The electrical submersible pump (ESP) is one of the most widely used artificial lift methods in the petroleum industry. Although not recommended to be used in sand production well, ESP is still applicable in high producing well with a minimal percentage of solid concentration. Besides, the temporarily produced fracture sand from the proppant backflow can also severely reduce ESP boosting ability in weeks or months. Therefore, it is crucial to study the wear in ESP stages under sandy flow condition. Various erosion equations and models were developed for different materials and affecting factors. However, the predictions of these erosion models in ESPs need to be evaluated to make a proper selection. Comparisons of wear patterns and erosion rates were conducted using the computational fluid dynamics (CFD) software ANSYS. In order to validate the simulation results, an experimental facility was designed and constructed to study the sand erosion process in an ESP. Stages were painted to obtain erosion patterns, and the weight loss was measured. Six erosion models were implemented in the simulations to select the most accurate one in predicting ESP erosion rates. Then, three ESPs, including two mixed-type pumps and one radial-type pump, were modeled to study the effect of pump types with the selected erosion model. Finally, the steady-state discrete phase model (DPM) erosion simulations were carried out to investigate particle density and size effects.

Differential pressure (DP) between stages of multi-stage diffuser type centrifugal pump of an electric submersible pump (ESP), system was measured in a two-phase flow environment. The test results were used to improve existing numerical models and bubble size models for better prediction and to improve the design process. This paper details the comprehensive testing and model validation process for ESP pump stages. Various sizes and types of ESP pump stages were tested in a high-pressure two-phase flow loop. Each stage performance was monitored using high-precision DP transducers. Test results were used to calibrate numerical simulation in two-phase flow. The correct bubble size for individual test conditions was calculated based on the test results. The bubble size calculation was crucial for good model prediction in two-phase flow. After several types and sizes of ESP stages were validated, same bubble size model could be used on other pump stages that have similar sizes and designs. Presented in this paper are the actual gas volume fraction (GVF) handling capabilities of typical ESP stages, including the effects of flowrate, GVF, pump speed, inlet pressure, flow mixing and inlet/outlet effects. After the lab test was complete, computational fluid dynamics (CFD) was used to investigate the hydraulic performance of the same stages and conditions tested in the lab. Simulation results were compared with test results and optimized for the correct bubble size of the secondary phase. The current bubble size model predicted very well for the lower GVF combined with large flow rate pump. The low flow pump had DP errors in excess of 10% for GVF values over 40%. The validated numerical model can be used to improve pump modeling and enable improvements in pump design that could increase the gas-handling capabilities of pump stages of similar size and styles.

Computational Fluid Dynamics (CFD) is used to investigate the hydraulic performance of a centrifugal pump within the electrical submersible pump (ESP) unit in single-phase flow. The geometry consists of a three-stage centrifugal pump with an impeller and a diffuser in each stage. The stage performance is influenced by the inlet and outlet conditions of the stage, and therefore, three stages were modeled. The simulations were run at 3,500 RPM for various flow rates within the operating range. The k-ε turbulence model and the shear stress transport (SST) turbulence model were used to compare the capabilities of the model on performance predictions. Simulations were run in steady and unsteady flow conditions with a single vane and a full pitch model. Hydraulic performance such as efficiency, pump head, and break horse power (BHP) obtained from numerical analysis were compared with the test results to validate the CFD model. The comparison results revealed that the CFD overpredicts the pump head and underpredicts the BHP by 5 to 10%. The discrepancy between measurements and predictions are reasonable because the hydraulic leakage and bearing power losses are not modeled in CFD. The overall predicted efficiency is higher than the measurements because of overpredicted head and underpredicted BHP. Comparing numerical simulations with different turbulent models showed no significant difference between the k-ε model and the SST model. The steady/ unsteady flow comparison also showed similarity in the hydraulic performance near the best efficiency point. For design purposes, steady flow simulation with a single vane and the k-ε model were used to cut computational time.

Electrical submersible pump (ESP) is one of the most widely used artificial lift methods in the petroleum industry. It is crucial to study the wear in ESP stages with sand production, which can severely reduce pump performance and life span. Usually, experiments and simulation studies were conducted for simple flow geometry such as direct impingement and pipe elbow. Various erosion equations and models were developed for different material and affecting factors. However, the predictions of these erosion models for complex flow geometry need to be evaluated in order to make a proper selection. This study will compare the wear patterns and erosion rates of six different erosion models in three ESPs by using commercial Computational Fluid Dynamics (CFD) software ANSYS Fluent. The results will offer engineers a brief guidance of erosion model selection for complicated flow domain. In this paper, stages of three ESPs, DN1750, TE2700 and Flex31, are modeled. For each pump, the flow domain of two stages are selected and high-quality structured meshes, comprising around 1.2 to 1.8 million hexahedral grids per stage, are generated by ICEM or Turbogrid. Six erosion models, Ahlert (1994), Haugen (1995), Zhang (2007), Oka (2007), Mansouri (2014) and DNV (2015), are simulated under pump best efficiency point. Among six selected erosion models, Ahlert (1994) gives a much higher wear rate than others, while DNV predicts lowest, Besides, the impact angle functions show that all models, except Haugen (1995), assume steel to be a ductile material. Furthermore, the erosion pattern, location, and distribution of all three pumps are different, which indicates different solid particle handling capabilities and failure reasons of radial type and mixed type ESPs.

As one of the most widely used artificial lift methods, electrical submersible pumps (ESPs) have been improved gradually since the 1910s. However, its performance and run life are affected by many problems such as gas lock, high viscosity fluid, corrosion, and erosion. With the development of horizontal well drilling and multistage hydraulic fracturing, sand production from unconsolidated sandstone and proppant backflow often cause severe damage to ESPs resulting in reduced operating lifespan. Measuring wear in an ESP pump and monitoring performance degradation is not only very difficult in field cases, but also in experimental studies. The results are precious for understanding the wear mechanism inside an ESP as well as guiding the ESP design and simulation. At the same time, vibration and performance data can provide significant guidance to ESP failure diagnosis, which can potentially reduce the time and cost of well service and extend ESP run life. Wear processes inside an ESP can be classified by different modes of mechanisms. Erosive wear can be observed in the primary flow channel of the impeller (rotor) and diffuser (stator). Particle strike shroud surfaces and the scratched material is flushed away by fluids. Various semi-mechanistic erosion equations are available to be coupled with Computational Fluid Dynamics (CFD) to predict the erosion in ESPs. In the secondary flow region, balance chamber and sealing rings, particles are presented between the stator and the rotating rotor. Therefore, abrasive wear is believed to dominate the wearing process. Unlike erosion, abrasion is more complicated and abrasion equations are highly depended on geometries, physical mechanism and load between particle and target surface. In this study, a sand wear test flow-loop is designed and constructed to investigate wear in ESPs. Performance degradation, Erosion pattern, abrasion rate, and stage vibration of an ESP were recorded in a 64-hour sandy flow test.