Intelligent Stage Selection Method for Refracturing Based on the Type-2 Fuzzy Logic System

To solve the problems of low-ratio high-contribution stages in the horizontal well initial fracturing and complex factors affecting refracturing effects in tight reservoirs, analytic hierarchy process, grey correlation, BP neural network, and fuzzy logic methods are comprehensively applied to establish a multi-level Type-2 fuzzy logic system (T2-FLS): Based on the improved analytic hierarchy process–grey relational method, the weight of different geological engineering factors on the productivity contribution of the candidate refracturing stages was calculated. The BP neural network system was used to establish the refracturing productivity prediction model, and the value of each influencing factor was divided into four grades in the T2-FLS. The membership function of the T2-FLS and comprehensive quantitative evaluation criterion of candidate well interval for refracturing is established. The productivity improvement in different refracturing candidate wells and different refracturing candidate stages of the same horizontal well after refracturing was compared and analyzed. For relatively heterogeneous reservoirs, the incomplete fracturing stage with high contribution rate of primary fracturing to productivity has the highest refracturing potential. For relatively homogeneous reservoirs, the stimulation potential of refracturing in the unstimulated section is higher than that in the simulated section. The high and sub-high potential stages of the high comprehensive potential candidate wells should first be selected for refracturing, and then select the high potential stages in the low comprehensive potential candidate wells. After reasonable refracturing stage selection, the productivity of some refractured horizontal wells can even be increased by more than 10% compared with the initial production.