1. Introduction
Desalination has rapidly expanded and evolved into a vitally important water source over many regions of the world. With this exponential regional growth of desalination comes significant economic, environmental and social impacts (or sustainability pillars) [
1]. As a consequence, there is the increasing recognition by global, national, regional and institutional entities that the sustainability pillars should be applied to large physical and social development endeavours with desalination as being no exception [
2].
Based on the report of the World Commission on Environment and Development, sustainability can be defined as “development that meets the needs of the present without compromising the ability of future generations to meet their own needs” [
3]. From this report, wider sustainability definitions have been developed, which include the triple-bottom-line concept covering environmental, social and economic factors [
4]. This approach considers each of the pillars of equal importance in the decision-making process [
5,
6,
7,
8,
9]. In fact, up to seven sustainability pillars can be considered depending on the analysis and context in which the pillars are used [
6,
10].
Desalination is an energy-intensive process with various environmental, economic and social impacts. Therefore, desalination should be assessed according to its sustainability on these factors. However, the most commonly used sustainability framework generally focuses on one of the above-mentioned aspects [
11]; for example, the well-established environmental life cycle analysis [
12]. However, since the life cycle analysis does face limitations and model uncertainties [
12,
13], it ought to be integrated with a social-economic analysis for a better representation of sustainability [
14]. Nevertheless, these types of integrated research are limited in the literature [
11]. As an illustration, in research, the life cycle analysis was integrated with a water cost study focusing on a sea water reverse osmosis (SWRO) plant in Perth, Australia [
15]. Nonetheless, this form of economic and environmental integration is more common than the integration of environmental and social analysis, as the corresponding evaluation tends to be more complicated and complex [
11]. For example, Afgan et al. [
16] considered four desalination options based on a Decision Support System Shell. However, there was no consideration as to the social factors and impacts of desalination. Therefore, desalination assessments may only incorporate a few economic parameters [
15,
17], or just provide a detailed cost–benefit analysis [
18,
19], all lacking the integration of all the three environmental, economic and social sustainability factors. Thus, it is recommended to apply an integrated sustainability assessment framework to consider all factors of desalination. This allows for the assessment of current or emerging desalination technologies for their relative sustainability [
11]. Furthermore, Ibrahim et al. [
11] suggested an integrated universal support framework incorporating several sustainability factors as subsets of four main sustainability components, including environmental, economic, social and technical factors.
A methodology for the evaluation of the sustainability pillars was proposed by Lior [
2]; the method includes a relatively straightforward sustainability analysis of reverse osmosis (RO) desalination plants including a small number of calculation metrics. The methodology included equations for the formulation of a composite sustainability index in relation to relevant design and operational parameters. In this case, the methodology allows for a mathematical analysis including optimization and sensitivity evaluations. The method included the selection and calculation of metrics, in addition to weighting and aggregation, creating a sustainability indicator through sensitivity analysis between the choice of weights and the combined environmental and social impact factors [
1].
For most regions, an environmental impact assessment is generally a legislative requirement prior to construction and operation of a proposed desalination plant. Fuentes-Bargues [
20] presented an extensive environmental impact assessment on desalination works. Decision-makers tend to favour multi-criteria decision analysis methodologies such as variations of fuzzy logic applications and existing matrix mathematics, as these support research into the feasibility of new desalination projects. Common examples have been described [
21,
22,
23,
24,
25]. Various indicators covered in these studies are based on economic, environmental/topographical, technological and social factors. Though, modifying existing operating plants to perform better in all facets of sustainability is imperative. Nonetheless, many desalination plants are still cutting costs to the detriment of environmental health, unless the practices are prohibited under reinforced and effective governmental water management legislation. Regional government agencies (e.g., the Ministry of Environment and Water in the UAE) can provide incentives for existing plants to adopt new sustainable practices due to research advancements. Additionally, regulatory bodies could encourage and adopt an effective ranking system to either buy water subsidies for founding a sustainable supply chain or publish information related to sustainable performance of works [
26]. This will provide incentives to plant operators to enhance sustainable works’ efficiency.
A ranking system such as the multi-criteria approach is a method that creates a list of sustainability indicators derived from a corresponding assessment. For example, Afgan and Darwish [
16] ranked various desalination technologies with indicators in relation to economic influences, while considering fossil energy consumption to demonstrate the need of using this energy source sparingly. Chang [
27] adopted an ecological indicator method to a seawater desalination bed to evaluate the impacts of this water production system within a vulnerable marine ecosystem. To improve the current practise of life-cycle assessments for environmental evaluation of desalination works, Zhou et al. [
28] evaluated 30 desalination reports. They concluded that life-cycle assessments, in terms of their feasibility and reliability, contribute to the uncertainty in the evaluation outcome.
The selection of sustainability indicators as they relate to context is not considered to have a dynamic nature, i.e., an indicator framework is fixed, regardless of if the individual criterion fits the boundary conditions of plant operation. Therefore, it is essential to adopt a modelling framework that permits a certain degree of adjustment. This can close the gap between model developer and practitioner. This level of flexibility will be more appealing to decision-makers. Artificial intelligence (AI) techniques have been adopted to support the sustainability concept within technical issues. Among the several approaches to adopt AI techniques for system evaluation, artificial neural networks (ANN) and fuzzy logic (FL) systems have been prominently adopted in the area of sustainability [
24,
29]. For example, neural networks have been utilized by Abdeljawad et al. [
30] to forecast key water parameters such as salt concentration to evaluate reverse osmosis plant performance along the Gaza Strip, Palestine. Additionally, an ANN has been used by Mashaly et al. [
29] for assessing and optimizing solar performance under hyper arid environments. Furthermore, Kant and Sangwan [
31] developed models utilizing ANN and support vector regression (SVR) methods to evaluate power consumption. During the model validation, it was found that the ANN yielded better results than the SVR model, emphasising the advantage of using ANN.
Gauging sustainability includes overcoming the barrier of being able to convert a holistic and intangible component to something that is quantifiable and tangible. Therefore, fuzzy logic is more predominant within multi-criteria decision-making tools as well as decision support systems that measure progress in terms of sustainability. Fuzzy logic supersedes these barriers by implementing the use of fuzzy set theory for the precise reason that this methodology is widespread among decision-makers for the development of models [
32].
Gagliardi et al. [
33] outlined a model to determine city sustainability with regard to urban planning, using a weighted fuzzy logic approach. Ghadimi et al. [
34] also applied a comparable method to evaluate sustainability from start to finish. The benefit of these investigations was that expert knowledge is communicated within the evaluation framework, allowing for logical changes that the system may require to achieve enhanced performance in all aspects of sustainability. Ghassemi and Danesh [
24] assessed the performance of desalination units compared to a set of indicators that was classified into environmental, technical and economical components using a hybrid-fuzzy multi-criteria decision analysis method. Reverse osmosis technologies were superior compared to multi-stage flash distillation plants, predominantly due to their compliance to renewable energy techniques.
As a ranking system necessitates evaluation by experienced practitioners in the sector for which the model is utilised, the techniques in the aforementioned research required quantification using sustainability performance figures, which are frequently problematic to attain. Fuzzy logic overcomes this barrier and permits linguistic evaluation, allowing contributions from various experts. To implement the desired ranking technique, methodologies should provide flexibility concerning the applied indicator set, recognizing the uncertainty in data and mimicking the human cognitive ability to allocate scores to the works under assessment. This is essential for a holistic sustainability assessment, with its associated wide-ranging influences.
Therefore, the aim of this study is to score the sustainable performance of desalination plants utilizing a fuzzy logic ranking framework based on a holistic indicator set that captures the performance of the plants. The current research features two UAE desalination plants as case studies. The modelling framework has been structured to exclusively score the performance of operational desalination plants.
2. Study Area and Case Studies
The Gulf Cooperation Council (GCC) countries and Yemen cover an area of approximately 2.8 million km
2 consisting primarily of arid and desert landscapes. On average, the precipitation is less than 100 mm/year in this region and, as a consequence, surface water sources are scarce [
35]. This problem is exacerbated by the deep groundwater levels in the region (i.e., the amount of groundwater removal exceeding natural inflow), which makes the water scarcity values to be estimated at <500 m
3/capita/year [
36]. Furthermore, water consumption in the region is one of the highest in the world [
37], with an average water consumption rate of between 300 and 759 L per person per capita compared to the USA and China, with 580 L and 90 L per person per capita, respectively [
38]. Additionally, water is highly subsidised in the Gulf region, with some consumers paying less than 5% of water production in some countries [
39]. Consequently, significant backing has been given to the construction of desalination infrastructure to battle the increasing water demand in the Gulf region. For example, the UAE is currently one of the largest desalinated water-producing countries in the world, with a capacity at about 1776 million m
3/year, despite its relatively small landmass and population [
40]. Seawater desalination remains one of the most reliable alternative sources of water in the Gulf region since its inception in the 1950s [
41].
Desalinated water technologies continue to gain momentum in the Gulf region due to emerging technology and innovative research that allows for the development of more energy-efficient plants with lower operational costs. For example, research and development has reduced the excessive capital costs associated with plant construction, contributing to the reduction of the overall unit cost of desalinated water [
42]. Nonetheless, the by-product of the desalination process is often criticized for its adverse environmental marine impacts in addition to its energy-intensive methods, both in terms of construction and operation [
43].
The desalination industry in the Gulf is commanded by two important factors; namely, by the presence of already established plants, and the proposal of new desalination works projects. While there is considerable research and literature referencing the selection process of appropriate desalination technologies, under a certain set of boundary conditions or sustainability assessments of several desalination processes [
44], there are few reports concerning the current operation of desalination plants. Therefore, it is important to consider and adopt a sustainable assessment methodology to assess and score the performance of existing desalination plants.
Desalination plants are assets requiring a considerable investment, which is recovered over a long timeframe. Therefore, the current scenario in the GCC foresees existing plants whose design was carried out based on now-obsolete energy, environmental requirements and the need for new plants. At the same time, the desalination technology is rapidly changing, with more energy-efficient and environmentally friendly processes being designed.
In this research, two generic and randomly selected anonymous UAE case studies have been utilized to demonstrate the model. Plant X adopted multi-stage flash distillation fuelled by natural gas from a cogeneration power plant and plant Y adopted reverse osmosis fuelled by natural gas obtained from a power plant.
6. Discussion
This section aims to discuss the potential of using fuzzy logic in sustainability assessment based on the results of the present study and related published work. In the last decade, several researchers attempted to employ fuzzy logic for this purpose. In a study conducted by Phillis and Andriantiatsaholiniaina [
68], fuzzy logic operations were represented as powerful tools for compensating the lack of full knowledge in existing methods of sustainability measurement.
Other related studies emphasized the necessity of applying fuzzy propositions as capable alternatives for evaluating strong sustainability of ecosystems [
69,
70], ecology [
71], and environmental systems [
72]. The use of fuzzy logic in assessment of industrial sustainability, which is a very challenging task, has also been reported in the literature [
73,
74]. Moreover, in light of the current topic of this paper, several publications have studied the fuzzy-based approaches for evaluating sustainability in various areas of hydraulics and hydrosiences such as soil-water interaction [
75], self-purifying capacity of rivers [
76] and groundwater [
77]. The findings of the study show that the fuzzy-based models can considerably help decision-makers in their assessment of the intended problem.
In the current study, we proposed a methodology for sustainability ranking of desalination plants using the Mamdani Fuzzy Logic Inference Systems. In the developed model, the various rule bases used for the individual fuzzy inference system model adopt the linguistic individual score, integrates it and subsequently transforms it to a more adequate form of holistic quantified information utilized for ranking. If the input values for ‘economic’, ‘environment’ and ‘society’ increase, the corresponding output ‘ranking’ also rises, as seen in
Figure 8. Furthermore, depending on the preference of the user, the proposed model can easily be made more complex by adding weights to the selected criteria depending on the specific case study context.
Several studies [
16,
24] indicate superior sustainability performance of reverse osmosis processes over multi-stage flash distillation. However, it should be taken into consideration that the corresponding research is based on a designed framework to adopt the most appropriate desalination units for a specific context. In contrast, the current research proposed a model software structure to judge the sustainability performance of current desalination plants.
The validation of the findings overcome the barriers to accomplish sustainability in desalination works. In the perspective of the UAE case study country, the outcomes promote a required step change of the desalination market by promoting additional membrane separation techniques, which currently constitute, for example, only 12% of the desalination plants in the UAE.
The methodology described in this study can provide decision-makers with a tool to derive information from dissimilar databases. The fuzzy ranking framework amalgamates various expert knowledge to yield a single indicator used for sustainability ranking for easy evaluation between the proposed planning scenarios.
As a result, the decision-makers can find areas to be improved by investments to promote the sustainability of the desalination plant. The surface views of the model provide an easy tool to understand how the single ranking was derived.
The proposed model was validated using data available in the literature for the UAE. However, validating the model with real data would enrich the outcome. To demonstrate the superiority of this approach, the proposed methodology should be repeated for other regions to take into consideration other local factors as it is expected that sustainability values given by experts vary for other countries and hence may change the overall sustainability rank.