1. Introduction

In the coming decades, the energy demand is expected to grow rapidly (Langè et al., 2015). Natural gas is a prime, clean, safe, and most useful form of energy (Kazemi et al., 2016), which has been widely used as an industrial and indoor fuel. There are impurities in natural gas which cause serious problems such as fouling and corrosion in pipelines, freezing, plugging, and erosion (Al-Lagtah et al., 2015; Gutierrez et al., 2016; Muhammad and GadelHak, 2015; Qiu et al., 2014). H₂S and CO₂ are the major impurities in natural gas which must be removed to bring natural gas to the market (Alhseinat et al., 2015; Cho et al., 2015). Chemical absorption by alkanol amines, bio scrubber, adsorption, membrane, and the oxidation of H₂S in an iron-chelate process have been used for the removal of these contaminates over years (Ameli Forough, 2012; Tavan et al., 2016). Amine absorption process is the most commonly used acid gas removal technology (Niu and Rangaiah, 2014). This process allows acidic gases to be dissolved in a solvent and released by regeneration at a later stage (Devold, 2013; Jassim, 2016). Amines are some compounds of ammonia (NH₃) which have replaced one or more of their hydrogen atoms with hydrocarbon groups (Arthur Kidnay, 2006). Depending on the composition and operation conditions of the raw gas streams to be treated, aqueous amines are used to meet the sweet gas specifications (Vahid Abkhiz, 2014).

Furthermore, the amines capacity for the absorption of acidic gases is an important characteristic (Momeni and Riahi, 2014). Mono-ethanol amine (MEA), di-glycol amine (DGA), di-ethanol amine (DEA), di-isopropanol amine (DIPA), and methyl di-ethanol amine (MDEA) are commonly used in the natural gas sweetening process (Shokouhi et al., 2015).

The selection of a suitable amine solution for sweetening depends on the feed condition and process objectives (Zahid et al., 2017). It can reduce capital and operating costs while providing more flexibility in achieving specific purity requirements. The specification of the required purity with respect to acidic gases (CO₂ and H₂S), equipment sizing, and operating costs are the primary concern in the gas sweetening. Amine circulation rate, reboiler/condenser size and duty, and corrosion problems are important factors to be considered in the selection of the proper amine. In the amine sweetening unit, 50 to 70% of the initial investment is related to the magnitude of the solvent circulation rate, and 10 to 20% of the initial investment is dependent on the regeneration energy requirement. In addition, about 70% of the operating costs, excluding labor, result from regeneration. Since the selection of the proper amine can greatly reduce both the regeneration energy requirement and solution circulation rate, choosing the amine or a combination of amines best suited to the conditions can have a dramatic impact on the overall costs associated with a sweetening unit (Astarita, 1983; John Polasek, 2006).

The selection of amine solvent in the process of gas treating is considered to be an important decision in gas treating companies. To this end, in the current work, MADM techniques were used for amine solvent selection for the first time. In the next section, a summary of MADM relevant literature is described.

2. Relevant literature

Multi-criteria decision-making (MCDM) is a discipline aimed at supporting decision makers which are faced with numerous and conflicting alternatives to make an optimal decision. To achieve this purpose, two critical questions should be unlocked: preference structure and weights. MCDM methods have been considered in many fields, some of which include economic, social, medical, and technical issues (Nădăban et al., 2016). To facilitate systematic research in the field of MCDM, Hwang and Yoon suggested that MCDM problems can be classified into two main categories, namely multiple attribute decision making (MADM) and multiple objective decision making (MODM), based on the different purposes and different data types (Gwo-Hshiung Tzeng and Huang, 2011). Different methods of decision analysis are illustrated in Figure 1, among which TOPSIS and AHP methods are highlighted.

Figure 1

Decision analysis methods.

The TOPSIS method is currently one of the most frequently used methods for MCDM. Its original version was intended for dealing with real-valued data. The Technique for Order Preferences by Similarity to an Ideal Solution (TOPSIS) method was proposed by Hwang and Yoon in 1981. The main idea came from the concept of the compromise solution to choose the best alternative nearest to the positive ideal solution (optimal solution) and the farthest alternative from the negative ideal solution (inferior solution); then, the best one of sorting is chosen, which will be the best alternative (Hwang and Yoon, 1981). As a practical application of this method, we can address the works of Ansarifar et al.(2015), Barros and Wanke (2015), and Chen et al. (2015).

A new interval type-2 fuzzy multiple-attribute decision making model is developed using TOPSIS and decision making trial and evaluation laboratory (DEMATEL) by Baykasoglu et al. (Baykasolu and Glck, 2017), Hatami et al. (Hatami-Marbini and Kangi, 2017), and Runik et al. (Rudnik and Kacprzak, 2017). They have presented a general overview about the development of fuzzy TOPSIS methods. Nadaban et al. have also mentioned several works presenting some applications of fuzzy TOPSIS such as location problem, supplier selection, and sustainable and renewable energy (Nădăban et al., 2016).

In this study, for deciding on the amine solvent selection, a systematic decision-making approach was proposed. The proposed approach will be explained in the next section.

3. The proposed approach

Hierarchical structure creation, calculation of criteria weights by AHP, and selection by TOPSIS are three phases of the selection of the amine solvent. The different phases of the proposed approach are presented in Figure 2.

In the first phase, criteria and alternatives are identified using technical information gathered from Ilam Gas Treating Company; criteria are divided into two categories: physical and process, and twelve physical indexes and nine process indexes are detected. The physical criteria include 1-molecular weight, 2-boiling point, 3-freezing point, 4-critical constants, 5-density, 6-weight, 7-relative density, 8-specific heat, 9-thermal conductivity, 10-latent heat of vaporization, 11-flash point, and 12-toxicity. The process criteria are 1-acid gas pickup (m³(GPA)/L at 38 °C), 2-acid gas pickup (mol/mol amine), 3-lean solution residual acid gas, 4-rich solution acid gas loading, 5-solution concentration, 6-approximate reboiler heat duty, 7-reboiler temperature, normal operating, 8-heats of reaction, and 9-absorption. MEA, DEA, DIPA, DGA, and MDEA are intended as alternatives. The decision hierarchical structure is depicted in Figure 3.

Figure 2

The proposed approach.

Figure 3

Decision hierarchical structure.

In the second phase, the weights of the criteria are calculated by the analytic hierarchy process (AHP), developed by Thomas L. Saaty in 1980. An advantage of the AHP is that it is designed to handle situations in which the subjective judgments of individuals constitute an important part of the decision process. Basically, the AHP is a method of (1) breaking down a complex, unstructured situation into its component parts; (2) arranging these parts, or variables into a hierarchic order; (3) assigning numerical values to subjective judgments on the relative importance of each variable; and (4) synthesizing the judgments to determine which variables have the highest priority and should be acted upon to influence the outcome of the situation. The major steps of AHP include:

1. developing a graphical representation of the problem in terms of the overall goal, the criteria, and the decision alternatives;
2. specifying his/her judgments about the relative importance of each criterion in terms of its contribution to the achievement of the overall goal;
3. indicating a preference or priority for each decision alternative in terms of how it contributes to each criterion;
4. using a mathematical process to synthesize the information (including consistency checking) and provide a priority ranking of all the alternatives and criteria, given the information on relative importance and preferences.

Finally, in the third phase, alternative ranking is performed. Due to the different quantities of criteria (crisp and interval data), ranking procedure is performed twice. At first, amine solvents are ranked by the traditional TOPSIS method with the crisp data, and then the alternatives are ranked by the modified TOPSIS method with the interval data; ranking average is considered as the final ranking. In the following, the traditional and modified TOPSIS methods are presented.

3.1. The traditional TOPSIS method with the crisp data

Suppose that there are n criteria (C₁, C₂, …, C_n) which are taken into consideration among m alternatives (A₁, A₂, …, A_n) in an MCDM problem. Let  be the performance of the alternative i at the criterion j. The basic principle of the TOPSIS method is that the best alternative should have the shortest distance from the ideal solution and the farthest distance from the negative-ideal solution. Its procedure is given below:

1. Calculate the normalized value .

1. Calculate the weighted normalized value .

where,  is the weight of the criterion.

1. Determine the positive-ideal solution  and the negative-ideal solution .

where, S_B and S_C denote the set of benefit criteria and the set of cost criteria respectively.

1. Calculate Euclidean distance.

1. Calculate the relative closeness to the ideal solution.

1. Rank the preference order.

For ranking, we can choose the best alternative with the maximum value of the relative closeness. According to the other works (Khezerloo et al., 2011; Levin, 2004; Sevastianov, 2007), the modified TOPSIS method with the interval data is described.

3.2. Modified TOPSIS method with interval data

1. Calculate the normalized interval value as given below:

1. Calculate the weighted normalized interval data  defined by:

1. Identify the positive ideal solution and the negative ideal solution as follows:

1. Calculate the Euclidean distance.

1. Calculate the relative closeness Cl_i as reads:

4. Case Study: Ilam Gas Treating Company

The amine solvent selection in Ilam Gas Treating Company (IGTC) is performed using the proposed approach. The flow sheet of the plant is presented in Figure 4, and the composition and feed condition of the plant is tabulated in Table 1

Figure 4

Amine gas sweetening flow sheet.

Table 1

Amine gas sweetening feed composition and conditions.

The crisp and interval data about the physical and process criteria are listed in Tables 2 and 3.

Table 2

Physical criteria of the crisp data.

Table 3

Process criteria of the interval data.

The normalized values of the data are calculated in Tables 4 and 5.

Table 4

Normalized value of the crisp data.

Table 5

Normalized value of the interval data.

Physical index weights are determined by using the AHP method as follows:

Experts proposed process index weights as follows:

The positive-ideal solution A⁺ and the negative-ideal solution A^- are tabulated in Tables 6 and 7.

Table 6

A⁺ and A^- of the physical criteria.

Table 7

A⁺ and A^- of the process criteria.

The Euclidean distance, the relative closeness to the ideal solution, and the rank of the preference order for the physical criteria are presented in Table 8.

Table 8

Rank the preference order of physical criteria.

The Euclidean distance and the relative closeness to the ideal solution for the process criteria are summarized in Tables 9 and 10.

Table 9

Euclidean distance .