Research And Evaluation Of Instability Mechanism Of Rainfall-induced Tailings Dam Based On Bayesian

With the development of my country’s mineral resources entering an unprecedented period of development,the scale of mine waste discharge is increasing.A tailings dam is a dam-like structure for stacking mine waste,which contains a large amount of heavy metal ions.The safe operation of tailings dams has always been the core of mine safety research,and the research on the stability of tailings dams is one of the most important research directions.Once the dam body is unstable and damaged,it will not only cause serious harm to the safety of life and property of the downstream people,but also cause incalculable pollution to the environment.Rainfall events are one of the factors leading to the instability of tailings dams.Evaluating the stability of tailings dams during rainfall events is crucial to the safe operation of mines.Taking the Luomukeng tailings dam in Jiangxi Province as the research background,a1:100 tailings dam model test was established to simulate the instability state of tailings at various spatial locations under different rainfall intensity.The physical parameters of tailings are obtained through geotechnical tests:particle grade parameters(D ₅₀,C _u,C _c),moisture content,wetting line,density,internal friction angle and cohesion,and the tailings unit is calculated based on the infinite slope model slip resistance and sliding force.The variation law of tailings parameters was analyzed from the two dimensions of rainfall and height,and the interaction law between tailings parameters was studied through Spearman correlation analysis.Due to the variable rainfall infiltration mechanism in the dam body,the link between rainfall events and dam instability is uncertain.In order to reduce this uncertainty,Bayesian analysis was used to predict the tailings dam under different rainfall.Stability,a Bayesian-based tailings dam stability evaluation model is established.Use the information gain to get the early warning value of each parameter and use it as an index to calculate the prior probability of the physical parameter.Combined with the Leaky Noisy-or gat extended model to calculate the conditional probability,building Bayesian Models with Ge NIe,the spatial position of the tailings dam under different rainfall conditions is got through Bayesian analysis.instability probability.The model links the experimental parameters with the uncertainty of the tailings dam instability mechanism,which further improves the accuracy of the objective evaluation.A new Bayesian-based tailings dam stability assessment model is established,which clarifies the instability mechanism of tailings dams and predicts the state of tailings dams at different spatial locations under different rainfall events.The self-learning of the structure and parameters is realized,which has significant advantages in Bayesian reasoning,and can be reversed through the results to trace the cause of the dam damage.