Study On Quantitative Analysis And Prevention Method Of Failure Risk For Offshore Well Control Manifold

During drilling and production,if overflow or kick accident occurs,and appear improper operation of the killing well or failure of the well control equipment,etc,the blowout accident may be occurred easily.Well control manifold is a special manifold in the well control equipment.Its function is to balance the pressure between wellbore and formation by spraying and injecting drilling fluid to ensure the smooth operation of oil and gas drilling.In view of the current research on well control manifold is focused on qualitative analysis such as risk identification and safety evaluation,this paper studies on quantitative analysis and prevention method of failure process for well control manifold.Based on the API standard,system composition and pressure control process of well control manifold,combined with literature research,a fault tree model is established to identify the failure factors and it is concluded that erosion wear is the main form of the perforation failure.The erosion theory and erosion evolution process applicable to well control manifold are summarized,and the factors affecting erosion are analyzed from three aspects: particle,fluid and material.Based on Fluent software,a finite element model of well control manifold erosion wear is established,the characteristics of velocity and pressure distribution and the trajectory of solid particles under different working conditions are analyzed;the main position of the solid particles on the erosion of the well control manifold and the size of the erosion area are discussed;the influence of different pipe diameter,bending angle,flow velocity,particle diameter and mass flow rate on the maximum erosion rate is compared.The research indicates that the erosion position of the well control manifold is concentrated on the outer arch side of the elbow,and the diameter,bending angle,particle diameter and mass flow rate have a great influence on the erosion rate.In order to realize the rapid prediction of the maximum erosion rate for the well control manifold,combined with the machine learning method,two prediction models of BP neural network and support vector machine are established.Based on orthogonal experiment,the most characteristic training samples are obtained,and the genetic algorithm is introduced to optimize key parameters.The results show that the average prediction error of GA-BP neural network is 14.7%,while the average prediction error of GA-SVM is only 6.9%.Therefore,GA-SVM can better achieve a reasonable prediction of the maximum erosion rate of well control manifolds.Finally,based on erosion wear,the failure model of well control manifold is constructed,and the failure criterion to calculate the failure time under the action of erosion and internal pressure is proposed.By optimizing the structure of the well control manifold to change the erosion position and weaken the erosion,then practical risk control measures are formulated.