Surrogate Model of Shale Stress Based on Plackett-Burman and Central Composite Design

Summary Multifactor analysis and accurate prediction of dynamic stress in shale reservoirs are of great practical significance for designing hydraulic fracturing. In this paper, a surrogate model for the rapid prediction of shale stress is proposed by considering the geomechanical heterogeneity and multiscale seepage mechanism of shale gas. The Plackett-Burman method is used to compare the influence of different parameters on shale stress, and significant parameters are selected as decision variables for establishing a surrogate model. The surrogate model for predicting stress is obtained by central composite design fitting, and the interaction of significant factors on shale stress is studied. The results show that after considering the heterogeneity, the minimum horizontal stress fluctuation range is 20.25 to 44.03 MPa and the maximum horizontal stress fluctuation range is 26.46 to 49.77 MPa in the area controlling hydraulic fracture. The initial reservoir pressure, as well as the length and width of hydraulic fractures, are the key factors influencing reservoir stress. The analysis of variance demonstrates that the proposed method is effective for predicting shale stress. The research results are helpful for gaining a deeper understanding of the evolution mechanism of dynamic stress fields in shale reservoirs and provide guidance for treatment design and dynamic optimization of gas wells.