Numerical Simulation and Performance Analysis of DesanderDuring Tight Gas Provisional Process

Abstract

Tight gas wells in Southwest oil and gas fields have significant production and high sand output intensity. The sand out of the wellhead has a certain erosion effect on the downstream pipeline, the equipment, and affects the normal production. This paper models and simulates the desander used at the wellhead according to the real parameters of the tight gas wellhead, and explores the effects of gas production, pressure, temperature, sand particle size, water content, and other factors on the desander’s sand removal efficiency. This paper combines the principle of fluid dynamics to analyze the internal mechanism of the effect trend and according to the simulation results uses the Pearson correlation coefficient quantification of the effect of each operating parameter to explore the optimal boundary condition parameters applicable to the desander. From the simulation results, it can be seen that the separation efficiency of the desander is the highest when the gas production rate is 4 × 10⁴ m³/d, the pressure is 7 MPa, and the lower the working temperature is, the larger is the gravel particle size. Combined with the sand management problems occurring in the field of tight gas wells, suggestions are made for the optimization of the operating parameters and structure of the desander, which will provide a basis for supporting the rapid production and large-scale beneficial development of tight gas fields.

1. Introduction

1.1. Motivation

Tianfu gas field, located in the central part of Sichuan basin, is currently the most important production area of tight gas of the PetroChina Southwest Oil and Gas Field Branch. In 2022, 809.25 × 10⁸ m³ of predicted reserves were submitted in the section of sand. The plan to submit the natural gas control geological reserves in 2023 is 947 × 10⁸ m³, of which the proved reserves amount to 25.447 billion cubic meters, which has a large potential for development. Zitong block tight gas well has a design of 10 pilot wells and a test mining scale of 65 × 10⁴ m³/d. There is a platform well station around the discharge, stable production, three decreasing stages of ground construction, and the discharge stage uses an integrated discharge skid. The wellhead flow pressure is 26.6~28.8 MPa, the flow temperature is 25 °C~40 °C, and the natural gas treatment demand focuses on separation, desanding, dehydration, and dehydrocarbonization.

According to the production distribution of wells and stations, the current situation of the regional pipeline network, and the production and marketing situation, Zitong block tight gas field adopts the main process program of “multi-well gas gathering, gas-liquid mixing and transmission, J-T valve dewatering and dehydration and dehydrochlorination, and LNG recovery”. The gas from each well is heated, throttled, desanded, and metered, and then mixed with gas and liquids and transported to the downstream station for processing, with the natural gas sold locally and the LNG exported, and the condensate hauled to the downstream sales unit. In the process of tight gas field extraction, due to the complex geological conditions of the gas reservoir, a large number of solid particles may appear in the wellbore along with the flow of natural gas, and failure to effectively remove the sand particles will lead to the wear and tear of the downhole equipment and decreased production capacity, which may cause damage to the gas reservoir [1]. At the same time, it will also have a significant effect on the surface gathering and transportation system, such as pipeline sand accumulation and blockage, pipeline wear, pipeline corrosion, pipeline valve damage, etc. [2]. The Zitong block tight gas production process, due to the poor sand removal effect, has an obvious existing sand plugging effect, sand discharge storage management is missing, and there are pipeline equipment sand cleaning and unblocking difficulties and other problems [3]. At present, most of the sand particles in the production process of gas wells during the discharge period cannot be precipitated in time and then carried with the gas and water flow into the downstream pipeline and equipment, and only a trace amount of sand particles are separated out in the desander. The sand production of gas wells is analyzed by manually taking samples from the sand discharge port of the desander, there is a lack of online real-time monitoring tools and means, and the sampling situation does not match with the actual sand production. Therefore, effective sand management is essential for the sustainable and economic development of tight shale gas resources.

Some field stations use a single-cylinder horizontal desander, which is poured into the bypass production when the cartridge is cleaned, and there is no secondary separation equipment, during which the sand production will enter the downstream pipelines and field stations. Sand produced from gas wells enters downstream pipelines and field stations and is deposited in pipelines, equipment, and instruments. This will cause malfunctions such as an increase in the differential pressure, a decrease in pipeline transmission efficiency, and shutdown of automatic control valve protection. At the same time, during the sand discharge process, the gas well production liquid will flow out with the sand particles, and there is no specific equipment or measures to recover and store the sand discharged from the desander. The mixed liquid also contains a trace amount of condensate, which does not meet the management requirements when placed in the open air, and the liquid is recovered by manually lifting and placing it in the tanker loading area. The sand samples from the site are shown in Figure 1. In this paper, a large amount of sand was produced at the wellhead during the discharge period of tight gas wells in Zitong block. Fluent 2022 R1 software was used to simulate and model the desander used on site, analyze the boundary parameters of the existing desanding process, evaluate the adaptability of the desanding device, and propose improvement measures for the sand discharge problem on site.

1.2. Literature Review

In the production process of tight gas wells, there is a pressure difference between the inside and outside of the well wall, which will prompt the sand in the formation at the bottom of the gas well to come out with the pressure. At the same time, the wall of the gas well or the internal rock layer will be loosened and dislodged due to abrasion and erosion after a long operation period. This will also lead to the sand coming out of the gas well. The sand problem is common in the production process of tight gas wells, which seriously affects the production efficiency and equipment life.

In order to solve the sand problem in gas wells, wellhead desanders have become key equipment. As early as the 1990s, Krebs petroleum technology company and Shellexpro company developed and used wellhead desanders. The Mozley wellspin desander developed by Natco has good relocatability, can be easily transferred between different gas wells, and has been adopted by many users. The integrated desanding system developed by Merpro (UK) is more comprehensive, not only separating and transferring, but also providing an on-line cleaning function, and even realizing direct offshore desanding. These early developments have laid the foundation for the development of wellhead desanders. The current challenges for desanders include the separation of ultra-fine particles, integration with existing systems, and implementation of advanced monitoring and real-time data analysis for predictive maintenance.

Although the existing wellhead desander has solved the problem of gas well sand discharge to a certain extent, it still has many shortcomings. The inner wall of some desanders is easily worn, which leads to decreased separation effect and shortened service life. The sand removal methods of some desanders are defective, such as manual sand removal and pressure relief sand removal during shutdown, which are not only inefficient, but also affect the continuity of production. The precision of sand removal also needs to be improved, and it is difficult for some of the existing sand removers to meet the needs of environments with large sand content.

Numerous scholars have researched the effect factors of the separation efficiency of desanders through various simulation methods and given suggestions for structure or parameter optimization. Basyouny. A [4]. focused on numerical simulation and experimental validation of two-phase flow in a horizontal separator, aiming to probe deeply into the physical properties of multiphase flow in a gravity separator. Alves et al. [5] evaluated the modularization of the separation performance of a mini-hydrocyclone to analyze the effect of geometrical parameters (e.g., cone angle, length of cylindrical section, diameter of the underflow opening) and operational parameters on the separation efficiency. Liu et al. [6] demonstrated the explicit law of the effect of pressure, sand grain size, mass flow rate, density, and shape factor on the erosion rate by CFD simulation in shale gas wells. Liang et al. [7] carried out experimental research on offshore platform section plug-flow traps and proposed an internal component optimization scheme to optimize the sand control performance by installing a cyclone or baffle inside the trap, which provides a new idea for sand control in offshore platform production equipment. Kou et al. [8] proposed a novel cyclone separator with a conical inner core. The effect of its internal particle motion behavior and key structural parameters on the separation performance is analyzed by numerical simulation. The design basis is provided for improving the separation efficiency of multiphase flow. Jing et al. [9] proposed the desanding scheme of water doping and viscosity reduction combined with cyclone separation for the sand separation problem in offshore thick oil extraction. The desanding characteristics of the cyclone under different working conditions and structural parameters were thoroughly researched by numerical simulation of computational fluid dynamics (CFDs). Gorobets, A. V. [10] explored the separation performance of desanding and de-oiling hydrocyclones processing three-phase feeds (oil, particulate silica, and water), focusing on the effect of oil particle aggregates (OPAs) on the separation efficiency. Liang et al. [11] examined the hazards of the sand problem on downhole tools when gas wells are flooded to resume production. In particular, the traditional downhole cyclone is insufficiently researched for the separation of 10 μm fine particles, so an axial cyclone with helical fins was designed. The aim is to improve the separation efficiency of fine particles.

Although existing researchers have simulated and modeled the desander through various simulation methods, they have not quantified the effect of each factor on the desander efficiency in a fine way to analyze the magnitude of the effect of each factor. Moreover, most of the current research has focused more on experimental simulations in the laboratory, without research on the specific gas wellhead, and with a lack of real production data at the gas well site. Therefore, this paper simulates the desander flow field on the basis of collecting actual data from the gas field site, and according to the simulation data analysis, quantifies the importance of each effect factor, and puts forward optimization suggestions for the desander operating parameters.

1.3. Contributions

In this paper, research on the adaptation of a gas wellhead desander for a tight gas field is conducted, and the main contributions are as follows:

(1) Combined with the actual on-site production of gas wells in a tight gas field, the wellhead desander device structure of a specific gas well is disassembled, and the actual on-site production data are collected for simulation modeling;

(2) A fluent simulation model is established to investigate the effects of flow rate, pressure, temperature, and sand particle size on the desanding efficiency, and reveal the internal mechanism of the change in desanding efficiency;

(3) The effect of flow rate, pressure, temperature and other factors are quantitatively analyzed using Pearson’s correlation coefficient, and optimization suggestions are put forward for the use of an on-site desander.

2. Systematic Description

2.1. The Role of Desanders in Sand Management

Sand management strategies can broadly be categorized as exclusionary (preventing sand from entering the wellbore) or inclusionary (managing sand that is produced to the surface) [12]. While downhole methods like gravel packs and screens aim for exclusion, surface-based separation technologies are essential components of inclusionary strategies. Among these, desanders play a pivotal role in separating solid particles from multiphase production fluids (gas, oil, and water) at the surface.

Desanders are specifically designed to remove sand and other solids, thereby protecting valuable downstream equipment from erosion and blockages. By ensuring a cleaner fluid stream, they help maintain production throughput, improve the quality of produced hydrocarbons, enhance the reliability and safety of operations, and reduce maintenance frequency and costs. The timely removal of abrasive solids minimizes wear on critical components, extends equipment service life, and ensures the overall stability and efficiency of the production system [13].

The core principle behind cyclonic desanders is the exploitation of centrifugal force to achieve phase separation [14]. The process typically involves the following steps:

(1) Tangential Inlet: The multiphase production fluid enters the cylindrical section of the desander tangentially at a specific velocity [15,16]. This tangential entry imparts a strong rotational or swirling motion (vortex) to the fluid mixture.

(2) Centrifugal Separation: The swirling motion generates high centrifugal forces, which act more strongly on the denser solid particles (sand) than on the lighter fluid phases (gas, oil, water). This forces the solid particles radially outward towards the inner wall of the cyclone [17].

(3) Solids Transport and Collection: Driven by gravity and the downward component of the swirling flow, the separated solids move down along the cyclone wall and the conical section. They eventually collect in an accumulator vessel or sand collection hopper located at the bottom of the unit. Automated flushing systems are often employed to periodically remove the accumulated solids.

(4) Fluid Outlet: The lighter, cleaner fluid phases (gas and liquids) migrate towards the central core of the vortex. Here, the flow direction reverses, and the desanded fluid moves upward and exits the cyclone through a central pipe known as the vortex finder (overflow).

2.2. Factors Influencing Desander Efficiency

The separation performance of a cyclonic desander is a complex function of operating conditions, fluid properties, particle characteristics, and the desander’s design. Key influencing factors identified in the reviewed materials include the following:

(1) Flow rate/velocity: This is a critical parameter as it directly determines the magnitude of the centrifugal forces generated within the cyclone. There is typically an optimal operating range for flow rate and inlet velocity. If the velocity is too low, the centrifugal force may be insufficient to effectively separate the smaller particles. Conversely, if the velocity is too high, it can lead to increased turbulence, potential re-entrainment of separated particles into the overflow stream, and significantly increased erosion rates. While a higher velocity generally increases efficiency (decreases cut size) up to a point 26, optimizing inlet velocity is crucial for balancing efficiency and erosion. One study on an optimized design cited inlet velocities below 10 m/s as beneficial for minimizing erosion while maintaining good efficiency for particles > 20 µm.

(2) Particle characteristics: Separation efficiency is highly dependent on particle size. Larger, heavier particles are separated much more readily than smaller, lighter ones. Efficiency drops significantly for fine particles, often below 20 µm or even 50 µm. Many high-efficiency claims are qualified by a minimum particle size. The minimum particle size effectively separable in practical field systems is often considered to be around 10–20 microns [18]. The maximum particle size is limited by the desander’s internal orifices to prevent plugging, typically around one-third of the underflow orifice diameter [19].

The density difference between the solid particles and the surrounding fluid is the fundamental driving force for centrifugal separation. Higher density solids are separated more easily than lower density solids.

(3) Pressure Drop: The pressure drop across the cyclone (typically measured between inlet and overflow) represents the energy consumed to drive the separation process. Generally, increasing the pressure drop leads to higher internal velocities, stronger centrifugal forces, and thus improved separation efficiency (lower D50). However, the efficiency gains diminish at very high pressure drops, while energy consumption and erosion rates increase significantly. An optimal operating pressure drop range is therefore recommended. It is also noted that increasing the pressure difference can increase the fluid velocity in the overflow pipe, potentially facilitating the escape of particles [20].

3. Model Building

3.1. Simulation Equation

(1)

Fundamental equations of fluid mechanics

As the application software of CFD, the core of Fluent software calculation is the N-S equation [21]. In this paper, the pressure calculation of the flow field of desander is mainly achieved by solving the continuity equation and the N-S equation. The continuity equation is shown in Equation (1) and the equations of momentum conservation for the fluid in the three directions of x, y, and z are shown in Equations (2)–(4).

$${\displaystyle \frac{\partial p}{\partial t}}+{\displaystyle \frac{\partial (\rho u)}{\partial x}}+{\displaystyle \frac{\partial (\rho \nu )}{\partial y}}+{\displaystyle \frac{\partial (\rho w)}{\partial z}}=0$$

(1)

$${\displaystyle \frac{\partial \left(pu\right)}{\partial t}}+div\left(pu\overrightarrow{V}\right)=-{\displaystyle \frac{\partial \rho }{\partial x}}+{\displaystyle \frac{\partial {\tau }_{xx}}{\partial x}}+{\displaystyle \frac{\partial {\tau }_{yx}}{\partial y}}+{\displaystyle \frac{\partial {\tau }_{yx}}{\partial z}}+F_x$$

(2)

$${\displaystyle \frac{\partial \left(pv\right)}{\partial t}}+div\left(pv\overrightarrow{V}\right)=-{\displaystyle \frac{\partial \rho }{\partial y}}+{\displaystyle \frac{\partial {\tau }_{xy}}{\partial x}}+{\displaystyle \frac{\partial {\tau }_{yy}}{\partial y}}+{\displaystyle \frac{\partial {\tau }_{zy}}{\partial z}}+F_y$$

(3)

$${\displaystyle \frac{\partial \left(pw\right)}{\partial t}}+div\left(pw\overrightarrow{V}\right)=-{\displaystyle \frac{\partial \rho }{\partial z}}+{\displaystyle \frac{\partial {\tau }_{xz}}{\partial x}}+{\displaystyle \frac{\partial {\tau }_{yz}}{\partial y}}+{\displaystyle \frac{\partial {\tau }_{zz}}{\partial z}}+F_z$$

(4)

(2)

Turbulence modeling

Although the large eddy simulation has a certain accuracy for desander simulation, its method has relatively high requirements for computers. So the RANS method, which has lower requirements for computers, is used for simulation. And the applicable turbulence models for desander simulation include the improved model and the RSM model [22]. The $k-\epsilon$ model is based on the vortex viscosity assumption proposed by Boussinesq, which introduces the concepts of end kinetic energy and end kinetic energy dissipation in the turbulence process, and the relationship between the Reynolds pressure and the velocity gradient is shown in Equation (5).

$$-\rho u_i^{\prime }{u}_i^{\prime }=u_t\left({\displaystyle \frac{\partial u_i}{\partial x_j}}+{\displaystyle \frac{\partial u_j}{\partial x_i}}\right)-{\displaystyle \frac{2}{3}}\left(\rho k+u_t{\displaystyle \frac{\partial u_i}{\partial x_i}}\right){\delta }_{ij}$$

(5)

where ${\delta }_{ij}$ is the strain rate of the tensor; $u_t$ is the turbulent eddy viscosity coefficient; and $k$ is the turbulent kinetic energy.

The turbulent kinetic energy is calculated as shown in Equation (6).

$$k={\displaystyle \frac{u_i^{\prime }{u}_i^{\prime }}{2}}={\displaystyle \frac{1}{2}}\left(u^{\prime 2}+v^{\prime 2}+w^{\prime 2}\right)$$

(6)

The energy dissipation rate ε for small eddies, $k-\epsilon$ is equal to the energy transfer for large-scale eddies, and the model assumes a turbulent dissipation rate as shown in Equation (7):

$$\epsilon ={\displaystyle \frac{u}{\rho }}\left({\displaystyle \frac{\partial u_i^{\prime }}{\partial x_k}}\right)\left({\displaystyle \frac{\partial u_i^{\prime }}{\partial x_k}}\right),u_t=\rho C_{\mu }{\displaystyle \frac{k^2}{\epsilon }}$$

(7)

However, the standard $k-\epsilon$ model is not applicable to the strong cyclonic distortion and the RNG $k-\epsilon$ model should be used for the desander because the swirl dominated flow in the RNG $k-\epsilon$ model is commonly used in the analysis of strong cyclonic flow. The RSM model takes more into account the anisotropy of turbulence viscosity, which is more accurate than the $k-\epsilon$ model. But the $k-\epsilon$ model performs better for some of the reflux problems, and the computational cost of the RSM model is high [23]. Therefore, the RNG model is used for the simulation of desander in this work for the corresponding simulation analysis.

(3)

Boundary condition setting

The wall of the desander is considered to be slip-free, but the near-wall surface is larger than the turbulent shear pressure due to the molecular viscosity caused by the low Re number. The above three turbulence models cannot be calculated directly, and the general treatment methods are the wall function method and the turbulence model with low Re. The wall function method is relatively simple to deal with, not needing to encrypt the grid, the wall region uses semi-empirical formulas to deal with, and only the first layer of the wall mesh has the corresponding requirements. The defined dimensionless parameters are shown in Equation (8).

$$u^{+}={\displaystyle \frac{u}{u_r}}, {\tau }_w=\rho u_{\tau }^2, y^{+}={\displaystyle \frac{\Delta y\rho u_{\tau }}{\mu }}$$

(8)

Since the flow in the desander is unpressurized, the inlet is the velocity inlet. The outlet generally has an outflow boundary and pressure outlet boundary. Generally the liquid/liquid desander diversion ratio is the operation parameter, and the liquid/solid desander diversion ratio is the performance parameter. In a liquid–liquid–solid desander, the outlet is not connected to the atmosphere and the flow size can be adjusted, so the outflow boundary conditions are more in line with the actual situation.

3.2. Field Parameters

The on-site desander is installed in the natural gas pipeline containing sand. The desander is equipped with a filter cartridge in the cylinder, which is reliably supported and connected to the inner wall of the cartridge, with a bolt holder in the upper part and connected to the natural gas outlet in the lower part. When the natural gas containing sand and gravel passes through the device, it is filtered by the filter cartridge, the sand and gravel are deposited in the cartridge of the device without affecting the normal transportation of natural gas (water), and the natural gas is discharged from the desander through the annular space after filtration. The separation efficiency ≥ 60 μm, 99.5%. The structural parameters of the desander are shown in

Table 1. Structural parameters of filtered desander.

Structure Parameters	Sizes
Inlet diameter/(mm)	89 × 14
Outlet diameter/(mm)	89 × 14
Height from inlet center point to bottom of desander/(mm)	106.5
Height from outlet center point to bottom of desander/(mm)	417
Desander cylinder diameter/(mm)	273 × 30
Total height of filter desander/(mm)	1283
Number of cartridges/(pcs)	1
Inlet type/(radial inlet/tangential inlet)	Radial direction
Outlet type/(radial outlet/tangential outlet)	Radial direction

.

3.3. Device Modeling

(1)

Three-dimensional modeling

The gas is heated at the wellhead by an electric induction heating device, throttled by an electric adjustable throttle valve, and desanded by a medium-pressure desander. Then the gas and liquid are separated and metered by a separating and metering skid. Finally, they are mixed and transferred to Zitong 3 wells by clearing the pipe out of the station valve manifold skid out of the station. According to the desander skid modeling simulation in Section 3.2, to analyze the effect of different flow rates, pressure, temperature, particle size, water production, and other physical parameters on the separation efficiency of the desander, the desander simulation model is shown in Figure 2a.

(2)

Gridding

To obtain an accurate solution, polyhedral meshes need to be generated and used for simulation. The desander was meshed using Workbench Meshing and a tetrahedral mesh was selected for numerical simulation and performance evaluation of the desander. After importing the mesh into Fluent, the mesh quality is necessary to determine the convergence of the calculations. Therefore, a local encryption operation is performed on the mesh with low mesh quality. After generating the mesh, the cyclone desander is named and selected, the gas inlet is named as the inlet, and the gas outlet is named as the outlet. The surface part is the wall, and the mesh delineation results are shown in Figure 2b.

(3)

Grid-independent verification

Irrelevance verification is a process that ensures that after a certain number of meshings are achieved, a further increase in the number of meshes has a negligible effect on the simulation results. If the number of meshes is too low, it indicates that the meshing of the fluid domain is not detailed enough, which may lead to a significant decrease in the accuracy of the simulation results. However, simply pursuing an increase in the number of meshes does not imply an improvement in mesh quality [24]. Too many meshes will not only slow down the simulation solution but also may make the computation more complicated. Therefore, a variety of factors need to be considered when choosing the number of meshes. For the same type of model, multiple divisions with different numbers of meshes are performed and simulation calculations are carried out separately. By comparing the results of these simulations, it can be determined in which range of grid numbers the simulation results are similar and thus determine this range as the optimal number of grids. This method can find the number of meshes that can guarantee the simulation accuracy while maintaining the computational efficiency.

In order to ensure that the quality of the mesh does not affect the calculation results as much as possible, it is necessary to verify the mesh irrelevance of the generated mesh. The specific method is as follows: five sampling points a, b, c, d, e are set at different locations on the cross-section of Z = 1200 m. The velocities at the locations of the five sampling points on the mesh are verified by changing the number of meshes to detection to observe the effect of grid changes on the computational accuracy of different nodes. Four different numbers of grids are used during the validation period, numbered from low to high according to the number of grids, which are M1, M2, M3, and M4 and the corresponding numbers of grids for different models are 136,340, 233,627, 283,302, and 324,099, respectively. As shown in Figure 2c, the flow velocities of the sampling points on the cross section of Z = 1200 m in the models M1~M4 are compared, and finally the total number of grid cells is chosen to be 324,099. M4 grid with a total number of grid cells of 324,099 is selected for subsequent simulations.

The quality metrics of the final delineated grid are shown in Figure 3. The cell mass of the grid is mostly in the range of 0.75–1, the aspect ratio is mostly in the range of 2–3, the grid skewness is mostly in the range of 0–0.13, and the orthogonal mass is mostly in the range of 0.88–1. The smallest cell size of the mesh is 7 mm, which is located in the cylinder and overflow pipe part of the cyclone desander, and the cell size of all other parts is 10 mm.

4. Desander Adaptability Evaluation

The model of desander established by Fluent software was used to evaluate the adaptability of desander. Since the presence of liquid water at the inlet of the desander may have an effect on the separation efficiency, the inlet fluid is set to be gas–solid two-phase and gas–liquid–solid three-phase in this paper. Inlet flow rate, pressure, temperature, sand content, sand particle size, and water production parameters were selected to investigate the separation efficiency of the desander [25]. For the sake of simplicity and accuracy of the result statistics, the mass flow rate of the different particle groups is taken as the same value. Entering the desander in a face-injection manner perpendicular to the inlet interface, the number of particles involved in the iteration, the number of trapped particles, and the number of escaped particles were counted at the end of the iteration. The separation efficiency of the desander with different structures was calculated by Equation (9). The simulation scheme of desander is formulated as shown in

Table 2. Desander simulation scheme.

Number	Influencing Factors	Effect Factor Values	Simulation Conditions
1	Flow	2 × 104 m3/d	Pressure: 6 MPa Temperature: 20 °C Sand size: 250 μm Water: 35 m3/d
		4 × 104 m3/d
		6 × 104 m3/d
		8 × 104 m3/d
		1 × 105 m3/d
2	Pressure	3 MPa	Flow: 6 × 104 m3/d Temperature: 20 °C Sand size: 250 μm Water: 35 m3/d
		4 MPa
		5 MPa
		6 MPa
		7 MPa
3	Temperature	−20 °C	Flow: 6 × 104 m3/d Pressure: 6 MPa Sand size: 250 μm Water: 35 m3/d
		−10 °C
		0 °C
		10 °C
		20 °C
		30 °C
4	Sand size	50 μm	Flow: 6 × 104 m3/d Pressure: 6 MPa Temperature: 20 °C Water: 35 m3/d
		100 μm
		250 μm
		500 μm
		1000 μm
5	Water yield	5 m3/d	Flow: 6 × 104 m3/d Pressure: 6 MPa Temperature: 20 °C Sand size: 250 μm
		10 m3/d
		20 m3/d
		35 m3/d
		50 m3/d
		100 m3/d

.

$$\eta ={\displaystyle \frac{M_1-M_2}{M_1}}$$

(9)

where, $\eta$ is the desander sand removal rate, %; $M_1$ is the desander inlet sand cumulative flow, g; and $M_2$ for the desander outlet sand cumulative flow, g.

4.1. Effect of Gas Production

Under the conditions of water content and no water content, the inlet flow rate of the desander is set as 2~10 × 10⁴ m³/d. The particle trajectory of the internal flow field of the desander is shown in Figure 4, and the separation efficiency of the desander is analyzed according to the Fluent simulation results. The Fluent simulation results show that there is a close relationship between the desander inlet flow rate and the separation efficiency. When the inlet flow rate increases, the fluid flow rate increases so that the residence time of the water inside the desander is greatly shortened [26]. The shorter residence time leads to insufficient time for the sand particles to settle, especially for the fine particles, which cannot be effectively settled within a limited time, and thus fail to be completely separated.

At the same time, the increase in flow velocity may also trigger the phenomenon of turbulence. The appearance of turbulence will seriously damage the normal distribution of the cyclone field inside the desander. Sand particles in the desander mainly rely on centrifugal force to realize the separation, and the normal distribution of the cyclonic field is the key to ensure that the centrifugal force is effective. Once the cyclone field is destroyed, the separation effect on sand particles under the action of centrifugal force will be greatly reduced. Water content has a certain effect on the separation efficiency of the desander, the gas production is the same, and the separation efficiency of the desander under the water content conditions is slightly higher than under the conditions of no water, but overall are with the increase in gas production and decline. Under the condition of water content, when the inlet flow rate is 2 × 10⁴ m³/d, the efficiency of the desander is 100%, when the flow rate increases to 6 × 10⁴ m³/d, the separation efficiency decreases to 98.26%, and when the flow rate increases to 10 × 10⁴ m³/d, the separation efficiency decreases to 90.58%.

According to the simulation results, it can be seen that when the gas production is in the range of 4–10 × 10⁴ m³/d, the separation efficiency decreases with the increase in gas production. Because the production is negatively correlated with the separation efficiency, when selecting the grit remover, the production demand in the actual production should be fully considered. According to the estimated production, the desander with a processing capacity slightly larger than the actual production should be selected to avoid reducing the separation efficiency due to the short residence time of sand particles in the desander caused by too high a production.

4.2. Effect of Pressure

Set the working pressure of the desander to 3~7 MPa, and the particle trajectory of the internal flow field of the desander is shown in Figure 5. Increasing the working pressure, within the appropriate range, the pressure increase will increase the fluid flow rate and rotational speed, which will enhance the centrifugal force effect. And the increase in centrifugal force will help the sand particles to be separated from the water, so that the larger particles can be separated out more easily. The sand removal and separation efficiency increased from 91.20% to 98.98% when the working pressure of the filtered desander in this work was increased from 3 MPa to 7 MPa under water containing conditions.

If the pressure exceeds the design range of the desander, it may cause the flow field to be unstable, affecting the settling path of the sand particles. Additionally, excessive pressure may cause the eddy current or turbulence phenomenon, so that the separated sand particles are remixed into the fluid, thus reducing the separation efficiency. It may also cause some fragile sand particles to break up and generate smaller particles, which are difficult to separate. The scouring force of the fluid and sand particles on the inner wall of the equipment increases, accelerating the wear and tear of the equipment and affecting its long-term separation performance.

When the working pressure is in the range of 3–7 MPa, the separation efficiency of the desander increases with the pressure. Although the effect of pressure on the separation efficiency is relatively small, it is positively correlated with the separation efficiency under both water-free and water-containing conditions. In the actual selection, this can be combined with the overall pressure of the system and appropriate use of higher pressure to improve the separation efficiency. However, too high a pressure may bring other engineering problems, such as equipment pressure requirements, and therefore needs to be assessed comprehensively.

4.3. Effect of Temperature

The working temperature of the desander was elevated from −20 °C to 30 °C, and the particle trajectory of the internal flow field is shown in Figure 6. The separation efficiency first decreases and then increases under water-containing conditions and gradually decreases under water-free conditions. The flow rate of the gas increases significantly as the temperature rises. The rise in temperature endows the gas molecules with more energy, which makes their movement more active, which in turn leads to the acceleration of the gas flow rate. In this case, the sand particles in the high-speed gas flow coercion, creating more power through the filter cylinder, so that the original can be effectively intercepted and the sand particles have more opportunities to pass, thus reducing the separation efficiency of the desander.

When the temperature continues to increase, the oil phase is distributed as discrete droplets in the water-continuous phase under water-containing conditions, as there is still some oil production at the gas wellhead. The viscosity of the aqueous continuous phase is much lower than that of the oil phase, the density difference between the oil droplets and the aqueous phase remains positive. The driving force of the solid particles migrating toward the wall under centrifugal force is 25% higher than that in the case of the pure oil phase. Oil and water droplets as “flexible particles” absorb turbulent pulsation energy, so that the disordered motion of the main flow field is weakened. The elastic effect of the interface between the phases inhibits the generation of small-scale vortices and reduces the escape of particles due to turbulent diffusion. As a result, the desanding efficiency rises.

On the contrary, in the absence of water, the increase in temperature also leads to a decrease in the viscosity of the fluid, which means that the internal friction within the fluid is reduced. This change weakens the interaction force between the sand particles. When the interaction force between sand particles is weakened, their dispersion in the fluid increases, making it difficult for them to aggregate and form larger particle clusters, which are then difficult to separate. Therefore, the weakening of the interaction force between the sand particles as the temperature rises under the condition of no water content further affects the separation effect of the desander, leading to a reduction in the separation efficiency.

When the working temperature is in the range of −20 to 30 °C, the separation efficiency decreases with the increase in temperature. If the working environment temperature is high, it is necessary to choose the desander that can maintain stable performance in the corresponding temperature range. In addition, consider equipping the temperature regulating device in order to maintain the appropriate desanding temperature.

4.4. Effect of Sand Size

The particle trajectories of the flow field inside the desander are shown in Figure 7 for grit sizes of 50 μm and 1000 μm. The change in the grit particle size has an important effect on the separation performance of the desander, which is manifested by the fact that the increase in the grit size usually significantly improves the separation efficiency. Larger particle size sand particles due to their own inertia are stronger; in the process of cyclone or gravity separation inside the desander, compared with small particle size sand particles, they are more likely to deviate from the mainstream direction of the fluid, and thus be guided to the separation area. This is because in the fluid environment, the centrifugal force or gravity effect on the large-size sand particles is more significant. The strong external force makes the large-size sand particles have a faster settling speed, and they can quickly move to the specific separation area, which makes it easier to be successfully separated and captured by the desander.

In contrast, sand with a smaller grain size is subjected to relatively high resistance in the fluid due to its own lighter mass. This resistance seriously impedes the settling movement of small-sized sand particles, so that their settling speed is greatly slowed down. In some cases, the small-size grit particles will even be carried directly with the flow of the fluid and cannot be effectively separated from the fluid, which ultimately leads to a reduction in the separation efficiency of the desander. According to the simulation results, when the average gravel particle size is 50 μm, the separation efficiency of the desander is 66.72%, increasing the average gravel particle size, the separation efficiency of the desander is firstly improved and then decreased. And the separation efficiency is the highest when the average gravel particle size is 250 μm, which is 98.26%.

4.5. Effect of Water Content

When the water content is 5 m³/d, 35 m³/d, and 100 m³/d, the particle trajectory in the desander is shown in Figure 8. When the water content is 20~35 m³/d and 50~100 m³/d, the larger the water content, the higher the separation efficiency. When the water content is 5~20 m³/d and 35~50 m³/d, the larger the water content, the lower the separation efficiency. At the beginning of the moderate increase in water production, the elevated fluid flow rate enhances the cyclonic field or centrifugal force of fluid flow inside the desander, and the separation effect on the sand particles under centrifugal force or gravity is enhanced [5]. In addition, an appropriate increase in flow rate prevents particles from accumulating at the bottom of the device and promotes the effective migration of sand particles to the separation zone, thus improving the separation efficiency. When the water production continues to rise beyond the optimal flow rate range designed for the equipment, the residence time of the fluid inside the desander is significantly shortened. The sand particles are carried to the outlet by the fluid before they settle sufficiently, resulting in a decrease in the separation efficiency. At the same time, too high a flow rate will lead to internal flow field disorder, the stability of the cyclonic field or settling field is damaged, and may even trigger turbulence phenomenon, so that it has been separated from the sand particles remixed into the fluid. In addition, fine particles in the high flow rate make it more difficult to overcome the fluid resistance for effective separation, thus further reducing the separation efficiency.

5. Main Controlling Factors of Separation Efficiency

According to the simulation in the previous section, it can be seen that the gas volume, pressure, temperature, sand particle size, and other factors have a certain effect on the desanding efficiency. Due to the complexity of the gas well desanding mechanism, taking into account the production cost and the working condition range of instrumentation, it is impossible to ensure that each control parameter is within the range of the highest desanding efficiency in the actual production process. For this reason, it is necessary to explore the degree and size of the effect of various influencing factors on the efficiency of desanding and analyze the main control factors affecting the efficiency of desanding so as to provide a basis for the control of parameters of desanding in on-site production. This includes the control of gas volume, pressure, temperature, sand particle size, and performance of separation efficiency correlation analysis.

Since the data of yield, temperature, and pressure in this research are continuous variables and have more obvious linear relationship, Pearson correlation analysis [27] was used in this research. The Pearson correlation coefficient is a statistical measure of the degree of linear correlation between two variables and is calculated as Equation (10).

$$r=\frac{{\displaystyle \sum \left(x_i-\overline{x}\right)\left(y_i-\overline{y}\right)}}{\sqrt{{\displaystyle \sum {\left(x_i-\overline{x}\right)}^2{\left(y_i-\overline{y}\right)}^2}}}$$

(10)

$x_i$, $y_i$ is the observed value of the two variables and $\overline{x}$, $\overline{y}$ is the average of the variables.

For the two conditions of water content and no water content, the parameter correlation analysis was carried out using 21 sets of data on the separation efficiency of the desander with different production, pressure, temperature, and grain size conditions in this work [7]. The heat map of the correlation coefficient of the separation efficiency is shown in Figure 9 (correlation threshold: |r| > 0.7 (high), 0.3 < |r| ≤ 0.7 (medium), |r| ≤ 0.3 (low)). Filter desander separation efficiency is affected by gas production, pressure, temperature, gravel particle size, water content, and other factors. After analysis, it can be seen that whether in the conditions of water content or not, the parameters of production, pressure, temperature, and particle size will have an effect on the separation efficiency of the desander. And the degree of influence shows a certain law: the degree of influence of the particle size is the largest, followed by the production, followed by the temperature, and the degree of influence of the pressure is relatively the smallest.

Specifically, in the absence of water conditions, the sand particle size and working pressure changes and separation efficiency show a positive correlation. That is as the sand particle size increases or the working pressure increases, the separation efficiency of the desander will increase. This is because with the larger particle size of the sand particles, its inertia is larger, easier to separate from the fluid in the process of desanding. And the increase in working pressure may change the flow state of the fluid, making the sand particles easier to be separated and captured. On the contrary, temperature shows a negative correlation with the change in output and separation efficiency. When the temperature increases, the physical properties of the fluid may change, for example, the viscosity decreases, resulting in a change in the movement characteristics of the sand particles in the fluid, which is more likely to flow with the fluid and difficult to be separated, thus decreasing the separation efficiency; the increase in the output means that the amount of fluid passing through the desander per unit of time is increased, and the fluid flow rate is accelerated, which causes the sand particles to shorten the retention time in the desander and decreases the probability of separation, which in turn leads to the decrease in separation efficiency.

It is noteworthy that the water content condition gave results consistent with the no water content condition. Again, the sand particle size and working pressure were positively correlated with the separation efficiency, and the temperature and output were negatively correlated with the separation efficiency, and the order of the influence of each parameter on the separation efficiency was still particle size > flow > temperature > pressure. This consistency indicates that the mechanism of the influence of these parameters on the separation efficiency of the desander has a certain stability and universality under different water content conditions.

6. Conclusions

(1) Adaptability analysis of the desander and determination of its optimal use boundary parameters can effectively improve its application performance in the tight gas field site. The separation efficiency of the gas well field desander in this research is positively correlated with the sand particle size and working pressure and negatively correlated with the temperature and production rate. Under the condition of no water content, the effect of separation efficiency is grain size > production > temperature > pressure. Under the condition of water content, the effect of separation efficiency is grain size > production > temperature > pressure, and the separation efficiency is the highest when the water content is 35 m³/d.

(2) In view of the fact that the sand particle size has the greatest effect on the separation efficiency, in practical application scenarios, if the known sand particle size is large, it should be selected to take full advantage of the easy separation of large sand particle-size characteristics in that type of desander. If the sand particle size range is wide, the desander needs to have a better adaptability to ensure that the different particle sizes of sand have a high separation efficiency. By optimizing the operating parameters of the desander, it is possible to ensure the best balance between the removal efficiency of sand particles in the gas well fluid and energy consumption.

(3) Based on the problems related to desander management at the site of this gas well, the following suggestions are made in terms of optimizing the design of the desander in the future. In terms of materials, more wear-resistant and corrosion-resistant materials are required to improve the impact abrasion resistance of the key components of the desander and extend the service life of the equipment to the desander. In terms of automation, the development of an intelligent desanding system with automatic desanding, real-time monitoring, and other functions, to improve the desanding efficiency and equipment operation stability. In terms of desanding precision, in-depth research is required on the desanding mechanism to improve the internal flow field structure, in order to enhance the separation ability of tiny sand particles.