Three‐dimensional visualization experimental study on proppant transportation and placement in the Hi‐way channel hydraulic fracturing technique

The pattern of placement of the proppant in the fracture determines the stimulation and success of the surgery. Using a three‐dimensional visual parallel plate fracture model capable of simulating fracture height, length, and variable fracture width, we experimentally investigate the impact of fiber loading, fracture fluid viscosity, sand concentration, pulse time, perforation pattern, and injection rate on proppant transport, placement pattern, and access rate. Experimentally and visually, the migration process and the morphology of the proppant placement in the fracture were assessed. The formation laws of stable support columns and well‐shaped high conductivity fracture channels in fractures have been elucidated. Experimental results show that increasing the fiber concentration improves the channel rate and the proppant mass is large and stable. The channel rate increases with the viscosity of the fracturing fluid, and it is difficult to form a perfect channel when the viscosity is small. If the constructive displacement is too slight or too large, it will not favor the formation of a Hi‐way channel in the fracture. The optimal design displacement is 4.5 m3/min. When the proppant concentration is low, the passage rate is large, but the stability of proppant mass is poor under the scour of large displacement, the conductivity is low under the high closure pressure, the proppant concentration is too high, the suspension is difficult, the settlement rate is large, and it is difficult to form a large passage. The current column and the passage rate are larger for pulse times of 15–20 s. Cluster perforations are better than large continuous perforations. The higher the number of clusters in the same number of holes, the higher the channel rate. The formation laws of stable support pillars and well‐shaped high conductivity fracture channels in fractures have been elucidated, which can be useful for efficient stimulation of oil and gas reservoirs.


| INTRODUCTION
Hydraulic fracturing technology is the main exploitation of low permeability reservoir stimulation, [1][2][3][4][5] to realize the traditional hydraulic fracturing technology is full of proppant in the fracture, put forward in 2010 by Schlumberger company technology by changing the artificial fracture fracturing proppant in the shop form, put conventional evenly scattered across into heterogeneous, Artificial fractures are supported by many "pillars" like bridge piers, [6][7][8][9] and smooth "channels" are formed between the pillars, and many "channels" are connected to form a network, thus realizing the form of large fractures containing many small fractures, which greatly improves the oil and gas seepage capacity. Since 2010, this technology has been successfully applied to more than 4000 fracturing operations in Argentina, the United States, Russia, Mexico, India, Egypt, and the low-porosity and ultra-low permeability oil field and Sulige gas field in China, with remarkable stimulation effects. [10][11][12][13][14] Transportation and placement of proppant particles is a key factor in the efficient development of unconventional reservoirs through fracking technology. Proppants formed by hydraulic fracturing are laid down to form oil and gas percolation channels, which ultimately improve oil and gas production. Various factors affect the current transport process and the shape of the placement, especially the shape of the crack. Many experimental studies have focused on the shape of the fracture, such as assuming that the fracture is straight or tortuous [15][16][17] Zhang et al. proposed that proppant transportation and distribution are critical elements of hydraulic fracturing technology for the effective development of unconventional and low-permeable reservoirs. 18 The hydraulic fractures are closed on particle beds and generate conductivity pathways, influencing the eventual hydrocarbon productivity. Various factors influence the transport process, especially the fracture shape. 19 Experimental results and mine-back evidence show that the actual fracture is not straight, but tortuous. [20][21][22] Several experimental studies have focused on understanding particle transport within a straight fracture. Kern et al. initially developed a narrow slot formed by two parallel Plexiglas plates, 23 and then Babcock, et al. subsequently improved it to represent a vertical planar fracture. 24 These experiments show that once an initial bed is formed by particle settlement near the entrance, fluidization and sedimentation govern the particle transport to elongate the dune with an equilibrium height toward the flow direction. 25 Wang et al. derived two explicit formulae for predicting the equilibrium heights of a particle bed with the experimental slot at the STIM-LAB facilities. 26 These formulas have been widely used to calibrate numerical simulation models and compare them with experimental results for a straight slot. [27][28][29] Experiments have also been conducted in planar core samples and Plexiglas plates with rough surfaces, 30-34 demonstrating a tree-like settling pattern during proppant transport. Liu 35 and Fjaestad et al. 36 also conducted experiments using straight slots to investigate whether resuspension can enhance particle transport deeper into the fracture. Qu et al. 37 conducted an experimental investigation in a narrow, tortuous slot with two 90°bends while flowing four particle types, and confirmed that particle transport mechanisms in a tortuous slot are more complex than in a straight slot. which provided fundamental insights into understanding particle-fluid mixture flow in a tortuous fracture.
Researchers have conducted extensive studies on the adaptability and field applications of high-speed channel fracturing techniques. Results and understanding of fracture placement parameters and high-speed channel formation results were obtained, and the proposed placement distribution rules were clarified for different fiber additions, fluid viscosity, and fracture sand embankment construction parameters. However, the effects of engineering parameters such as pulse time and perforation patterns on proppant placement have not been systematically and thoroughly studied for fracturing fluid viscosity, fiber addition, sand concentration, and displacement. At the same time, limited by instrument and device conditions, there have been few simulation studies of visualization of proppant placement shapes taking into account bearing capacity.
To address the above issues, we have systematically simulated the transport and placement of the proppant during fracture using a three-dimensional (3D) visual parallel plate model, which can simulate fracture height, fracture length and variable fracture width. The impact of six engineering parameters on the proppant placement, such as the viscosity of the fracturing fluid, the pump displacement, and the pulse time, were evaluated experimentally and visually. The formation laws of stable support columns and well-shaped high conductivity fracture channels in fracture have been clarified, providing an experimental basis for the optimization design of high conductivity hydraulic fracturing schemes.

| THE TECHNOLOGY PRINCIPLE
The channel fracturing technique produces artificial fractures supported by clusters of proppant stacks. The proppant does not act as a conductive medium but as a supporting column to prevent fracture closure or rupture of the surrounding channel walls. In a conventional propane pack, the desired properties are propagated throughout the artificial fracture, all propellant particles are in contact with each other, and the fluid flow is confined to the pores between the propellant particles. The discontinuous proppant pack consists of proppant polymer blocks or proppant segments that form a discrete network of Hi-way channels. Oil and gas flow inside the fracture through open channels between the proppant columns. Using a surface device, the fiber containing the proppant is pulsed to achieve a nonuniform placement of the proppant to achieve an open channel. Fibers were added to prevent the property from settling during placement and closing. Compared to a uniform proppant pack, the conductivity of the channel is significantly increased. Figure 1 shows a visualization of the strut distribution with Hi-way flow channels between the proppant stacks.

| Multicluster perforation process
The multicluster perforation process of flow-limited fracturing is used for high-channel fracturing. It performs uniform multicluster perforations over a long section with the same phase and hole density as regular perforations. Its purpose is to form multiple short liquid inlets on the casing to enable the function of a screen pipe. When the fluid in the tubing is injected into the proppant slug at high speed, the split effect will naturally occur on the casing, forming multiple independent fluid flows into the formation, which is conducive to the formation of independent proppant "column" in the fracture, the proppant is more evenly distributed in the fracture height and the geometric shape of the passage is more regular.

| Alternate pulse slug pumping mode
The injection of high-speed channel fracturing fluid is consistent with a conventional fracturing process. The main difference is that the property is injected as a pulsed slug in the sand-carrying fluid phase, alternating between a profile of the property and a profile of the pure liquid as the proppant concentration is gradually increased. In the preliquid stage, the gel liquid or slippery water can be pumped, and in the proppant slug stage, the gel mixed fiber can be injected to ensure the stable support "column." At the end of the construction, a continuous current slug needs to be followed to have a stable and uniform proppant filling layer at the joints. The slug pumping process is conducive to the formation of "channels" in the fractures. The pure liquid pushes the previous section of proppant into the formation to form a section of proppant "pillar" zone. Due to the isolation of the pure liquid in the middle, there is a certain space between the "pillars" of the proppant vacuum zone. After the liquid breaks up and flows back, a network of multiple channels is formed. The slug pumping process facilitates the formation of channels in the fracture, where the pure fluid pushes the previous properties into the formation, creating a proppant strut. The proppant vacuum region leaves space between the struts due to the isolation of the pure liquid in the middle. After the liquid is broken, a network of channels is formed. The key difference between channel fracturing and conventional fracture pumping procedures is the inclusion of a proposed gel pulse separation in the short pulse technique.

| Laboratory equipment
The large 3D visualized proppant placement setup consists of a flow meter, sand mixing tank, submersible pump, tank, inlet ball valve, fracture and outlet valves. A flow meter was installed at the entrance and an exit valve at the exit. The visualization plate of the device is a toughened Plexiglass with a maximum bearing capacity of 10 MPa, which further increases the pumping pressure capacity of the visualization device. It can model the 3D shape of the fracture. See Table 1 for specific simulation parameters (Figures 2 and 3).

| The experimental principle
The Reynolds similarity criterion is used to ensure that the Reynolds numbers in the joints are consistent with those in the field construction during the experiment. The effect of each experimental parameter on the proposed placement morphology was investigated by combining different experimental parameters and optimizing the parameters best suited for this size of fracture with the goal of high passage rates. The formula for calculating the channel rate is: where HR is the channel rate, %. A u is the area of unfilled proppant in the propped fracture, m 2 . A p is to support the total wall area of the fracture, m 2 .

| Manual test method for proppant paving area
Set up a layer of thin paper, divide a uniform grid, each grid is 1 or 0.5 cm 2 in length and width, cover the visualization plate with paper, the proppant placement form is covered by white paper with grid, so that the proppant placement form can be described, and then flatten the paper with proppant placement form, read the grid number, so as to obtain the proppant placement area, read the number of covered squares, the more dense the squares, The more accurate the proppant placement shape area is.

| Fracturing fluid viscosity
0-50 mesh ceramic proppant was used in the experiment, and the proportion of fiber added was 5 kg/m 3 . The experimental scheme was shown in Table 2 and Figure 4. Figures 5-7 show that as the viscosity of the fracturing fluid increases, the settling rate decreases and the proposed clusters become relatively dispersed during settling, forming favorable channels and increasing the final passage rate. The viscosity of the base fluid is small, the property settling rate is fast, it is difficult to form a large stable proppant population, and the passage rate is low. As the fracture fluid viscosity increased to 250 mPa s, some of the proppant clumps passed through the plate in full suspension and deposited in the pipeline. This phenomenon becomes more severe as the viscosity is further increased.

| Fiber added amount
Pulse fracturing does not require the addition of fibers in the prefluid phase, the displacement phase, and the final injection phase. In the add-on phase, fibers are added throughout the process, which requires the fibers to be well-dispersed in both the properties and the fracturing fluid. The addition of fibers is critical for channel fracturing. The addition of fiber stabilizes proppant clumps, reduces the risk of proppant clump dispersal, significantly reduces the proppant settling rate, and improves current mobility. Experiments were performed using 30-50 mesh ceramic proppants with a displacement of 2.0 m 3 /h. The specific experimental schemes are listed in Table 3 and Figure 8.
The experimental results are shown in Figures 9-11. As can be seen in Figures 9-11, fiber addition has a significant effect on the property placement morphology. At low fiber additions, the property settling rate is fast, the propellant mass is small, and the channeling rate is low. As the number of fibers increases, the ability of the fibers to stabilize the proppant mass increases, leading to better channel formation, considerably increased property column support capacity, and larger channel rates. A clear increase in the channel rate with increasing fiber concentration was found by comparing the channel rate with increasing fiber concentrations of 1.0, 3.0, and 5.0 kg/m 3 . It can be seen that fiber addition has a large impact on the channel rate, but the larger the fiber addition, the better. During the experiment, when the fiber addition exceeds 6 kg/m 3 , the holes become blocked and injection is difficult.

| Proppant addition
Proppant added amount is too little channel rate should be higher, but the stability and flow conductivity of the propping agent group are poorer, the settlement of proppant concentration is overly high, too quick and it is difficult to form the channel, under the same viscosity,   proppant concentration increased, the settling velocity is accelerated, visible as proppant concentration increases, the need to continuously improve the fracturing fluid suspension performance is more conducive to increase the rate of the channel, Pulsed fracturing typically uses fiber additions to achieve this goal. In the experiments, 30-50 mesh ceramic proppants with a fiber fraction of 5.0 kg/m 3 were used. The experimental scheme is given in Table 4. Figures 12-15 show the flow rate values for the proposed formation channel for different fiber additions. The experimental results show that the overall channel rate gradually decreases with increasing proppant concentration under certain conditions with additional experimental parameters. At the proposed dose of 350.0-600.0 kg/m 3 , the channel rate remained stable between 30% and 60%. When the proposed dose exceeds 600.0 kg/m 3 , the channel rate is less than 20%. The pattern of proppant placement in the fracture varies significantly for different concentration properties. At low concentrations, proppant levitation is strong, proppant clusters can form, and permeability is extreme. Due to thin supporting columns, stability is poor at extreme displacement washout and conductivity maintenance is poor at extreme closure pressure. The concentration of the proppant was too elevated, the suspension capacity of the fiber-containing fracturing fluid was insufficient, and the settling rate was too rapid to form an effective channel. Therefore, the proppant concentration should be moderate. The static sand properties of fibercontaining fracturing fluids should be measured before the measurement to optimize the fiber concentration and propose to increase the success rate of pulsed sand fracturing.

| Construction of displacement
Simulations of the effect of constructive displacement on the mobility and distribution of properties have shown that in conventional fracturing the current is dispersed and enters the fracture as a single particle, whereas in channel fracturing it enters the fracture as a mass. In the experiments, 30-50 mesh ceramic proppants with a fiber fraction of 5.0 kg/m 3 were used. The Sand concentration was 400.0 kg/m 3 and the proppant injection time was 225 s. The specific experimental schemes are listed in Table 5. According to the Reynolds similarity principle, different field shifts are converted into laboratory shifts for the experiment. The field construction displacements were 3.0, 4.0, and 5.0 m 3 /min, corresponding to experimental displacements of 33, 50, and 65 L/min. The experimental results (Figures 16-19) show that the channel rate increases first and then decreases with displacement. The experimental shift is 3.0 m 3 /h, which is the maximum channel rate reached with a field shift of 4.5 m 3 /min. It is currently difficult to fill the entire fracture height at both low and high displacements. It   can be seen that the shift has a significant effect on the property placement morphology. When the displacement is too large, the erosion force of the propellant mass migration becomes large and the propellant mass is scattered. When the flow rate is too low, the proppant clumps are heavily dispersed in the perforations, and the proppant clumps settle quickly and are subsequently washed and squeezed by the fluid.

| Pulse time
The injection times were 15, 20, and 25 s, respectively. Experiments were performed with 30-50 mesh ceramic proppants at a pumping rate of 2.0 m 3 /min, a fiber addition ratio of 5.0 kg/m 3 , and a sand concentration of 400.0 kg/m 3 .
The fracturing fluid viscosity is 225 mPa s, and the specific experimental scheme is shown in Table 6 and Figure 20.
The experimental results in Figures 20-23 show that due to the short learning time, in the range of 25 s, it is difficult to study quantitatively in laboratory experiments. To see the effect of pulse injection time on the placement of the proppant in the pulsed fracturing technique, the final proposed placement morphology was observed after repeated injection over several stages. Within a certain range of times, as the pulse injection time decreases, the independent column is larger, the channel shape is better, and the channel rate is larger.

| Perforation layout method
To investigate the impact of multiple cluster perforations on proppant placement and channel rates in pulsed   Table 7.
As can be seen in Figures 24-27, large sections of injected liquid enter linearly under successive punctures, which are difficult to separate and lead to minute channels. Perforating fluid under nonuniform into clusters, naturally appear shunt effect, forming numerous independent fluids into the formation, prompting proppant within the seam form independent of the support pillars, and on the fracture height distribution more uniform, channel geometry more rules, the channel rate is greater, the greater number of clusters and the number of the hole the same situation, better form channel effect, therefore, cluster perforation is commonly used in pulse fracturing.

| CONCLUSIONS
The transport and placement of the proppant during fracturing was modeled using a 3D visual parallel plate model, and the impact of six engineering parameters on proppant placement, such as viscosity of the fracturing fluid, pump displacement, and pulse time, were evaluated experimentally and visually. The formation laws of stable support columns and well-shaped high conductivity fracture channels in fractures have been elucidated.
The passage rate increases with the viscosity of the fracturing fluid. When the viscosity of the fracturing fluid is small, the proppant is prone to settle during construction, resulting in poor channels. When the viscosity of Combining the analytical results for the channel rates, we conclude that the experimental shift is 50 L/min and the corresponding field construction shift is 4.5 m 3 /min, which is the optimal shift.
The propranolol dose has a significant effect on the shape of the current placement and the channel rate value. At low concentrations, the channel rate is large, but the supporting columns are thin, the stability is poor under extreme displacement washing, and the conductivity is poor under extreme closure pressure. The proppant concentration is extremely high, the proposed suspension is difficult, the settling rate is rapid, it is difficult to form a large channel and the channel rate is small, 350-500 kg/m 3 is better in the conditions studied here.
Since the pulse injection time is reduced, the independent column is larger, the channel shape is better, and the channel rate is larger. Compared to the channel rate, the pulse time is better in the 15-20 s range.
The perforation pattern has a significant effect on the morphology of the current placement and the penetration rate. Clustered perforations are better than largescale continuous perforations. The higher the number of clusters, the higher the passage rate for the same number of holes.