Intelligent Prediction of Acid-fracturing Performance in Carbonates Reservoirs

Abstract Acidizing joint hydraulic fracturing is a widely used technique to increase the fracture conductivity in carbonate reservoirs stimulation. Predicting acid fracturing performance for optimum fracturing job design requires a detailed understanding of acid-rock reactions, rock strength and the stress applied to the rock and their effect on the fracture conductivity. The available models have many prediction limitations to acidized fracture conductivity with closure stress. Artificial intelligence is suggested to obtain a more precise prediction for acid-fracture conductivity. Artificial neural network (ANN) and adaptive network- based fuzzy inference system (ANFIS) are used to develop intelligent models capable of delivering an accurate design to acid-fracture conductivity. Published experimental data from the literature is used to train the models. 70% of the data was used for training and 30% was used in testing. The results showed that both ANN and ANFIS models outperformed the currently available models. ANFIS subtractive clustering with 0.4 cluster radius showed the best match to experimental data with 1.36% average percentage error and 0.998 R2.