Balance and Correlation Analysis of Oilfield Injection-Production System Based on Data Mining

Abstract

Appropriate water injection volume is the most basic parameter required to maintain stable formation pressure and ensure the development effect of water flooding in oilfields. However, the determination of appropriate water injection volume has always been a major problem in oilfield water injection management. Based on the Grey relational algorithm, this paper determines the optimal connected injection-production well group through the study of the dynamic relationship between oil and water wells. To specifically predict water injection, a Sparrow search algorithm optimisation model based on Sine mapping is proposed. A Sine-SSA-BP algorithm was devised to predict water injection volume and both the improved algorithm and the original BP algorithm were applied to real-world data to assess their predictive accuracy. The prediction results of the Sine-SSA-BP algorithm were found to be closer to the true value than the results of the original BP algorithm, and the average error percentage is reduced by 23.86%. Therefore, the new algorithm can predict and calculate the water injection volume more accurately. The research content of this paper can provide a theoretical basis for advising adjustment measures in the block to slow down the rise of water content, maintain stable production, and improve the efficiency of the mechanical mining system.