An Adaptive and Fast Simulated Annealing Algorithm for Fracture Network Modeling in Naturally Fractured Reservoirs

Knowing the orientation and spatial configuration of fractures in naturally fractured reservoirs is necessary for the prediction of the flow in those reservoirs. Among the fracture network modeling methods، simulated annealing algorithm (SA)، due to its capability in solving large problems and finding a global optimum، is well-known. This paper proposes a new adaptive and fast simulated annealing algorithm for modeling naturally fractured reservoirs. This algorithm improves computation performance without degrading solution quality by incorporating a method for the estimation of the initial value of the temperature like parameter، T0، using an adaptive Markov chain length، NT (inner iterations) and suggesting a new fast and adaptive annealing schedule. Moreover، we discuss some aspects of a special objective function and proof this issue that why the minimum of this objective function occurs when the final configuration is a two orthogonal fracture sets with the same number of fractures in each set. Finally، we extend this objective function and modify it to develop an objective function that could generate conjugate fractures with any arbitrary intersection angle.