Acoustic Logging Characteristic And Identification Method Study For Mid-deep Gas Formation

As mid-deep gas formation has the character of low porosity, low permeability, the log response characteristics commonly not evident than other formation. Though it usually can show high resistance property, it's hard to identify whether it is oil formation or gas formation. Acoustic wavetrain logging has aboundant information, including compressional and shear wave velocity, amplitude, attenuation coefficient, and so on. Therefore, this thesis utilizes rock physics theory to judge the character of acoustic well-log parameters as the reservoir is saturated with gas, and extracts some characteristic parameters which are sensitive to mid-deep gas formation, finally uses BP ANN technique to identify mid-deep gas formation.First, this thesis summarizes some well-log response property and general distinguishing methods of shallow and mid-deep gas formation, then using double-phase theory and Biot-Gassmann equation to do forward modeling of some acoustic characters such as compressional and shear velocity, poisson ratio, bulk modulus when the reservoir is filled with gas. In order to research the influence of gas on attenuation, this article utilizes three models (Biot, BISQ and Dvorkin) to calculate inverse quality factor respectively, and process actual XMAC data using spectral ratio method, results show spectral ratio method is the most effective method to calculate attenuation.On the basis of above research, this thesis does some research on applying BP ANN technique to reservoir identification, first analyses sensitivity of well-log response character of gas formation, combining the sensitive ones with acoustic character parameters as the input vector of ANN, take the identifying reservoir types as output vector, the result is effectiveness.