Research on a New Method for Intermittent Production of Horizontal Wells in Low-Permeability Gas Field

Abstract

With the development of drilling technology in recent years, an increasing number of horizontal wells have been widely used in low-permeability gas fields. Although horizontal wells are much more productive than vertical wells, fluid accumulation can occur when the formation energy drops and the gas flow rate in the wellbore is not sufficient to remove the loaded fluid from the wellbore. Intermittent production is a good method for preventing liquid loading, but thus far, there is no difference between the open and closed working system of horizontal wells and that of vertical wells, and there is a certain misunderstanding here. In this work, experiments were conducted on the opening process and closing process of a horizontal well, and it was found that the loaded fluid in the horizontal wellbore is a large source of water. It enters the vertical section during the opening process, thereby raising the liquid level and storing the fluid loaded in the vertical section during the closing process, which is difficult to unload. Combined with the experimental results, the production dynamics of a horizontal well with intermittent production was analyzed, and the well opening process and well closing process were divided into four stages. On this basis, a new multifrequency well opening method for intermittent production of horizontal wells was proposed to unload the liquid in horizontal wellbore. The field application case shows that this method can effectively eliminate the drawbacks caused by using conventional methods and increase the average gas production in one cycle by 46%.

1. Introduction

Liquid loading is a serious problem in gas wells in low-permeability gas fields. During the early stage of production, gas wells produce at a relatively high rate. As the reservoir pressure decreases, gas well production decreases, resulting in a gas flow rate in the wellbore below the critical velocity, which prevents liquid from being completely discharged from the wellbore. The lower the gas reservoir permeability, the more severe the phenomenon of liquid loading [1,2].

The fluid that accumulates in the wellbore can come from free water in low-permeability gas reservoirs, or it can come from condensate dissolved in the natural gas. Many papers have been published discussing the problem of liquid loading in gas wells. Turner et al. proposed a model for predicting the critical gas velocity for vertical wells through experimental results [3].

$$u_c=6.6{\left[\frac{\sigma ({\rho }_1-{\rho }_g)}{{\rho }_g{}^2}\right]}^{0.25}$$

(1)

where u_c is the critical gas velocity, m/s; σ is the gas–liquid surface tension, N/m; ρ_l is the liquid density, kg/m³; and ρ_g is the gas density, kg/m³.

Barnea et al. proposed a new model through experiments, and the model predicted the transition of two-phase flow patterns with different inclination angles [4]. Belfroid et al. used a large amount of experimental data to modify the existing prediction model, taking into account the influence of inclination, and modified the Turner model to improve the prediction accuracy [5]. Luo et al. showed the influence of the nonuniform film thickness around the circumferential position of the pipe on the liquid loading through experimental data, modified the Barnea model, and verified the accuracy of the model through a large number of experiments [6]. Alsaadi et al. carried out experiments using a 3 in ID pipe with a deviation of 60°–80° from the vertical direction. The results showed a significant effect of the deviation angle on the onset of liquid loading [7]. Shekhar et al. considered the influence of the pipe diameter and inclination angle based on Belfroid’s model, revised the model, and compared the results with laboratory and field data, demonstrating that the model is more accurate at predicting liquid loading [8]. Cleide Vieira et al. used air and light oil to test fluid velocities and fluid reversal points through inclined pipes. A 60 mm diameter test pipe was positioned at angles of 30°, 45° and 60° from horizontal. The experiments showed that the liquid film model was in good agreement with the experimental results [9].

However, there are several methods that can be used to unload gas wells, such as a gas lift, velocity string, foam and intermittent production [10,11,12,13]. The intermittent production method involves a cyclic process of shutting in the gas well to allow it to restore pressure and then switching the well to production. This is one of the simplest and least expensive ways to unload gas wells, and it can also be used together with foam or a gas lift. An intermitter may be required to control the opening and shutting times in each production cycle. A number of studies have conducted extensive research on intermittent production. Arachman et al. compared three different techniques—gas lift, intermittent production and velocity string—and applied them to low-pressure gas condensate wells with large bore completions. Arachman concluded that choosing the best method among these techniques requires consideration of economic factors and reliability and a trade-off between operational and capital costs [14]. Chen et al. established an optimal production system for intermittent gas wells and determined the well opening mode among different wells. A series of remote control equipment that can control the closing and opening of gas wells using pressure was developed to replace manual equipment in the Sulige gas field [15]. Siripatrachai et al. built intermittent shut-in models for each well, realized the automation of intermittent shut-in, optimized intermittent shut-in conditions and improved the production efficiency for each well [16]. Westende et al. conducted a series of reservoir modeling experiments and vertical production tubing experiments to determine the impact of the reservoir on liquid loading. They utilized edge computing through various parameters of the reservoir to improve the intermittent production efficiency [13]. Gómez combined a deep learning forward model of historical training data with global optimization particle swarm optimization to develop a framework suitable for single and multiple wells [17].

These studies on intermittent production regimes did not distinguish between vertical wells and horizontal wells. It is not easily noticed that there is an important factor that can affect the effectiveness of intermittent production called the wellbore storage effect during gas well start-up, which is much more pronounced in horizontal wells than in vertical wells. Many papers have studied and modeled the wellbore storage effects on production dynamics. Liu et al. established a mathematical model describing the flow of slightly compressible fluid through a double-porosity and double-permeability system including wellbore storage. The exact solutions for various combinations of reservoir parameters were determined [18]. Hatzignatiou et al. investigated the effect of wellbore storage on the pressure data in a solution gas drive reservoir. It could be seen that increasing the effect of wellbore storage decreased the sand face flow rate for a given value of the oil flow rate [19]. Tong et al. investigated a relaxation model of non-Newtonian viscoelastic flow in a fractural reservoir with wellbore storage. Their results indicate that wellbore storage distorts the transient pressure dynamics of wells during flow [20].

In this study, we carried out experiments on the process of closing and opening a horizontal well and explored the main differences between horizontal wells and vertical wells in terms of liquid loading. We carried out field pressure testing of a horizontal well and analyzed the effect of the wellbore reservoir on the liquid loading dynamics of the horizontal well by combining transient production data and transient pressure data. Based on experimental and test studies, we propose a new opening method for the intermittent production of horizontal wells, named the multifrequency well opening method, which can eliminate the wellbore storage effect. Dynamic production results show that the new opening method can largely match reservoir seepage and wellbore tubular flow with each other and enhance the recovery of intermittent production horizontal wells.

2. Horizontal Well Opening and Closing Process Experiment

2.1. Experimental Apparatus

The purpose of this experiment was mainly to observe the dynamic changes in liquid loading in horizontal wells during the closing and opening process without measuring the pressure along the well section. The experimental apparatus consisted of air and water supply systems and a tubular test section made of transparent PVC pipe with a diameter of 60 mm. To reflect the undulating trajectory of the horizontal wellbore in the reservoir, four 1 m long pipe sections were combined at different inclination angles and, finally, connected with one 1 m long vertical pipe section. A schematic diagram of the experimental facility is shown in Figure 1. Horizontal wells in tight gas reservoirs, generally, adopt hydraulic fracturing to increase production, which will generate multiple inflow points along the horizontal wellbore. This experiment focuses on observing the inflow phenomenon from the horizontal section to the vertical section; therefore, the effect of this aspect was ignored, and only one fluid entry point was set at the end of the pipe.

Water is pumped from the water pump to the flow loop through a plunger pump, which can deliver a water flow rate up to 10 m³/h. Air is delivered to the gas storage tank by the compressor, pressurized to 2 MPa, and then delivered to the flow loop through the storage tank. The compressor can provide a maximum flow rate of 5.6 m³/min. We designed a gas–liquid mixing nozzle that allows water and air to be fully mixed as they enter the flow loop. A water flow meter and gas flow meter are used to accurately measure the flow rates of the gas and liquid phases, respectively. The experimental apparatus was also equipped with a visualization system using two cameras. All of our experiments were conducted in a low-pressure system, and the outlet pressure was set at atmospheric conditions.

2.2. Experimental Method

The experiments on the opening process of the horizontal wells focused on the following two aspects: During the horizontal well opening process, what effect will it have on the liquid loading in the vertical section after the gas in the reservoir starts to flow into the horizontal wellbore? During the horizontal well closing process, how does fluid loading in the horizontal section affect shutting down the well?

In order to observe these phenomena, the following experimental steps were developed in this paper:

(1) Horizontal well opening process experiment

The horizontal wellbore experimental flow loop was connected and injected with a sufficient amount of water to bring the fluid level to 50% of the vertical well section. Then, the plunger pump was kept closed, and the liquid flow rate was zero. Turning on the air compressor and plunger pump at a slower rate, an water injection volume of 20 m³/d and air injection volume of 4000 m³/d was gradually reached to simulate the effect of gradual gas production from low-permeability gas reservoirs on wellbore flow during the horizontal well opening process.

(2) Horizontal well closing process experiment

The horizontal wellbore experimental flow loop was connected and injected with water at an injection volume of 20 m³/d and air at an injection volume of 4000 m³/d to bring the water in the vertical wellbore to a stable state in which the gas can carry the water stably. Subsequently, the gas compressor and plunger pump were shut down, and the characteristics of the fluid distribution in the horizontal wellbore after normal production and shutdown were observed, respectively.

2.3. Results and Discussion

(1) Horizontal well opening process result

The experimental result after water injection in the horizontal well is shown in Figure 2a. After opening the air compressor and gradually adjusting the gas volume from small to large, it was found that the liquid column in the vertical section increased to 70% of the vertical well section under the impetus of the gas, as shown in Figure 2b. Then, the gas started to cross the liquid to form a bubble flow, slug flow and annular flow, in turn, until it started to unload the liquid normally. In other words, in the initial stage of the horizontal well opening process in a low-permeability gas field, the gas entering the wellbore will push the liquid in the horizontal wellbore toward the vertical wellbore and push up the liquid level in the vertical section. This is not a good phenomenon, as it generates a certain back pressure at the bottom of the well, resulting in an instantaneous increase in the bottom hole pressure and then affecting gas well production. This situation actually occurs in both vertical and horizontal wells, but because the horizontal section is longer and a large amount of liquid exists in the undulating horizontal wellbore, the opening process is much more affected by rising fluid levels than in vertical wells.

(2) Horizontal well closing process result

The experimental result of the horizontal well with normal production is shown in Figure 3a. Under the condition of a stable state, there is a certain amount of fluid accumulation in the “undulating” wellbore of the horizontal section, but the distribution of fluid accumulation in the wellbore is uneven. There is a certain amount of fluid on both the left and right sides of the “undulating” wellbore, and the amount of fluid on the right side (30% of the horizontal wellbore) is significantly greater than that on the left side (10% of the horizontal wellbore). After entering the vertical section of the horizontal well, the fluid is carried out of the wellbore by the air in the form of an annular flow type. It was also found that an increase in the gas volume, while keeping the fluid volume constant, led to a decrease in the liquid loading in the “undulating” wellbore and vice versa.

After the horizontal well was opened normally for a period of time, the liquid plunger pump and gas compressor were turned off, and the results are shown in Figure 3b. As can be observed, during the closing process, as the gas volume and liquid volume decrease sharply, the various flow patterns in the different well sections disappear, and the liquid loading in the vertical well section drops rapidly and disappears. The liquid loading in the wellbore mainly accumulates in the right end of the “undulating” wellbore of the horizontal well (80% of the horizontal wellbore), and the left end of the “undulating” wellbore accumulates less liquid (30% of the horizontal wellbore).

(3) Discussion

From the horizontal well opening process and closing process experiments, it can be seen that because of the existence of a long and undulating horizontal section, the presence of a large amount of fluid accumulation in the horizontal section significantly affects the horizontal well discharge production, which is significantly different from that of a vertical well. During the well opening process, the gas in the horizontal wellbore pushes the fluid into the vertical wellbore first and pushes up the liquid level in the vertical wellbore, thus creating some back pressure on the bottom of the well. During the well closing process, the fluid in the vertical wellbore is again stored in the horizontal wellbore with certain distribution characteristics. The large amount of liquid loaded in the horizontal wellbore is like a large “water source” that affects the production of horizontal wells. What kind of negative impact this has on the production dynamics of horizontal wells and how to eliminate this negative impact needs to be studied in the context of the production dynamics of horizontal wells in tight gas reservoirs.

3. Dynamic Analysis of Intermittent Production

The reservoir conditions in low-permeability gas reservoirs are poor, and there is a significant wellbore storage effect during the opening of horizontal gas wells for intermittent production. The presence of the wellbore storage effect leads to an obvious incongruity between wellbore fluid unloading and formation gas production. We intercepted an instantaneous casing pressure and instantaneous gas production curve of a horizontal well, named A1, in a low-permeability gas field during an intermittent production cycle, as shown in Figure 4.

It can be seen in Figure 4 that the working system of horizontal well A1 is 8 h open and 5 h closed. The production dynamics of A1 from shut-in to start-up can be divided into four stages.

(1) During the shut-in stage, the entire wellbore is in a closed state, and the liquid level in the wellbore is in a relatively static stage. Because of the pressure balance in the oil well, there is a certain liquid level difference between the tubing and the casing annulus.

(2) During the early opening stage, the gas in the wellbore is rapidly discharged at the moment of well opening because of the wellbore storage effect, and the liquid loading in the tubing is unloaded one time. The gas production reaches a maximum instantaneously and then decreases rapidly from the instantaneous gas production curve. Because of the low permeability of the formation, the formation has not yet reacted, and it is still in a “static state”. As shown in Figure 4, there is a brief period of no production at the wellhead after gas production drops to zero.

(3) During the late opening stage, the formation begins to gradually produce gas, which pushes the liquid of the horizontal section to enter the vertical tubing and then pushes up the liquid level in the vertical section (this phenomenon is consistent with the experimental observations). Then, the casing pressure increases gradually.

(4) With increasing time, the gas production rises. The gas pushes the liquid in the horizontal section into the vertical tubing and gradually begins to carry the liquid normally so that the bottom flowing pressure gradually decreases, and the casing pressure reduces and gradually tends to be stable.

We summarize the four stages of intermittent production of horizontal wells in a low-permeability gas field in Figure 5.

The biggest difference between horizontal well liquid loading and vertical well liquid loading is that a horizontal well has a horizontal section stretching thousands of meters. In the intermittent production process, a horizontal wellbore stores a large amount of water, which can be called the “water source” of the horizontal wellbore.

During the early opening stage of the horizontal well, the liquid loaded in the vertical wellbore can be quickly unloaded at the moment of the well opening, but this part of the liquid unloading is only limited to the depth of the tubing. It is difficult to unload the liquid loading in the horizontal wellbore. Because of the large amount of the “water source” in the horizontal wellbore, it is easy to refill the vertical wellbore with liquid. During the late opening stage, gas begins to be produced, and the liquid accumulation in the horizontal wellbore is pushed into the tubing by the produced gas, resulting in backpressure, which can reduce gas production.

4. New Method for Intermittent Production of Horizontal Wells

Based on the experimental results and the analysis of production dynamics, we found that the liquid in the horizontal wellbore is the key factor affecting the intermittent gas production of horizontal wells in a low-permeability gas field, and it is difficult to eliminate this liquid, because the oil tubing generally can only be put down to the entry window point of the horizontal well at the deepest point. The inspiration for this paper is that since the wellbore storage effect can produce a large instantaneous production to empty the fluid in the tubing at the moment of opening, we can use this wellbore storage effect to empty as much liquid as possible from the “water source” of the horizontal wellbore by opening the well multiple times.

In this paper, a new multifrequency well opening system is proposed to replace “one-time” opening with “multifrequency” opening. The opening process is shown in Figure 6, which divides the process of liquid unloading into five main stages.

(1) Well shut-in stage.

(2) Instantaneous opening of the well at one time, using the wellbore reservoir effect to unload liquid in the tubing.

(3) Shutting down the well to restore pressure.

(4) At the moment of the second well start-up, the wellbore storage effect is used to unload the liquid that has been pushed into the vertical wellbore by the “water source” of the horizontal wellbore again.

(5) Normal production, where gas production gradually stabilizes.

To verify the effectiveness of the new method, a field test was conducted for a horizontal well (A1) in a low-permeability gas field. According to the basic principle of the new method, we propose the following intermittent production cycle: open the well for 0.5 h and then close it for 1 h, repeat twice until the full opening at the third hour, and then keep the original working system after the full opening. The instantaneous casing pressure and instantaneous gas production curve of A1 are shown in Figure 7.

It can be seen in Figure 7 that after three cycles of well opening formed three times in the liquid unloading process, the “water source” in the horizontal wellbore was fully evacuated, and the formation capacity had been fully released after entering normal well opening production. On the one hand, the instantaneous production curve shows that the transient no-production phenomenon formed by the “one-time” opening disappeared from the production curve, and the instantaneous casing pressure curve shows that the rising trend of the casing pressure after the normal production slowed down significantly, i.e., the liquid loading phenomenon weakened significantly.

In order to compare the new multifrequency well opening method with the conventional one, an average gas production in one cycle calculation formula for intermittent production was proposed.

$$q_{g-average}=\frac{{\displaystyle \sum _{i=1}^n{q}_{gi}}}{t}$$

(2)

where q_g-average is the average gas production in one cycle, 10⁴ m³; q_gi is the gas production of the ith opening stage in one cycle, 10⁴ m³; n is the number of opening stages in one cycle, dimensionless; and t is the time for one cycle.

Table 1. Production effect of different well opening methods.

	Type	Gas Production (104 m3/d)	Percentage of Production Increase (%)
Average gas production in one cycle	“One-time” opening method	1.026	-
	“Multifrequency” opening method	1.498	46

displays the production effect of different well opening methods. After “mutifrequency” openings, the average gas production of the A1 well was much better than that of the “one-time” opening, with a 46% increase in production. By adding two more opening cycles and unloading the fluid in the horizontal section several times, the impact of the “wellbore reservoir effect” and the “water source” of the horizontal wellbore could be greatly mitigated.

5. Conclusions

Horizontal wells are a common well type for low-permeability gas reservoirs, and they can significantly increase gas production. As the formation energy decreases, intermittent production is an effective method for liquid unloading. However, the current “one-time” opening method is not conducive to the full utilization of energy and has obvious disadvantages.

The horizontal well opening process and closing process experiments show that the large amount of loaded liquid in the horizontal well has a great influence on the production of horizontal well production. In the opening stage, the gas pushes the loaded liquid in the horizontal wellbore into the vertical section and pushes up the liquid level in the vertical section, and in the closing stage, the loaded liquid in the vertical section will enter the horizontal section and is difficult discharge.

The dynamic data on intermittent production of horizontal wells show that the intermittent production of horizontal wells goes through four different stages. When the wellbore storage effect is released at the moment of well opening, the formation starts to produce gas slowly, driving the loaded fluid in the horizontal wellbore into the vertical section, generating a large back pressure at the bottom of the well and a significant rise in the set pressure, before entering the normal production stage.

Through the combined analysis of the experimental results and production dynamics, this study proposed a new method of multifrequency well opening combining experimental research and theoretical analysis, which can take advantage of the wellbore storage effect and continuously reduce the impact of the “water source” of the horizontal wellbore by opening several times. The field application case shows that the method has a good production effect and can play a significant role in increasing the production of intermittent production horizontal wells. Compared to the conventional method, the new method can increase the average gas production in one cycle by 46%.