Multidimensional Risk Assessment in Sustainable Coal Supply Chains for China’s Low-Carbon Transition: An AHP-FCE Framework

Abstract

Driven by the global energy transition and the pursuit of “dual carbon” goals, sustainability risks within the coal supply chain have emerged as a central obstacle impeding the low-carbon transformation of high-carbon industries. To address the critical gap in systematic and multidimensional risk assessments for coal supply chains, this study proposes a hybrid framework that integrates the analytic hierarchy process (AHP) with the fuzzy comprehensive evaluation (FCE) method. Utilizing the Delphi method and the coefficient of variation technique, this study develops a risk assessment system encompassing eight primary criteria and forty sub-criteria. These indicators cover economic, operational safety, ecological and environmental, management policy, demand, sustainable supply, information technology, and social risks. An empirical analysis is conducted, using a prominent Chinese coal enterprise as a case study. The findings demonstrate that the overall risk level of the enterprise is “moderate”, with demand risk, information technology risk, and social risk ranking as the top three concerns. This underscores the substantial impact of accelerated energy substitution, digital system vulnerabilities, and stakeholder conflicts on supply chain resilience. Further analysis elucidates the transmission mechanisms of critical risk nodes, including financing constraints, equipment modernization delays, and deficiencies in end-of-pipe governance. Targeted strategies are proposed, such as constructing a diversified financing matrix, developing a blockchain-based data-sharing platform, and establishing a community co-governance mechanism. These measures offer scientific decision-making support for the coal industry’s efforts to balance “ensuring supply” with “reducing carbon emissions”, and provide a replicable risk assessment paradigm for the sustainable transformation of global high-carbon supply chains.

1. Introduction

The global energy system is at a pivotal juncture characterized by profound transformation. While countries expedite their commitments to carbon neutrality under the Paris Agreement, a pronounced tension exists between persistent dependence on traditional energy sources and the imperative for sustainable development [1]. As the cornerstone of the industrial revolution, coal remains essential to the global electricity supply and heavy industry. However, the environmental and social externalities arising from its entire supply chain life cycle—from mining and processing to cross-border transport and end-use combustion—have precipitated systemic crises [2]. According to the International Energy Agency (IEA), the coal sector was responsible for 37.4 gigatons of CO₂ emissions in 2024, representing approximately 43% of total fossil fuel-related emissions. Furthermore, issues such as land degradation, groundwater contamination, and biodiversity loss—particularly from open-pit mining—are especially severe in resource-dependent regions of Asia and Africa. Despite robust growth in renewable energy capacity, coal remains indispensable for grid stability and industrial processes, such as steel production, thereby hindering its immediate phase-out [3]. Consequently, reconstructing a sustainable coal supply chain that reconciles the three pillars of energy security, economic viability, and ecological resilience has become a critical challenge for researchers and policymakers alike.

The sustainability challenges facing the traditional coal supply chain stem from multifaceted fractures across technological, economic, and systemic dimensions. From a technological perspective, although low-carbon innovations such as carbon capture and storage (CCUS) and high-efficiency low-emission (HELE) combustion have been extensively investigated, their commercialization is constrained by substantial initial capital investment and a lack of sufficient cross-sectoral coordination [4]. From an economic standpoint, supply chain stakeholders lack adequate incentives to internalize environmental costs [5], while regional disparities in carbon pricing exacerbate trade frictions and market distortions [6]. Socially, externalities—including the elevated incidence of occupational diseases in mining communities and particulate matter pollution along international transport corridors—are frequently overlooked in conventional decision-making frameworks [7]. Additionally, global energy geopolitics (e.g., the surge in coal import substitution following the Russia–Ukraine conflict) and policy uncertainties surrounding coal phase-out (e.g., Germany’s postponement of coal power retirement) have further complicated long-term investment and risk management strategies within the supply chain [8]. Against this backdrop, it is imperative to develop a dynamic risk assessment framework that identifies critical risk nodes, quantifies multidimensional interactions, and equips stakeholders with tiered response strategies.

The current research on energy supply chain risk management demonstrates significant imbalance. Owing to pronounced geopolitical sensitivity and price volatility, the oil and gas sectors have facilitated the development of sophisticated quantitative risk assessment models [9]. In contrast, the research on renewable energy supply chains primarily addresses raw material constraints and technological iteration risks, particularly those associated with wind and photovoltaic power generation [10]. However, analysis of coal supply chain risks remains largely restricted to isolated perspectives, e.g., environmental research focuses on carbon emissions and ecological footprints [11]; economic studies examine price elasticity and market imbalances [12,13]; while sociological inquiries emphasize labor rights and community conflicts [14]. This fragmented approach impedes a comprehensive understanding of risk transmission mechanisms within the coal supply chain and fails to adequately address the compounded challenges of policy compliance, climate-related physical risks, and technological substitution, particularly in the context of transnational operations. Moreover, the research on emerging economies such as China, India, and South Africa remains notably limited, despite these regions’ high coal dependency, weaker regulatory structures, and distinctive risk profiles compared with those in developed countries. To bridge these critical gaps, this study proposes the following core research questions:

RQ1. How can multidimensional risk interactions and their relative weights within sustainable coal supply chains be systematically quantified?

RQ2. Which emerging risk nodes present the greatest threats to supply chain resilience amidst the energy transition, and through which mechanisms do they operate?

RQ3. How can a dynamic risk assessment framework be developed to provide actionable governance pathways for coal supply chains under China’s “dual carbon” goals?

In response, this study develops a hybrid framework that integrates the AHP and FCE methods to assess multidimensional risks in the sustainable coal supply chain within the context of the global energy transition. Core risk factors are identified using both the coefficient of variation and Delphi methods, resulting in a risk assessment system encompassing eight primary and forty secondary indicators. By leveraging AHP-derived quantitative indicator weights, key risk transmission mechanisms are revealed using empirical data from China’s A Coal Group. The principal innovation of this research lies in the development of a hybrid AHP-FCE framework, which overcomes the limitations of traditional single-dimensional risk analysis. Through the construction of a quantitative assessment system consisting of forty sub-indicators across eight core risk dimensions—including economic, safety, and environmental risks—this study systematically elucidates, for the first time, the multidimensional and emerging risks confronting coal supply chains during the sustainable transition in developing countries. Furthermore, it explicates the interrelationships and critical impact pathways among various risk factors. This replicable assessment paradigm provides essential methodological support for the low-carbon transformation of high-carbon industries globally.

The structure of this paper is as follows: Section 2 systematically reviews the theoretical development of sustainable supply chain management and energy supply chain risk assessments, and critically examines the limitations of existing methodologies. Section 3 presents the research framework and the development of the risk assessment indicator system. Section 4 describes the design and validation process of the AHP-FCE model. Section 5 details the application of the AHP-FCE model for weighting indicators and evaluating enterprise risk. Section 6 discusses the assessment results and management implications. Section 7 concludes with key research contributions and policy recommendations. Through this study, we aim to promote a shift in coal supply chain management from “end-of-pipe governance” to “source prevention”, establish a replicable paradigm for risk assessment in the just transition of global high-carbon industries, and advance the coordinated achievement of climate governance and sustainable development goals.

2. Literature Review

2.1. The Integration Framework of Sustainable Supply Chain Risk Management (SSCRM) and MCDM

The first systematic conceptualization of sustainable supply chain risk management (SSCRM) can be traced to Carter and Rogers (2008) [15] in the International Journal of Physical Distribution & Logistics Management. Drawing on the triple bottom line framework, they developed an integrated model encompassing environmental, social, and economic risks, emphasizing the necessity of managing the interactive effects among these dimensions. Specifically, environmental risks include resource depletion, emissions, and compliance costs with climate policies; social risks are associated with labor safety, community conflicts, and ethical concerns arising from limited supply chain transparency; economic risks pertain to market volatility, technology investment, and insufficient supply chain resilience. Their research further underscores the importance of precise risk assessment as a prerequisite for effective risk management; targeted strategies require quantifying risk priorities and transmission pathways across the triple bottom line [16]. Subsequent studies have expanded both the methodological toolkit and application scenarios for SSCRM.

In recent years, multi-criteria decision-making (MCDM) methods have demonstrated significant practical value in sustainable supply chain risk management (SSCRM). Amid growing risks of supply chain disruptions—exacerbated by globalization, the sharp decline in international trade during the COVID-19 pandemic, and escalating environmental and social pressures—MCDM has emerged as an essential tool for addressing complex risk trade-offs through its structured approach and multidimensional evaluation capabilities [17]. Research indicates that MCDM methods are widely applied across sectors such as agriculture, pharmaceuticals, and energy—from green procurement to circular economy initiatives—enabling organizations to build resilient supply chain networks by quantifying risk probabilities and impact intensities [18]. For instance, the integration of intuitionistic fuzzy BWM and VIKOR models facilitates the precise identification of safety risks in lithium battery recycling [19], while grey DEMATEL has been effectively used to map flood risk transmission in food supply chains [20], underscoring the robust support that MCDM methods offer for advancing sustainability objectives.

The application of individual MCDM methods demonstrates their fundamental value in risk assessment. The analytic hierarchy process (AHP) and its fuzzy extension (FAHP) serve as foundational tools due to their hierarchical decomposition and weight calibration capabilities. For example, in the pharmaceutical supply chain, Moktadir et al. used FAHP to quantify the interactive effects of raw material shortages and regulatory changes, with a consistency ratio validating model reliability [21]. Similarly, Ganguly and Kumar employed classic AHP to construct a risk indicator hierarchy for the automotive supply chain, identifying supplier financial stability as a critical factor [22]. Soyer et al. developed a fuzzy hesitant methodology based on cognitive mapping, demonstrating that economic, social, and environmental risks negatively impact supply chain performance, with supplier-related, operational, and demand-side risks deemed most critical [23]. Additionally, DEMATEL’s causal network modeling has elucidated the transmission of technological risks leading to delivery delays in electronics supply chains [24], while TOPSIS has provided risk priority rankings for cold chain logistics based on proximity to ideal solutions [25], and TOPSIS’s static assessment cannot adequately capture dynamic risk evolution [26].

However, the complexity and uncertainty inherent in sustainable supply chain risk management (SSCRM) render single MCDM methods inherently limited. For instance, AHP’s disregard for criterion dependencies may lead to distorted weight distributions, while TOPSIS’s reliance on static distance measurements makes it challenging to capture the dynamic evolution of risks [26]. Furthermore, issues such as assumptions of independence, insufficient handling of fuzziness, and difficulties in capturing dynamics have prompted researchers to adopt paradigm fusion strategies. By integrating complementary MCDM methods, these strategies aim to overcome bottlenecks and significantly enhance model adaptability and interpretability. To address information more comprehensively and objectively, researchers combine weight determination methods—such as CRITIC, which emphasizes differences in objective data, or BWM, which focuses on consistency in decision makers’ preferences—with comprehensive ranking methods like TOPSIS, VIKOR, and WASPAS [27,28,29]. This effectively addresses the limitations of single methods in balancing weight scientificity and alternative rankings. Fuzzy logic has emerged as a critical “adhesive” for addressing linguistic fuzziness and uncertainty. Researchers integrate fuzzy sets, such as trapezoidal intuitionistic fuzzy sets, with traditional MCDM methods [18,28,30], enabling the effective transformation and quantification of decision makers’ subjective judgments and overcoming the limitations of traditional methods in handling qualitative information. In response to the complex interdependencies and dynamics of risk factors, fusion approaches transcend traditional static independence assumptions. For example, coupling DEMATEL with ANP to construct risk network relationship models [31] corrects the independence limitations of AHP-like methods and enhances the accuracy of assessing the global impact of supply disruption risks. Similarly, combining fuzzy FMEA with GRA enables dynamic identification of process failure modes [26]. Meanwhile, emerging methods are increasingly being integrated into SSCRM, with the cross-innovation of fuzzy logic and intelligent algorithms becoming a dominant trend. By using outputs from MCDM processes, such as weights determined by AHP, as inputs for intelligent optimization algorithms like particle swarm optimization (PSO) [32] or machine learning models [33], researchers achieve multi-objective optimization and uncover complex nonlinear risk patterns, thereby addressing the optimization search and big data pattern recognition limitations of traditional MCDM methods.

Table 1. Summary of previous studies.

Author	Risk Dimension	Application Method	Research Subject
Moktadir et al. [21]	Environment, technology, knowledge and support, society, financial	Grey DEMATEL	Leather industry
Ganguly et al. [22]	Supply risk, financial risk, information risk, manufacturing risk	Fuzzy AHP
Soyer et al. [23]	Supply risk, demand risk, operational risk, competition related risk, political and legal risk, environmental and societal risk, economical risk	Hesitant FCM	General supply chain
Wang et al. [25]	Internal risk, external risk, logistics risk, information risk	Improved entropy method combined with TOPSIS	Fresh produce
Mangla et al. [26]	supplier perspective, organization perspective, customer perspective, reverse logistics network design	Fuzzy FMEA	Plastic manufacturers
Rostamzadeh et al. [27]	Environmental risks, organizational risks, sustainable supply risks, sustainable production risks/manufacturer, sustainable distribution risks, sustainable recycling risks, IT-related risks	Comprehensive fuzzy TOPSIS-CRITIC	Oil industry
Moktadir et al. [29]	Social dimension, environmental dimension, economic dimension, technical dimension, institutional dimension	Pareto analysis and BWM	Leather Industry Supply Chain
Alshehri et al. [28]	Economic risk, management policy risk and environmental risk. Sub-risks are industrial emissions, market dynamics, management policy failure, financial constraints, and credit uncertainty.	Hesitant AHP and hesitant WASPAS	Saudi Arabia manufacturing
Amin et al. [30]	Supply and procurement risks, distribution risks, organizational risks, environmental risks	FUZZY VIKOR-CRITIC	Logistics industry
Tarei et al. [31]	Logistics (delivery systems), petroleum product quality, crude oil supply, customer orders, and legal/political regulations	DEMATEL-AHP	India’s oil supply chain
Feng et al. [32]	Economic risk, quality risk, technical risk, accident risk, social risk, environmental risk, management decision risk, cooperation risk	GABP and PSO-BP	Grape supply chain
Xu&Tian [34]	Stakeholder responsibility, legal responsibility, resource and environmental responsibility, community and government responsibility, philanthropic responsibility	AHP-FCE	Automotive industry
Giannakis et al. [35]	Propose endogenous risks and exogenous risks from the three dimensions of economy, society, and environment, and conclude that endogenous risks are the most important	FMEA	Industrial manufacturing and textiles
Song et al. [36]	Operational risks, economic risks, environmental risks, and social risks	DEMATEL	Telecommunications companies
Valinejad et al. [37]	Technical sustainability, economic sustainability, social sustainability, environmental sustainability, institutional sustainability	Failure mode and effects analysis methodology	Internet service providers
Erdil et al. [38]	Environmental problems, economic problems, social problems	FMEA	Clothing and textile industry
Xu et al. [39]	Operational risk, social risk, environmental risk	Risk assessment space and materiality analysis	Clothing and automotive industries
Abdel-Basset et al. [40]	Financial risk, supply risk factors, environment risk, operational risk, control and plan risk, information-/IT-related risks	TOPSIS-CRITIC	Telecommunications equipment company
Zhang et al. [41]	Supply operational risk, economic risk, environmental risk, social risk	BWM and linguistic value soft set theory	Energy industry chain
Elmsalmi et al. [42]	Manufacturing risk, risk of upstream and downstream partners, product design risk, supplier risk	MICMAC
Chen et al. [43]	Economic dimension, technological dimension, social dimension, demand dimension, policy dimension, supply dimension, market dimension, natural environment dimension	FDM and ISM	Telecom industry
Kalantari et al. [44]	Time delay risk, exchange rate fluctuations, raw material quality, and production quality	Multi-objective cross-entropy algorithm	Food supply chain
Hashim et al. [45]	Endogenous: workplace diversity, corporate strategy, operational plan, disruptions, internal stakeholder pressures; exogenous: economic conditions, regulatory compliance, natural calamities, external stakeholder pressure	FMEA and Pareto analysis	Textile industry
Hsu et al. [46]	External risks include man-made disasters, environmental disasters, natural disasters, and market trend risks. Internal risks include strategic management risks, information system risks, supplier and customer risks, internal unforeseen risks, internal business risks, and employee risks.	Integrated HOQ-MCDM	Electrical appliance manufacturers
Zhu et al. [47]	Environmental risk, organizational risk, technical risk, manage risk, cooperation risk.	WBS-RBS and BP neural network	Prefabricated building supply chain

summarizes the various MCDM research methods adopted by previous studies across different themes and highlights the risk dimensions relevant to this paper.

2.2. Risk Management of Sustainable Coal Supply Chain

The coal supply chain comprises a variety of participants, including resource exploration companies, coal producers, distributors, transportation enterprises, and end users. Internal influencing factors encompass coal resources, production, transportation, storage, and consumption, while external factors involve the natural environment as well as political, economic, and policy conditions [48,49]. The structure of the sustainable coal supply chain considered in this study is illustrated in Figure 1.

The coal supply chain is characterized by a strong dependence on geological conditions and resource distribution. The Asia-Pacific, European and Eurasian, and North American regions hold the largest coal reserves, with the United States, China, Russia, and Australia collectively accounting for 70.3% of global reserves [50]. This high resource concentration leads to extended supply chains, with transportation costs comprising 25–40% of end-user prices [51]. For instance, the average transportation distance on China’s “west-to-east coal transport” railway exceeds 1000 km, and transportation losses are substantially higher than those for oil and gas pipelines, further increasing supply chain vulnerability [52]. Furthermore, the coal supply chain generates significant environmental externalities throughout its life cycle [53]. Coal mining can cause soil and water loss up to 3.2 m³ per ton; coal washing consumes 2.8 m³/ton of water, with heavy metals in mine water exceeding safe levels in 34% of cases [2]. The combustion phase alone is responsible for 40% of global energy-related CO₂ emissions, intensifying the push for green supply chain transformation and the internalization of environmental costs [54]. The coal supply chain is also highly sensitive to energy policies and carbon regulations. For example, China’s dual control policy reduced coal’s share of energy consumption from 56.8% in 2020 to 54.2% in 2023, while the EU’s carbon border adjustment mechanism could elevate export costs to Europe by 18–24% [6]. Such policy variability has increased supply chain payback uncertainty and raised risk premiums by 2.3 percentage points [10]. Finally, the coal supply chain is marked by strong path dependency. The average operational lifespan of coal-fired power plants is 32 years, with the penetration rate of carbon capture, utilization, and storage (CCUS) technologies below 5% [55]. High transformation costs and long equipment upgrade cycles (8–10 years) create “carbon lock-in”, making the transition to low-carbon alternatives more challenging and hindering decarbonization efforts [8].

The structural contradictions within the coal supply chain suggest that, under the impetus of the “dual carbon” initiative, its sustainable transformation requires more than technical optimization; it demands a systematic overhaul encompassing resource reallocation, internalization of environmental costs, and institutional coordination [48]. Logistics vulnerabilities arising from geographical dependence [52], compliance risks linked to environmental externalities [54], and transformation delays resulting from technological lock-in [8] are deeply interconnected, forming a complex multidimensional risk landscape. The following section systematically reviews the core risks and their mechanisms during the coal supply chain’s transition to sustainability, drawing on the relevant SSCRM literature.

2.2.1. Economic Risks

Economic risks in the coal supply chain primarily arise from the strong coupling between macroeconomic cycles and the industry’s capital structure. The coal supply chain’s cross-regional network characteristics render it an “amplifier” of economic volatility. During economic downturns, inventory backlogs propagate through the price transmission mechanism, resulting in a vicious cycle of demand contraction, overcapacity, and debt default [48]. Furthermore, tightening restrictions by international financial institutions on high-carbon assets force coal enterprises to seek alternative financing through the “green screening” processes of capital markets. These structural financing constraints elevate capital costs and diminish firms’ financial flexibility, thereby hindering investment in technological transformation [2]. Zhao et al. further demonstrated that policy uncertainty suppresses R&D investment through an “investment wait-and-see” effect, causing stagnation in the commercialization of low-carbon technologies and perpetuating a negative feedback loop of technology lock-in and capital withdrawal [5].

2.2.2. Risks of Safe Operation

Safety risks in the coal supply chain arise from the mismatch between technical systems and geographical environments. Rioux et al. highlighted that geological heterogeneity, especially in deep mining operations, limits the effectiveness of standardized safety regulations, leading to a nonlinear increase in the probability of mine disasters such as gas outbursts with greater mining depths [52]. Wang et al. identified “technical fragmentation” during mining technology upgrades—characterized by incompatible interface standards between new and legacy equipment and fragmented data protocols—as a major contributor to increased complexity in human–machine interactions, escalating operational errors from isolated incidents to systemic failures [53]. Additionally, Silvia et al. demonstrated that, in highly mechanized environments, excessive reliance on automated systems can degrade workers’ risk perception, resulting in a cognitive trap of “skill degradation and risk misjudgment” [2].

2.2.3. Ecological and Environmental Risks

The transmission mechanism of environmental risks in the coal supply chain is fundamentally rooted in the institutional absence of ecological value accounting. Wang et al. demonstrated that ecological damage costs in mining areas are not incorporated into supply chain pricing, allowing enterprises to externalize environmental costs to the public and incentivizing extensive mining practices [54]. Wang et al. [50] found that current environmental governance primarily relies on “end-of-pipe” controls, neglecting source reduction at the mining stage. This results in diminishing marginal returns from pollution reduction technologies and limits their effectiveness. Additionally, the unclear allocation of ecological damage responsibilities across administrative jurisdictions has led to a “tragedy of the commons”, in which the emission reduction efforts of individual enterprises are offset by regional environmental degradation, thereby perpetuating “free-rider” behavior and weakening institutional incentives for pollution control.

2.2.4. Managing Policy Risks

Policy risks in the coal supply chain arise from conflicts and interactions among multi-level governance objectives. Policy implementation theory suggests that the strict requirements of central environmental supervision often conflict with local governments’ fiscal dependence on coal, resulting in a “goal–means” mismatch. Local authorities frequently resort to “selective law enforcement” to balance economic growth and emission reduction mandates, increasing the uncertainty of corporate compliance costs. Liu et al. showed that the EU carbon border adjustment mechanism (CBAM) transmits “carbon cost transfer” pressures to developing countries through international trade, forcing domestic exporters to bear implicit carbon costs; the absence of an integrated international carbon market further amplifies price distortion effects [6]. Zhang et al. demonstrated that stricter industry access standards reshape the supply chain network via “node screening”, eliminating non-compliant marginal suppliers, strengthening the bargaining power of core firms but simultaneously reducing overall supply chain flexibility [8].

2.2.5. Demand Risk

The driving mechanism of demand-side change in the coal supply chain lies in the synergistic effects of energy substitution and technological disruption. According to demand elasticity polarization theory, advancements in clean energy technologies not only directly reduce coal demand but also alter the elastic properties of the energy system, rendering traditional demand forecasting models ineffective. Additionally, geographical segmentation in extended supply chains amplifies the bullwhip effect, as transportation delays and information asymmetry cumulatively distort demand signals during cross-regional transmission. This “signal attenuation–response lag” mechanism leads to persistent inventory mismatches. Furthermore, the energy transition strategies of key customers pose existential threats to the coal supply chain structure; for example, when steel producers shift to hydrogen-based ironmaking technologies, the foundational demand for coking coal is fundamentally eroded through a “demand structure rupture” mechanism.

2.2.6. Sustainable Supply Risk

The evolution of supply risk in the coal supply chain is shaped by the interplay of resource scarcity and institutional constraints. Wang et al. proposed a “resource political economy” framework, highlighting that the spatial concentration of high-quality coal makes these regions focal points of geopolitical competition; the imposition of trade barriers compels firms to incur additional transaction costs through a “supply network reconstruction” mechanism [50]. Climate change further threatens supply chain resilience via the “extreme weather–logistics disruption” chain reaction, with railway transport vulnerabilities intensifying exponentially under abnormal climatic conditions [56]. Additionally, Zhang et al. emphasized that the “compliance cost transfer” of green procurement policies alters supplier competition: small and medium-sized coal mines, unable to absorb certification costs, are marginalized, causing the supply network to shift from a distributed to a core–periphery structure [8].

2.2.7. Information Technology Risks

The risk mechanism of digital transformation in the coal supply chain arises from systemic conflicts associated with technology integration. The decentralized architecture of blockchain fundamentally contradicts the sector’s hierarchical control structures, and the inflexible execution of smart contracts may exacerbate issues related to limited operational agility [57]. Additionally, the “path dependence” of machine learning algorithms impedes their ability to adapt to nonlinear changes in the energy landscape, while training biases in historical data systematically underestimate the sensitivity of demand forecasting models to clean energy substitution. Moreover, protocol heterogeneity among IoT devices creates “data islands”, and discrepancies in data standards across supply chain stages hinder information integration, ultimately reducing rather than enhancing overall supply chain collaboration efficiency.

2.2.8. Social Risks

The accumulation of social risks in the coal supply chain reflects underlying contradictions in benefit distribution during industrialization. The concept of “social license cost” suggests that resistance from mining communities increases project operating expenses via the “NIMBY effect–compensation game” mechanism. Furthermore, a disconnect between mechanization and occupational health protection creates a “technology–institutional” gap, e.g., while automation has reduced accident rates, outdated dust protection standards continue to expose workers to chronic health risks [28]. Additionally, the “public opinion resonance” mechanism enabled by social media magnifies the impact of environmental incidents beyond geographical boundaries, with reputational damage accelerating financial crises through a positive feedback loop of “stigmatization–capital withdrawal”.

Existing research has extensively applied integrated MCDM methods—such as DEMATEL-ANP and fuzzy TOPSIS-CRITIC—to supply chain risk assessment. However, these studies primarily address sectors like manufacturing and food, devoting limited attention to the specific characteristics of the coal supply chain. The coal supply chain possesses unique features that distinguish it from other industries, including strong geographical dependence, high environmental externalities, pronounced policy sensitivity, and technological lock-in effects. Notably, these industry-specific factors have not been systematically incorporated into the evaluation frameworks of the previous literature. In terms of methodological application, while fuzzy set theory has been widely utilized, the unique strengths of fuzzy comprehensive evaluation (FCE)—particularly its capability to deliver intuitive risk ratings—have not yet been fully integrated with the coal supply chain’s multi-level risk architecture.

By synthesizing an analysis of risk characteristics specific to the coal supply chain with multi-criteria decision-making (MCDM) methods, the AHP-FCE approach demonstrates strong adaptability in addressing multi-level fuzzy risk quantification challenges. This approach is especially effective in accommodating expert subjective evaluations and static weight allocations. Moreover, it is well-suited to small-sample expert questionnaire data while maintaining lower model complexity. Accordingly, this study employs the AHP-FCE model to conduct a robust evaluation of multidimensional risks in the sustainable coal supply chain and to examine their underlying interaction mechanisms.

3. Research Framework and Evaluation Index System

3.1. Research Framework

This study follows the process illustrated in Figure 2 and consists of five systematic steps to identify, prioritize, and evaluate risks in the sustainable coal supply chain. First, a combination of the Delphi method and the coefficient of variation is applied to identify the core risk factors, as outlined in Section 2. Second, the most appropriate risk assessment method is determined based on a comparative analysis of MCDM approaches applied to sustainable supply chain risk assessment, also discussed in Section 2. Third, the AHP is employed to rank the relative importance of each risk factor. Fourth, FCE is utilized to quantify qualitative risk information and assign risk levels using membership functions. Finally, based on the assessment outcomes, targeted risk mitigation strategies are proposed for the key risk factors identified in the case group of the sustainable coal supply chain.

3.2. Evaluation Index System

3.2.1. Preliminary Selection of Evaluation Indicators

The sustainable transformation of the coal supply chain entails multifaceted risks, thereby necessitating a robust risk assessment system capable of delivering accurate, objective, and scientifically validated evaluations of both overall supply chain performance and its individual segments. The primary objective is to provide actionable guidance for the sustainable development of the coal industry’s supply chain. In this study, multiple rounds of expert evaluation were carried out, and indicator selection was informed by a comprehensive analysis of the unique characteristics associated with coal supply chain transformation and sustainable supply chain risk management (SSCRM). Drawing upon the literature review in Section 2, eight key risk dimensions were identified as follows: economic risk, operational safety risk, ecological and environmental risk, management and policy risk, demand risk, sustainable supply risk, information technology risk, and social risk.

For the sub-criteria selection process, the principles of scientific validity, comprehensiveness, and independence were upheld by employing literature keyword frequency analysis, expert consultation, and theoretical assessment. The reference sources included the following: (1) SSCRM research papers, and (2) practical indicators utilized in the coal industry’s supply chain risk management, such as major accident hazard determination standards and coal mine safety regulations issued by the Ministry of Emergency Management of China. As a result, a preliminary evaluation system comprising forty-four risk factors distributed across eight categories was established (

Table 2. Coal enterprise sustainable supply chain risk assessment indicators (preliminary).

Primary Criteria	Sub-Criteria	References
Economic Risk	Economic Cycle Risk	Moktadir et al. [21]; Alshehri et al. [28]; Feng et al. [32]; Giannakis et al. [35]; Song et al. [36]; Valinejad et al. [37]; Abdel-Basset et al., 2020 [40]; Zhang et al. [41]
	Financing Risk
	Financial Constraints
	Solvency Risk
	R&D Investment
Safety Operation Risk	Natural Disasters	Ganguly et al. [22]; Soyer et al. [23]; Mangla et al. [26]; Alshehri et al. [28]; Feng et al. [32]; Hsu et al. [46];
	Safety Production Training
	Mining Technology Upgrade
	Equipment Failure
	Employee Operation Risk
Ecological Risk	Environmental Governance	Soyer et al. [23];Alshehri et al. [28]; Feng et al. [32]; Zhang et al. [41]; Hashim et al. [45]; Hsu et al. [46]; Zhu et al. [47]
	Mine Ecological Protection
	Pollution Reduction Risk
	Energy Consumption
	Environmental Accidents
Policy Risk	Regulatory Compliance	Alshehri et al. [28]; Feng et al. [32]; Hashim et al. [45]; Zhu et al. [47]
	Legal Risk
	Local Government Supervision
	Coal Industry Access Restrictions
	Corporate Sustainable Development Strategy
	Partner Risk
Demand Risk	International Energy Market Price Fluctuations	Soyer et al. [23]; Xu et al. [39]; Kalantari et al. [44]; Hsu et al. [46]
	Coal Industry Cyclicality
	Clean Energy Substitution
	Bullwhip Effect Risk
	Key Customer Disruption
Sustainable Supply Risk	Supplier Legal Compliance	Soyer et al. [23]; Rostamzadeh et al. [27]; Alshehri et al. [28]; Amin et al. [30]; Zhang et al. [41]; Elmsalmi et al. [42]
	Supplier Green Commitment
	Supplier Quality Management
	Core Supplier Disruption
	Procurement Cost Risk
	Coal Resource Depletion Risk
	Capacity Constraint Risk
	Logistics and Transportation Risk
IT Risk	Green Innovation Risk	Rostamzadeh et al. [27]; Moktadir et al. [29]; Feng et al. [32]; Abdel-Basset et al. [40]; Hsu et al. [46]; Zhu et al. [47]
	Artificial Intelligence Risks
	Digital Supply Chain System Stability
	IoT Device Reliability
	Information System Security
	Data Timeliness
Social Risk	Employee Health and Safety	Alshehri et al. [28]; Moktadir et al. [29];Feng et al. [32]; Zhang et al. [41];Hashim et al. [45]
	Employee Compensation and Benefits
	Community Relations Maintenance
	Public Opinion Risk

).

3.2.2. Optimization of Evaluation Indicators

To further refine the evaluation indicators, the Delphi method combined with the coefficient of variation approach was employed to construct a risk assessment index system for the sustainable supply chain of coal enterprises [58]. Ten experts—representing coal enterprise management, as well as the fields of safety engineering, mining engineering, and management from universities—participated in multiple rounds of anonymous consultation (

Table 3. Introduction of experts involved in confirming identified risks.

Number	Workplace	Position (Level)	Experience	Work Areas
E1	Coal Enterprises	Assistant to General Manager	20+	Business Management and Coordination
E2	Coal Enterprises	Head of Planning Section	12+	Corporate Strategic Planning
E3	Coal Enterprises	Sales Company Manager	20+	Supply Chain Management
E4	Coal Enterprises	Union Chairman	15+	Protecting Workers’ Rights
E5	Coal Enterprises	Production Team Leader	10+	Production Front Line
E6	Universities	Professor	30+	Coal Mine Safety
E7	Universities	Professor	15+	Intelligent Safety Production
E8	Universities	Associate Professor	10+	Coal Enterprises Emergency Management
E9	Universities	Professor	20+	Supply Chain Management
E10	Universities	Associate Professor	12+	Green Supply Chain Management

). Indicators were evaluated using a Likert five-point scale (1–5). The screening of indicators was conducted using the mean and coefficient of variation (CV) derived from expert scoring data. The mean value of each indicator (Xi) reflected its perceived importance, with a predefined threshold of ≥ 3 to ensure relevance. Additionally, the standard deviation (Si) was used to compute the coefficient of variation (CV = Si/Xi), which quantified the consistency of expert opinions (threshold CV ≤ 0.3). Unlike relying solely on standard deviation, the CV provided a more robust measure by accounting for the relative variability of the indicator, thereby improving the accuracy of importance assessment.

Detailed expert scores are provided in

Table 4. Expert scoring mean and coefficient of variation.

Primary Criteria	Sub-Criteria	AV	CV
Economic Risk	Economic Cycle Risk	4.4615	0.148000215
	Financing Risk	4.2308	0.196621295
	Financial Constraints	4	0.176776695
	Solvency Risk	4.1538	0.133607173
	R&D Investment	4.3077	0.17433882
Safety Operation Risk	Natural Disasters	4.6154	0.140916172
	Safety Production Training	4.7692	0.091876635
	Mining Technology Upgrade	4.3846	0.175184641
	Equipment Failure	4.3846	0.148333828
	Employee Operation Risk	4.9231	0.056364636
Ecological Risk	Environmental governance	3.0769	0.357058653
	Mine Ecological Protection	4.3846	0.175184641
	Pollution Reduction Risk	4.3077	0.17433882
	Energy Consumption	4.6923	0.10242836
	Environmental Accidents	3.9231	0.264532157
Policy Risk	Regulatory Compliance	4.4615	0.116250611
	Legal Risk	4.3846	0.175184641
	Local Government Supervision	3.7692	0.245893699
	Coal Industry Access Restrictions	3.9231	0.219865501
	Corporate Sustainable Development Strategy	4.3846	0.148333828
	Partner Risk	3.5385	0.318354953
Demand Risk	International Energy Market Price Fluctuations	4.4615	0.148000215
	Coal Industry Cyclicality	4.3077	0.17433882
	Clean Energy Substitution	3.8462	0.233708128
	Bullwhip Effect Risk	3.6923	0.231559023
	Key Customer Disruption	4.0769	0.211571132
Sustainable Supply Risk	Supplier Legal Compliance	4.4615	0.148000215
	Supplier Green Commitment	2.8923	0.341559023
	Supplier Quality Management	3.8462	0.179001763
	Core Supplier Disruption	4.4615	0.174051495
	Procurement Cost Risk	4.2308	0.141620045
	Coal Resource Depletion Risk	4.6154	0.16642427
	Capacity Constraint Risk	4.0769	0.233986412
	Logistics and Transportation Risk	3.9231	0.163215932
IT Risk	Green Innovation Risk	3.9231	0.243159543
	Artificial Intelligence Risks	2.991	0.333853008
	Digital Supply Chain System Stability	3.8462	0.233708128
	IoT Device Reliability	4.0769	0.157058653
	Information System Security	4.2308	0.171423507
	Data Timeliness	4.3077	0.198478395
Social Risk	Employee Health and Safety	4.6923	0.134279428
	Employee Compensation and Benefits	4.4615	0.148000215
	Community Relations Maintenance	4.3846	0.148333828
	Public Opinion Risk	4.2308	0.21906555

. Based on statistical analysis, four indicators—environmental governance, partner risk, supplier green commitment, and artificial intelligence risk—were eliminated due to insufficient mean scores or excessive coefficient of variation. Ultimately, 40 core sub-criteria were retained. The finalized risk assessment index system for the coal sustainable supply chain is illustrated in Figure 3.

4. Method

4.1. Subjective Weights Based on AHP Method

This study applies the AHP method to determine the weight coefficients of sustainable supply chain risk indicators for coal enterprises. AHP is a multi-criteria decision analysis method that structures complex decision problems into multiple hierarchical levels and assigns weights based on pairwise comparisons of the elements at each level. Its core principle is to transform subjective expert judgments into quantitative weights, thereby providing a systematic and structured approach to complex decision making. Proposed by Thomas L. Saaty in the 1970s, AHP has been widely adopted across various fields, including economics, engineering, and management. The general steps of AHP are as follows:

Step 1: Construct the hierarchical model. The decision problem is decomposed into multiple levels: the goal level (representing the overall objective), the criteria level (comprising indicators required to achieve the goal), and the alternative level (representing specific decision options).

Step 2: Formulation of the judgment matrix. For each parent element, a judgment matrix $A={(a_{ij})}_{n\times n}$ is constructed through pairwise comparisons of elements. The matrix entries $a_{ij}$ quantify the relative importance of element i over element j using Saaty’s 1–9 scale (

Table 5. Scale meaning table.

Importance	Standard
1	Both factors have equal importance
3	One factor is slightly more important than the other
5	One factor is significantly more important than the other
7	One factor is relatively important
9	One factor is definitely more important than the other

2 4 6 8 is used for situations between adjacent scales to indicate the corresponding degree of importance.

); the matrix adheres to the reciprocity rule $a_{ji}=1/a_{ij}$ and positivity $a_{ij}>0$. The general form of the judgment matrix is shown in

Table 6. General form of judgment matrix.

A	A1	A2	An
A1	a11	a12	a1n
A2	a12	a22	a2n
An	an1	an2	ann

.

Step 3: Computation of the weight vector. The weight vector $w=(w_1,\dots ,w_n)$ for elements is derived from A using the column normalization and row averaging method:

Column sums:

$$s_j={\displaystyle \sum _{i=1}^n{a}_{ij}},j=1,\dots ,n$$

(1)

Normalized matrix:

$$\overline{a_{ij}}={\displaystyle \raisebox{1ex}{$a_{ij}$}\!\left/ \!\raisebox{-1ex}{$s_j$}\right.},i,j=1,\dots ,n.$$

(2)

Weight vector:

$$w_i={\displaystyle \frac{1}{n}}{\displaystyle \sum _{j=1}^n\overline{a_{ij}}},i=1,\dots ,n.$$

(3)

The weights satisfy ${\displaystyle {\sum }_{i=1}^n{w}_i=1}$ and $w_i>0$. This method approximates the principal eigenvector without eigenvalue decomposition.

Step 4: Consistency verification. Logical consistency of pairwise judgments is evaluated using:

Matrix/vector product:

$$Aw=\left(\begin{array}{l}{\displaystyle {\sum }_{j=1}^n{a}_{1j}{w}_j}\\ {\displaystyle {\sum }_{j=1}^n{a}_{nj}{w}_j}\end{array}\right)$$

(4)

Maximum eigenvalue (${\lambda }_{\max }$):

$${\lambda }_{\max }={\displaystyle \frac{1}{n}}{\displaystyle \sum _{i=1}^n\frac{{(Aw)}_i}{w_i}}$$

(5)

Consistency index (CI):

$$CI={\displaystyle \frac{{\lambda }_{\max }-n}{n-1}}$$

(6)

Consistency ratio (CR):

$$CR={\displaystyle \frac{CI}{RI}}$$

(7)

where RI is the random index for matrix size n; the value of RI is shown in

Table 7. Average random consistency index.

n	1	2	3	4	5	6	7	8	9
RI	0	0	0.59	0.90	1.12	1.24	1.32	1.41	1.45

. Judgments are consistent if CR < 0.10; otherwise, A must be revised.

Step 5: Overall hierarchical ranking. The overall ranking involves calculating the weights of elements at each level using the methods described above and aggregating these weights hierarchically from top to bottom. This process quantifies the relative importance and contribution of each element to the overall objective in relation to the upper-level criteria.

4.2. Comprehensive Evaluation Based on the FCE Method

The FCE method is a method used to determine the risk assessment system and judge the risk level of an enterprise. It combines the principle of weighting indicators in the hierarchical method, and constructs a matrix by calculating the factor set and determining the weights. Then, the matrix and the weights are fuzzy operated to obtain the final evaluation result.

The following are the key elements of the FCE method:

(1) Establishment of the factor set and evaluation set. Collect and define various factors that affect risk assessment and quantify or quantify them, denoted by U = {U₁, U₂, U₃, …, U_m}. The set of possible evaluation grades, defined as V = {V₁, V₂, V₃, V₄, V₅}.

(2) Determine the weight. Use the hierarchical method or other methods to assign weights to each factor to reflect its relative importance. The weight vector is denoted as: W = {W₁, W₂, …, W_m}.

(3) Construct the fuzzy relation matrix. To reflect the degree of membership of each indicator to each grade in the evaluation set, a fuzzy relation matrix is established:

$$R=\left[\begin{array}{l}{r}_{11}{r}_{12}\dots r_{1m}\\ r_{21}{r}_{22}\dots r_{2m}\\ \dots \\ r_{n1}{r}_{n2}\dots r_{nm}\end{array}\right]$$

(8)

where $r_{ij}$ represents the membership degree of indicator $u_i$ to the evaluation grade $v_j$, and ${\displaystyle {\sum }_{j=1}^m{r}_{ij}=1}$ for each $i$.

(4) Fuzzy operation. The fuzzy comprehensive evaluation result is obtained via weighted aggregation, most commonly using the weighted average operator:

$$Y=W•R=[b_1,b_2,\dots ,b_m]$$

(9)

$$y_j={\displaystyle \sum _{i=1}^n{w}_i}\cdot r_{ij},(j=1,2,\dots ,m)$$

(10)

where $Y$ is the final fuzzy comprehensive evaluation vector, $y_j$ reflects the comprehensive membership of the evaluated object to the evaluation level $v_i$.

(5) Determine the risk level. The final assessment can be made according to the principle of maximum membership (i.e., the grade corresponding to the maximum value in $Y$), or by calculating a comprehensive score via weighted summation if quantitative scores are assigned to each grade.

5. Comprehensive Evaluation of Coal Sustainable Supply Chain Risks

Based on the construction of the coal sustainable supply chain risk assessment system in Section 3 and the AHP-FCE comprehensive risk assessment method described in Section 4, the weight of each indicator is determined, and a comprehensive evaluation of the case study coal supply chain is conducted.

5.1. Calculation and Testing of Index Layer Weight

Step 1: Establishing a hierarchical model. Based on the Delphi method and the coefficient of variation method, a two-level evaluation index system is constructed, thereby forming a hierarchical model for risk assessment of the sustainable supply chain in coal enterprises, as illustrated in Figure 3.

Step 2: Construct the judgment matrix and perform the calculations. Ten experts from the risk identification process listed in Table 3, along with four additional scholars with extensive experience in coal industry supply chain research, are invited to assign weights to the various risk assessment indicators for the sustainable supply chain of coal enterprises. The “Weight scoring table for the risk assessment of sustainable supply chain of coal enterprises” is distributed for this purpose. The survey data are collated and analyzed, with experts rating the relative importance of indicators at the same level using the pairwise comparison method and a 1–9 scale, where higher values denote greater importance (Table 5). After the initial round of questionnaire distribution, responses are collected and aggregated, and the results are fed back to the experts for a second round of scoring. This process is repeated a third time. After three rounds of surveys, consensus among expert opinions is achieved, and the distribution and collection of questionnaires concludes, with all 14 questionnaires returned—yielding a 100% response rate, sufficient for analysis. Ultimately, the judgment matrix for the risk assessment index system of the sustainable supply chain of coal enterprises is established (Appendix A).

Step 3: Calculation of indicator weights. The sum of each column in the nine judgment matrices (Appendix A) is calculated (Formula (1)). Each element of the matrix is then divided by the corresponding column sum to obtain a normalized matrix (Formula (2)). The mean value of each row in the normalized matrix is calculated to determine the weight vector w (Formula (3)). The specific weight values for each indicator are presented in

Table 8. Initial weight coefficients of risk assessment indicators for sustainable coal supply chain.

Primary Criteria	Weight	${\mathit{\lambda }}_{\mathbf{max}}$	CI	RI	CR	Sub-Criteria	Weight	${\mathit{\lambda }}_{\mathbf{max}}$	CI	RI	CR	Consistency Check
B1	0.1158	8.3819	0.0546	1.41	0.0387	C1	0.3736	5.0849	0.0212	1.12	0.019	Yes
						C2	0.2486
						C3	0.0798
						C4	0.1755
						C5	0.1224
B2	0.2526					C6	0.1622	5.0755	0.0189	1.2	0.0169	Yes
						C7	0.2847
						C8	0.1075
						C9	0.0918
						C10	0.3537
B3	0.1499					C11	0.2306	4.0459	0.0153	0.9	0.017	Yes
						C12	0.1948
						C13	0.1492
						C14	0.4254
B4	0.0681					C15	0.3774	5.0493	0.0123	1.12	0.011	Yes
						C16	0.2171
						C17	0.0816
						C18	0.1234
						C19	0.2005
B5	0.0464					C20	0.3123	5.0943	0.0236	1.12	0.0211	Yes
						C21	0.2888
						C22	0.1251
						C23	0.0809
						C24	0.1929
B6	0.0886					C25	0.2126	7.1191	0.0198	1.32	0.015	Yes
						C26	0.1308
						C27	0.1612
						C28	0.0787
						C29	0.3028
						C30	0.0688
						C31	0.0451
B7	0.0353					C32	0.098	5.0426	0.0107	1.12	0.0095	Yes
						C33	0.0926
						C34	0.2033
						C35	0.274
						C36	0.3321
B8	0.2434					C37	0.329	4.061	0.0203	0.9	0.0266	Yes
						C38	0.329
						C39	0.2002
						C40	0.1418

.

Step 4: The consistency test. The maximum eigenvalue (${\lambda }_{\max }$) is obtained using Formulas (4) and (5), where n represents the order of the matrix. The consistency index (CI) is computed using Formula (6), while the random consistency index (RI) is referenced from Table 7, according to the order n of the matrix. The consistency ratio (CR) is subsequently calculated using Formula (7). All indicator CR values are found to be less than 0.1 (Table 8), indicating satisfactory consistency and confirming the validity of the weighting results.

5.2. Risk Assessment of Sustainable Coal Supply Chain

5.2.1. Case Study Company

Based on consultations with the fourteen experts referenced in the previous section (Table 3), and corroborated by field investigations, a prominent large-scale energy group is ultimately selected as the empirical research subject for evaluating sustainable supply chain risks in coal enterprises. As an industry leader, this group is committed to sustainable business practices, emphasizing “intelligent coal development” and “low-carbon transformation” as core strategies. The company actively responds to national “dual carbon” policies and serves as a pilot for green supply chain reform, undertaking initiatives such as enhancing energy efficiency, reducing emissions, strengthening environmental protection and resource management, and robustly safeguarding the rights of relevant stakeholders. According to both community feedback and media evaluations, the group enjoys a high level of public satisfaction. In addition to meeting domestic market demands, the group exports products internationally and is actively engaged in global energy and mining trade. The group’s supply chain demonstrates the essential characteristics of a sustainable supply chain and is currently in a stable phase of development. Therefore, this entity is selected as the representative case study for the present research.

In 2024, the group reported annual revenues of RMB 506.4 billion and a profit of RMB 45.2 billion, with raw coal output reaching 252 million tons. The organization employs over 140,000 registered personnel across 60 affiliated units. Its core business segments span the entire coal value chain, comprising upstream coal mining, resource exploration, and ecological restoration projects; midstream operations such as coal washing, chemical processing, and supply chain management; and downstream activities, including coal-fired power generation, steel production, and clean energy subsidiaries (solar and hydrogen energy). Importantly, intelligent coal production accounts for 95% of total output. The group has established twenty-seven green mines and six national-level green factories. The group has pioneered solid waste utilization through gas-fired power generation and coal gangue backfilling technologies, annually consuming 2.6 million tons of coal gangue. In 2024, green electricity trading volume reached 422 million kWh, with new energy installed capacity totaling 5 million kilowatts.

To optimize supply chain management, the group has implemented a supplier tiered management system, established a dedicated supply chain management company, and developed an online tracking system covering equipment, materials, components, and sales. The group actively fulfills its corporate social responsibilities, investing RMB 326 million in rural revitalization initiatives in 2024 and creating 5975 new employment opportunities. Production safety is also prioritized, with significant investment and the establishment of eight national safety culture demonstration enterprises. Governance is overseen by the board of directors under a comprehensive ESG framework, with social responsibility reports published for thirteen consecutive years and subject to third-party verification. The group’s supply chain structure is comprehensive, exhibiting multiple points of risk exposure, and thus provides a representative context for systematically analyzing risk transmission mechanisms and the coupling effects of sustainable supply chain risks within the coal industry.

5.2.2. FCE

The questionnaire survey in this study targets a large energy group, specifically sampling management personnel from its coal production enterprises, coal supply chain management entities, coal internal consumption companies, and other affiliated enterprises. A total of 100 questionnaires are distributed during the period from February to May 2024, and 80 responses are received, resulting in a response rate of 80%. After validity screening, 6 invalid questionnaires are excluded, yielding 74 valid responses. Among the respondents, 58 are management personnel, 31 hold undergraduate or higher degrees, and 50 have more than 10 years of work experience. All interviewees are core staff members within the large energy group A, ensuring a highly targeted and high-quality sample capable of capturing the informed judgments of in-group personnel regarding sustainable supply chain risks. This provides robust data support for constructing the fuzzy comprehensive evaluation model. The specific evaluation steps are as follows:

Step 1: Establishment of the factor set and evaluation set. A set of risk assessment factors is constructed based on the risk indicator system in Figure 3 and then the evaluation comment set is determined. The questionnaire results of the respondents on the sustainable supply chain risk indicators of large coal group A are used as the comment set. The comment set is V = {V₁ (High), V₂ (Moderately High), V₃ (Medium), V₄ (Moderately Low), V₅ (Low)}. The risk assessment questionnaire data are shown in the table. The detailed questionnaire results are presented in

Table 9. Data from the questionnaire on sustainable supply chain risk assessment for a large coal group.

Evaluation of Sub-Criteria	Comments					Statistics
	H	MH	M	ML	L
C1 Economic Cycle Risk	6	11	34	13	10	74
C2 Financing Risk	3	12	27	18	14	74
C3 Financial Constraints	10	17	27	13	7	74
C4 Solvency Risk	8	20	20	13	13	74
C5 R&D Investment	4	17	33	12	8	74
C6 Natural Disasters	5	17	19	20	13	74
C7 Safety Production Training	31	23	6	10	4	74
C8 Mining Technology Upgrade	18	31	10	9	6	74
C9 Equipment Failure	4	10	21	20	19	74
C10 Employee Operation Risk	5	11	21	19	18	74
C11 Mine Ecological Protection	18	22	19	9	6	74
C12 Pollution Reduction Risk	10	18	22	17	7	74
C13 Energy Consumption	1	9	18	26	20	74
C14 Environmental Accidents	2	14	25	17	16	74
C15 Regulatory Compliance	8	14	28	15	9	74
C16 Legal Risk	6	16	23	15	14	74
C17 Local Government Supervision	15	30	18	7	4	74
C18 Coal Industry Access Restrictions	10	29	22	7	6	74
C19 Corporate Sustainable Development Strategy	14	32	19	6	3	74
C20 International Energy Market Price Fluctuations	6	34	28	5	1	74
C21 Coal Industry Cyclicality	4	29	33	7	1	74
C22 Clean Energy Substitution	3	25	36	8	2	74
C23 Bullwhip Effect Risk	6	14	37	13	4	74
C24 Key Customer Disruption	5	13	40	9	7	74
C25 Supplier Legal Compliance	11	23	30	7	3	74
C26 Supplier Quality Management	11	22	32	7	2	74
C27 Core Supplier Disruption	4	21	32	9	8	74
C28 Procurement Cost Risk	8	15	34	10	7	74
C29 Coal Resource Depletion Risk	6	17	34	9	8	74
C30 Capacity Constraint Risk	7	19	33	10	5	74
C31 Logistics and Transportation Risk	6	12	37	11	8	74
C32 Green Innovation Risk	3	14	37	10	10	74
C33 Digital Supply Chain System Stability	6	16	38	9	5	74
C34 IoT Device Reliability	8	19	38	5	4	74
C35 Information System Security	10	20	35	7	2	74
C36 Data Timeliness	11	19	33	8	3	74
C37 Employee Health and Safety	10	23	28	7	6	74
C38 Employee Compensation and Benefits	6	15	42	7	4	74
C39 Community Relations Maintenance	6	20	40	4	4	74
C40 Public Opinion Risk	3	19	37	10	5	74

.

Step 2: Determine the weight. Extract the weight vector of risk assessment indicators from weight Table 8.

The weight of primary criteria:

W = (0.1158, 0.2526, 0.1499, 0.0681, 0.0464, 0.0886, 0.0353, 0.2434)

The weight of sub-criteria:

W₁ = (0.3736, 0.2486, 0.0798, 0.1755, 0.1224)

W₂ = (0.1622, 0.2847, 0.1075, 0.0918, 0.3537)

W₃ = (0.2306, 0.1948, 0.1492, 0.4254)

W₄ = (0.3774, 0.2171, 0.0816, 0.1234, 0.2005)

W₅ = (0.3123, 0.2888, 0.1251, 0.0809, 0.1929)

W₆ = (0.2126, 0.1308, 0.1612, 0.0787, 0.3028, 0.0688, 0.0451)

W₇ = (0.098, 0.0926, 0.2033, 0.274, 0.3321)

W₈ = (0.3290, 0.3290, 0.2002, 0.1418)

Step 3: Construction of fuzzy relation matrix. Construct an assessment matrix based on the questionnaire data in Table 9. The questionnaire data for each sub-criteria provide the frequencies of five rating levels (H, MH, M, ML, L), with a total sample size of 74. The membership of the sub-criteria is calculated by normalizing the frequencies; the frequency of each level is divided by the total sample size of 74. The frequency of sub-criteria C₁ (economic cycle risk): H = 6, MH = 11, M = 34, ML = 13, L = 10, membership calculation:

$$\begin{array}{l}{r}_{c_{11}}=6/74\approx 0.0811\\ r_{c_{12}}=11/74\approx 0.1486\\ r_{c_{13}}=34/74\approx 0.4595\\ r_{c_{14}}=13/74\approx 0.1757\\ r_{c_{15}}=10/74\approx 0.1351\end{array}$$

Therefore, the membership vector of C₁ is

$$[0.0811,0.1486,0.4595,0.1757,0.1351]$$

Similarly, the membership vectors are calculated for all 40 sub-criteria (C₁ to C₄₀), as shown below:

$$R_1=\left[\begin{array}{l}0.0811 0.1486 0.4595 0.1757 0.1351\\ 0.0405 0.1622 0.3649 0.2432 0.1892\\ 0.1351 0.2297 0.3649 0.1757 0.0946\\ 0.1081 0.2703 0.2703 0.1757 0.1757\\ 0.0541 0.2297 0.4459 0.1622 0.1081\end{array}\right]$$

$$R_2=\left[\begin{array}{l}0.0676 0.2297 0.2568 0.2703 0.1757\\ 0.4189 0.3108 0.0811 0.1351 0.0541\\ 0.2432 0.4189 0.1351 0.1216 0.0811\\ 0.0541 0.1351 0.2838 0.2703 0.2568\\ 0.0676 0.1486 0.2838 0.2568 0.2432\end{array}\right]$$

$$R_3=\left[\begin{array}{l}0.2432 0.2973 0.2568 0.1216 0.0811\\ 0.1351 0.2432 0.2973 0.2297 0.0946\\ 0.0135 0.1216 0.2432 0.3514 0.2703\\ 0.0270 0.1892 0.3378 0.2297 0.2162\end{array}\right]$$

$$R_4=\left[\begin{array}{l}0.1081 0.1892 0.3784 0.2027 0.1216\\ 0.0811 0.2162 0.3108 0.2027 0.1892\\ 0.2027 0.4054 0.2432 0.0946 0.0541\\ 0.1351 0.3919 0.2973 0.0946 0.0811\\ 0.1892 0.4324 0.2568 0.0811 0.0405\end{array}\right]$$

$$R_5=\left[\begin{array}{l}0.0811 0.4595 0.3784 0.0676 0.0135\\ 0.0541 0.3919 0.4459 0.0946 0.0135\\ 0.0405 0.3378 0.4865 0.1081 0.0270\\ 0.0811 0.1892 0.5000 0.1757 0.0541\\ 0.0676 0.1757 0.5405 0.1216 0.0946\end{array}\right]$$

$$R_6=\left[\begin{array}{l}0.1486 0.3108 0.4054 0.0946 0.0405\\ 0.1486 0.2973 0.4324 0.0946 0.0270\\ 0.0541 0.2838 0.4324 0.1216 0.1081\\ 0.1081 0.2027 0.4595 0.1351 0.0946\\ 0.0811 0.2297 0.4595 0.1216 0.1081\\ 0.0946 0.2568 0.4459 0.1351 0.0676\\ 0.0811 0.1622 0.5000 0.1486 0.1081\end{array}\right]$$

$$R_7=\left[\begin{array}{l}0.0405 0.1892 0.5000 0.1351 0.1351\\ 0.0811 0.2162 0.5135 0.1216 0.0676\\ 0.1081 0.2568 0.5135 0.0676 0.0541\\ 0.1351 0.2703 0.4730 0.0946 0.0270\\ 0.1486 0.2568 0.4459 0.1081 0.0405\end{array}\right]$$

$$R_8=\left[\begin{array}{l}0.1351 0.3108 0.3784 0.0946 0.0811\\ 0.0811 0.2027 0.5676 0.0946 0.0541\\ 0.0811 0.2703 0.5405 0.0541 0.0541\\ 0.0405 0.2568 0.5000 0.1351 0.0676\end{array}\right]$$

Step 4: Select the weighted average multiplication-bounded operator to calculate the comprehensive evaluation value:

$$\begin{array}{l}{Y}_1=W_1\ast R_1\\ =(0.3736,0.2486,0.0798,0.1755,0.1224)\ast \left[\begin{array}{l}0.0811 0.1486 0.4595 0.1757 0.1351\\ 0.0405 0.1622 0.3649 0.2432 0.1892\\ 0.1351 0.2297 0.3649 0.1757 0.0946\\ 0.1081 0.2703 0.2703 0.1757 0.1757\\ 0.0541 0.2297 0.4459 0.1622 0.1081\end{array}\right]\\ =(0.0767 0.1897 0.3935 0.1908 0.1491)\end{array}$$

Similarly, we can get:

$$\begin{array}{l}{Y}_2=W_2\ast R_2=(0.1852 0.2358 0.2057 0.2110 0.1622)\\ Y_3=W_3\ast R_3=(0.0959 0.2146 0.2971 0.2229 0.1694)\\ Y_4=W_4\ast R_4=(0.1296 0.2865 0.3183 0.1562 0.1095)\\ Y_5=W_5\ast R_5=(0.0656 0.3481 0.4525 0.0996 0.0341)\\ Y_6=W_6\ast R_6=(0.1030 0.2612 0.4410 0.1156 0.0793)\\ Y_7=W_7\ast R_7=(0.1199 0.2501 0.4786 0.1001 0.0514)\\ Y_8=W_8\ast R_8=(0.0931 0.2595 0.4903 0.0922 0.0649)\end{array}$$

In summary, the evaluation matrix of the first-level indicators is:

$$Y=\left[\begin{array}{l}0.0767 0.1897 0.3935 0.1908 0.1491\\ 0.1852 0.2358 0.2057 0.2110 0.1622\\ 0.0959 0.2146 0.2971 0.2229 0.1694\\ 0.1296 0.2865 0.3183 0.1562 0.1095\\ 0.0656 0.3481 0.4525 0.0996 0.0341\\ 0.1030 0.2612 0.4410 0.1156 0.0793\\ 0.1199 0.2501 0.4786 0.1001 0.0514\\ 0.0931 0.2595 0.4903 0.0922 0.0649\end{array}\right]$$

Step 5: Calculate the evaluation results to get the comprehensive fuzzy evaluation results:

$$A=W\ast Y=(0.1179 0.2444 0.3600 0.1602 0.1173)$$

The five evaluation levels V = {V₁, V₂, V₃, V₄, V₅} are assigned quantitative values of {100, 80, 60, 40, 20}, respectively. Using a simple weighted average,

$$0.1179\ast 100+0.2444\ast 80+0.36\ast 60+0.1602\ast 40+0.1173\ast 20=61.71$$

The comprehensive risk evaluation score is calculated as 61.71. The corresponding scores for each sub-criteria are presented in

Table 10. Risk assessment scores of sustainable supply chain of a large coal group.

A large-scale coal group sustainable supply chain risk assessment score (61.7130)	Primary Criteria	Score	Sub-Criteria	Score
	B1 Economic Risk	57.0771	C1 Economic Cycle Risk	57.2973
			C2 Financing Risk	52.4324
			C3 Financial Constraints	62.7027
			C4 Solvency Risk	59.1892
			C5 R&D Investment	59.1892
	B2 Safety Operation Risk	61.4103	C6 Natural Disasters	54.8649
			C7 Safety Production Training	78.1081
			C8 Mining Technology Upgrade	72.4324
			C9 Equipment Failure	49.1892
			C10 Employee Operation Risk	50.8108
	B3 Ecological Risk	56.8925	C11 Mine Ecological Protection	70.0000
			C12 Pollution Reduction Risk	61.8919
			C13 Energy Consumption	45.1351
			C14 Environmental Accidents	51.6216
	B4 Policy Risk	63.4079	C15 Regulatory Compliance	59.1892
			C16 Legal Risk	55.9459
			C17 Local Government Supervision	72.1622
			C18 Coal Industry Access Restrictions	68.1081
			C19 Corporate Sustainable Development Strategy	72.9730
	B5 Demand Risk	66.2291	C20 International Energy Market Price Fluctuations	70.5405
			C21 Coal Industry Cyclicality	67.5676
			C22 Clean Energy Substitution	65.1351
			C23 Bullwhip Effect Risk	61.3514
			C24 Key Customer Disruption	60.0000
	B6 Sustainable Supply Risk	63.8610	C25 Supplier Legal Compliance	68.6486
			C26 Supplier Quality Management	68.9189
			C27 Core Supplier Disruption	61.0811
			C28 Procurement Cost Risk	61.8919
			C29 Coal Resource Depletion Risk	61.0811
			C30 Capacity Constraint Risk	63.5135
			C31 Logistics and Transportation Risk	59.1892
	B7 IT Risk	65.7402	C32 Green Innovation Risk	57.2973
			C33 Digital Supply Chain System Stability	62.4324
			C34 IoT Device Reliability	65.9459
			C35 Information System Security	67.8378
			C36 Data Timeliness	67.2973
	B8 Social Risk	64.4749	C37 Employee Health and Safety	66.4865
			C38 Employee Compensation and Benefits	63.2432
			C39 Community Relations Maintenance	65.4054
			C40 Public Opinion Risk	61.3514

.

6. Results and Discussion

6.1. Result

6.1.1. Overall Results Analysis

According to the risk assessment criteria outlined in the previous section, the overall risk level for the coal power group is classified as “moderate”. As can be seen from Figure 4, the ranking of the primary risk dimensions is as follows: demand risk > information technology risk > social risk > sustainable supply risk > management policy risk > operational safety risk > economic risk > ecological and environmental risk. Specifically, demand risk (66.23), information technology risk (65.74), and social risk (64.47) are identified as the most critical threats, jointly accounting for 47.2% of the total risk weight.

Specifically, demand risk is predominant due to market uncertainties associated with the ongoing transformation of the energy structure. Information technology risk reflects structural challenges related to system failures and data security in the context of supply chain digitalization. Social risk is closely linked to safeguarding the rights of the company’s employees and maintaining positive community relations. Notably, sustainable supply risk is shaped by both resource depletion and vulnerabilities within the supplier network, whereas management policy risk is mitigated by the effectiveness of the current industry regulatory framework. The comparatively lower levels of operational safety and economic risk are attributable to the company’s standardized safety management practices and the operational stability afforded by its state-owned capital structure. To optimize risk prevention and control, enterprises should prioritize dynamic demand-side monitoring, the upgrading of information infrastructure, and the establishment of effective stakeholder coordination mechanisms.

6.1.2. Analysis of Secondary Indicator Results

Figure 5 indicates that the economic risks faced by Coal Group A are primarily characterized by financing dependence (financial constraints and R&D investment limitations) and debt transmission risk (debt repayment pressure). Prioritization of fund allocation restricts the autonomy of R&D investment and market expansion. It is recommended that the group optimize its financing structure and debt management by establishing a diversified financing matrix and implementing a dynamic financial early warning system.

Figure 6 shows that the safety operation risks are mainly concentrated in two areas: gaps in the training system (insufficient understanding of safety norms) and lagging technology iteration (mining equipment efficiency). Risk arises from (1) inadequate employee safety awareness and emergency response capabilities, leading to operational risks; and (2) a mismatch between the standardization level of mining technology and equipment update cycles, increasing engineering geological risks. To systematically mitigate safety risks, it is advised to develop a tiered safety training system and implement a technology iteration evaluation mechanism.

Figure 7 reveals that ecological and environmental risks center on the efficiency of ecological restoration (land reclamation rates) and pollution control (intensity of the “three wastes” emissions). Risk transmission occurs through (1) exceeding environmental carrying capacity due to mining activities, and (2) cumulative pollution from delayed implementation of end-of-pipe treatments and insufficient clean production processes. It is recommended to establish a comprehensive environmental control mechanism based on the pressure–state–response model, alongside promoting clean production transformation for effective environmental risk management.

Figure 8 shows that management policy risk features the following dual constraints: absence of strategic planning (insufficient ESG execution) and regulatory pressure (intensity of government–enterprise relations). Risks are transmitted via (1) an insufficient match between the company’s sustainability strategy and stakeholder expectations, delaying green transformation; and (2) time lags between regulatory changes and corporate compliance adaptation, increasing institutional transaction costs. To address these, an ESG strategic framework and a collaborative governance system with government entities are recommended.

As illustrated in Figure 9, demand risk manifests as market structure homogeneity (reduced international energy price elasticity) and enhanced substitution effects (higher clean energy penetration). Risk generation is driven by (1) the interaction between energy price cyclicality and supply chain bullwhip effects, and (2) policy-induced energy consumption shifts reducing demand rigidity. Structurally hedging demand-side risks requires establishing a demand elasticity adjustment system and an energy substitution buffer mechanism.

Figure 10 indicates that sustainable supply risk is concentrated in supplier network vulnerability and capacity rigidity constraints. Risk is transmitted through (1) decoupling of supplier financial health and ESG performance, increasing supply chain disruption risk; and (2) combined impacts of production equipment efficiency thresholds and strict environmental regulations limiting capacity growth. To address this, a supplier resilience assessment matrix and a framework for improving capacity elasticity are recommended.

Figure 11 highlights that information technology risk is primarily associated with data chain disruption and network vulnerabilities. The risk mechanism includes (1) incompatibility between information heterogeneity and dynamic decision-making needs across supply chain nodes, amplifying the bullwhip effect; and (2) the combined impact of system vulnerabilities and human error, increasing the risk to data asset security. Effective mitigation involves building an intelligent decision support system and a layered network security defense framework.

Finally, Figure 12 demonstrates that social risks are focused on the occupational exposure index (health and safety thresholds) and negative externalities impacting community relations (relationship maintenance costs). Risks are transmitted via (1) failures in safety management and occupational health systems, increasing both accident and occupational disease rates; and (2) conflicts between mining-related ecological footprints and community development demands, raising the risk of losing social license to operate. A multidimensional approach—establishing an occupational health and safety resilience system and a co-governance platform for social responsibility—is recommended for effective social risk management.

6.2. Discussion

This study elucidates the multidimensional risk profile of sustainable coal supply chains, providing evidence that both corroborates and challenges existing theoretical frameworks. The deployment of the integrated analytic hierarchy process–fuzzy comprehensive evaluation (AHP–FCE) model not only validates the efficacy of multi-criteria decision-making (MCDM) techniques in addressing complex risk interdependencies but also extends the methodological frontiers of conventional risk assessment paradigms.

The combination of AHP and FCE addresses the historical limitation of coal supply chain risk research, which has frequently been restricted to single-dimensional analyses (e.g., environmental or economic risks). By quantitatively determining hierarchical weights for forty sub-indicators across eight risk dimensions, this study demonstrates that the application of fuzzy logic can effectively resolve ambiguity in expert judgment, particularly in contexts characterized by interdependent risks such as policy compliance and technological obsolescence. The robustness of the weighting process is substantiated by consistency ratios (CR < 0.1) across all judgment matrices, which aligns with Ganguly and Kumar’s findings regarding the application of FAHP in automotive supply chains [22]. Furthermore, this research advances beyond the static confines of conventional AHP frameworks [26] by incorporating a dynamic feedback mechanism through a three-round Delphi process, thereby achieving methodological innovation.

The prominence of demand risk and information technology risk underscores the disruptive impacts of energy transitions and digitalization on coal supply chains. While this finding aligns with Zhang et al. [8] concerning shifts in demand elasticity resulting from clean energy integration, the present study further reveals vulnerabilities in IT infrastructure—a dimension that has been largely overlooked in the previous coal sector research. Conversely, the relatively low economic risk score (57.08) contrasts with the findings of Alshehri et al. [28] in the Saudi manufacturing context, suggesting that the financial resilience of state-owned enterprises mitigates market volatility but may reinforce technological path dependency. The considerable exposure to sustainable supply continuity risks and challenges regarding the “social license to operate” highlights an urgent need for strategic realignment. For instance, the observed weak correlation between supplier ESG performance and financial health (as revealed by supply chain cluster analysis) underscores the necessity of implementing a blockchain-based traceability system to reinforce compliance [59]. Furthermore, the coexistence of advanced safety training initiatives and persistent operational risks indicates a systemic disconnect between procedural adherence and on-site execution—a gap that could be addressed through the adoption of real-time IoT monitoring frameworks.

Although the case of company A provides nuanced micro-level insights, the state-owned and large-scale nature of the enterprise may limit the generalizability of these findings to small, medium-sized, or privately owned coal companies. Additionally, the static nature of the AHP–FCE approach constrains its ability to capture the dynamic evolution of risks, such as those arising from climate-induced logistics disruptions. Future research should integrate system dynamics modeling to simulate nonlinear risk transmission pathways [56] and broaden case studies to encompass cross-border supply chains (e.g., the Indonesia–Australia coal corridor), thereby enhancing the global applicability and robustness of the proposed framework.

7. Conclusions

Driven by the ongoing global energy transition and the pursuit of carbon neutrality, sustainability risks inherent in coal supply chains have emerged as a core challenge constraining the decarbonization of high-carbon industries. This study conducts an empirical investigation using Group A, a leading coal enterprise in China, as a representative case. By integrating the AHP and FCE methods, a comprehensive risk assessment framework is developed, encompassing eight primary dimensions and forty secondary indicators: economic, operational safety, ecological and environmental, management policy, demand, sustainable supply, information technology, and social risks.

The results indicate that the overall risk score for the case enterprise is 61.71, corresponding to a “medium” risk level. Among the evaluated dimensions, demand risk (66.23), information technology risk (65.74), and social risk (64.47) are identified as the most significant contributors. Specifically, demand risk is primarily driven by the volatility of international energy prices and market substitution effects resulting from clean energy alternatives; information technology risk predominantly arises from the instability of digitalized supply chain systems and reliability issues in Internet of Things (IoT) devices; and social risk is reflected in insufficient employee health protection measures and escalating community relations maintenance costs. Furthermore, financing constraints within economic risks (62.70), delayed equipment upgrades in operational safety risks (49.19), and deficiencies in end-of-pipe controls within ecological and environmental risks (51.62) further illuminate the structural contradictions hampering technological, regulatory, and market coordination across coal supply chains.

Based on these results, the following policy recommendations are proposed: First, optimize demand-side management by establishing shock absorption mechanisms for energy substitution, underpinned by elastic forecasting models; hedge against international price volatility through market diversification (for example, by expanding into emerging Southeast Asian markets); and accelerate research and development of low-carbon technologies such as hydrogen-based steelmaking to better anticipate structural disruptions in demand. Second, enhance digital risk management by promoting the integration of blockchain technology and industrial internet solutions; develop decentralized data-sharing platforms (such as supply chain traceability systems enabled by smart contracts); and strengthen cybersecurity infrastructure, including mandatory deployment of intrusion detection systems (IDSs) and quantum encryption at critical junctures. Third, establish a co-governance mechanism for social risks by incorporating occupational health metrics (such as dust exposure limits and incidence rates of occupational diseases) into the ESG disclosure framework and dynamically linking ecological compensation to local economic development through “mining community benefit sharing agreements”. Finally, promote coordinated policy innovation by improving the internalization of environmental costs; pilot the linkage between coal resource tax and ecological restoration funds; and integrate compliance costs associated with the EU carbon border adjustment mechanism (CBAM) into pricing models for export-oriented supply chains to mitigate cost-shifting risks arising from international trade frictions.

Future research directions should include comparative analyses of risk heterogeneity in coal supply chains across emerging economies (e.g., China, India, and South Africa); quantitative assessments of the impacts of geopolitical and resource endowment variations on risk transmission pathways; application of machine learning models and real-time operational data to develop dynamic risk warning and simulation platforms, enabling the capture of nonlinear effects of extreme climate events on logistics networks; and coupling life cycle assessment (LCA) with complex network theory to simulate the synergistic impacts of policy interventions and technological disruptions, thus elucidating threshold effects and critical junctures in risk propagation.

Although this study centers on large state-owned coal enterprises in China, the AHP-FCE hybrid methodological framework established herein provides a robust scientific foundation for advancing risk governance in global high-carbon supply chains. Future research should further incorporate emerging risk categories and expand empirical analyses to encompass small- and medium-sized coal mines, as well as transnational supply chains, thereby enhancing the universality and contemporaneity of the model.