Fracture Aperture Prediction Method based on Hierarchical Expert Committee Machine in Tight Clastic Reservoir

Summary Fracture aperture is an important parameter to evaluate the quality of fracture controlled tight clastic reservoir. The well logs were always used to predict the fracture aperture, but some linear regression methods do not match well with complex logging data due to the characteristics of low porosity and low permeability in tight clastic reservoir. The machine learning method can improve prediction accuracy, but it always generates unstable prediction models. A static committee machine (CM) can reduce errors and uncertainties by combining multiple learners, but the weight of integrating learners is difficult to determine. In order to promote the accuracy and efficiency of weight calculation, the CM is improved by using analytic hierarchy process (AHP) and joint neural network (JNN) model. Based on the prediction performance of each expert network, the hierarchical expert committee machine (HECM) model is formed by adding the hierarchical network module adaptively. The experiment shows that HECM model can reduced the relative error of the prediction results, and increased the stability of the prediction model. The HECM model provides a new method for fracture aperture prediction in tight clastic reservoir by mining potential information of input logging data.