Layered production prediction of oil wells based on ConvLSTMA

Abstract

ABSTRACT With the continuous development of oilfield, oilfield production prediction is becoming more and more important. Most of the current methods are based on machine learning prediction methods and predict the oil production of a single well with relatively low accuracy. In this paper, we focus on dividing single wells into specific layer segments and propose a layered production prediction method for oil wells based on ConvLSTMA model with convolutional neural network, long- and short-term memory network, and fine-grained attention mechanism. The method can effectively utilize stratified historical data to predict future oilfield production and can be adapted to various complex production prediction scenarios. Through experiments, it can be seen that the model performs better than existing models in predicting oil production data, reducing the MSE index by about 2%, MAE index by about 3%, and R2 coefficient by 2%.