Predicting aqueous phase trapping damage in tight reservoirs using quantum neural networks

Abstract

Formation damage associated with aqueous phase trapping (APT) often occurs during drilling wells using water-based fluids in tight reservoirs. Prediction of a reservoir’s APT severity is of great importance, since well productivity can be improved through proper prediction and consequent attempts to reduce formation damage. In this paper, the mechanism for APT occurrence is analyzed. Different factors affecting APT are evaluated and selected to develop a neuron network model for APT prediction, which is based on the information processing method of biological neurons and quantum neural algorithm. The model proposed in this paper is quantum neural network (QNNs) model, which is considered to have an advantage over previous models in terms of the internal algorithm. The model can be used to predict the severity of APT in tight sandstone formations quantitatively. This model has been applied in one pilot area in Jinlin oilfield, China. The results show very good accuracy in comparison with the experimental data.