A hybrid stability analysis bypassing method for compositional simulation

Stability analysis has become the standard method to test the phase appearance in compositional reservoir simulation. This method is very computationally intensive and consumes most of the CPU time in simulation because every single-phase grid block needs to perform this test at every outer iteration in solving global equations (the iterations to solve pressure, saturation and composition). To reduce this part of simulation time, some stability analysis bypassing methods are proposed. Two main methods are the neighborhood bypassing (NB) method and the shadow region bypassing (SB) method. Based on these two methods, a hybrid bypassing (HB) method is presented to improve the efficiency and accuracy of the current two-phase stability analysis bypassing methods. The NB method has limited applications and a comprehensive comparison of these two methods has not been given, so the most effective method is yet unknown. We implement these three bypassing methods in a parallel isothermal compositional simulator and provide a detailed comparison of the performance of these three methods for different reservoir grid models, hydrocarbon fluids and field operations. The results show that the performance of these methods is very case-dependent, and the drawbacks of the NB and SB methods are clearly demonstrated. It is also shown that the HB method always performs better than the SB method and better than or equally compared to the NB method in most cases. Based on the performance of these methods, we give the recommendation that which method should be used in a certain condition.

Automated summary

Stability analysis is the standard method for testing phase appearance in compositional reservoir simulation. To reduce computational time, the neighborhood bypassing (NB) method and the shadow region bypassing (SB) method are proposed, and a hybrid bypassing (HB) method is presented. Experimental results show that the HB method performs better than the SB method and better than or equally to the NB method in most cases. Based on the performance of these methods, recommendations are given for their use in specific conditions.