Method Research For Well Logging Evaluation Of Tight Gas-bearing Sandstone In T Gas Field Of Changqing

With the further exploration and development,logging evaluation is faced some problems,such as complicated pore structure,low gas saturation,non-obvious log response characteristics and reservoir fluid identification of T block on Changqing gas field.In this paper,the system that could calculate reservoir parameters,identify fluid and productivity prediction is formed combining sedimentary characteristics and four-property relationship research on T gas field,which could provide support for the later development of gas field.In this paper,made the T block as research object,detailed study of four-property relationship and pore structure.And on this basis,we build the model of shale content,porosity,permeability,saturation,which used single-correlation analysis,multivariate fitting after probabilistic neural network classification,layer spot method.And the new way could improve interpretation accuracy is studied by divided the first result into three categories of bigger than the real,smaller than and in the acceptable range.The reservoir physical property lower limit is also researched using well test intersection method,porosity-permeability cross plot method,and tail flick method.The lower limit of porosity is 0.05%,and permeability lower limit value is 0.034 mD.Based on testing data and conventional log data,we mainly use three porosity,p-wave equivalent elastic modulus method,correlation coefficient method,the lateral induction coupling measurement method,gas index method identify the fluid.The result shows that the fluid identification must integrate multiple way for higher heterogeneity in low porosity and low permeability reservoir,and the way combined the lateral induction coupling measurement method and gas index method has better application effect on low resistance measurement.Combined with the actual gas production data and logging data,the reservoir productivity prediction is computed using logging parameters which have better correlation with daily production by multivariate fitting.Sand structure evaluated by lithologic homogeneous coefficients is put forward to indicate productivity,that the coefficients are computed by well logging curve smoothness.Dividing the level of daily production capacity and using probabilistic neural network predict the level of productivity could have a better application effect.