This paper summarizes the work of the exploratory phase of a Joint Industry Project (JIP) investigating operational instabilities and ‘no flow’ events observed in a large number of Steam Assisted Gravity Drainage (SAGD) wells produced with ESPs, where, while the ESP is operating, flow to surface suddenly stops. The impact of these events to the SAGD operations has included lost production due to downtime and reduced drawdown, additional stresses to the ESP due to operating without flow for extended periods and repeated shut-downs and restarts and some system failures. The primary objectives of this exploratory phase were to better understand the mechanisms and key factors responsible for these no flow events, to identify possible mitigation actions, which may be in the wellbore trajectory, ESP landing position, ESP system component designs, or operating practices. This work was structured as a series of progressive, inter-related tasks, using a combination of analytical models and Computational Fluid Dynamics (CFD), to systematically examine the components of an ESP completion and to narrow-in on the primary contributing factors. The approach was adopted in an effort to assess if the instabilities were due to flow conditions upstream of the ESP, past the ESP, through the intake or in the initial stages of the pump. For each step of the analysis, the output boundary conditions of an upstream analysis became the input boundary conditions of the corresponding downstream assessment. First, a combination of one-dimensional (1D) multiphase flow and multiphase CFD models were used to characterize the fluid conditions and flow behaviour in the lateral and heel of the well below the ESP. Second, flow past the ESP motor was examined using CFD models to examine the impact of motor heating on subcool reduction and steam vapour generation. CFD simulations were also used to examine fluid separation and flow into a bottom feeder ESP intake to assess the amount of non-condensable gas (NCG) entering the ESP and if the pressure drop through the intake was sufficient to cause significant steam vapour flashing. Finally, a representative SAGD ESP stage was analyzed using both a broad suite of analytical surge models to assess the stability of the pump given the gas-liquid fraction entering the pump, as well as CFD simulations of the rotating stage. The focus of the CFD assessment within the entrance and first stage of the pump was to determine whether vapour flashing was occurring within the ESP stage and the apparent impact on the stage’s ability to generate head. The results from this exploratory phase of the JIP indicated that vapour flashing within the ESP impeller due to insufficient required Net Positive Suction Head (NPSHr) appeared to be one of the dominant mechanisms causing no flow events, as opposed to NCG or steam entering the pump from the intake. Future work for this JIP includes validation of the CFD results using lab test data, further CFD simulations of other ESP stage designs and at a wider range of operating conditions, and examination of alternate ESP designs that may allow for production at lower subcool and lower NPSHr values.
SAGD operators are actively installing Inflow Control Devices (ICDs) in their SAGD production wells to enable production at low or ‘negative’ subcool values in order to maximize drawdown, increase fluid production rates and optimize SAGD economics. Operating the wells under higher vapour conditions could expose the ICDs to high velocity steam carrying sand particles, causing erosion and liner failure. A staged approach has been developed for assessing the relative erosion risk of candidate ICDs, including a stage one qualitative assessment, followed by analytical erosion calculations and Computational Fluid Dynamics (CFD) modeling. In the first stage qualitative assessment, factors such as the anticipated fluid flow path, relative fluid velocity, and ICD housing design are used as inputs to estimate relative erosion rates, identify ICD surfaces where erosion rates are anticipated to be high and to rank ICD candidates based on relative erosion risk. The evaluation of potential erosion risk included consideration of both the predicted erosion severity and the probable consequences of erosive wear, which could include loss of the desired flow control response or damage to the integrity of the body or housing of the ICD. The result of the first stage is a short-list of the top-ranked ICD design candidates. In the second stage, analytical erosion models are then used to quantify erosion rates on identified target surfaces. Flow velocities and impact angles identified in the first stage are used with analytical flow relationships, such as equations for expansion jets downstream of the flow control elements, as inputs to selected erosion models. The result of this second stage is an updated short list of top-ranked ICD designs. Finally, coupled solid particle and fluid multiphase CFD simulations are conducted on a small set of ICD design candidates to obtain detailed results on the location and severity of erosion within the flow control element and on all the surfaces of the designs. The results of this stage are specific identification of susceptible erosion surfaces and rates, providing information for the selection of the appropriate ICD design. At each stage, the number of ICD candidate designs under consideration is reduced. In this manner, simpler methods of analysis (qualitative assessment) may be readily applied to a large number of devices, while more intensive modeling approaches (multiphase CFD) are reserved for a smaller set of design candidates. This approach provides insights into erosion risk based on successively more in-depth analysis methods, including particle paths and erosion locations in the CFD analysis stage which may not be identified using higher-level analytical methods. These results of the CFD analysis could then be used to help improve accuracy when applying analytical erosion models.
This paper summarizes the work of a client project investigating an "unexpected flow phenomenon" observed for an operational downhole jet pump installed in a deep, high-pressure, light hydrocarbon-condensate well in the Duvernay region of Western Alberta. The jet pump appeared to allow the well to produce multiphase condensate, water, and Non-Condensable Gases (NCGs) without the injection of power fluid. Due to a lack of understanding of the cause of this phenomenon, the well production rate could not be predicted for future installations. The objective of this work was to understand the underlying mechanisms causing this flow occurrence, and subsequently use the findings to optimize the artificial lift pump without the use of the injection system. This study was structured into three tasks: creating a custom fluid model for the Duvernay well; importing the custom model into Computational Fluid Dynamics (CFD) simulations to model the flow through the jet pump; and verifying the accuracy of the simulations with a coupled wellbore analysis of the Duvernay system. The fluid model was created with an in-depth Pressure-Volume-Temperature (PVT) analysis using the chemical composition of the production fluid determined from field samples. The Duvernay fluid model consisted of density, viscosity, and specific heat relationships as a function of the local temperature and pressure. The fluid model was implemented into three-dimensional (3D) multiphase flow CFD simulations of the existing jet pump to characterize the flow. The results showed that the jet pump nozzle created a sharp pressure drop triggering the hydrocarbon mixture to flash from a supercritical fluid phase to a gas-liquid mixture resulting in a gas-lift effect that produced flow to surface. A one-dimensional (1D) radial wellbore analysis was conducted for a large range of production flow rates at the current field wellhead pressure to generate a well outflow curve. The discharge pressure of the CFD results was compared to the wellbore pressure at the pump depth to verify the results of the simulations. Using further CFD simulations, the pump was optimized by changing the nozzle and diffuser designs to reduce downstream turbulence and improve discharge pressure recovery. Lastly, a parametric study was conducted using CFD for multiple mass flow rates and nozzle diameters to create a semi–empirical model to predict production rates for any given pump size. This model was used for a second case study to test feasibility in future Montney region wells, in which optimal pump specifications were sized based on the region's downhole properties. The results of this case study showed that this new optimized pump has the capability to produce flow rates in wells that do not naturally produce condensate, with a predictive model having been developed to choose the ideal pump geometry specifications to maximize the outflow.
Major challenges are foreseen in quantitative risk assessment of ILI detected crack-related features for thin-wall pipelines due to disproportionate ILI sizing uncertainties relative to pipe wall thickness. Therefore, the likelihood of defects growing into through-wall cracks, leading to product leakage, even at relatively low operating pressures needs to be considered in thin-wall pipelines. To support quantifying the risk associated with operating such pipelines, leak rate simulations were conducted to help with the release consequence assessment and risk ranking of ILI reported crack features to design an appropriate mitigation plan. Finite element analysis (FEA) and computational fluid dynamics (CFD) methods were used to determine the physical characteristics of through-wall cracks and the resulting leakage rates. The study highlights that, for a given fluid, the threshold for leak occurrence and the leakage rate depend primarily on the crack geometry and the operational pressure. CFD simulation results for the sensitivity cases modelled in this study showed that the leak rate can become very significant as the crack opening and internal pressure increase. These CFD results were then compared with the results obtained from a closed-form analytical model. It was determined that the analytical model started to deviate from the CFD results as the internal pressure increased and the crack opening became larger. This was explained by the fact that the analytical model was intended to be used for single-phase flow under laminar, isothermal conditions. Since its applicability to the turbulent flow regime has not been established, the deviation between the CFD results and the analytical results suggests that the use of the analytical model in the turbulent flow regime could greatly underestimate the leak rate. In addition, the importance of the design of experiment, and proper modeling of turbulence and crack surface roughness in the leakage rate estimation was demonstrated.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.