Answers are grounded in this paper and cite their sources.
During the oil extraction process, wax accumulation on the surface of pumping units reduces oil production efficiency and significantly shortens the service life of equipment. Wax deposition is closely related to temperature; therefore, accurately predicting temperature changes during crude oil extraction is a key step in foreseeing wax deposition problems in advance. In this study, the temperature time-series data collected from actual oilfields were analyzed. Considering that such temperature data exhibit complex and nonstationary characteristics, which pose challenges to traditional prediction models, a residual hybrid long short-term memory network (R-LSTM) was adopted to construct the prediction model. Among it, the residual structure can effectively alleviate the gradient vanishing problem during network training, ensuring that the model can efficiently capture the long-term dependencies in the temperature time series. The R-LSTM model proposed in this study is used to predict the medium-term temperature of oil wells within the next 12 time steps. Comparative experimental analysis with traditional models [recurrent neural network (RNN), LSTM, and gated recurrent unit (GRU) models] and a hybrid model [a convolutional neural network with self-attention (CNN-SA)] shows that the mean absolute percentage error (MAPE) of the proposed model is 0.52%, and its prediction accuracy is significantly better than that of other comparative models.
No takes yet. Share an insight, caveat, or question.
Hou et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: