A Method of Oil Well Production Prediction Based on PCA-GRU
Hongtao Hu, Jinrong Feng, Xin Guan · 2019
Accurately predicting oil well production is an important task in the process of oilfield development. It is the key to formulate oilfield development plans and achieve reasonable development goals. In order to improve the prediction accuracy of oil well production, this paper proposed a recurrent neural network prediction model based on Principal Component Analysis and Gated Recurrent Unit to accurately predict the oil well production quantity. The prediction model is built with the dynamic changing pattern of production data with respect to time series. Firstly, analyze and select the main factors that affect the production in the process of oil well production, then reduce the dimension of the selected factors through the principal component analysis method to eliminate the correlation between the influencing factors. Secondly, take the main factors after dimension reduction as the neurons to the network input layer. Finally, establish the PCA-GRU oil well production neural network prediction with python. Experimental results show that this model provides a more accurate, more stable and more valid prediction result compared with other oil prediction models, proving that it is applicable to industry oil well production.