Background Deep coal reservoirs are characterized by well-developed microstructures and natural fractures, as well as strong heterogeneity in reservoir petromechanical properties and in-situ stress field distribution. With the progress of coalbed methane (CBM) production, the formation pressure and in-situ stress within the drainage area of CBM wells evolve continuously, which further intensifies the complexity of stress distribution. Severe fracturing channeling between new and existing wells is predominantly triggered by in-situ stress variations induced by large-scale hydraulic fracturing and long-term production; therefore, accurate characterization of the dynamic evolution laws of in-situ stress during post-fracturing and production processes is of critical importance for well placement of new wells, fracturing design optimization, and anti-channeling control during fracturing operations.

Methods In this study, taking the Daning-Jixian Block in the Ordos Basin as the study case, we established a 4D in-situ stress evolution prediction model for deep CBM reservoirs during long-term production based on the geological characteristics of the study area. By constructing an initial 3D geomechanical model, a fracturing-induced stress field model, and a production-induced stress field model, we realized high-precision dynamic characterization of the stress field from the initial in-situ state to long-term production state, and verified the accuracy of the proposed model with field microseismic monitoring and production monitoring data.

Results and Conclusions  The main results and conclusions are drawn as follows: (1) The difference between the maximum and minimum principal stresses in the target well block decreases after fracturing, and the horizontal stress difference increases gradually with production progress; during 24 months of production, the minimum horizontal in-situ stress around the well pad decreases by 2.10 to 5.75 MPa, the in-situ stress orientation shows a significant variation with an average deflection of 4.12°, and a larger stress reduction amplitude corresponds to better fracture network connectivity, a larger supply range of formation energy, and higher production potential of the reservoir. (2) The study shows that a horizontal well trajectory at an angle of 70° to 110° with the maximum principal stress direction combined with a well spacing of 400 m delivers the optimal reservoir stimulation effect, which is conducive to maximizing the Estimated Ultimate Recovery (EUR). (3) An optimized zipper fracturing operation strategy is defined: for areas with unbalanced in-situ stress, zipper fracturing is performed on adjacent wells, specifically, the well in the low-stress zone is fractured first to balance the in-situ stress after stimulation, followed by the well in the high-stress zone, to achieve sufficient propagation and coalescence of the fracture network; for areas with similar in-situ stress levels, taking a 3-well pad as an example, the two lateral wells are fractured first, after which the variation amplitude of the stress difference on both sides of the middle well is consistent to realize balanced in-situ stress, and then the fracturing parameters of the middle well are optimized to achieve the target of fracture network coalescence. (4) A refined perforation cluster division method based on in-situ stress magnitude is proposed in this study, which provides solid technical support for the fracturing and efficient development of deep CBM wells.