Abstract: Aimed at the complexity and diversity of the oil and gas reservoir in tight clastic rock, the authors propose a method to inverse the apparent density of fracture, that is, first velocity is inversed using a hybrid method of genetic algorithms (GA) and neural network algorithms, then apparent density data of fracture gained from FMI, finally the GA method is used again to gain the apparent density profile of fracture. The 2nd member of Xujiahe formation of a gas reservoir in west Sichuan is taken as a case, on the base of logging analysis, the apparent density of fracture with nonlinear method is reversed using the FMI data, and the apparent density profile of fracture gained. The result has higher resolution and fit for the practice, it indicates that this method is an active try to predict fractured reservoir.

Key words: tight clastic rock, gas reservoir, apparent density of fracture, high resolution, nonlinear inversion