Oil Well Production Prediction Model Based on Attention-BiLSTM

Predicting oil well production is an important task in oilfield production. LSTM model is usually used to predict oil well production, but the output of LSTM model at the current moment is determined by the input at the previous time, ignoring the influence of the input at the subsequent time. In order to further improve the accuracy of oil well production prediction, this paper proposes an oil well production prediction model based on bidirectional long short-term memory network (BiLSTM) and attention mechanism. It uses BiLSTM as the backbone of the model, which can simultaneously consider the characteristics of oil well production data in the past and future, and perform recursion and feedback on it, so the prediction result is better than that of unidirectional LSTM. At the same time, the attention mechanism is used to assign weights to each time step of the BiLSTM model, which further improves the prediction accuracy of the model. Through experiments, the R2 coefficient of determination of the Attention-BiLSTM model proposed in this paper reaches 0.81, which is higher than the accuracy of the unidirectional LSTM and the BiLSTM model without the introduction of the attention mechanism. It provides a new method to accurately predict oil well production and guide oilfield development decisions.