Percolation Mechanism And Development Law Of Volcanic Reservoirs In Xinjiang

At present,Xinjiang volcanic reservoirs can be divided into two types considering natural energy: existence or absence of bottom water.In order to study the problems of low well production,fast water breakthrough and rapid increase of water cut after water breakthrough.For no bottom-water volcanic reservoirs,the non-linear seepage mechanism and stress sensitivity are studied from the perspective of seepage mechanism.On this basis,the compound flow model of fractured horizontal wells in volcanic reservoirs is established,and the sensitivity of parameters is analyzed,and the relationship between different parameters and productivity is obtained.For bottom-water volcanic reservoirs,according to actual production data and geological parameters,the factors affecting the productivity of fractured horizontal wells in volcanic reservoirs are studied,and the grey relational degree evaluation method is used to rank them,so as to optimize the main control factors of productivity.In the actual production process,productivity is affected by many factors,it is difficult to find the direct relationship between factors and productivity from the actual production data.Therefore,a conceptual model is established by numerical simulation,and the development law of volcanic reservoir is studied by single variable control method.On this basis,the productivity prediction is studied,and the support vector machine algorithm is improved by machine learning theory.Firstly,the dimension of input variables is reduced by principal component analysis.Then,the parameters of support vector machine algorithm are optimized by genetic algorithm and particle swarm optimization.The results show that the improved support vector machine algorithm can accurately predict the productivity of fractured horizontal wells in Xinjiang volcanic reservoirs.