Structural optimization of downhole axial-flow cyclone separation pipe string

Single-well injection-production technology (SWIPT) is an effective solution for the economic development of high-water-cut oilfields, with downhole oil-water separation playing a crucial role in ensuring its efficient operation. In order to meet the 5.5-inch casing size requirements of onshore oilfields and enhance the separation performance of single-well injection-production technology, a dual-pump suction single-well injection-production technology (DPS-SWIPT) was innovatively proposed, and an axial-flow cyclone separation pipe string (AFCSPS) was designed. The significance of structural parameters affecting the AFCSPS separation performance is calculated based on the Plackett-Burman (PB) design method. Response surface methodology (RSM) and artificial intelligence (AI) algorithms were employed to optimize the structural parameters of the cyclone separation pipe strings. A multivariate response mathematical model has been developed to quantitatively describe the relationship between structural parameters and separation efficiency. A comparison of the optimization methods reveals that the sparrow search algorithm back propagation neural networks-genetic algorithm (SSABP-GA) optimization algorithm yields the best results, with the value of separation efficiency reaching 99.93 %. The separation efficiency of the hydrocyclone remained above 95.38 % across various operating and physical parameters, indicating that the optimal structure demonstrates a high degree of adaptability within the studied range. Numerical simulations and experimental results show good agreement for oil core morphology and separation efficiency, within an average efficiency error of 1.48 %. The study can provide support and reference for the application of the equipment in downhole oil-water separation technology.