decommissioning experts.

Combined approaches are particularly effective in dealing with issues such as decommissioning which require bringing together the goals of different stakeholder groups and handling a wide range of different types of data. The methods used in these models are often an adaptation of the multicriteria decision analysis (MCDA) method. Most studies on decommissioning optimization fall into this category due to its suitability for cases where data are not measurable or non-existent.

A decision model rarely uses only quantitative or qualitative methods. In many cases, they are combined. Such approaches are used in the multiparameter decision model and include, but are not limited to, the following methods [5]: “decision tree”; goal programming; semi-quantitative and qualitative methodologies; analytic hierarchy process (AHP); elimination and choice translating reality (ELECTRE); multiattribute utility theory (MAUT); mixed-integer programming, environmental benefit analysis; oracle multicriterial general assessment of decommissioning (OMEGA); preference ranking organization method (PROMETHEE); strengths, weakness, opportunities, and threats (SWOT analysis); political, economic, social, technology, legal, and environmental (PESTLE) approach; multicriteria decision method (MCDM); and other methods.

2.3.1. Methodology Based on Material and Energy Flow Analysis

One of the first studies in this area was the use of material and energy flow analysis to evaluate the effectiveness of two decommissioning options compared to a leaving-in-place option for a large fixed steel platform in the North Sea. Twelve decommissioning options were considered where data not available for analysis were obtained from platform owners on a confidential basis with estimates derived from two past decommissioning projects. The results showed a useful relationship between partial and full decommissioning options that could be extended to other platforms. In addition, the wide range of issues included in the model makes it useful as a guide for future work. However, the information needed for analysis is only available as a result of the extensive platform research that takes place just before decommissioning and is often hidden from the public by operators to protect their reputation. Thus, this approach is not easily replicated for use on other platforms.

2.3.2. Net Present Value Analysis and Weighted Evaluation Method

The experience of using the knowledge of decommissioning experts is of interest in developing a hybrid approach that combines the weighted evaluation (WE) method and net present value (NPV) analysis to assess the non-financial and financial implications of the four decommissioning options. This approach is essentially a variation of the analytical hierarchy method (which is the MCDA method) and involves pairwise comparisons of criteria to determine weights and evaluate decommissioning options based on their individual performance with reference to these criteria. The scores obtained are divided by the net present value of the options being evaluated to obtain a benefit-to-cost ratio, and this ratio is then used to rank the [3].

Mathematically, the procedure is expressed by Equations (1) and (2):

$$A_i=V_{i=1,n}{{\displaystyle \sum }}_{j=1}^m{W}_j{S}_{ij}$$

(1)

$$BT{C}_i=V_{i=1,n}\frac{A_i}{NP{V}_i}$$

(2)

where A is the weighted evaluation of the decommissioning option; W is the weight of the importance of the criterion; S is the performance rating of the option with reference to the criterion; and BTC and NPV are the benefit-to-cost ratio and the net present value of the decommissioning option, respectively.

To develop a more reproducible analysis and emphasize the importance of public opinion in evaluating decommissioning options, the approach used in this model has been refined by taking into account industry experience. In addition, the use of expert opinion also suggests that as more decommissioning experience is gained, the evaluation results will become more accurate, despite the subjective nature of opinions. However, it remains unclear how discussions with decommissioning experts were synthesized to provide weight and performance indicators. Therefore, this method also has the opportunity to better structure the process of gathering more reliable expert knowledge.

2.3.3. Multicriteria Decision Analysis

To evaluate decommissioning options based on the opinion of decommissioning experts, it is proposed to use the MCDA method, also called multicriteria decision making. The MCDA method involves assigning scores of 1 and 0 to decommissioning options depending on their characteristics according to decision criteria and ranking each option based on the sum of its scores across all criteria.

This method was applied to an evaluation of options for decommissioning an offshore platform in California, where the results showed that leaving the platform in place was the preferred option [6,7]. The method presented also indicates the relevance of the criteria weights and the viability of using expert opinion when evaluating decommissioning options using the MCDA method. However, the MCDA method is oversimplified as it is a simple pass/fail analysis that provides insignificant information about the relative performance of the options being evaluated.

In addition, conducting a complete evaluation of decommissioning options with all the criteria used in their work is a complex task requiring more than 30 evaluations from each of the experts.

2.3.4. Multiattribute Utility Theory

Various publications evaluating options for decommissioning 27 platforms in California suggest using an MCDA method called multiattribute utility theory (MAUT) [6,7]. The MAUT method is based on utility theory and involves the independent determination of each of the criteria scores. The information for the analysis was obtained from the statistical database for the considered platforms. This model has been widely accepted and has influenced government departments to adjust their regulatory policies to accommodate decommissioning options. In this article, quantitative parameters such as cost were scaled to achieve comparability by interpolation, and the weights of the criteria were changed to suit the preferences of the decommissioning participants.

As a result, it became possible to perform sensitivity analysis using Equation (3):

$$U=\frac{X_i-X_{worse}}{X_{best}-X_{worse}}$$

(3)

However, the criteria used are mainly environmentally oriented and do not take into account other equally important aspects of decommissioning, such as safety and technical feasibility. In addition, due to the complexity of projects, the MAUT method requires the collection of a large amount of preliminary data before it can be applied for decommissioning.

The authors of the method still report gaps in data despite the fact that over a 20-year period 27 platforms were evaluated based on only two decommissioning options. Therefore, replicating this approach in other regions with many more platforms would be costly.

2.3.5. Analytic Hierarchy Process

In addition to the above methods, it is proposed to use the analytic hierarchy process (AHP) to evaluate platform decommissioning options based on structural parameters and optimal project planning [1]. Information for analysis is obtained by interviewing decommissioning experts and is adjusted to reduce inconsistencies in their responses.

This approach is well structured and minimizes errors when comparing expert opinions. However, without taking into account environmental impacts or public opinion, the decommissioning aspects considered are limited, with the weights of the criteria being fixed. In practice, the weights of the criteria are variable as they reflect the preferences of the stakeholders evaluating decommissioning options. Therefore, the analysis can be more reliable if other important criteria are included and flexible weights are used.

2.3.6. Analysis of Optimization Methods for Decommissioning Options

A summary of the reviewed approaches to optimizing decommissioning options, with their decision criteria considered as subsets of criteria for comparative evaluation of options in the case of UK decommissioning, is shown in

Table 3. Existing approaches to optimize decommissioning options.

Methods Used	Approach Type	Criteria Considered	Source	Strengths	Weaknesses
Material and energy flow analysis; financial flows analysis	Combined	Safety—health and safety; Environment—relative energy use and emissions, rate of extraction of materials from existing structure, clean seabed, conservation of non-renewable resources, impact on the marine environment, impact of resource extraction, impact of landfill; Public opinion—jobs in the UK, impact on the fishing industry, impact on fish stocks and other marine life; Economics—financial costs.	Li, Y. et al. [5]	Wide range of decision criteria	Does not take into account personnel safety and the technical feasibility of each option; difficult to reproduce due to the difficulty of obtaining the necessary data
Net present value (NPV) analysis and weighted estimation (WE) method	Combined	Safety—safety, liability for the future; Environment—environmental impact; Technical feasibility; Public perception—public demands; Economic—income generation. * List also includes regulatory requirements	Eke, E. et al. [1]	Easily repeatable; wide range of decision criteria; criteria are weighted	Expert opinion is poorly structured
Multicriteria decision analysis (multicriteria approval)	Combined	Safety—health and safety; Environment; Public opinion—social and economic problems, additional stakeholder concerns; Economic—financial.	Fowler et al. [3]	Easy to understand; wide range of decision criteria	It is difficult to compare similar options because there are only two evaluation results; requires a lot of opinions
Ecosystem-based management (EBM) evaluation	Quantitative	Environment—habitat extent and structural complexity, biodiversity, trophic interactions, dispersal, spawning, secondary production, benthic production, rare species; Public perception—esthetic values, marine-harvested species (commercial), marine-harvested species (recreational), biomedical.	Alvarado, K. et al. [8]	Based on empirical data and minimizes the range of data; detailed analysis of environmental impact options	Limited range of decision criteria; difficult to obtain data on some criteria
4D and 5D building information modeling (BIM)	Quantitative	Technical—graphs; Economic—cost, resources.	Bernstein, B.B. et al. [6]	Based on empirical data; supports project management	Limited range of decision criteria; difficult to obtain data on some criteria
Multicriteria decision method (multiattribute utility theory)	Combined	Safety—compliance; Environment—impact on marine mammals and birds, impact on benthos, air quality, water quality; Economic—economic costs; Acceptability for society—marine resources and fish biomass, impact on ocean access.	Cinelli, M. et al. [2]	Wide range of decision criteria; criteria are weighted	Limited number of decommissioning options under consideration (full and partial decommissioning); difficult to obtain data on some criteria
Multicriteria decision method (analytic hierarchy process)	Combined	Safety—structural integrity; Technical—platform type, weight management, logistics requirements; Economic—weight management.	Bressler, A. et al. [7]	Expert opinion is structured; easily replicated	Limited range of decision criteria

[1].

A predominant theme in the reviewed optimization models is the use of expert opinion to facilitate the analysis of decommissioning options. The lack of empirical data related to offshore decommissioning presents a serious problem for optimizing the process of selecting decommissioning options, so the use of expert opinion seems justified. However, we must not forget that human judgment is subjective by its very nature.

Therefore, it is important to progressively integrate decommissioning data as it becomes available into the approach used for decision making. The most important characteristics of a reliable model of decommissioning optimization options are defined below:

• The option should take into account both qualitative and quantitative criteria;
• If the use of expert opinion is unavoidable, then caution should be taken when collecting opinions systematically using a well-structured technique to minimize the subjectivity of human judgment;
• The multicriteria decision method is an adequate model for choosing options.

3. Model of Ranking and Selection of Criteria for Assessing the Decommissioning of Offshore Structures (Platforms)

The justification of decisions using the theory and methods of multicriteria analysis is becoming more and more relevant.

Based on a test case that reflects the practice of decommissioning offshore platforms in the state of California (USA), the logic and algorithm of calculations are reproduced when choosing the best options for decommissioning facilities. The presented calculation scheme is quite universal and can be used on other objects.

In order to select the optimal version of the evaluation criteria, we will consider below three options for the model of projects for the decommissioning of offshore structures (platforms): D₁—full decommissioning; D₂—partial decommissioning; D₃—leaving in place (artificial reef).

To do this, we first decide on the choice of appropriate criteria.

For the ranking and selection of criteria, a technique can be applied [9,10] based on the fact that for a comparative criteria analysis of the situation in the criteria space, two subspace indicators, S and D, are introduced into consideration, which, like the criteria space, are subsets of the m-dimensional Euclidean subspace (m is the number of criteria) S ∈ R^m, D∈ R^m. S is the subspace where it is desirable to have the values of the criteria that characterize the object after the decision making (scenario, control action).

In cases where the desired state is specified by coordinates rather than intervals, the subset S may consist of a single point S0. D is a subset of points that, according to the manager’s estimates, defines the current state of the object about which the decision is made.

The value of the jth criterion and the relationship of this value with the physical parameters for subsets S and D can be reflected using base scales [9,10] similar to the scale shown in Figure 2. This figure shows an example of the formation of a base scale using the values of the gas production indicator as an example.

It is possible to use an example when various development options are considered for some virtual field. Within these options, the expert selects the maximum and minimum production volumes and forms the boundaries within which all possible gas production volumes are located for all selected development options.

Let us assume [11,12] that the base scale for our example is divided into 5 equal segments (intervals), which are assigned a score in points, for example, 1, 2, 3, 4, 5. However, we can notice that the task of choosing a scale—5, 10, 20, or 100—is not simple, since the choice largely depends on the physical parameters of the problem being solved and the experience and intuition of the manager. It is in this case that the subjectivity of the assessments made by the decision maker (DM) is manifested. An example of constructing such a scale is shown in Figure 3.

The calculation of the values of indicators (criteria) is carried out according to the base scale, as in the scheme of Figure 3.

B is the dimensional (true) value of the criterion (in our case, this is the value of gas production volumes); X is the desired (dimensionless—in points) value of the criterion in the base scale.

From the proportion of the ratio of segments in Figure 3, we have:

$$X-mi{n}_1=\frac{B-mi{n}_2}{{\Delta }_2}{\Delta }_1$$

(4)

$$X=\frac{B-mi{n}_2}{{\Delta }_2}{\Delta }_1+mi{n}_1$$

(5)

$${\Delta }_1=max1-min1, {\Delta }_2=max 2-min2$$

(6)

The results obtained show that X is a projection of B onto the base scale.

Figure 3 shows the ratio of indicators of the base scale and the scale for evaluation criteria in terms of “the more, the better”. For the indicators of “the fewer, the better” scale, which should be the inverse, min₂ should correspond to max₁.

Having thus determined the scores of each indicator on the base scale, depending on the degree of their importance, the expert can also assign them a status of importance in points, as shown above in Figure 2 (from 1 to 5), for example.

The significance of the jth criterion [9,10,11,12] (its “status importance”), K_j, is determined by some function of the values of the jth criterion in areas D and S. The values of the jth criterion K_j, taking into account the values of $K_j^D$ and $K_j^S$ in the areas D and S, are:

$$K_j={\gamma }_j{F}_j\left(K_j^D, K_j^S\right)$$

(7)

The specific form of the function F_j can be expressed as the difference between $K_j^S$ and $K_j^D$, showing how much and how to improve the situation, or as their quotient, showing how many times it is necessary to improve the situation. The coefficient γ_j is also determined based on the personal experience and knowledge of the leader or expert.

The F_i sequence, being in fact an integral assessment of K_i from the point of view of a manager (DM), gives a series of decreasing criteria importance and, at the same time, shows the expert where to focus attention. The K_i sequence differs significantly from the F_i sequence in that it can be used to reduce uncertainty while reducing the set of criteria in decision making or goal selection.

$$\alpha \left(n\right)=\frac{{{\displaystyle \sum }}_{i=1}^n{K}_i}{{{\displaystyle \sum }}_{i=1}^N{K}_i}$$

(8)

where N is the number of criteria considered, and n is the maximum number of criteria in the reordered sequence that will be taken into account by the manager when making decisions.

We can apply this methodology to solve the problem of ranking and selecting criteria for evaluating the decommissioning of offshore structures (platforms), which are presented in

Table 4. Ranking and selection of criteria for assessing the decommissioning of offshore structures (platforms).

No.	Types of Criteria	${\mathit{K}}_{\mathit{j}}^{\mathit{D}}$	${\mathit{K}}_{\mathit{j}}^{\mathit{S}}$	${\mathit{K}}_{\mathit{j}}^{\mathit{S}}$$-{\mathit{K}}_{\mathit{j}}^{\mathit{D}}$	αj in %	${\mathit{K}}_{\mathit{j}}=( {\mathit{K}}_{\mathit{j}}^{\mathit{S}}$$-{\mathit{K}}_{\mathit{j}}^{\mathit{D}}) {\mathsf{\alpha }}_{\mathit{i}}$	Place of Kj
B1	Economic indicators
C1	Discounted cash flow	3	5	3	5	13	XI
C2	Internal rate of return	3	5	3	5	13	XIII
C3	Payback period	3	5	3	5	13	XIV
C4	Economic limit test	5	9	5	5	23	I
C5	Economic risk	3	5	3	5	13	XV
C6	Profitability index	3	5	3	5	13	XVI
C7	Costs	4	8	4	5	20	III
C8	Funding opportunities	4	7	4	5	18	VIII
C9	Provision of benefits and tax holidays	3	6	3	5	15	IX
B2	Risks
C10	Potential reduction in marine ecosystem sustainability	3	5	3	5	13	XVII
C11	Technical capability (feasibility)	4	8	4	5	20	VI
B3	Environmental and subsoil protection
C12	Air and water pollution	4	8	4	5	20	V
C13	Natural disasters	3	5	3	5	13	XVIII
B4	Organizational, technical, and social indicators
C14	Field viability for 10 years	3	5	3	5	13	XIX
C15	Existing legal framework for implementation	4	7	4	5	18	VII
C16	Degree of interest of participants (investors and authorities)	3	6	3	5	15	X
C17	Safety (including ice conditions)	4	8	4	5	20	IV
C18	Safety of personnel, interests of the local population, etc.	5	9	5	5	23	II
C19	Accessibility to the platform (ocean)	3	5	3	5	13	XX
C20	Possibility of determining the effect of recycling	3	5	3	5	13	XII

.

Let us assume that the conversion of Table 4 indicators from real (dimensional) values to scores using the base scale method has already been completed. In this case, in Table 4, $K_j^D$ is the score of the characteristic of the criterion at the initial stage of decommissioning; $K_j^S$ is the value of the same characteristic taking into account the possible measures taken; and α_j is the relative weight of the criterion according to the decision maker. This table has the following form.

4. ESG Extended Program

Figure 4 shows a schematic overview of the ESG program.

The term ESG was proposed in 2005 with the advent of the Principles for Responsible Investment (PRI), which are based on the notion that environmental, social, and governance issues, such as climate change, human rights, and executive compensation, can affect the performance of investment portfolios and that, in order to ensure a more resilient global financial system, they should be considered in investment decisions along with more traditional financial factors. This term also includes social and corporate initiatives, and all three aspects together can influence the entire organization of an oil and gas company.

In recent years, the environmental, social, and governance goals and programs of ESG, a successor to Corporate Social Responsibility (CSR), have become an important focus for oil and gas companies. Environmental concerns and the resulting drive for an energy transition with the growth of renewable energy sources have taken on greater urgency. The full range of issues addressed by ESG programs is increasingly becoming a priority for industry leaders.

However, within the framework of the ranking model and the selection of criteria for assessing the decommissioning of offshore structures (platforms), the ESG program presented in Figure 4 needs to be expanded to include economic indicators: discounted cash flow, internal rate of return, payback period, economic limit test, economic risk, profitability index, costs, funding opportunities, and provision of benefits and tax holidays; and risks: potential reduction in marine ecosystem sustainability and technical capability (feasibility). These indicators are included in Table 4.

Based on the results of the calculations, this table shows that 15 out of 20 considered criteria are covered; in the table, they are shown by filling and make up 75% of the “importance status” of all criteria. To evaluate the effectiveness of field development options, the decision maker may decide to select a group that has 75% of the “weight” of all criteria or 15 criteria as the main ones. Usually, in practice, a group that has 70–77% of the “weight” of all criteria is used to make a decision.

However, the decision maker can be guided by other considerations [9,10,11,12], for example, the “Pareto principle” (80/20 rule), in which case the number of criteria will be reduced to four: α₄ = 20%. These will be the criteria that won the first three places: economic limit test, safety of personnel, interests of the local population, and costs. The choice of the level of separation always depends on the decision makers—their competence, experience, and intuition—and on the extent to which they consider it appropriate to take into account certain criteria, their completeness, and their informativeness.

5. Analytic Hierarchy Process (AHP) for Multicriteria Risk Assessment of Decommissioning Projects for Offshore Structures (Platforms)

For a clear example, let us consider a scenario where it is known that for the development of some projects for the decommissioning of offshore structures (platforms), three favorable technological options for their implementation are identified by experts (DM) [13], as above: D₁, D₂, D₃. Let us assume that from a variety of gas field development efficiency indicators (

Table 5. Criteria types and options for decommissioning of offshore structures (platforms).

Level 1 (A).	Decommissioning of offshore structures
Level 2 (B).	Criteria
Level 2 (B1).	Economic indicators
Level 2 (B2).	Risks
Level 2 (B3).	Environmental and subsoil protection
Level 2 (B4).	Organizational, technical, and social indicators
Level 3 (C).	Types of criteria
Level 3 (C1).	Discounted cash flow (DCF)
Level 3 (C2).	Internal rate of return (IRR)
Level 3 (C3).	Payback period (PBP)
Level 3 (C4).	Economic limit test (ELT)
Level 3 (C5).	Economic risk
Level 3 (C6).	Costs
Level 3 (C7).	Funding opportunities
Level 3 (C8).	Provision of benefits and tax holidays
Level 3 (C9).	Technical capability (feasibility)
Level 3 (C10).	Air and water pollution
Level 3 (C11).	Existing attractive legal framework for implementation
Level 3 (C12).	Degree of interest of participants (investors and authorities)
Level 3 (C13).	Safety (including ice conditions)
Level 3 (C14).	Safety of personnel, interests of the local population, etc.
Level 3 (C15).	Possibility of determining the effect of recycling
Level 4 (D).	Decommissioning options for offshore structures
Level 4 (D1).	Full decommissioning
Level 4 (D2).	Partial decommissioning
Level 4 (D3).	Leaving in place

), the manager (DM) chose indicators (criteria) that make up 75% of all those considered in this table. The manager grouped these criteria as follows.

The manager will have to choose the best development option [13] from the above set of performance indicators. For a visual solution of the task, we can depict the following hierarchical structure (Figure 5). In Figure 5, the B letters indicate the types of criteria (indicators) of the efficiency of field development, the C letters indicate specific criteria, and the D letters indicate options for the decommissioning of offshore structures.

Let us make a matrix of paired comparisons of criteria types (Table 5), asking the expert comparing them the question: “How much more likely is the realization (updating) of event A (a variant of the criterion) than the realization of event B (another variant of the criterion) if these specific conditions (criteria, indicators, etc., describing this task) are met?” To quantify the degree of what is more likely—the realization of A or B, it is necessary to use the fundamental Saaty rating scale from 1 to 9 of the AHP method [13].

This 1–9 scale looks as follows:

• 1—Equal importance (two compared factors (objects) contribute equally to the objective);
• 3—Somewhat more important (experience and judgment slightly favor one object over the other);
• 5—Much more important (experience and judgment strongly favor one object over the other);
• 7—Very much more important (experience and judgment very strongly favor one object over the other; its importance is demonstrated in practice);
• 9—Absolutely more important (the evidence favoring one object over the other is of the highest possible validity);
• 2, 4, 6, 8—Intermediate values (for example, 2 is of slightly more significant importance).

The matrix of judgments is compiled in such a way that if the priority of the ith object over the jth is bij, then the priority of the jth object over the ith is 1/bij, where bii = 1 and bii ≠ 0.

An example of a matrix of pairwise comparisons according to the criteria of Level 2 of the hierarchy may look as follows (

Table 6. Matrix of pairwise comparisons (Level 2).

A	AB1	AB2	AB3	AB4	∑	πi
AB1	1.0	2.0	1.0	1.0	5.0	0.286
AB2	2.0	1.0	1/2	1.0	4.5	0.257
AB3	1.0	1.0	1.0	1.0	4.0	0.229
AB4	1.0	1.0	1.0	1.0	4.0	0.229
λmax = 4.4; consistency index (CI) = 0.14

).

In this matrix, A is the global criterion (optimum decommissioning of the steel jacket); C1, …, C15 are the types of criteria (Table 5).

In this matrix, π_i is the vector of priorities (subjective probabilities of the realization of options (criteria)); π_i is calculated as the sum of the row elements divided by the sum of all matrix elements; and ${\displaystyle \sum }{\pi }_i=1$.

In order to be able to believe the matrix of pairwise comparisons, it is necessary to calculate its consistency index: CI = (λ_max − n)/(n − 1), where n is the dimension of the matrix of pairwise comparisons, and λ_max is the maximum value of the normalized eigenvector of the pairwise comparison matrix, which is calculated by the following algorithm, which is close to the Saaty algorithm (or it is possible to use the corresponding toolbox of the corresponding mathematical package):

The elements of the first column are added and the result is multiplied by π1;

The elements of the second column are added and the result is multiplied by π2;

The elements of the third column are added and the result is multiplied by π3;

The elements of the fourth column are added and the result is multiplied by π4;

The results obtained are added up and λ_max is obtained, that is:

AB1 = 1.0 + 2.0 + 1.0 + 1.0 = 5.0 × 0.29 = 1.429;

AB2 = 2.0 + 1.0 + 1.0 + 1.0 = 5.0 × 0.26 = 1.286;

AB3 = 1.0 + 0.5 + 1.0 + 1.0 = 3.5 × 0.23 = 0.800;

AB4 = 1.0 + 1.0 + 1.0 + 1.0 = 4.0 × 0.23 = 0.914;

λ_max = 1.429 + 1.286 + 0.800 + 0.914 = 4.429, approximately 5 (due to the approximation of the Saaty algorithm), which corresponds to an absolutely consistent skew-symmetric matrix: CI = (4.429 − 4)/(4 − 1) = 0.14.

The matrix of Table 6 can be considered consistent, since the pairwise comparison matrix (skew-symmetric matrix) is considered to be absolutely consistent if CI = 0–0.15 [9]. Inconsistency occurs when an expert, for example, when comparing three objects, writes down P1 = 6P3, P2 = 4P3, i.e., P1 = 20P3, but P1 = 6P3 is written down, which violates the transitivity of estimates.

Next, we compile four matrices of pairwise comparisons (expert judgments), prioritizing options in relation to each criterion (

Table 7. Matrix of pairwise comparisons (Level 3).

B1	B1C1	B1C2	B1C3	B1C4	B1C5	B1C6	B1C7	B1C8	∑	πi
B1C1	1.0	2.0	3.0	2.0	1.0	1.0	1.0	2.0	13.0	0.171
B1C2	0.5	1.0	1.0	1.0	2.0	2.0	2.0	3.0	12.5	0.164
B1C3	0.3	1.0	1.0	2.0	1.0	2.0	2.0	1.0	10.3	0.136
B1C4	0.5	1.0	0.5	1.0	3.0	3.0	1.0	2.0	12.0	0.158
B1C5	1.0	0.5	1.0	0.3	1.0	1.0	1.0	2.0	7.8	0.103
B1C6	0.5	0.5	0.5	0.3	1.0	1.0	2.0	1.0	6.8	0.090
B1C7	1.0	1.0	0.3	1.0	0.5	1.0	1.0	1.0	6.8	0.090
B1C8	1.0	0.5	1.0	1.0	0.3	1.0	1.0	1.0	6.8	0.090
λmax = 9.0; CI = 0.14

,

Table 8. Matrix of pairwise comparisons (Level 3).

B2	B2C9	∑	πi
B2C9	1.0	1.0	1.000
λmax = 1.0; CI = 0.0

,

Table 9. Matrix of pairwise comparisons (Level 3).

B3	B3C10	∑	πi
B3C10	1.0	1.0	1.000
λmax = 1.0; CI = 0.0

and

Table 10. Matrix of pairwise comparisons (Level 3).

B4	B4C11	B4C12	B4C13	B4C14	B4C15	∑	πi
B4C11	1.0	2.0	3.0	2.0	1.0	9.0	0.303
B4C12	0.5	1.0	1.0	2.0	2.0	6.5	0.219
B4C13	0.3	2.0	1.0	0.5	1.0	4.8	0.163
B4C14	0.5	0.5	2.0	1.0	1.0	5.0	0.169
B4C15	1.0	0.3	1.0	1.0	1.0	4.3	0.146
λmax = 5.6; CI = 0.14

).

Let us compile a matrix of priorities (

Table 11. Matrix of priorities (C).

C	B1	B2	B3	B4				D
C1	0.171	0.000	0.000	0.000	X		=	0.049
C2	0.164	0.000	0.000	0.000				0.047
C3	0.136	0.000	0.000	0.000				0.039
C4	0.158	0.000	0.000	0.000				0.045
C5	0.103	0.000	0.000	0.000				0.029
C6	0.090	0.000	0.000	0.000		0.286		0.026
C7	0.090	0.000	0.000	0.000		0.257		0.026
C8	0.090	0.000	0.000	0.000		0.229		0.026
C9	0.000	1.000	0.000	0.000		0.229		0.257
C10	0.000	0.000	1.000	0.000				0.229
C11	0.000	0.000	0.000	0.303				0.069
C12	0.000	0.000	0.000	0.219				0.050
C13	0.000	0.000	0.000	0.163				0.037
C14	0.000	0.000	0.000	0.169				0.039
C15	0.000	0.000	0.000	0.146				0.033

) from the vectors of criteria priority columns in relation to their own type of criteria and multiply it on the right by the vector column of priorities of criteria types in front of each other in relation to the main (complex) criterion D.

Next, 15 matrices of judgments are compiled (

Table 12. Matrix of judgments.

C1-8	CnD1	CnD2	CnD3	∑	πi
CnD1	1.0	2.0	3.0	6.0	0.529
CnD2	0.5	1.0	2.0	3.5	0.309
CnD3	0.3	0.5	1.0	1.8	0.162
λmax = 3.0; CI = 0.01

,

Table 13. Matrix of judgments.

C9	CnD1	CnD2	CnD3	∑	πi
CnD1	1.0	2.0	2.0	5.0	0.441
CnD2	0.5	1.0	3.0	4.5	0.397
CnD3	0.5	0.3	1.0	1.8	0.162
λmax = 3.2; CI = 0.09

,

Table 14. Matrix of judgments.

C10	CnD1	CnD2	CnD3	∑	πi
CnD1	1.0	3.0	2.0	6.0	0.529
CnD2	0.3	1.0	2.0	3.3	0.294
CnD3	0.5	0.5	1.0	2.0	0.176
λmax = 3.2; CI = 0.09

and

Table 15. Matrix of judgments.

C11-15	CnD1	CnD2	CnD3	∑	πi
CnD1	1.0	2.0	2.0	5.0	0.441
CnD2	0.5	1.0	3.0	4.5	0.397
CnD3	0.3	0.5	1.0	1.8	0.162
λmax = 3.2; CI = 0.08

).

Next, the corresponding matrix of priorities is compiled, which is calculated by multiplication; then, on the right, it is multiplied by the vector column of priorities of the criteria C1–C15 to each other, obtained in the previous step, which allows us to calculate the priorities of development options.

This matrix of priorities, as a result, will have the following form (

Table 16. Matrix of priorities (D).

D	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12	C13	C14	C15
D1	0.5	0.5	0.5	0.5	0.5	0.5	0.5	0.5	0.4	0.5	0.4	0.4	0.4	0.4	0.4
D2	0.3	0.3	0.3	0.3	0.3	0.3	0.3	0.3	0.4	0.3	0.4	0.4	0.4	0.4	0.4
D3	0.2	0.2	0.2	0.2	0.2	0.2	0.2	0.2	0.2	0.2	0.2	0.2	0.2	0.2	0.2

).

This matrix is multiplied on the right by the vector column of criteria priorities.

The first element of the priority vector D₁, D₂, D₃ is obtained as:

D₁: 0.529 × 0.049 + 0.529 × 0.047 + 0.529 × 0.039 + 0.529 × 0.045 + 0.529 × 0.029 + 0.529 × 0.026
+ 0.529 × 0.026 + 0.529 × 0.026+ 0.441 × 0.257 + 0.529 × 0.229 + 0.441 × 0.069 + 0.441 × 0.050
+ 0.441 × 0.037 + 0.441 × 0.039 + 0.441 × 0.033 = 0.487.                

D₂: 0.309 × 0.049 + 0.309 × 0.047 + 0.309 × 0.039 + 0.309 × 0.045 + 0.309 × 0.029 + 0.309 × 0.026
+ 0.309 × 0.026 + 0.309 × 0.026+ 0.397 × 0.257 + 0.294 × 0.229 + 0.397 × 0.069 + 0.397 × 0.050
+ 0.397 × 0.037 + 0.397 × 0.039 + 0.397 × 0.033 = 0.348.                

D₃: 0.162 × 0.049 + 0.162 × 0.047 + 0.162 × 0.039 + 0.162 × 0.045 + 0.162 × 0.029 + 0.162 × 0.026
+ 0.162 × 0.026 + 0.162 × 0.026+ 0.162 × 0.257 + 0.176 × 0.229 + 0.162 × 0.069 + 0.162 x 0.050
+ 0.162 × 0.037 + 0.162 × 0.039 + 0.162 × 0.033 = 0.165.                

A_D1 = 0.487; A_D2 = 0.348; A_D3 = 0.165.                        

The analysis shows that the most preferable option is the elimination of D₁. Accordingly, the risk of non-fulfillment of this version of the project (its multicriteria assessment) is P_D1 = 1 − 0.487.

In addition to the models of risk assessment in the narrow sense discussed above—the risk assessment of projects and assessment of the decommissioning of offshore structures (platforms)—knowledge of risks is often required in a broad sense, that is, taking into account all kinds of losses (most often financial), especially when assessing technological, organizational, environmental, social, economic, and other risks. The theory that makes it possible to do this is called the move by nature (statistical decisions).

6. Conclusion and Policy Implications

There is a significant arsenal of methods to effectively support abandonment decisions under conditions of uncertainty, many of which are successfully used in decision making related to the abandonment of wells and the decommissioning of fields. These methods are divided into quantitative, qualitative, and combined methods, with their numerous varieties, advantages, and disadvantages, and the corresponding area of use.

To select the best options for the decommissioning of offshore oil platforms, it is proposed to use the AHP, which includes mathematical modeling methods used to solve complex decision-making problems to determine the best alternatives or course of action.

The use of AHP presents a suitable method for analyzing multiple-criteria decisions for the problem under consideration due to its well-structured procedure and ease of use. This method minimizes the subjectivity of human judgment.

The use of the AHP method with variable weights of criteria provides useful information on the implications of platform decommissioning for any option being evaluated; it can help decision makers determine the best decommissioning option.

In addition, the results show that stakeholder preferences regarding which decision criteria are most important can change the outcome of the assessment of decommissioning options.

The analysis performed in previous studies [1,8] suggests that partial removal of the steel jacket from the Hermosa platform is the best option for its decommissioning, regardless of stakeholder preferences. However, this result is specific to the California offshore zone and may be different for other object locations, such as the United Kingdom Continental Shelf in particular.

To conduct a multicriteria analysis, it was proposed for the first time to use a four-level hierarchical structure within the framework of the ranking model and selection of criteria for assessing the decommissioning of offshore structures (platforms) taking into account environmental, social, and corporate governance (ESG).