Assessment of Sustainable Maintenance Strategy for Manufacturing Industry

Abstract

This study creates a framework to aid in the sustainability of maintenance strategies. The framework was created using expertise from the industry and academia. Using this knowledge, three multi-criteria tools were chosen for the maintenance strategies evaluation. The tools include grey relational analysis (GRA) techniques, additive ratio assessment (ARAS), and step-wise weight assessment ratio analysis (SWARA). In a production system, they were used to assess four planned maintenance strategies. The strategies are periodic maintenance (S1), meter-based maintenance (S2), predictive maintenance (S3) and prescriptive maintenance (S4). The ARAS approach was used to obtain the strategy rating for the various requirements. This study used the SWARA method to determine the requirements’ importance using an intuitionistic fuzzy triangular number. The ARAS results were combined using the GRA method. This study observed that the criteria utilised to choose a maintenance strategy for equipment depend on the information collected from six specialists in a manufacturing organisation. For instance, it was discovered that S3 was the maintenance approach that best suited the system’s technical needs. At the same time, S2 was found to be less effective. The economic needs analysis showed that S1 is the maintenance strategy that is most appropriate for the system, while S3 is the least appropriate. S1 is the most appropriate maintenance method for the system, given the social requirements, whereas S2 is the least effective. According to the results of the environmental requirements, S2 is the best maintenance plan for the system, while S4 is the worst. According to the GRA approach, the system’s best and least appropriate maintenance strategies are S2 and S4, respectively.

1. Introduction

Since many variables determine whether maintenance planning is successful, it is a complicated undertaking. To modify and capture the requirements of their short- and medium-term maintenance policies, stakeholders, are, therefore, continuously examining their maintenance strategy. The burden of maintenance managers increased due to this necessity as they look for models to help them make wise decisions regarding their systems. Hence, manufacturing industries continually review maintenance strategies to meet daily operational needs. A manufacturing system’s effectiveness is decided by its design’s calibre and a suitable maintenance plan to prevent failure and collapse. Production costs include a sizable amount of maintenance costs, and this percentage has rapidly increased as automation, and sophisticated technologies have improved. A systematic maintenance strategy reduces equipment failures and prevents expensive production shutdowns [1]. As a result, maintenance practices are becoming more popular in the manufacturing sector, emphasising system dependability and availability and lengthening the usable life of the equipment. Any asset’s life cycle includes a key stage called maintenance. It is essential to achieving an organisation’s long-term objectives.

Due to its effects on performance, overall cost-effectiveness, quality, availability, on-time performance, safety, and environmental requirements, maintenance has gained importance. Some of the maintenance terminologies used in the literature to describe maintenance strategies include approaches, techniques, tasks, policies, methodologies, and concepts. It is crucial to pick a maintenance strategy to meet environmental and safety standards, increase availability, and extend asset life while reducing expensive breakdowns. Facilities and equipment are now more complicated and dependable because of the rapid development of new technology. Preventive maintenance and condition-based maintenance strategies replace the old maintenance philosophy of breakdown and scheduled maintenance [2]. It is critical to have a planned framework for optimising and prioritising maintenance activities because it is estimated that between 18 and 30 per cent of additional maintenance costs are incurred owing to inadequate maintenance planning [3,4].

To achieve operational and business objectives, maintenance which is a method for boosting the availability and dependability of a company’s assets, is inevitable [5]. Maintenance can be carried out in various methods depending on several variables, including cost, danger, and the seriousness of consequences. The three main categories of maintenance strategies currently employed in the building maintenance sector are corrective maintenance [6,7,8], preventive maintenance [9,10,11], and predictive maintenance [12].

Preventive maintenance is an effective technique to decrease or eliminate equipment failure and boost reliability. With today’s manufacturing systems becoming more sophisticated than previous ones, much research has been conducted on preventive maintenance. The results imply that system maintenance is necessary to prevent equipment failure. Performing preventative maintenance on equipment at integer multiples for a predetermined period is known as periodic maintenance, sometimes known as time-based maintenance [11]. According to a usage-based maintenance mode that some academics proposed, equipment maintenance decisions should incorporate monitoring technologies and consider the possibility that the system would occasionally be inactive or only partially operational. Studies on time-based or usage-based maintenance have only briefly examined a few prognostic systems.

In unexpected circumstances, developing a maintenance schedule well in advance could be challenging. As a result, most equipment may be maintained even when it still has a significant amount of useful life left, incurring high maintenance costs. As a result, several preventative maintenance strategies fall short of the fundamental operational requirements of contemporary business. Therefore, this study proposed a framework incorporating sustainability indicators for planned maintenance strategy selection.

2. Methodology

The advantages of employing an equipment maintenance strategy are considered when considering equipment’s technical, environmental, and social sustainability requirements. From those mentioned above, this study analyses maintenance strategies from techno-economic, social-economic, and environmental viewpoints (Figure 1). The cost of implementing an equipment maintenance strategy is determined using sustainability standards for equipment. This issue opened up a possibility for applying the SWARA, ARAS, and GRA methods to rank the equipment maintenance strategies from various angles. The following nomenclatures are used to describe the parameters in the conceptual framework.

2.1. Interval-Valued Triangular Fuzzy Numbers (IVTFN)

Ref. [13] proposed the fuzzy number set theory. A fuzzy number is a convex fuzzy set with membership values between 0 and 1, defined by a range of real numbers. One of its applications is judging decision-makers based on hazy and ambiguous information. In these circumstances, linguistic variables are widely utilised to denote judgments in intricate and ill-defined problems. In order to address the shortcomings of the fuzzy technique and represent real-world problem scenarios, ref. [13] offered several enhancements, including n-type fuzzy sets. Since it is pretty challenging to calculate this value precisely, the membership value is represented as an interval of real integers. In this regard, ref. [14] introduced interval-valued fuzzy sets as a generalised concept of fuzzy numbers. Ref. [15] described an interval-valued triangular fuzzy number as:

$$\overline{A}=\left[{\tilde{A}}^{\mathrm{L}},{\check{\mathrm{A}}}^{\mathrm{U}}\right]=\left[\left({\mathrm{a}}_{\mathrm{l}}^{\prime }, {\mathrm{a}}_{\mathrm{m}}^{\prime }, {\mathrm{a}}_{\mathrm{u}}^{\prime }; {\mathrm{w}}_{\mathrm{A}}^{\prime }\right), \left({\mathrm{a}}_{\mathrm{l}}, {\mathrm{a}}_{\mathrm{m}}, {\mathrm{a}}_{\mathrm{u}}; {\mathrm{w}}_{\mathrm{A}}\right)\right]$$

(1)

where ${\tilde{A}}^{\mathrm{L}} \mathrm{and} {\check{\mathrm{A}}}^{\mathrm{U}}$ represent lower and upper triangular fuzzy numbers, respectively; ${\tilde{A}}^{\mathrm{L}}\subset {\check{\mathrm{A}}}^{\mathrm{U}}$; ${\mathsf{\mu }}_{\tilde{A}}\left(x\right)$. represents the membership function and the degree to which an event may be a member of $\tilde{A}$; ${\mathsf{\mu }}_{{\tilde{A}}^{\mathrm{L}}}\left(x\right)={{\mathrm{w}}^{\prime }}_{\mathrm{A}}$ and ${\mathsf{\mu }}_{{\tilde{A}}^{\mathrm{u}}}\left(x\right)={\mathrm{w}}_{\mathrm{A}}$ are the lower and upper membership functions, respectively.

If ${\mathrm{w}}_{\mathrm{A}}={{\mathrm{w}}^{\prime }}_{\mathrm{A}}=1$, then fuzzy triangular numbers are obtained by normalised interval-valued triangular fuzzy numbers [16].

2.2. SWARA Method Based on Interval-Valued Triangular Fuzzy (IF) Numbers

Several multi-criteria decision-making processes demand criteria weights. Weights assigned to criteria may be arbitrary or ill-defined. In addition to computing weighted criterion values, decision-makers may utilise other methods for assessing the importance of criteria (weights) [17,18,19].

Ref. [20] suggested the SWARA (Step-wise Weight Assessment Ratio Analysis) approach for weighing criteria subjectively based on the views and judgment of decision-makers and experts [21,22]. The method’s key characteristic is its ability to gauge experts’ or stakeholders’ opinions regarding the weighting of several criteria. Due to its ease of use and short time commitment, this method is subjective. The following steps describes the SWARA-IVTFN method [23]:

Prioritisation of Criteria: During this stage, each decision-maker must arrange the final criteria for ranking the choices according to importance in decreasing order. The following is how the relative weight of the criteria (S) is determined: the relative importance of each criterion is determined in proportion to the criterion with the greatest rank. Equation (2) is used to determine each criterion’s fuzzy coefficient [24]:

$${\tilde{k}}_j=\{\begin{array}{c}\tilde{1}, j=1\\ {\tilde{s}}_j+\tilde{1}, j>1\end{array}$$

(2)

where denotes criterion j, ${\tilde{k}}_j$. denotes the fuzzy coefficient for criterion j and ${\tilde{s}}_j$. denotes the criterion j significance.

To calculate the initial weight for the criteria: This stage requires that the initial weight of each criterion is obtained using Equation (3).

$${\tilde{q}}_j=\{\begin{array}{c}\tilde{1}, j=1\\ \frac{{\tilde{q}}_{j-1}}{{\tilde{K}}_j}, j>1\end{array}$$

(3)

where ${\tilde{q}}_j$. denotes the calculated weight for criterion j.

To obtain the final normalised weights (W): Finally, decision-makers are aided by Equation (4) to obtain the final normalised fuzzy weight for the criteria:

$${\tilde{\mathrm{w}}}_j=\frac{{\tilde{q}}_j}{{{\displaystyle \sum }}_{k=1}^n{\tilde{q}}_k}$$

(4)

where ${\tilde{\mathrm{w}}}_j$. denotes the normalised fuzzy weight for criterion j.

Furthermore, the defuzzification process has been used as Equation (5) to express the final weights of the criteria:

$$\mathrm{gm}\left(\tilde{B}\right)=\frac{l+l^{\prime }+m+u^{\prime }+u}{5}$$

(5)

where $\mathrm{gm}\left(\tilde{B}\right)$. denotes the crisp weight for criterion j; l and u denote lower and upper membership functions, respectively, of an interval-valued fuzzy number; l’ and u’ denotes the lower and upper non-membership functions of an interval-valued fuzzy number; m denotes the core of an interval-valued fuzzy number.

2.3. Additive Ratio Assessment (ARAS) Method Based on Interval-Valued Triangular Fuzzy Numbers

Decision-makers’ knowledge is used in solving complicated economic problems by various MCDM methodologies [25]. As one of the unique and valuable methodologies in the field of MCDM, Zavadskas and Turskis introduced the ARAS method in 2010 [26]. The following are the steps of the ARAS-ITF technique [27]:

Forming a decision matrix and determining the optimal performance rating for each criterion Equation (6):

$$\tilde{X}=\left[\begin{array}{ccc}x11& \cdots & x1m\\ ⋮& \ddots & ⋮\\ xni& \cdots & xnm\end{array}\right]$$

(6)

where the decision matrix is given as m x n – m denotes the alternatives and n denotes criteria; ${\tilde{x}}_{0j}$; is the optimal value of criterion j. The symbol ‘~’ above each letter rresents an interval-valued fuzzy set.

Equation (7) is used to compute the best performance rating for each interval-valued criterion:

$${\tilde{x}}_{0j}=\left[\left(l_{0j}, {l^{\prime }}_{0j}\right),m_{0j},\left({u^{\prime }}_{0j}, u_{0j}\right)\right]$$

(7)

Other symbols are defined as follows:

$$l_{0j}=\{\begin{array}{c}\begin{array}{c}max\\ i\end{array}{l}_{ij}; j\in {\Omega }_{max}\\ \begin{array}{c}min\\ i\end{array}{l}_{ij}; j\in {\Omega }_{min}\end{array}$$

(8)

$${l^{\prime }}_{0j}=\{\begin{array}{c}\begin{array}{c}max\\ i\end{array}{l^{\prime }}_{ij}; j\in {\Omega }_{max}\\ \begin{array}{c}min\\ i\end{array}{l^{\prime }}_{ij}; j\in {\Omega }_{min}\end{array}$$

(9)

$$u_{0j}=\{\begin{array}{c}\begin{array}{c}max\\ i\end{array}{u}_{ij}; j\in {\Omega }_{max}\\ \begin{array}{c}min\\ i\end{array}{u}_{ij}; j\in {\Omega }_{min}\end{array}$$

(10)

$$m_{0j}=\{\begin{array}{c}\begin{array}{c}max\\ i\end{array}{m}_{ij}; j\in {\Omega }_{max}\\ \begin{array}{c}min\\ i\end{array}{m}_{ij}; j\in {\Omega }_{min}\end{array}$$

(11)

$${u^{\prime }}_{0j}=\{\begin{array}{c}\begin{array}{c}max\\ i\end{array}{u^{\prime }}_{ij}; j\in {\Omega }_{max}\\ \begin{array}{c}min\\ i\end{array}{u^{\prime }}_{ij}; j\in {\Omega }_{min}\end{array}$$

(12)

2.4. Grey Relational Analysis

The Grey Relational Analysis Method creates single values for a multi-criteria problem [28,29]. This approach can solve issues involving intricate interactions between different criteria or elements. The Taguchi technique and grey relational analysis were coupled by [28] to handle the multi-response simulation challenge. Their analysis showed that the proposed strategy produced results that did not considerably deviate from the literature methods for this issue (TOPSIS, response surface method, dual-response system, and scatter search).

The sustainability criteria are normalised to ensure comparability of the data used to produce the grade-related grades for the sustainability standards [30]. Equations (13) and (14) are used to determine normalised values of the benefit- and cost-based sustainability criteria, respectively.

$$x_{ijk}=\frac{y_{ik,\max }-y_{ijk}}{y_{jk,\max }-y_{ik,\min }} \mathrm{if} j=\mathrm{benefit}\text{-}\mathrm{based} \mathrm{sustainability} \mathrm{criteon}$$

(13)

$$x_{ijk}=\frac{y_{ijk}-y_{ik,\min }}{y_{jk,\max }-y_{ik,\min }} \mathrm{if} j=\mathrm{cost}\text{-}\mathrm{based} \mathrm{sustainability} \mathrm{criteon}$$

(14)

where y_ijk and x_ijk denotes the actual and normalised values, respectively, for sustainability index i regarding criterion j based on maintenance strategy’s k.

The grey relational coefficient of the sustainability criteria is determined based on Equation (15).

$$\gamma \left(x_{ojk},x_{ijk}\right)=\frac{{\Delta }_{\min }+\zeta \Delta \max }{{\Delta }_{ijk}+\zeta \Delta \max }$$

(15)

$${\Delta }_{ijk}=\left|x_{ojk}-x_{ijk}\right|$$

(16)

where $\zeta$ represents a distinguishing coefficient [24]; $\gamma \left(x_{ojk},x_{ijk}\right)$ denotes the grey relational coefficient for sustainability index i regarding criterion j based on maintenance strategy’s k; $x_{ojk}$ denotes the referenced value for criterion j.

The grey relational grade for a maintenance strategy is expressed as Equations (17) and (18).

$$G_{ik}^b={\displaystyle \sum _{j=1}^P{w}_{ij}\gamma \left(x_{ojk},x_{ijk}\right)} \mathrm{if} j \mathrm{is} \mathrm{a}\mathrm{benefit}\text{-}\mathrm{based} \mathrm{criterion}.$$

(17)

$$G_{ik}^c={\displaystyle \sum _{p=p+1}^n{w}_{ij}\gamma \left(x_{ojk},x_{ijk}\right)} \mathrm{if} j \mathrm{is} \mathrm{a}\mathrm{cost}\text{-}\mathrm{based} \mathrm{criterion}.$$

(18)

$$x_k=\frac{{\displaystyle \sum _{i=1}^K{G}_{ik}^b}}{{\displaystyle \sum _{i=1}^K{G}_{ik}^c}}$$

(19)

where $G_{ik}^b$ and $G_{ik}^c$ represents the benefit and cost values, respectively, of a sustainability index i regarding a maintenance strategy k and $x_k$ denotes the cost-benefit value for performance index a maintenance strategy k.

3. Case Study

The case study for this study was a manufacturer of non-alcoholic beverages whosebusiness is situated in Lagos, Nigeria. Currently, the company has a dedicated maintenance department and runs two production shifts. The maintenance department is in charge of maintaining the production lines that make canned and bottled beverages. The department uses in-house staff to carry out planned maintenance tasks in accordance with the company’s maintenance policy. Periodic training programs are provided to the maintenance personnel to increase their productivity. After equipment audits, this study found that four planned maintenance strategies could be used to increase the company’s equipment sustainability. The strategies are periodic maintenance (S1), meter-based maintenance (S2), predictive maintenance (S3), and prescriptive maintenance (S4) (

Table 1. Maintenance strategies.

Strategy	Description
Periodic maintenance (S1)	Is a strategy that mandates routine maintenance to be carried out while the asset is in use at predetermined periods.
Meter-based maintenance (S2)	Is carried out as a result of a meter reading trigger. A meter counts the hours a piece of equipment has been in operation, the distance it has been driven, the number of parts produced, or an operational factor such as pressure or flow rate.
Predictive maintenance (S3)	Is based on the equipment’s condition rather than average or predicted life statistics. This strategy aims to directly monitor a machine when it is in normal operating condition to predict the breakdown before it occurs.
Prescriptive maintenance (S4)	It leverages the internet of things, machine learning, and artificial intelligence to provide detailed advice on maintaining equipment. It incorporates methods for making assumptions, testing and retesting data, and analysing historical data.

).

During the implementation of the suggested conceptual framework, energy consumption, water pollution, soil pollution, air pollution, noise control, and vibration control were considered environmental sustainability criteria. Labor, administrative and overhead costs, outsourcing, materials, spare parts, and penalty charges are the factors for economic sustainability. Social sustainability requirements include tenant satisfaction, tenant relations, psychosocial well-being, safety, compliance with laws, and security [31]. In contrast, technical sustainability criteria include equipment longevity, dependability, availability, efficiency, reparability, and service quality (

Table 2. Sustainability requirements.

Technical Requirements	Economic Requirements
Equipment longevity (c11)	Labour cost (c21)
Equipment reliability (c12)	Overhead cost (c22)
Equipment availability (c13)	Outsourcing cost (c23)
Equipment efficiency (c14)	Materials cost (c24)
Service quality (c15)	Spare parts cost (c25)
Equipment reparability (c16)	Penalty cost (c26)
Social requirements	Environmental requirements
Occupants’ health and safety (c31)	Water pollution (c41)
Human traffic (c32)	Soil pollution (c42)
Job satisfaction (c33)	Air pollution (c43)
Employee development (c34)	Noise control (c44)
	Vibration control (c45)

).

Based on Table 2, the maintenance strategies are evaluated using the linguistic variables in

Table 3. Linguistic terms for the maintenance strategies.

Linguistic Terms	Fuzzy Numbers
Very high (VH)	(0.75, 0.8); 0.9; (0.9, 0.9)
High (H)	(0.55, 0.6); 0.7; (0.8, 0.85)
Moderate (M)	(0.35, 0.4); 0.5; (0.6, 0.65)
Low (L)	(0.15, 0.2); 0.3; (0.4, 0.45)
Very low (VL)	(0.1, 0.1); 0.1; (0.2, 0.25)

. On the other hand,

Table 4. Linguistic terms for the sustainability criteria importance.

Linguistic Terms	Fuzzy Numbers
Equally important (EI)	(0.75, 0.8); 0.9; (0.9, 0.9)
Moderately less important (MLI)	(0.55, 0.6); 0.7; (0.8, 0.85)
Less important (LI)	(0.35, 0.4); 0.5; (0.6, 0.65)
Very less important (VLI)	(0.15, 0.2); 0.3; (0.4, 0.45)
Much less important (MLI)	(0.1, 0.1); 0.1; (0.2, 0.25)

presents linguistic variables used to evaluate the significance of the sustainability criteria in Table 2. The proposed framework’s implementation is based on industry understanding of maintenance systems. The data needed for the framework’s execution was generated by this study using the expertise of six groups of specialists [32]. A well-structured questionnaire was employed to gather data from the specialists during the data collection procedure. There are two parts to the questionnaire. Information regarding the significance of the sustainable needs is given in the first element of the questionnaire, whilst data on the strategies’ ratings are found in the second. The questionnaire includes a succinct introduction and purpose statement. This study ensures that the questionnaire is brief and contains no presumptive questions. It contains impartial questions while including open-ended questions and a rating scale.

4. Results and Discussion

Two viewpoints are taken into consideration when analysing the results of the tactics. The ARAS approach was utilised to generate the necessary information in the first viewpoint, which examined the methods from a requirement standpoint. The second viewpoint examined the techniques from the sustainability standpoint; GRA was utilised to produce the necessary data.

4.1. Technical Requirement Results

Based on the evaluation received from the manufacturing system specialists, the maintenance plans are examined. This study first established the significance of the sub-criteria included in this requirement. The experts’ evaluation of the technical requirements is shown in

Table 5. Technical assessment of the criteria.

Criterion	E1	E2	E3	E4	E5	E6
C11	LI	MLI	VLI	EI	EI	EI
C12	MLI	LI	VLI	LI	LI	EI
C13	MLI	VLI	MLI	EI	LI	LI
C14	MLI	VLI	LI	EI	MLI	EI
C15	VLI	EI	EI	VLI	MLI	VLI
C16	MLI	MLI	MLI	EI	MLI	MLI

. When the SWARA approach was implemented, this information was transformed into an intuitive fuzzy number. The SWARA outputs for the technical criteria in this study are in

Table 6. Technical requirement weights.

Criterion	l	l’	m	u’	u
C11	0.3142	0.3292	0.3541	0.3755	0.3918
C12	0.2349	0.2388	0.2424	0.2426	0.2470
C13	0.1734	0.1713	0.1649	0.1588	0.1542
C14	0.1250	0.1202	0.1099	0.1024	0.0949
C15	0.0888	0.0825	0.0756	0.0709	0.0657
C16	0.0636	0.0580	0.0532	0.0497	0.0463

. When the interval-valued fuzzy numbers in Table 6 were reduced to single values using Equation (5), it was discovered that c₁₁ and c₁₅ are the most and least critical technical criteria, respectively.

The experts’ responses to the maintenance plans under the technical criteria are shown in

Table 7. Technical assessment of the maintenance strategies.

Criterion	E1	E2	E3	E4	E5	E6	E1	E2	E3	E4	E5	E6
	S1						S2
C11	L	L	M	L	VH	M	L	M	H	VH	L	L
C12	VH	L	VH	VH	H	L	L	H	H	L	H	L
C13	H	H	H	L	VH	L	H	M	VH	VH	M	L
C14	H	VH	VH	L	M	H	M	VH	VH	L	M	H
C15	M	L	VH	VH	H	L	VH	L	H	VH	M	VH
C16	VH	L	M	H	M	M	L	H	H	M	H	L
	S3						S4
C11	VH	VH	VH	H	L	M	H	VH	M	H	L	VH
C12	VH	VH	L	H	VH	VH	M	VH	M	M	L	L
C13	M	L	M	L	M	L	VH	H	H	M	L	M
C14	H	H	M	VH	M	M	VH	L	VH	VH	L	L
C15	L	L	VH	L	L	H	VH	H	VH	H	VH	M
C16	M	L	M	H	H	H	L	H	VH	L	L	L

. These data were translated to ITF values and combined to create the aggregated decision matrix. This study normalised the matrix and created the weighted normalised matrix by combining the normalised decision matrix with the criteria weights (

Table 8. Weighted normalised technical IF values.

Criterion		S1	S2	S3	S4		S1	S2	S3	S4
C11		0.0489	0.0541	0.1082	0.1030		0.0594	0.0648	0.0971	0.1079
C12		0.0569	0.0385	0.0871	0.0569		0.0634	0.0448	0.0746	0.0560
C13	l	0.0404	0.0434	0.0341	0.0493	l’	0.0443	0.0473	0.0266	0.0532
C14		0.0289	0.0271	0.0312	0.0275		0.0305	0.0287	0.0287	0.0323
C15		0.0179	0.0219	0.0158	0.0243		0.0185	0.0222	0.0135	0.0283
C16		0.0145	0.0133	0.0164	0.0127		0.0148	0.0137	0.0148	0.0148
C11		0.0708	0.0759	0.1062	0.1012
C12		0.0664	0.0498	0.0764	0.0498
C13	m	0.0443	0.0467	0.0295	0.0443
C14		0.0289	0.0275	0.0275	0.0260
C15		0.0179	0.0209	0.0139	0.0229
C16		0.0141	0.0133	0.0141	0.0116
C11		0.0832	0.0832	0.1084	0.1008		0.0840	0.0888	0.1101	0.1089
C12		0.0655	0.0479	0.0734	0.0559		0.0647	0.0567	0.0712	0.0545
C13	u’	0.0391	0.0447	0.0336	0.0414	u	0.0409	0.0414	0.0310	0.0409
C14		0.0259	0.0259	0.0253	0.0253		0.0248	0.0237	0.0245	0.0218
C15		0.0172	0.0194	0.0140	0.0203		0.0158	0.0175	0.0134	0.0190
C16		0.0137	0.0119	0.0126	0.0115		0.0119	0.0118	0.0124	0.0102

). Figure 2 shows that S3 is the best suitable maintenance approach based on the ARAS outputs. S2, on the other hand, is the system’s least effective maintenance method (Figure 2).

4.2. Economic Requirement Results

Table 9. Economic assessment for the criteria.

Criterion	E1	E2	E3	E4	E5	E6
C21	MLI	EI	VLI	EI	MLI	VLI
C22	LI	MLI	MLI	VLI	EI	EI
C23	MLI	EI	EI	VLI	EI	VLI
C24	LI	VLI	MLI	MLI	LI	LI
C25	LI	MLI	EI	VLI	EI	EI
C26	EI	EI	MLI	LI	EI	LI

lists the experts’ opinions on the significance of the economic criteria. the SWARA approach for requirements evaluation was implemented usingng this data,. The linguistic values in this table were translated to ITF values for processing in this investigation.

Table 10. Economic requirement weights.

Criterion	l	l’	m	u’	U
C21	0.3159	0.3288	0.3495	0.3728	0.3839
C22	0.2337	0.2381	0.2428	0.2466	0.2481
C23	0.1655	0.1640	0.1586	0.1549	0.1530
C24	0.1352	0.1308	0.1209	0.1097	0.1046
C25	0.0908	0.0850	0.0787	0.0713	0.0679
C26	0.0589	0.0534	0.0495	0.0447	0.0426

lists the final results for the economic requirements. Equation (5) was used to convert the interval-valued fuzzy numbers in Table 10 to single values, and it was found that c₂₁ and c₂₆ are the most and least significant economic criteria, respectively.

The data shown in

Table 11. Economic assessment of the maintenance strategies.

Criterion	E1	E2	E3	E4	E5	E6	E1	E2	E3	E4	E5	E6
	S1						S2
C21	M	VH	M	VH	VH	M	H	H	H	VH	VH	H
C22	L	M	VH	M	VH	VH	VH	M	L	VH	VH	H
C23	M	L	H	H	VH	H	VH	M	H	VH	L	VH
C24	M	L	M	VH	H	M	L	H	H	L	VH	M
C25	H	H	M	VH	L	H	L	L	H	L	H	L
C26	VH	VH	H	VH	H	VH	M	VH	M	H	M	VH
	S3						S4
C21	VH	L	L	VH	VH	H	VH	L	VH	M	M	H
C22	H	VH	M	H	VH	M	M	M	VH	H	M	H
C23	VH	H	H	M	H	L	H	L	H	VH	H	H
C24	VH	H	VH	H	VH	L	M	VH	M	H	M	M
C25	H	H	L	M	H	H	VH	L	VH	VH	H	VH
C26	M	M	M	VH	VH	M	H	VH	M	L	M	H

serves as the foundation for the empirical study of the maintenance strategies based on the economic criterion. This study transformed the linguistic data in the table into ITF values for analysis. A normalised decision matrix for the strategies was created by averaging and normalising the values. For this decision-making issue, we created a weighted normalised decision matrix with this new matrix and the data in Table 11—see

Table 12. Weighted normalised values for the economic requirements.

Criterion		S1	S2	S3	S4		S1	S2	S3	S4
C21		0.0881	0.0786	0.0727	0.0765		0.0842	0.0758	0.0891	0.0797
C22		0.0669	0.0628	0.0513	0.0576		0.0629	0.0594	0.0594	0.0563
C23	l	0.0492	0.0432	0.0359	0.0334	l’	0.0445	0.0395	0.0445	0.0356
C24		0.0414	0.0414	0.0210	0.0233		0.0372	0.0372	0.0274	0.0289
C25		0.0219	0.0374	0.0153	0.0108		0.0195	0.0312	0.0208	0.0136
C26		0.0130	0.0171	0.0118	0.0112		0.0109	0.0141	0.0150	0.0133
C21		0.0859	0.0784	0.0902	0.0950
C22		0.0614	0.0584	0.0584	0.0646
C23	m	0.0411	0.0372	0.0411	0.0391
C24		0.0325	0.0325	0.0251	0.0307
C25		0.0181	0.0265	0.0191	0.0150
C26		0.0100	0.0125	0.0131	0.0139
C21		0.0868	0.0930	0.0953	0.0977		0.0961	0.0859	0.1004	0.1015
C22		0.0598	0.0598	0.0612	0.0659	u	0.0642	0.0615	0.0595	0.0628
C23	u’	0.0401	0.0345	0.0401	0.0401		0.0389	0.0380	0.0389	0.0372
C24		0.0284	0.0300	0.0244	0.0269		0.0274	0.0274	0.0235	0.0262
C25		0.0169	0.0223	0.0184	0.0136		0.0160	0.0208	0.0162	0.0149
C26		0.0098	0.0112	0.0112	0.0127		0.0093	0.0107	0.0112	0.0113

for more details about the weighted normalised decision matrix. When this data was processed using the ARAS approach, it is discovered that S1 is the best maintenance strategy for the system (Figure 3); in contrast, S3 is the worst.

4.3. Social Requirement Results

Table 13. Social assessment of the requirements.

Criterion	E1	E2	E3	E4	E5	E6
C31	VLI	MLI	EI	MLI	EI	LI
C32	LI	LI	VLI	LI	LI	MLI
C33	EI	VLI	LI	MLI	MLI	MLI
C34	MLI	MLI	LI	LI	MLI	EI

lists the linguistic evaluations of the social prerequisites for the maintenance plan selection dilemma. This study produced ITF values for the social criteria (

Table 14. Social requirement weights.

Criterion	l	l’	m	u’	u
C31	1.0826	1.1192	0.3964	0.4175	0.4372
C32	0.8049	0.8025	0.2654	0.2670	0.2659
C33	0.6510	0.6381	0.2044	0.1925	0.1854
C34	0.4615	0.4402	0.1338	0.1231	0.1116

) using the information in Table 13. These linguistic values were transformed to ITF values while evaluating the requirements’ crisp values. When the interval-valued fuzzy numbers in Table 14 were converted to single values using Equation (5), it was discovered that c₃₁ and c₃₄ were the most and least significant social criteria, respectively.

The data in

Table 15. ITF for the social requirements.

Criterion	E1	E2	E3	E4	E5	E6
C31	VLI	MLI	EI	MLI	EI	LI
C32	LI	LI	VLI	LI	LI	MLI
C33	EI	VLI	LI	MLI	MLI	MLI
C34	MLI	MLI	LI	LI	MLI	EI

are the empirical study of the maintenance strategies under societal requirements. The verbal data were transformed into ITF values and then analysed. The values were combined and normalised to create the normalised decision matrix for the strategies. Based on the normalised values and the data in Table 14 and

Table 16. Weighted normalised social IF values.

Criterion		S1	S2	S3	S4		S1	S2	S3	S4
C31		0.2760	0.2760	0.2654	0.2654		0.2500	0.2500	0.3500	0.2692
C32	l	0.2071	0.2224	0.1251	0.2502	l’	0.1928	0.2057	0.1469	0.2571
C33		0.1708	0.1501	0.1238	0.2063		0.1533	0.1362	0.1442	0.2044
C34		0.1137	0.1064	0.0916	0.1499		0.0986	0.0928	0.1052	0.1435
C31		0.0859	0.0859	0.1123	0.1123
C32	m	0.0599	0.0632	0.0474	0.0948
C33		0.0461	0.0417	0.0438	0.0729
C34		0.0283	0.0269	0.0298	0.0488
C31		0.0888	0.0939	0.1174	0.1174	u	0.1000	0.1000	0.1186	0.1186
C32	u’	0.0648	0.0631	0.0533	0.0857		0.0614	0.0665	0.0532	0.0847
C33		0.0431	0.0401	0.0431	0.0663		0.0438	0.0414	0.0419	0.0583
C34		0.0290	0.0268	0.0290	0.0383		0.0248	0.0245	0.0259	0.0364

, Table 16 displays the weighted normalised decision matrix for the strategies. S4 is the best method for the case study, as shown in Figure 4. Conversely, S2 is the system’s least effective method.

4.4. Environmental Requirement Results

The data shown in

Table 17. Environmental assessment for the criteria.

Criterion	E1	E2	E3	E4	E5	E6
C41	MLI	VLI	MLI	LI	MLI	EI
C42	EI	EI	MLI	VLI	EI	MLI
C43	MLI	VLI	MLI	MLI	MLI	MLI
C44	VLI	VLI	EI	MLI	LI	EI
C45	EI	LI	LI	LI	VLI	MLI

provides the foundation for examining maintenance tactics under environmental requirements. To use the SWARA approach, we transformed this data into ITF data. The SWARA outputs for the environmental requirements are shown in

Table 18. Environmental requirement weights.

Criterion	l	l’	m	u’	u
C41	0.3221	0.3320	0.3488	0.3712	0.4005
C42	0.2310	0.2332	0.2353	0.2398	0.2590
C43	0.1956	0.1950	0.1922	0.1850	0.1599
C44	0.1463	0.1417	0.1322	0.1205	0.1094
C45	0.1050	0.0981	0.0916	0.0835	0.0712

. The most and least significant environmental criteria were c₄₁ and c₄₅, respectively, when the interval-valued fuzzy numbers in Table 18 were transformed to single values using Equation (5). The data in

Table 19. Environmental assessment for the maintenance strategies.

Criterion	E1	E2	E3	E4	E5	E6	E1	E2	E3	E4	E5	E6
		S1								S2
C41	VH	H	H	M	L	VH	VH	M	M	M	VH	M
C42	H	M	H	H	H	VH	M	H	L	L	L	H
C43	L	VH	VH	VH	M	M	L	M	M	M	VH	L
C44	M	L	M	VH	VH	H	L	L	L	VH	L	H
C45	H	H	H	M	M	L	VH	VH	L	L	L	H
			S3							S4
C41	H	M	L	H	M	L	L	L	VH	H	VH	VH
C42	VH	H	H	H	L	L	VH	L	VH	H	H	M
C43	M	VH	VH	L	M	VH	H	M	VH	L	VH	M
C44	H	H	H	L	VH	L	VH	L	H	M	L	VH
C45	M	VH	L	H	H	L	L	L	M	M	M	M

provides the foundation for the empirical study of the maintenance strategies under the environmental criterion. This study transformed the linguistic data in the table into ITF values for analysis. The total values for the maintenance strategies are shown in

Table 20. Weighted normalised environmental IF values for the strategies.

Criterion		S1	S2	S3	S4		S1	S2	S3	S4
C41		0.0831	0.0888	0.0858	0.0644		0.0767	0.0815	0.1087	0.0652
C42		0.0516	0.0896	0.0473	0.0426		0.0492	0.0806	0.0591	0.0443
C43	l	0.0481	0.0710	0.0373	0.0392	l’	0.0452	0.0641	0.0452	0.0405
C44		0.0354	0.0540	0.0285	0.0285		0.0321	0.0467	0.0343	0.0286
C45		0.0284	0.0284	0.0209	0.0273		0.0241	0.0241	0.0241	0.0259
C41		0.0795	0.0837	0.1060	0.0795
C42		0.0499	0.0748	0.0582	0.0524
C43	m	0.0438	0.0584	0.0438	0.0461
C44		0.0296	0.0401	0.0312	0.0312
C45		0.0213	0.0213	0.0213	0.0278
C41		0.0888	0.0846	0.1110	0.0867		0.0940	0.0983	0.1109	0.0972
C42		0.0546	0.0710	0.0609	0.0533		0.0569	0.0761	0.0647	0.0612
C43	u’	0.0430	0.0528	0.0430	0.0462	u	0.0381	0.0452	0.0381	0.0385
C44		0.0269	0.0347	0.0307	0.0283		0.0255	0.0316	0.0258	0.0267
C45		0.0210	0.0198	0.0204	0.0223		0.0164	0.0176	0.0170	0.0202

. These values were normalised to create the normalised decision matrix for the technical requirements to analyse the strategies.

The data in Table 18 and Table 20 were combined to create the weighted normalised matrix. The data in Table 20 was processed using ARAS, while the ARAS outputs for the strategies based on the environmental criteria are shown in Figure 5. The conclusion that S2 is the best maintenance method is logical. On the other hand, S4 is the least suitable maintenance strategy for the system (Figure 5).

4.5. Sustainability Results

The GRA techniques combined the ARAS data (

Table 21. GRA data.

Criterion	S1	S2	S3	S4
Technical (C1)	0.1706	0.1652	0.1922	0.1898
Economic (C2)	0.1827	0.1814	0.1768	0.1776
Social (C3)	0.2619	0.2594	0.2617	0.3369
Environmental (C4)	0.1661	0.2032	0.1855	0.1653

). The decision-making problem’s final outputs were produced based on the weighted criteria in

Table 22. Criteria importance.

Criterion	Importance
C1	0.3841
C2	0.2778
C3	0.1970
C4	0.1411

. Figure 6 demonstrates that S2 is the most effective system maintenance method. On the other hand, S1 was determined to be the system’s least effective maintenance plan (Figure 6).

5. Conclusions

This study has shown how vital sustainability is when deciding on the best maintenance plan for equipment. This was accomplished by outlining various criteria for building sustainability needs. The analysis of several preventive maintenance procedures regarding equipment sustainability needs was achieved using a framework that included SWARA, ARAS, and GRA. The performance of the proposed framework took into account four maintenance strategies: periodic maintenance (S1), meter-based maintenance (S2), predictive maintenance (S3) and prescriptive maintenance (S4). The framework’s applicability—was demonstrated using an illustration example. Six specialists were also engaged in assessing the maintenance system techniques for a manufacturing organisation. The following findings were reached:

• S3 is the most appropriate maintenance strategy for the system based on the technological requirements, while S2 is the least effective method.
• The economic needs analysis showed that S1 is the maintenance strategy that is most appropriate for the system, whereas S3 is the least appropriate.
• S1 is the most appropriate maintenance method for the system, given the social requirements, while S2 is the least effective.
• The results of the environmental requirements indicated that S2 is the best maintenance plan for the system, while S4 is the worst.
• According to the GRA approach, the system’s best and least appropriate maintenance strategies are S2 and S4, respectively.

The design specifications for the selection problems are what drive the suggested framework. System specifications could be included in it to expand its applicability. Our upcoming research efforts will address this issue.