A New Method to Predict Water Breakthrough in Intelligent Completion Horizontal Well

Abstract

With the oilfield production entering a later period, the yield will be seriously affected if the problem is not addressed. This article first determines through analysis that the evaluation parameters for horizontal well water output are reservoir permeability and porosity. Second, the relationship between the real-time monitoring data and the above parameters was defined, along with the collected raw data of temperature, pressure, and flow rate. Using different artificial intelligence algorithms for historical reconstruction again, the error of the BP neural network calculation is minimum and the permeability, porosity, and water yield possibility are within 0.8%. Further improvement of the BP neural network algorithm and establishment of a new neural network prediction model were achieved. Then, the water production situation of a certain horizontal well was predicted, and the calculation error was 0.6%. It shows that water entry was forecast to occur 42 months after production. Finally, we designed a reasonable water control and oil increase plan that can extend the anhydrous oil recovery period by over 900 days. The result shows that the prediction method proposed in this article is reasonable and feasible; the predicted results of effluent are accurate; and the designed water control plan is effective. It can slow the rate of water breakthrough in horizontal wells and improve oilfield recovery efficiency.