Research on Water Content Prediction Method Based on LSTM Neural Network

Abstract

In the development of oilfields, water cut is an important indicator to reflect the development of oilfields. At present, most oilfields are derived by mathematical formulas or predicted by a single neural network, and the accuracy is not high. In order to improve the prediction accuracy of water content, this paper proposes a water content prediction model PSO-LSTM, which uses particle swarm algorithm to optimize the hyperparameters of long-term and short-term memory neural network. Taking the long-term and short-term memory neural network as the backbone of the model, the PSO algorithm is used to find its hyperparameters. Through experiments, the prediction accuracy R2 of the PSO-LSTM model proposed in this paper can reach 0.91. Compared with the LSTM neural network model and the BP neural network model, the accuracy is the highest, and the expected goal is achieved.