Mechanistic modeling of gas effect on Multi-stage Electrical submersible pump (ESP) performance with experimental validation

• Published in: Chemical engineering science 2022-04, Vol.252, p.117288, Article 117288
• Main Authors: Zhu, Haiwen, Zhu, Jianjun, Zhang, Hong-Quan
• Format: Article
• Language: English
• ISSN: 0009-2509; 1873-4405
• DOI: 10.1016/j.ces.2021.117288

•A mechanistic gas–liquid two-phase ESP performance model is proposed in this study.•The model estimates gas bubble size, flow pattern, drag coefficient, etc., to predict the ESP head.•Two types of two-phase tests, namely surging test and mapping test, were conducted for three ESPs.•For the first time, the model is validated with three different types of ESPs with acceptable accuracy (20% error).•The proposed gas bubble size equation can also be used in CFD simulations. The performance of widely used Electrical Submersible Pump (ESP) is significantly affected by gas entrainment, which is a commonly encountered phenomenon in the petroleum industry. The boosting pressure of an ESP gradually degrades with the increase of inlet gas volumetric fraction (IGVF). The flow becomes unstable, and a breakdown occurs when the flow pattern switches from bubble flow to intermittent flow. Therefore, an accurate ESP model is necessary to help design and operate the ESP system under gassy flow conditions. The gas–liquid two-phase tests of different ESPs were extensively carried out at the Tulsa University Artificial Lift Projects (TUALP) to investigate the complex flow behaviors. A mechanistic model was proposed for the gas–liquid flow inside a rotating ESP, which captures the gas–liquid two-phase flow characteristics, including in-situ gas void fraction, flow pattern transition, bubble size, boosting pressure, etc. In the new model, the pump head is calculated by subtracting recirculation head loss, turning head loss, friction head loss, and leakage head loss from the Euler head. The best-match flowrate QBM approach is introduced, and the recirculation head loss is generated by the mismatch between the velocity of Qin and QBM. The drag force coefficient correlations are selected based on flow patterns and bubble sizes. Then, the mixture density and in-situ gas void fraction are calculated according to the force balance on gas bubbles. The proposed model is validated by experimental data of a 5-inch radial-type ESP (TE2700), a 5-inch mixed-type ESP (GC6100), and a 4-inch mixed-type ESP (MTESP) of the TUALP. The predicted ESP boosting pressures and flow patterns agree well with the corresponding experimental measurements.