Fuzzy comprehensive evaluation of rock burst risk based on combined subjective and objective weighting Authors: Jing Zhang, Beifang Wang, Yuanhao Lv, Xinyi Liu Published: 2026 ABSTRACT AbstractDuring deep coal mining, the rock mass is subjected to high in-situ stress and strong mining disturbances. Its mechanical properties are more sensitive than those of shallow rock mass, which can easily induce rock burst disasters, seriously restricting the safety and efficient production of the mine. To address this problem, this study takes the 718 working face of Hongyang No. 3 Mine as the engineering background, integrates theoretical analysis, mathematical modeling, and field application to deeply analyze the main factors of rock burst. And establishes a rock burst risk evaluation mathematical model based on subjective and objective combination weighting. The research results show that the main control factors for rock burst include natural geological factors, mining technology factors and organizational management factors, with weights of 0.62322, 0.23949 and 0.13729 respectively. Among these, rock burst tendency, coal pillar width, and prevention and control investment are identified as the most influential factors. Based on the analytic hierarchy process and the entropy weight method, the subjective and objective weights are combined, and the comprehensive weights are obtained. On this basis, a multi-level fuzzy comprehensive evaluation model for rock burst risk is constructed. After evaluation, the 718 working face is classified as a medium rock burst risk level, which is consistent with field observations, thereby verifying the reliability and applicability of the proposed model. This study effectively integrates expert experience, test results, and actual on-site conditions, overcomes the one-sidedness of a single weighting method, and thereby provides a scientific basis for the quantitative risk evaluation and optimization of rock burst prevention strategies under similar mining conditions.Keywords: Deep mining, Rock burst, Combined weighting, Fuzzy comprehensive evaluationSubject terms: Energy science and technology, Engineering, Natural hazards, Solid Earth sciences Introduction With increasing mining depth in coal mines, the original rock stress level by the surrounding rock has significantly increased. Under the combined influence of high ground stress and intense mining disturbances, the stress environment becomes increasingly complex, leading to a higher incidence of rock burst disasters. Moreover, the mechanisms and inducing factors of rock burst exhibit pronounced nonlinear and multi-factor coupling characteristics. Extensive studies have been conducted on the occurrence mechanisms and risk evaluation of rock burst from various perspectives. In terms of the mechanism of rock burst occurrence, Li et al.1 proposed a far-middle-near-field division standard for the influence range of mining activities, revealed the influence laws of mining disturbances on roadway stability, and also established a theoretical model for the occurrence of rock burst considering damage effects. Dou et al.2 systematically studied the mechanism of rock burst during horizontal subsection mining based on the special mining condition of steeply inclined and extremely thick coal seam. Pan et al.3,4 explored the occurrence laws of rock burst, classified the main influencing factors leading to the occurrence of rock burst, and proposed prevention and control countermeasures. Chen et al. 5 used advanced numerical calculation methods to deeply analyze the influence of bedrock pressure on roof movement in front of the work, and established a theoretical model of the mechanism of rock burst in mining roadways under hard roof conditions. Yan et al. 6clarified the underlying mechanisms of rock bursts in negative coal pillar roadways induced by dynamic load disturbances, and propose a methodology for identifying rock burst hazards. In terms of the evolution of rock burst risk, the three-dimensional geological model built by Jiang7 et al. based on borehole data, combined with grid optimization and FDEM numerical simulation, achieved fine delineation of rock burst risk areas. Cui8 et al. employed a spatial scanning method to divide the area into grids, and proposed an intelligent early warning grading method to effectively identify impact risk areas based on microseismic data. Wang et al.9 analyzed the occurrence laws of rock burst from three aspects: the influence distance of the fault structure, the spatial structure under the influence of the fault, the spatial structure of the overburden, and the mutual disturbance of multiple hard roof. Zhang10 et al. employed the advancing distance of the working face as a prediction basis to reveal the potential spatial correlation of the occurrence of rock burst. In terms of rock burst risk evaluation, Agrawal11 et al. analyzed the variability of coal mechanics, elasticity and reservoir properties caused by lithologic heterogeneity, resulting in the probability of rock burst. Rong12 et al. adopted the elastic energy density calculation method and used residual energy to grade the burst risk of coal seam, providing a calculation method for rapid evaluation of burst risk before mining and production. Tian13 et al. proposed a rock-burst disaster factor that takes into account mining stress and ground stress, which confirmed the possibility of rock-burst occurring in deep soft coal seams without dynamic load. Shi et al.14 used explanatory structural modeling and decision-making testing and evaluation laboratory methods to systematically identify and quantitatively analyze the disaster-causing factors of coal mine rock burst accidents. Based on a systematic analysis of hundreds of coal mines with rock burst, Tan et al.15 proposed four types of objective main control factors and three types of artificial main control factors that lead to rock burst. Zeng et al.16 determined the fuzzy evaluation index for the factors influencing the risk of rock burst and established a rock burst risk evaluation model. Zhu et al.17 used the fuzzy comprehensive evaluation method and the analytic hierarchy process to establish a rock burst risk evaluation method based on the geodynamic environment. Zhang 18 used factor analysis method to systematically identify the factors affecting rock burst, determined the main risk sources and their action mechanisms, and built a comprehensive evaluation system including two first-level indicators and seven second-level indicators. Although substantial progress has been made in the risk prediction and evaluation of rock burst, several limitations remain in existing studies. First, many evaluation systems mainly focus on geological or mining factors, while the combined effects of natural conditions, mining layout, and prevention management are not sufficiently considered. Second, some weighting methods rely heavily on expert judgment, which may introduce subjectivity, whereas purely data-driven methods may ignore engineering experience and the physical meaning of qualitative indicators. Third, the hierarchical coupling relationship among multi-source factors under deep high-stress mining conditions has not been adequately quantified. To overcome these limitations, this study proposes a multi-level fuzzy comprehensive evaluation model based on combined subjective–objective weighting. The analytic hierarchy process is used to reflect expert knowledge, while the entropy weight method is introduced to characterize data variability, thereby improving the rationality, objectivity, and applicability of rock burst risk evaluation. Engineering background Hongyang No. 3 Mine is located at the junction of Sujiatun District and Dengta City, Shenyang City, Liaoning Province. The coalfield area is about 30 km2 and the geological reserves reach 158 million tons. The main coal mining layers are coal 3, coal 7, coal 12–1 and coal 13, and the formation inclination angle is 3° ~ 6°. It is classified as a typical deep coal mine prone to rock burst. The 718 fully mechanized working face is located on the east side of the goafs of the 702, 704 and 706 working faces. The working face adopts a comprehensive mechanized one-time mining full height mining method. The strike length is 601 m, the dip length is 180 m, and the average burial depth is 900 m. The overall structure of the working face is a monoclinal structure, with moderate geological conditions, developed faults, and fault activity and dense distribution significantly weaken the integrity of the coal and rock mass, forming stress concentration areas and energy accumulation zones, which can easily induce rock burst. The main mining coal seam of No. 718 Working Face is a composite coal seam, consisting of coal 7-1, coal 7-2 and coal 7-3, with thicknesses of 0.30 m , 0.45 m and 1.28 m respectively. The coal seam is well stratified, has poor integrity, and shows weak burst tendency. The direct roof of the coal seam is black mudstone, with an average thickness of 11.2 m, and subsides with mining. The main roof is fine sandstone with hard lithology and an average thickness of 7.6 m. It responds significantly to roof collapse and mining disturbances, and the roof has medium rock burst tendency. During mining, problems such as broken surrounding rock of the roadway and sinking of the roof have occurred, as shown in Fig. 1. Therefore, it is urgent to carry out a comprehensive evaluation of the risk of rock burst in the 718 fully mechanized working face, determine the risk level, and formulate corresponding prevention and control measures. Deformation of surrounding rock in roadway. Analysis of influencing factors inducing rock burst Based on a comprehensive analysis of mining activities at the 718 working face, the main control factors for rock burst risk evaluation can be summarized into three categories: natural geological factors, mining technology factors and organizational management factors. Natural geological factors Among the natural geological factors, the most important factors are rock burst tendency, roof rock structure, geological structure and mining depth. Rock burst tendency The tendency of rock burst risk is mainly determined by the physical and mechanical characteristics of the coal and rock itself, among which key parameters such as compressive strength and tensile strength have a significant impact on the mechanical behavior of the coal and rock. In general, the higher the strength and elastic modulus of the coal body, the more elastic energy it can accumulate during the loading process, and the potential risk of rock burst will increase accordingly. At present, rock burst tendency is commonly evaluated using four key indicators: uniaxial compressive strength, elastic energy index, impact energy index, and dynamic failure duration. Based on these indicators, coal seams can be classified into four categories: strong, medium, weak, and no rock burst tendency. The No. 7 coal seam of Hongyang No. 3 Mine has medium rock burst tendency, under certain stress conditions and mining disturbances, there is still the possibility of inducing rock burst. Roof rock structure The structural characteristics of the roof rock strata, especially the hard and thick sandstone roof developed above the coal seam, are one of the important controlling factors affecting the occurrence of rock burst. The mechanism is that hard and thick sandstone has strong elastic energy accumulation and burst release characteristics under mining disturbance. When the roof breaks, bends or slips and fails, the stored elastic energy is instantly released, producing strong vibration, thereby inducing roof coal seam type or roof type rock burst, as shown in Fig. 2. The occurrence of rock burst is mainly controlled by the structural combination and mechanical properties of the rock strata within 100 m above the coal seam. Among them, the high-strength and large-thickness sandstone layer plays a leading role in the accumulation and release of rock burst energy. Mechanism of rock burst induced by hard thick sandstone roof. According to the value selection rules and calculation formula (1) of geological evaluation indicators in the comprehensive index method for impact risk evaluation, the theoretical value of the characteristic parameter of roof thickness in the area of 718 working face is Lst = 90.2, which is greater than 90, indicating that the roof of the working face has strong rigidity and energy storage conditions, which has a significant impact and promotion on the occurrence of rock burst disasters. where: n -The number of rock formations within 100 m above the coal seam; hi -The thickness of the ith layer within the roof of 100 m; ri -The weak plane coefficient of the ith layer. Geological structure Geological structures such as faults and folds significantly influence the stress distribution and deformation characteristics of coal and rock masses. These structures not only alter the mechanical integrity of the rock mass but also lead to the accumulation of residual tectonic stress. Under the superposition of mining-induced stress, local stress concentration and energy accumulation are likely to occur, thereby increasing the risk of rock burst. Practice has shown that rock burst events often occur concentrated in structural stress abnormal areas such as syncline axes, near faults, sudden change zones of coal seam inclination and fold development areas, as shown in Fig. 3. Relationship between occurrence of rock burst and geological structure. The overall 718 working face is monoclinal structural characteristics, with stable formation dip but significantly affected locally by fault activity. Three main faults are developed in the area: Fy1, Fy2 and Fh13, with drops of 5 m, 5 m and 3 m respectively. In addition to the above-mentioned main faults, the working face is also controlled by multiple fault zones such as Fy6, Fy3, Fh2, and Fh4 , forming a superimposed area of structural stress, which greatly increases the risk of rock burst. According to the standards of complexity of various geological factors, the complexity of geological structures is divided into: relatively simple, moderately complex, more complex, and complex based on the number and development degree of faults and folds, and the rock enclosure and the degree of damage to coal bodies. After grading and quantifying it, the risk grading standards for geological structure complexity are shown in Table 1. Table 1Risk grading standards for geological structure complexity.Complex degree of geological structureSimpleModerately complexMore complexComplexLevelLevel I(0.25)Level II(0.5)Level III(0.75)Level IV(1.0) Risk grading standards for geological structure complexity. Level I (0.25) Level II (0.5) Level III (0.75) Level IV (1.0) Mining depth With the increase of mining depth, the self-weight stress of the overlying rock mass increases significantly, and the in-situ stress level borne by the coal and rock mass increases simultaneously, which leads to an enhanced accumulation of elastic energy within it. When the stress increases to the ultimate bearing capacity of the coal body, the coal body is more prone to instability and failure, thus inducing rock burst. At this time, the probability of occurrence of rock burst and its energy release intensity increase with the increase of burial depth, showing a clear positive correlation , as shown in Fig. 4. Statistical analyses indicate that rock burst risk exhibits a staged relationship with mining depth. When the depth is less than 350 m, rock burst rarely occurs; between 350 and 500 m, the probability gradually increases; and beyond 500 m, the risk rises sharply. At depths exceeding 800 m, the coal-rock system typically enters a high-energy instability state. The 718 working face, with an average burial depth of approximately 900 m, is therefore highly susceptible to rock burst. Relationship between mining depth and rock burst risk. Mining technical factors The most important factors among mining technical factors are coal pillar width, goaf and roof management methods. Coal pillar width Coal pillar width directly affects the distribution and concentration of stress around the working face. If the coal pillar is too narrow, its bearing capacity is insufficient, making it prone to shear failure or crush instability under high stress conditions. Conversely, excessively wide coal pillars hinder stress release and may result in severe stress concentration zones. In particular, under conditions such as isolated or peninsula coal pillars, multi-directional stress superposition becomes significant. The interaction between coal pillars and surrounding goaf boundaries further intensifies stress concentration, increasing the likelihood of sudden instability in coal and rock masses 19, as shown in Fig. 5. Schematic diagram of coal pillar stress distribution form. Combined with the basic roof of the 718 working face is 7.6 m fine sandstone, and based on the reserved deformation size of the surrounding rock of the tunnel and the control requirements for bolt failure, the width of the coal pillar reserved for the 718 working face is 8 m. On-site stress monitoring shows that the coal pillar is in a superimposed state of two-way bearing pressures, and the peak vertical stress reaches 22.3 MPa. The 718 working face is adjacent to the goaf of 702, 704, 706, and 708. Due to the fact that the protective coal pillars of each stop-line do not form a good alignment relationship, the stress distribution in the area is uneven and the formation of multiple high stress concentration zones significantly increase the rock burst risk. Goaf The spatial relationship between a working face and adjacent goafs can generally be classified into four types: solid coal working face, one side of goaf, two sides of goaf, and an island-type working face with goaf on three or more sides. For a solid-coal working face, the stress field is mainly controlled by the original rock stress, and the stress distribution is relatively uniform. When goafs exist on one or more sides, the pressure-relief effect of the goaf and the residual abutment pressure interact, resulting in a pronounced stress gradient around the working face. Under conditions with goafs on both sides or three or more sides, the coal mass is subjected to superimposed multi-directional abutment pressures, which readily leads to the formation of high-stress concentration zones. During the advancement of the working face, roof collapse and goaf compaction cause stress redistribution, transferring and superimposing concentrated stress toward the coal wall and the front of the working face. Meanwhile, mining-induced dynamic stress interacts with the residual static stress around the goaf, accelerating local stress accumulation. Once the stress exceeds the ultimate strength of the coal-rock mass, rock burst may be induced. This effect is particularly significant in island-type working faces, where the surrounding goafs limit the available pressure-relief space and hinder effective stress release, thereby markedly increasing rock burst risk, as shown in Fig. 620. Therefore, accurately characterizing the distribution of goafs and its influence on the stress field is essential for rock burst prediction and prevention. Superposition of pressure for excavating into goaf. The 718 working face is located on the east side of the goafs of the 702, 704, and 706 working faces. Under the combined influence of adjacent goafs, fault structures, and roadway intersections, multiple high-stress concentration zones are formed, which further increases the likelihood of rock burst. The relationship between the working face and adjacent goaf can be divided into: solid coal working face, goaf on one side, goaf on both sides, and goaf on three sides and more. The risk classification standards for the relationship between the working face and adjacent goaf are shown in Table 2. Table 2Grading standards for the relationship between working face and adjacent goaf.Relationship between working face and adjacent goaf areaSolid coal sideOne sideBoth sidesThree and more sidesLevelLevel I(0.25)Level II(0.5)Level III(0.75)Level IV(1.0) Grading standards for the relationship between working face and adjacent goaf. Level I (0.25) Level II (0.5) Level III (0.75) Level IV (1.0) Roof management method Roof management methods play an important role in controlling rock burst risk. When a hard main roof undergoes sudden failure, the elastic energy stored in the coal and rock mass is rapidly released, generating strong dynamic loads. This process can significantly increase the risk of rock burst. Therefore, the risk classification of roof management methods is carried out, as shown in Table 3. Table 3Risk grading standards for roof management methods.Roof managementNatural collapseReturn pillar roofDeep-hole blastingHigh pressure water injectionLevelLevel IV(1.0)Level III(0.75)Level II(0.5)Level I(0.25) Risk grading standards for roof management methods. Level IV (1.0) Level III (0.75) Level II (0.5) Level I (0.25) At present, the total collapse method is generally used to manage the roof in Hongyang No. 3 Mine. This method may easily cause the risk of rock burst when the roof lithology is hard and the thickness is large. Especially when the overlying rock layer contains a bottom conglomerate layer, the collapse process is often accompanied by the sudden fall of a large rock mass, resulting in a significant increase in local dynamic loads. In addition, the main roof of the No. 7 coal seam is fine sandstone with a thickness of about 7.6 m. When pressed, it is easy to form a large-scale suspended roof structure, resulting in a sharp increase in abutment pressure on coal pillars and induces local stress concentration, thereby further improving the possibility of rock burst risk. Organizational management factors Organizational and management factors mainly include prevention and control investment, prevention and control safety measures and prevention and control emergency plans. The grading standards are formulated with reference to the safety management regulations on rock burst prevention and control in the " Detailed Rules for the Prevention and Control of Rock Burst in Coal Mines", the " Safety Regulations in Coal Mine", and the relevant emergency management requirements of the " Measures for the Administration of Contingency Plans for Work Safety Accidents". Prevention and control investment Scientific and reasonable roadway support is an important engineering means to reduce the risk of rock burst. By optimizing the supporting parameters and structural forms, the overall strength and residual bearing capacity of the surrounding rock can be effectively improved, so that the coal and rock can still have certain bearing performance after failure, thereby delaying the energy release process, weakening the concentrated and instantaneous release of rock burst loads, and significantly reducing rock burst energy and strength. The grading evaluation standards are shown in Table 4. Table 4Classification standard for prevention and control investment.Evaluation indexLevelEvaluation criteriaPrevention and control investmentLevel I (0.25)Good support effect and large investment in prevention and controlLevel II (0.5)Good support effect and reasonable prevention and control investmentLevel III (0.75)The support effect is average, and the investment in prevention and control is insufficientLevel IV (1.0)Poor support effect, too little investment in prediction and prevention Classification standard for prevention and control investment. The investment in prevention and control of Hongyang No. 3 Mine is reasonable. In terms of support, the support effect is better by optimizing the support methods of the floor and two sides and improving the support intensity. In terms of monitoring and early warning, the mine conducted in-situ stress tests. Effectively improve the efficiency and pertinence of rock burst prevention construction. To sum up, the prevention and control investment in the 718 Working Face is relatively in place and belongs to Level II. Prevention and control safety measures Rock burst occurrence is primarily controlled by the imbalance between stress accumulation and the bearing capacity of coal and rock masses. Preventive safety measures are intended to reduce local stress concentration, promote stress transfer, and enhance the energy absorption capacity of the surrounding rock system. The effectiveness of these measures depends largely on whether they are implemented in accordance with the designed parameters and site-specific risk conditions. Meanwhile, zoning-based risk management can identify areas with different hazard levels, enabling differentiated control strategies and avoiding the unreasonable allocation of prevention resources. Personnel access restrictions further reduce exposure in high-risk zones and help minimize potential accident losses. Therefore, systematic, standardized, and traceable rock burst prevention measures can effectively reduce both the probability and intensity of rock burst disasters. The grading criteria for rock burst prevention and safety measures are shown in Table 5. Table 5Classification standard for prevention and control safety measures.Evaluation indexLevelEvaluation criteriaPrevention and control safety measuresLevel I(0.25)Optimize the mining layout, strictly delineate dangerous areas and manage them in different areas. Staff limit management is strictly enforced, and individual protective wear is completeLevel II(0.5)The measures comply with the requirements of operating procedures and can cover major dangerous areas. Staff limit management is not strictly enforced, and individual protective wear is incompleteLevel III(0.75)Only target known areas with strong rock burst, and insufficient attention is paid to non-dangerous or weakly dangerous areas. Staff limit management is not strictly enforced, and individual protective wear is incompleteLevel IV(1.0)There is no systematic document of anti-blast measures, and personnel and equipment in rock burst hazardous areas are operating in disorder, and there is no personal protection Classification standard for prevention and control safety measures. Level I (0.25) Level II (0.5) Level III (0.75) Level IV (1.0) The Hongyang No. 3 Mine has set up a special rock burst prevention office and is equipped with a professional and technical team with rich experience in rock burst prevention. Mine employees have a strong awareness of rock burst prevention and sound mine pressure monitoring and dynamic management systems. Comprehensive evaluation shows that the 718 working face has reached Level II in terms of rock burst prevention and safety protection measures. Prevention and control emergency plan Rock burst is sudden and instantaneous destructive, and can easily induce secondary disasters such as roof falls and gas gushes. No matter how complete monitoring and pressure relief measures are, the risk cannot be reduced to zero. The core value of emergency plans lies in: improving personnel response capabilities through drills, disaster avoidance routes and emergency broadcast systems beforehand; minimizing casualties through clear command structures and self-rescue and mutual rescue processes during the process; and preventing the expansion of the disaster through joint rescue afterwards. Mines lacking effective plans will inevitably fall into chaos once a rock burst occurs, and the consequences of the accident will be multiplied. Therefore, the emergency plan is the last line of defense to control losses and needs to be used as one of the rock burst risk evaluation indicators. The specific grading standards are shown in Table 6. Table 6Classification standards for emergency plans.Evaluation indexLevelEvaluation criteriaPrevention and control emergency measuresLevel I(0.25)The plan is complete and in compliance with standards. Clear disaster avoidance routes, emergency broadcast systems, and evacuation chambers are set up underground to operate normally. All employees understand the plan processLevel II(0.5)The plan meets the standards and has a clear command structure and response process. The refuge chamber is operating normally. The vast majority of employees understand the basic steps of the planLevel III(0.75)The content of the plan is outdated and has not been updated according to mining changes. A small number of employees basically understand the emergency plan processLevel IV(1.0)The plan is not targeted. No drills were organized and there was no emergency communication and broadcasting system. The refuge chamber is occupied or fails. Employees are unaware of the evacuation, alarm and rescue procedures after a rock burst occurs Classification standards for emergency plans. Level I (0.25) Level II (0.5) Level III (0.75) Level IV (1.0) The rock burst emergency plan of Hongyang No. 3 Mine was formulated in accordance with relevant requirements and contains complete response procedures. Underground escape routes are clearly arranged, and refuge chambers are maintained in normal operation. In addition, mine personnel are familiar with the emergency response process. Therefore, the 718 working face is evaluated as Level I in terms of rock burst prevention and emergency planning. Rock burst risk evaluation model Rock burst is a complex dynamic disaster resulting from the nonlinear interaction of multiple influencing factors. Traditional evaluation approaches based on single factor analysis or purely subjective or objective weighting methods are often insufficient to capture the coupling effects and uncertainties inherent in rock burst occurrence. To address these limitations, this study proposes a multi-level fuzzy comprehensive evaluation model based on combined subjective and objective weighting. The analytic hierarchy process is employed to determine subjective weights reflecting expert knowledge, while the entropy weight method is used to derive objective weights based on data variability. By integrating these two approaches, the proposed model improves both the reliability and objectivity of rock burst risk evaluation. Analytic hierarchy process Construction of hierarchical structure Fully considering natural geological factors, mining technology factors and organization and management factors, the multi-level fuzzy evaluation criterion layer and indicator layer of rock burst risk are obtained as shown in Fig. 7. Schematic diagram of hierarchical analysis. Construction of judgment matrix The single-level fuzzy evaluation model is the basis of the multi-level fuzzy evaluation model, and the multi-level fuzzy evaluation model is an extension of the single-level fuzzy evaluation model. In the case of secondary fuzzy evaluation, the single factor evaluation is generally the fuzzy evaluation at the previous level. Therefore, the single factor judgment matrix for multi-level fuzzy evaluation is: In the formula, rik = bjk, (i = 1,2,3,…,p). R is the judgment matrix; rij is the relative importance of indicator i relative to indicator j in a hierarchy for the risk of rock burst. Therefore, the second-level fuzzy evaluation is: Similarly, a multi-stage fuzzy evaluation model can be established. The single level fuzzy evaluation result is generally regarded as an evaluation at the next level, that is, a row vector R (u) for a single factor to evaluate its own matrix, so any evaluation matrix can be expressed as : For the given weight: The multi level comprehensive evaluation is: The relative importance of each element was quantified using a scale of 1 to 9, as shown in Table 7. Table 71–9 scale table.ScaleMeaning1Element i is equally important as element j3When comparing elements i with j, one is slightly more important than the other5When comparing elements i with j, one is significantly more important than the other7Comparing elements i with j, one is more important than the other9When comparing elements i with j, one is extremely important than the other2, 4, 6, 8The importance of the two elements lies between the above scale values otherReciprocalThe value of bji corresponding to bij 1–9 scale table. Weight calculation Calculate the maximum eigenvalue of the A and its standardized eigenvalue vector Wi, and then obtain the relative weights of indicators at this level. The specific calculation method is as follows: The product obtained by multiplying each row of elements 7 The product obtained by multiplying each row of elements 7 Then get the geometric average 8 Then get the geometric average 8 Calculate standardized feature vectors 9 Calculate standardized feature vectors 9 Calculate the maximum eigenvalue 10 Calculate the maximum eigenvalue 10 Consistency test Matrix consistency was tested using random consistency ratios CR. If the CR value is less than 0.1, the calculation result is acceptable. The random consistency indicator RI can be queried from Table 8. Table 8Mean random consistency index.n12345678910111213141516RI000.520.891.121.261.361.411.461.491.521.541.561.581.591.5943 Mean random consistency index. Calculate consistency indicators CI11 Calculate consistency indicators CI11 Calculate one-time ratio CR12 Calculate one-time ratio CR12 Entropy weight method Data normalization Suppose that the decision goal is an original matrix composed of m evaluation objects and n evaluation indicators: If the types of evaluation indicators are different, horizontal comparison cannot be made. Therefore, it is necessary to carry out dimensionless standardization on these data to offset their impact and facilitate calculation. For positive indicators: For negative indicators: Among them, zij is the normalized data, xmax is the maximum value, and xmin is the minimum value of the evaluation indicator. + 0.0001; (i = 1,2,3, …, n; j = 1,2,3,…,m ); ; (i = 1,2,3, …, n; j = 1,2,3, …, m ). Indicators normalization16 Indicators normalization16 The entropy value of the indicator.17 The entropy value of the indicator.17 The information entropy difference coefficient18 The information entropy difference coefficient18 The weight of the indicator19 The weight of the indicator19 Improved fuzzy comprehensive evaluation method The traditional fuzzy comprehensive evaluation method primarily relies on the analytic hierarchy process to determine the weight of indicators, relies too much on expert experience, lacks unified measurement standards, is highly subjective and arbitrary, and is influenced by differences in the characteristics of various methods, which easily leads to imbalance of indicator characteristics and the results obtained are too one-sided. The objective weighting method is based on data theory and can avoid human interference. Therefore, this study will use the subjective and objective comprehensive weighting method to improve the fuzzy comprehensive evaluation method. By integrating objective combination weights with traditional fuzzy comprehensive evaluation, the scientific nature of evaluation results is improved. The calculation process of the improving fuzzy comprehensive evaluation method is shown in Fig. 8. Improving the calculation process of fuzzy comprehensive evaluation method. Construction of evaluation model Establishment of factor set Establish an evaluation index factor set Y related to the evaluation object. In this study, natural geological influence factors, mining technology influence factors and organizational management influence factors are regarded as the first-level evaluation factor set: Y = {natural geological factors, mining technology factors, organizational management factors} = {y1,y2,y3}. Subsets of factors in the above first-level evaluation factors, such as rock burst tendency, roof management methods, prevention and control investment, etc., are taken as the second-level evaluation factor set: y1={ rock burst tendency, roof rock structure, geological structure, mining depth}, y2={ coal pillar width, goaf, roof management method }, y3={ prevention investment, prevention and control safety measures, prevention and control emergency plan }. Establishment of weight set During the evaluation process, evaluation indicators will have different degrees of impact on the evaluation objects. Therefore, corresponding weights should be applied to evaluation indicators according to different degrees of influence. In this study, the evaluation indicators will be weighted through the subjective and objective comprehensive weighting method, and the objective comprehensive weights calculated by the entropy method are combined with the subjective weights calculated by the analytic hierarchy process to obtain the final comprehensive weights for assignment 21. In order to ensure that the sum of squares of deviations between the combination weight and the subjective weight and the objective weight is minimized, a function is established: Take the derivative of formula (21). When the first-order derivative of the objective function is 0, the sum of the combination weight and the squares of the subjective and objective deviations is the smallest. At this time, α is 0.5, and the combination evaluation effect is the best. Then the final comprehensive weight of the indicator is: Determination of membership functions and membership degrees Definition of fuzzy set: The mapping is given on the universe of discourse, and the corresponding relationship between μ and X is: A is called a fuzzy set on X determined by μ, denoted D; and the degree to which elements in μ belong in A is called the membership function, denoted . Membership function is an important concept in fuzzy sets, and the size of membership directly measures the degree to which an element belongs to the fuzzy set. The closer the value is to 1, the higher the degree of belonging; the closer the value is to 0, the lower the degree of belonging. This study will comprehensively use qualitative and quantitative indicators to determine the membership function. Rock burst tendency C11 Establish the membership degree of rock burst risk tendency indicators, as shown in Table 9. Table 9Rock burst tendency and membership degree.Rock burst tendencyNo riskWeakMediumStrongUniaxial compressive strengthRc ≤ 7 MPa7 MPa < Rc ≤ 10 MPa10 MPa < Rc ≤ 14 MPaRc > 14 MPaElastic energy indexWET < 22 ≤ WET < 3.53.5 ≤ WET < 5WET ≥ 5Impact energy indexKE < 1.51.5 ≤ KE < 3.53.5 ≤ KE < 5KE ≥ 5Dynamic failure durationDT ≥ 500 ms300 ms < DT ≤ 500 ms50 ms < DT ≤ 300 msDT ≤ 50 msMembershipdegree0.20.40.60.8 Rock burst tendency and membership degree. Membership degree Roof rock structure C12 Establish the membership degree of roof rock structure indicators, as shown in formula 24. where: x-distance from hard thick rock formation with strength greater than 60 MPa to coal seam. Geological structure C13 Establish the membership degree of geological structure indicators, as shown in Table 10. Table 10Geological structure types and membership degree.Structure typeSimpleMedium complexSlightly complexComplexExtremely complexMembership degree0.20.40.60.81.0 Geological structure types and membership degree. Mining depth C14 Establish the membership degree of the mining depth indicator, as shown in formula 25. where: M-coal seam burial depth, m. Coal pillar width C21 Establish the membership degree of the coal pillar width index, as shown in formula 26. where: d-coal pillar width, m. Goaf C22 Establish the membership degree of mining depth indicators, as shown in Table 11. Table 11The relationship between working face and goaf and membership degree.Relationship with adjacent goaf areasSolid coal sideOne sideBoth sidesThree and more sidesMembership degree0.250.50.751.0 The relationship between working face and goaf and membership degree. Roof management method C23 Establish membership degree of roof management index, as shown in Table 12. Table 12Roof management method and membership degree.MethodSpontaneous cavingProp pullingDeep-hole blastingHigh pressure water injectionMembership degree1.00.80.60.4 Roof management method and membership degree. Prevention and control investment C31 Establish the membership degree of prevention and control investment grading indicators, as shown in Table 13. Table 13Classification and membership degree of prevention and control investment.Prevention and control investment levelLevel ILevel IILevel IIILevel IVMembership degree0.30.60.91.0 Classification and membership degree of prevention and control investment. Prevention and control safety measures C32 Establish the membership degree of grading indicators for prevention and control safety measures, as shown in Table 14. Table 14Classification and membership degree of prevention and control safety measures.LevelLevel ILevel IILevel IIILevel IVMembership degree0.20.50.81.0 Classification and membership degree of prevention and control safety measures. Prevention and control emergency plans C33 Establish the membership degree of prevention and control emergency plans, as shown in Table 15. Table 15Classification and membership degree of prevention and control emergency plans.LevelLevel ILevel IILevel IIILevel IVMembership degree0.250.50.751.0 Classification and membership degree of prevention and control emergency plans. Engineering application Based on the analysis of mine data of Hongyang No. 3 Mine and on-site measurement, 13 types of indicator data required for three different working face models in the mining area to which the evaluation area belongs were extracted, and the qualitative indicators were quantified. The results are shown in Table 16. Table 16Quantitative indicator values for selecting influencing factors.Work face informationQuantitative indicator value718 working face702 working face704 working faceUniaxial compressive strength3.533.283.43Elastic energy index2.124.22.13Impact energy index1.431.62.28Dynamic failure duration200222190Roof rock structure90.283.692.1Mining depth900892873Geological structure1.00.750.75Coal pillar width8.08.08.0Goaf0.50.50.75Roof management method0.250.250.25Prevention and control investment0.50.50.75Prevention and control safety measures0.50.250.25Prevention and control emergency plan0.250.250.25 Quantitative indicator values for selecting influencing factors. Combined with the engineering geology and mining technical conditions of the 718 working face, a multi-level fuzzy evaluation model with subjective and objective combination weights is used to evaluate the risk of rock burst of the working face. The specific evaluation results are shown in Table 17. Table 17Evaluation result of rock burst risk on 718 working face.Secondary indicatorsWeightThird-level indicatorsComprehensive weightMembership degreeScoreNatural geological factors0.62322Rock burst tendency0.5019130.600.665702Roof rock structure0.2566440.59Geological structure0.1267900.80Mining depth0.1146541.0Mining technical factors0.23949Coal pillar width0.4805360.77Goaf0.2265240.5Roof management method0.2929461.0Organizational management factors0.13729Prevention and control investment0.4290030.6Prevention and control safety measures0.2374440.5Prevention and control emergency plan0.3335480.25 Evaluation result of rock burst risk on 718 working face. After calculation, the weights of natural geological factors, mining technology factors, and organizational management factors are 0.62322, 0.23949, and 0.13729 respectively. The maximum eigenvalue = 3.018, CI = 0.009, CR = 0.017 < 0.1, passing the consistency test. Natural geological factors The subjective weight of natural geological factors evaluation indicators is y1 = {rock burst tendency, roof rock structure, geological structure, mining depth} = {0.52899,0.23390,0.150470,0.08662 }. = 4.246, CI = 0.082, CR = 0.091 < 0.1, this passes the consistency test. The objective weight of mining technical factors evaluation indicators is z1 = {rock burst tendency, roof rock structure, geological structure, mining depth} = {0.47483,0.27938,0.10311, 0.14268}. From the formula (22), the comprehensive weights of natural geological factors are {0.501913,0.256644,0.126790,0.114654}. Mining technical factors The subjective weight of mining technical factors evaluation indicators is y2 = {coal pillar width, goaf and roof management method} = {0.53896,0.29725,0.16378 }. = 3.009, CI = 0.005, CR = 0.009 < 0.1, this passes the consistency test. The objective weight is z2 = {coal pillar width, goaf and roof management method} = {0.42111,0.15579,0.42111}. From the formula (22), the comprehensive weights of mining technical factors are {0.480536, 0.226524,0.292946}. Organizational management factors The subjective weight of organizational management factor evaluation indicators is y3 = {prevention and control investment, prevention and control safety measures, prevention and control emergency plans} = {0.52467,0.14155,0.33376 }. = 3.054, CI = 0.027, CR = 0.051 < 0.1, this passes the consistency test. The objective weight is z3 = {prevention and control investment, prevention and control safety measures, prevention and control emergency plans } = {0.33333, 0.33333, 0.33333}. From the formula (22), the comprehensive weights of organizational management factors are {0.429003,0.237444,0.333548}. Referring to the evaluation data and literature, the rock burst risk is divided into five levels, and the classification standards are shown in Table 18. According to the comment set, the rock burst risk score of the 718 working face is 0.665702, and the risk level is C, which belongs to medium rock burst risk and is more consistent with the actual situation on site. Table 18Classification of rock burst dangerous conditions.Rock burst risk levelStateRange of scoresASafe < 0.45BWeak rock burst0.45–0.65CMedium rock burst0.65–0.85DStrong rock burst0.85–0.95EDangerous > 0.95 Classification of rock burst dangerous conditions. Therefore, based on the specific production conditions of the 718 working face, it is recommended to take the following rock burst prevention measures during the mining process: During the mining period, continuous and comprehensive stress monitoring and displacement monitoring should be implemented at the working face, including indicators such as bolt deformation, roadway deformation, and mine pressure. Through dynamic analysis and risk evaluation, prevention and control plans and key technical param are scientifically determined to achieve refined management of rock burst prevention and control. During the mining period, continuous and comprehensive stress monitoring and displacement monitoring should be implemented at the working face, including indicators such as bolt deformation, roadway deformation, and mine pressure. Through dynamic analysis and risk evaluation, prevention and control plans and key technical param are scientifically determined to achieve refined management of rock burst prevention and control. For the broken area of the surrounding rock of the roadway, technologies such as grouting reinforcement can be adopted to improve the integrity and mechanical strength of the surrounding rock and improve its deformation characteristics; the reasonable coal pillar width should be comprehensively determined based on the stress superposition area, goaf distribution, stop-mining line location, on-site stress monitoring and numerical simulation results. Through systematic stress control, the possibility of rock burst can be reduced and a guarantee for mining can be provided. For the broken area of the surrounding rock of the roadway, technologies such as grouting reinforcement can be adopted to improve the integrity and mechanical strength of the surrounding rock and improve its deformation characteristics; the reasonable coal pillar width should be comprehensively determined based on the stress superposition area, goaf distribution, stop-mining line location, on-site stress monitoring and numerical simulation results. Through systematic stress control, the possibility of rock burst can be reduced and a guarantee for mining can be provided. In order to reduce the roof subsidence of the working face, it is necessary to continue to increase investment in preventing and controlling rock burst, and adopt a multi-level joint support method of "anchor, net and cable" to improve the carrying capacity of the roadway support structure. A segmented support strategy is adopted in critical stages of the roof (such as the “square” stage) to ensure the safety and stability of the roadway. In order to reduce the roof subsidence of the working face, it is necessary to continue to increase investment in preventing and controlling rock burst, and adopt a multi-level joint support method of "anchor, net and cable" to improve the carrying capacity of the roadway support structure. A segmented support strategy is adopted in critical stages of the roof (such as the “square” stage) to ensure the safety and stability of the roadway. Conclusion Comprehensively analyze the geological and production conditions of the 718 working face, and propose the main control factors that induce rock burst. Among them, the natural geological influence factors include rock burst tendency, roof rock structure, geological structure and mining depth; the influence factors of mining technology include coal pillar width, goaf and roof management methods; the influence factors of organization and management include prevention and control investment, safety measures and emergency plans. Comprehensively analyze the geological and production conditions of the 718 working face, and propose the main control factors that induce rock burst. Among them, the natural geological influence factors include rock burst tendency, roof rock structure, geological structure and mining depth; the influence factors of mining technology include coal pillar width, goaf and roof management methods; the influence factors of organization and management include prevention and control investment, safety measures and emergency plans. Integrating the subjective and objective combination weighting method, a multi-level coal mine rock burst risk evaluation model based on the improved fuzzy comprehensive evaluation method was constructed, and the weights and membership degrees of main control factors were determined. The weight of rock burst tendency among natural geological factors is 0.501913; the weight of coal pillar width among mining technical factors is 0.480536; the weight of prevention and control investment among organizational and management factors is 0.429003. These three factors have the most significant impact on rock burst. Integrating the subjective and objective combination weighting method, a multi-level coal mine rock burst risk evaluation model based on the improved fuzzy comprehensive evaluation method was constructed, and the weights and membership degrees of main control factors were determined. The weight of rock burst tendency among natural geological factors is 0.501913; the weight of coal pillar width among mining technical factors is 0.480536; the weight of prevention and control investment among organizational and management factors is 0.429003. These three factors have the most significant impact on rock burst. The rock burst risk evaluation and calculation were carried out on the 718 working face. The rock burst risk score of the working face was 0.665702, and the risk level was C, which belonged to medium rock burst risk. The evaluation results are consistent with the conclusions of existing research reports, demonstrating the validity and applicability of the proposed model. The rock burst risk evaluation and calculation were carried out on the 718 working face. The rock burst risk score of the working face was 0.665702, and the risk level was C, which belonged to medium rock burst risk. The evaluation results are consistent with the conclusions of existing research reports, demonstrating the validity and applicability of the proposed model. Acknowledgements This work was supported by Project supported by the National Natural Science Foundation of China (Grant No. 52374092), and Basic Scientific Research Project of the Educational Department of Liaoning Province (Grant No. LJ212410147086). Author contributions J.Z.: Writing—Original draft preparation, Conceptualization, Methodology, Supervision. B.W.: Writing—review & editing, Funding acquisition. Y.L.: Data curation, Validation, Visualization. X.L.: Data analysis and data collection. All authors have read and agreed to the published version of the manuscript. Funding This work was supported by Project supported by the National Natural Science Foundation of China (Grant No. 52374092), and Basic Scientific Research Project of the Educational Department of Liaoning Province (Grant No. LJ212410147086). Data availability All data generated or analyzed during this study are included in this published article. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement All data generated or analyzed during this study are included in this published article.