To investigate the instability risk of waste dump slopes under the influence of rainfall seepage, a risk analysis coupling model was developed, integrating the bow-tie model with a fuzzy Bayesian network. This model encapsulates the instability risk factors associated with waste dumps, encompassing internal, environmental, and managerial factors. The bow-tie model for slope instability was constructed utilizing both fault tree and event tree models. Through the event tree analysis, four potential damage outcomes were identified: the emergence of potential hazards within the waste dump, progression of damage, increased water content and self-weight, and eventual slope instability. Consequently, four safety barriers are proposed to mitigate and manage damage and instability in waste dumps: risk identification and damage assessment, investigation of latent hazards such as structural defects and weak water-bearing surfaces within the waste dump, implementation of drainage systems, and remediation of waste dump damage. A Bayesian network model, grounded in the bow-tie model, was developed to characterize the polymorphism and uncertainty associated with factors contributing to slope instability. To address the probabilities of model nodes, a triangular fuzzy number and a modified weighted similarity aggregation method were employed. This approach mitigates the subjectivity inherent in expert assessments and overcomes the challenges of accurately quantifying various probabilities. Risk prediction was executed through forward inference using the Bayesian network model, while risk diagnosis was performed via reverse reasoning. Key factors influencing instability were identified through importance analysis. The study utilized a metal mine waste dump slope in Inner Mongolia as a case study, wherein enterprise experts were invited to provide scores. GeNIe software was employed to compute the prior probability, posterior probability, and the significance of each factor contributing to the risk of slope damage in the waste dump. A quantitative study was conducted to examine the relationship between damage and influencing factors in the context of waste dump slope backgrounds. The study identified key factors contributing to waste dump slope damage under rainfall seepage conditions, including rainfall-induced seepage erosion, the self-weight of the waste dump slope, internal defects, high water content, and the failure of emergency response mechanisms. Based on the analysis results, targeted instability control measures were proposed. The analysis of engineering case studies demonstrates that this model aligns with the approach of advancing safety production protocols. It effectively describes the impact and probability of various factors contributing to damage occurrence under rainfall seepage conditions. This model holds significant reference value for the risk management of geological disasters in waste dump slopes.
贝叶斯网络模型通过变量集合的联合概率分布表征多个变量之间的相互关系,能够有效解决传统可靠性分析方法的不足(赵帅等,2026)。模型中的节点表示系统变量,有向弧线表示因果变量之间的关系,以概率描述节点变量的不确定性,以条件概率表(Conditional Probability Table,CPT)描述节点关系,从而实现分类、诊断和预测(Zhang et al,2020)。
Mirzaei AliabadiM, PourhasanA, MohammadfamI,2020.Risk modelling of a hydrogen gasholder using Fuzzy Bayesian Network (FBN)[J].International Journal of Hydrogen Energy,45(1):1177-1186.
[2]
ZhangG H, ChenW, JiaoY Y,et al,2020.A failure probability evaluation method for collapse of drill-and-blast tunnels based on multistate fuzzy Bayesian network[J].Engineering Geology,276:105752.
AnRan, KongLingwei, ZhangXianwei,et al,2023.A multi-scale study on structure damage of granite residual soil under wettingdrying environments[J].Chinese Journal of Rock Mechanics and Engineering,42(3):758-767.
ChangZheng, HeXuzhuo, FanHanwen,2024.Research on the application of Bayesian Network in LNG marine transportation risk assessment[J].Journal of Safety and Environment,24(12):4541-4551.
GuoMingyang, ChenMiao, JiaLü,et al,2025.Risk and causation analysis of ship collision breakage into water based on Apriori-ISM-BN[J].Journal of Harbin Engineering University,46(9):1693-1700.
HuJingwen, NieWen, LuSong,et al,2023.A static Bayesian-based model for evaluating rainfall-induced tailings dam instability[J].Metal Mine,52(11):268-275.
KangEnsheng, MengHaidong, DuPengdong,et al,2023.Study on the resistivity variation characteristics of dump slope damage under rainfall conditions[J].Mining Research and Development,43(9):51-61.
KangHongpu, ZhangZhen, HuangZhizeng,2020.Characteristics of roof disasters and controlling techniques of coal mine in China[J].Safety in Coal Mines,51(10):24-33,38.
LiYaoyi, ChenGuoqing, ShiMinghan,et al,2024.Intelligent optimization method for slope morphology and stability of phos-phogypsum stack based on LM-LSO algorithm[J].Gold Science and Technology,32(5):882-893.
LiYulong, HouXiangyu,2022.Risk diagnosis and prediction of mega-project waste dump based on fault tree and Bayesian network integration[J].Journal of Systems & Management,31(5):861-874.
LianghaiLü, LiangYiyuan, ZhangHaobin,et al,2024.Research on dynamic risk assessment method of indoor gas leakage based on fuzzy DBN[J].Journal of Safety and Environment,24(4):1337-1345.
ShenJianhong, LiuShupeng,2023.Application of fuzzy dynamic Bayesian network in risk evolution analysis of deep foundation pit construction[J].Journal of Safety and Environment,23(12):4211-4221.
WangHua, LuoShaofeng,2023.Risk analysis of methanation unit system based on HAZOP and fuzzy extended bow tie(FEBT)model[J].Safety and Environmental Engineering,30(2):27-34.
WangHuiwen, LiuBin, LiDandan,et al,2024.A human factors reliability analysis method for levee safety using Bayesian networks[J].Water Resources and Hydropower Engineering,55():275-282.
WangRuipeng, WangJin’an,2022.Analysis on the influence factors of slope instability in west open-pit mine[J].Mining Research and Development,42(1):64-70.
XuXiaonan, NingXin, ZhuYuanpeng,et al,2025.Application of fuzzy comprehensive evaluation-BP neural network for electric fire risk assessment[J].Journal of Safety and Environment,25(1):1-10.
ZhangYaxian, HouZhongjie,2022.Back analysis of rock mass parameters based on support vector machine and Bayesian method[J].Yangtze River,53(6):186-192.
ZhaoShuai, ShiCongling, QianXiaodong,et al,2026.Reliability assessment of gas extinguishing system based on fault tree analysis and Bayesian network[J].Safety and Environmental Engineering,33(1):202-210.