A novel and efficient dimensionality-adaptive scheme for logging-while-drilling electromagnetic measurements modeling

Logging-while-drilling (LWD) electromagnetic (EM) measurements have been developed and increasingly used to place the oil/gas wells in formations for decades. Successfully implementing the technology in geosteering and reservoir evaluation highly depends on the efficiency and accuracy of data processing. The ability to accurately image the formation structures around horizontal wells has increased dramatically with advances in forward/inversion algorithms. However, the current fixed-dimension modeling methods are unable to balance the tradeoff between the efficiency and accuracy of the EM LWD simulation. For this reason, we propose a dimensionality-adaptive scheme that combines the advantages of low dimensional algorithms (semi-analytic) and high dimensional algorithms (2.5D) to enhance the efficiency and accuracy of the modeling. The new scheme is able to select the proper model dimension and invoke its corresponding algorithm automatically. By applying the developed size-adaptive sliding-window method, a complex 2D model can be converted to a series of 1D models for a single record point. The model control factor and the cross-dimension factor are further defined to quantitatively evaluate the model simplification feasibility. The former analyzes the formation structure complexity, while the latter estimates the simplification accuracy by comparing the results from the obtained 1D models. A 2D example is presented to showcase the effectiveness and accuracy of the dimensionality-adaptive scheme. Numerical results show that the defined factors are feasible to quantitatively evaluate the simplification rationality. The novel dimensionality-adaptive scheme is more efficient and accurate compared to the fixed-dimension algorithm, which can be used for both pre-drilling modeling and real-time inversion.