Article: Evaluation of factors affecting expansion of weak-base ASP flooding based on grey correlation analysis combined with BP neural network Journal: International Journal of Environment and Pollution (IJEP) 2026 Vol.76 No.6 pp.23 - 43 Abstract: This study proposes an innovative coupled evaluation framework that integrates grey relational analysis (GRA) and an improved backpropagation neural network (BPNN) to address the challenge of high prediction uncertainty caused by complex multifactor interactions in weak alkali, surfactant, and polymer (ASP) flooding. Firstly, a multidimensional dataset consisting of 12 geological and engineering parameters is constructed through systematic data preprocessing and feature extraction; subsequently, GRA is used to quantify the dynamic correlation between each factor and the swept volume expansion, enabling the screening of seven main control variables based on threshold values; then, a three-layer feedforward PNN is developed using momentum term and adaptive learning rate optimisation to accelerate convergence and enhance nonlinear mapping capability. The experimental results indicate that the proposed method accurately identifies porosity, permeability, and polymer concentration as key influencing factors. The average grey correlation coefficients are 0.86, 0.83, and 0.79, respectively. Inderscience Publishers - linking academia, business and industry through research

Title: Evaluation of factors affecting expansion of weak-base ASP flooding based on grey correlation analysis combined with BP neural network

Authors: Bo Sun; Jingming Fan; Min Wang; Yunfeng Zhang; Qinan Chen; Meiling Jiang

Addresses: Daqing Petroleum Administration Bureau, Northeast Petroleum University, Daqing, 163113, Heilongjiang, China ' The 3rd Oil Extraction Factory of Daqing Oilfield Company Limited, Daqing, 163113, Heilongjiang, China ' The 3rd Oil Extraction Factory of Daqing Oilfield Company Limited, Daqing, 163113, Heilongjiang, China ' Northeast Petroleum University, Daqing, 163113, Heilongjiang, China ' Geological Research Institute of Daqing Oilfield First Oil Production Plant, Daqing, 163113, Heilongjiang, China ' Daqing Petroleum Administration Bureau, Northeast Petroleum University, Daqing, 163113, Heilongjiang, China

Abstract: This study proposes an innovative coupled evaluation framework that integrates grey relational analysis (GRA) and an improved backpropagation neural network (BPNN) to address the challenge of high prediction uncertainty caused by complex multifactor interactions in weak alkali, surfactant, and polymer (ASP) flooding. Firstly, a multidimensional dataset consisting of 12 geological and engineering parameters is constructed through systematic data preprocessing and feature extraction; subsequently, GRA is used to quantify the dynamic correlation between each factor and the swept volume expansion, enabling the screening of seven main control variables based on threshold values; then, a three-layer feedforward PNN is developed using momentum term and adaptive learning rate optimisation to accelerate convergence and enhance nonlinear mapping capability. The experimental results indicate that the proposed method accurately identifies porosity, permeability, and polymer concentration as key influencing factors. The average grey correlation coefficients are 0.86, 0.83, and 0.79, respectively.

Keywords: weakly alkaline surfactant polymer flooding; data-driven; evaluation of influencing factors; GRA; grey relational analysis; BPNN; back propagation neural network.

DOI: 10.1504/IJEP.2026.153606

International Journal of Environment and Pollution, 2026 Vol.76 No.6, pp.23 - 43

Received: 30 Jul 2025
Accepted: 16 Dec 2025
Published online: 18 May 2026 *

Title: Evaluation of factors affecting expansion of weak-base ASP flooding based on grey correlation analysis combined with BP neural network

Keep up-to-date