Numerical simulation of three-dimensional airflow in a novel dual-feed rotor spinning box

Abstract

In order to regulate turbulence strength and determine airflow characteristics in a new dual-feed rotor spinning unit, the internal flow field is investigated. A computational fluid dynamics technique is employed to numerically study the three-dimensional model of the internal airflow in the new design. The effects of air velocity variation on turbulence strength, negative pressure, Re, and wall pressure distribution are investigated based on simulation data and previous studies. The results show that the turbulence strength and Re increased with increase in inlet air velocity. Pressure profiles inside the rotor varied significantly with positive pressure observed at the channel exits. Minimal inlet velocity maintains the flow field in the rotor interior below 100 m/s, which gives the ideal turbulence required to minimize yarn quality deterioration. The dual-feed rotor spinning unit showed more orderly streamline patterns with fewer vortices compared to the conventional one. The numerical simulation can provide insights on airflow studies and some guidelines for future prototyping and experiments to further improve the new design.

Keywords

dual-feed rotor spinning box standard k-ɛ turbulence model internal airflow three-dimensional numerical simulation

Airflow studies give a fundamental insight in understanding the concept of internal flows. Numerical analysis is a valuable tool in this regard; studies of high-speed airflow in open-end rotor spinning, air-jet texture, air filtration, and internal material flow studies have capitalized on the use of computational methods.^1–7 Internal and external airflows and air–particle interaction studies have been widely analyzed using computational simulation by the numerical methods approach.^1,7–10 Recent studies on airflow in rotor spinning show a significant correlation between the airflow characteristics, the spinning parameters, and consequently the yarn quality characteristics.^1,4 Open-end rotor spinning has some advantages, ranging from the low cost of production to the amenability to automation as compared to other yarn manufacturing systems. Despite the open-end spinning benefits, there are limited chances for effective yarn blending and guarantee of high-level fiber orientation to improve yarn properties. There have been significant advances in rotor spinning technology since its innovation in 1967.¹¹ Numerous studies have focused on advancement in spinning speed,² rotor and transfer channel design optimization,¹² studies of airflow characteristics,^4,13 and rotor spinning process parameter characterization.¹⁴

The dual-feed rotor spinning unit is a concept in rotor spinning that allows two slivers to be fed simultaneously, mainly to ease blending, fiber opening, and trash removal, with the main goal of improving fiber configuration in yarn. In 2007, Hajilari et al.¹⁵ reported the first dual-feed concept in rotor spinning. In their study, they introduced a technique of using two separate fiber feed systems in a modified SE-8 rotor spinning unit of Suessen rotor spinning by feeding the other viscose sliver on the trash removal zone. They reported improved fiber opening and configuration and better orientation, and the yarn properties, such as tenacity, extension, and work of rupture, improved compared to conventional rotor-spun yarn. Their dual-feed rotor spinning system limitation was its inability to extract trash because the trash removal zone was used to feed the other sliver. In 2014, Peyravi et al.¹⁶ introduced a dual-feed system with a knife edge and trash removal zone. This design had two feed rollers, a trash removal zone, and a single opening roller; the resulting yarns properties, such as tenacity, strain at peak, work of rupture, abrasion, and yarn hairiness, significantly improved compared to the conventional rotor-spun yarn. A new dual-feed rotor spinning unit patented by Wang et al.¹⁷ has two carding devices and transfer channels and allows feeding and processing of two slivers simultaneously up to the rotor wall where the fiber strands are integrated for yarn production.¹⁷

Lawrence and Chen^12,18 studied fiber trajectory and transfer channel optimization to investigate the relationship between rotor spinning parameters and fiber configurations. Their study optimized the conventional design by using high-speed photography to take images of long and short fibers transferred in different configurations. They optimized the airflow velocity, transfer channel design, and opening roller speed, among other parameters, to produce improved yarn properties. Their findings revealed that short fibers are easily drawn back to the mass of flow, while long fibers remained with their leading ends hooked around the pin-clothing. A narrow rectangular cross-section of the transfer channel was more beneficial to fiber straightening compared to a circular one. Other researchers have also studied the effect of negative pressure, geometrical structure at the inlet and outlet, opening roller speed, rotor velocity, rotor diameter, and the rotor groove dimensions on structural and mechanical properties of rotor-spun yarns using the computational fluid dynamics (CFD) technique.^1,4,14,19

CFD is a branch of fluid mechanics that employs numerical analysis and algorithms to calculate and explore problems that comprise of fluid flow phenomena. Commercial CFD software is used to perform calculations required to simulate the interaction of liquids and gases with surfaces defined by boundary conditions.²⁰ In 1996, Kong and Platfoot¹³ developed a two-dimensional (2D) CFD model to simulate the airflow pattern inside the transfer channel of a rotor spinning machine and analyzed the effect of channel design and airflow characteristics on the quality of yarns. Their study revealed that the strength and area of recirculation at the inlet of the transfer channel caused by the geometric variations and rotating opening roller are considerably related to airflow and, hence, the resultant yarn properties. In 2005, Cai²¹ developed a three-dimensional (3D) model based on one-way coupling method, and their results provide more insight on visualization and analysis of fiber flow inside the transfer channel. Lin et al.⁴ and Xiao et al.⁵ presented recent studies on airflows for a single transfer channel. They focused on the effect of rotor dimensions, rotor speed, transfer channel design, and airflow characteristics on the resulting quality of yarn. Their findings show that the airflow domain and airflow characteristics can give a deeper understanding of the relationship between the spinning parameters and yarn quality characteristics.

Despite the advantages of a dual-feed rotor spinning system, such as easy yarn blending, effective trash removal, improved fiber configurations and orientation, and better yarn properties, there has been no new research to investigate the airflow regime inside the rotor with two feed systems. Most of the previous work has focused on the conventional rotor spinning unit. Internal airflow study remains a challenging task that needs deeper understanding. Although it is still a complex phenomenon to characterize internal airflows and their influence on the quality of yarn, commendable achievements by current researchers using CFD have inspired our research.^{1,3–6,11,13} In typical rotor spinning, the opening roller speed, rotor speed, and material processed affect flow characteristics, such as the negative pressure and turbulence strength. The current study presents an air simulation experiment for 3D airflow in a dual-feed rotor spinning unit without material.

This study can give further guidelines for future prototyping and experiments to improve the dual-feed rotor spinning unit. We simulate the airflow field in the rotor, vary the inlet velocity on the two channels, and analyze the turbulent strength, velocity magnitude, and pressure distribution, and compare the results with the previous related studies and experimental data. The paper presents the theoretical model construction, solution strategy, results and discussion, simulation verification, and conclusions.

Theoretical model construction

The general principle of the conventional rotor spinning method is shown in Figure 1(a). The fibers are inserted to the feed roller in the form of slivers, which may contain more than 20,000 fibers in their cross-section. The input sliver is first opened and drafted by the opening roller, and the fibers are then transported via a tube to the rotor where the fiber strand is subjected to twist insertion by the torque generated by the rotating rotor. On the rotor, centrifugal forces compress the fibers onto the collecting surface of the rotor, which is then conveyed to the doffing tube and then wound to the cone package of the desired size.²² To characterize the flow domain, a clear understanding of the flow behavior from the opening roller to the rotor is paramount. In principle, the fibers leaving the opening roller become airborne, and their flow is assisted by an airstream that depends on the surface speed of the opening roller and the atmospheric pressure inside the rotor achieved by creating a partial vacuum zone by using a self-pumping effect. On average, there are 2–10 fibers²³ at the exit of the transfer channel; therefore, fiber movement can be modeled from the airflow field concept. The airflow domain inside the rotor spinning unit can be modeled by analyzing the air velocity and negative pressure, parameterization of turbulence, and Reynolds number.

Figure 1.

(a) Conventional rotor spinning unit. (b) Novel dual channel rotor spinning unit.

Turbulence characterization and Re

Specification of turbulence at the inflow boundary can be determined from turbulence intensity, turbulent viscosity ratio, or turbulence length scale parameters. Laminar flow prevails where viscous forces dominate the inertia forces; in CFD analysis, the relationship between these two forces is referred to as the dimensionless quantity Re (see equation (1)). In the rotor spinning unit, the approximate range of Re is 60,000–12,000 for commercial processing. The turbulence scale is the periodic variation and is more intricate to measure because most velocity profiles include many compounded scales of velocity instabilities and, so far, it is still challenging to decouple signals.²⁴ The viscosity ratio is defined as the ratio of the turbulent viscosity to molecular viscosity, and previous experiments have characterized viscosity ratio in fully turbulent flows on the equivalent Re.²⁵ If the turbulent viscosity ratio data is known, the Re and consequently the turbulence strength can be predicted (see Table 1).

Table 1.

Reynolds number (Re) scaled against turbulence viscosity ratio (β)

Re	3000	5000	10,000	15,000	20,000
β	11.6	16.5	26.7	34.0	50.1

When there is turbulence in a given flow, random fluctuations in velocity are created. If we consider longitudinal velocity (u) and vertical velocity (v) measured at a given point in a flow field, for a steady stream $u = \bar{u}$ and $v = \bar{v}$ for all time t, where $\bar{v}$ and $\bar{u}$ are the velocity at a time average. For turbulent flows, the velocity includes a mean and a turbulent component (see equation (2))

Re = \frac{ρ UL}{μ}

(1)

{u (t) = \bar{u} + u' (t) v (t) = \bar{v} + v' (t) mean turbulentflactuation}

(2)

Turbulent motions have random eddies; therefore, to give an account of turbulence strength, integration of the mean velocity in the continuum will provide the discrete mean velocity at equally spaced points at a given time (t + T), where T is longer than any turbulence time scale but shorter than the time scale for mean flow unsteadiness

{Meanvelocity \bar{u} = \int_{t}^{t + T} u (t) d t = \frac{1}{N} \sum_{1}^{N} u i Continuumterm discreteterm}

(3)

Turbulent fluctuation

{u' (t) = u (t) - \bar{u} : continuumterm u' = u i - \bar{u} : discretepoints}

(4)

Turbulent strength:

A larger $u rms$ indicates higher turbulence.

Dual-feed rotor spinning unit

In a dual-feed rotor spinning unit, there are two independent sliver feeding devices as well as two separate carding device components, as demonstrated in Figure 1(b). The principle and operation remain as in the conventional rotor spinning machine and the airflow principle in the two independent transfer channels is assumed to behave as per the previous studies.^4,13

Figure 2 shows a 3D geometric model of the computational domain of the dual-feed rotor spinning unit. The model was built on a Cartesian coordinate system, and the dimensions are shown in Table 2. With the origin positioned at the center of the rotor, the dual channel rotates along the vertical rotor axis in the direction of the longitudinal rotor axis. The channel exits are adjacently located, and once the fibers exit the channels, they are deposited on the rotor surface. Since the rotor rotates at high speed, the fibers from the two channels are blended to form a fiber strand along the rotor groove.

Figure 2.

Dimensional and geometric model of computational domain. 3D: three-dimensional.

Table 2.

Dimensions of the dual-feed rotor spinning unit

Name	Symbol	Dimension
Rotor outlet	h ₁	1 mm
Rotor height	h ₂	15 mm
Rotor diameter	D	40 mm
Transfer channel length	l	4 3 mm
Slide wall angle	θ	66°
Doffing tube diameter	d ₁	1 mm
Distance between transfer channels	d	72 mm

Governing equations and turbulence model

The process of conveying fibers from the opening roller to the rotor happens in a very short time. When the fibers in the tube become airborne, their motion through the transfer channel to the rotor wall happens in milliseconds; heat transfer can be regarded as negligible. However, this assumption is ideal when dealing with natural fibers, since synthetic fibers such as polyester have higher temperature excursions. A reasonable quantity of heat may be produced during the process. Re is used to characterize different flow regimes within a similar fluid, laminar flow occurs at low Reynolds numbers, and the flow is considered even and continuous due to the dominance of viscous forces. Turbulent flows occur at high Reynolds numbers and are characterized by chaotic eddies, vortices, and flow instabilities. The airflow inside the rotor is turbulent, and the speed in the rotor interior has been suggested to be below 100 m/s in order to avoid yarn breakages and for better fiber alignment and yarn characteristics.^4,14,18

The study adopted a standard k-ɛ model because near-wall effects and the characteristic swirl number were ignored; moreover, the simulation results were to be compared with previous studies simulated with the same model.⁴ The k-ɛ model is a semi-empirical model developed for high Reynolds number flow and turbulent flow analysis. This is the first approach to understand the flow regime in the dual-feed rotor spinning machine; for deeper understanding, comparison of the standard, re-normalization group (RNG), realizable k-ɛ and shear stress transport (SST) k-ɛ turbulence models can be studied at a later stage. The airstream inside the rotor spinning unit was presumed isentropic, viscous, incompressible, and turbulent flow.

The principal equations, which include the mass conservation equation and the momentum conservation equation, can be articulated as

\frac{\partial v}{\partial t} + \nabla . ρ . \vec{v}

(6)

where

ρ

is the fluid density and v is a velocity vector. This equation implies that the sum of the time-dependent change of density

ρ

and the 3D of the current density

ρ v

is zero

\frac{\partial (ρ u i)}{\partial t} + div (ρ u u i) = div (μ grad u i) - \frac{\partial p}{\partial x i} + S i

(7)

The standard k-ɛ two-equation turbulence model is useful in some engineering practices and research^7–9,26 to study fluid flows and dynamics. In our study, we were dealing with high Re, hence there was a justification to adopt this model. With the introduction of the turbulent viscosity variable $μ t$ (see equation (8)), which relates the turbulent kinetic energy k and the turbulent dissipation rate ɛ, the model constants k and ɛ can be obtained from equations (8)–(10)

μ t = ρ C μ \frac{k^{2}}{ɛ}

(8)

\frac{\partial (ρ k)}{\partial t} + \frac{\partial (ρ ku i)}{\partial x j} = \frac{\partial}{\partial x j} [(μ + \frac{μ t}{σ k}) \frac{\partial k}{\partial x j}] + G k - ρ ɛ

(9)

\frac{\partial (ρ ɛ)}{\partial t} + \frac{\partial (ρ ɛ u i)}{\partial x j} = \frac{\partial}{\partial x j} [(μ + \frac{μ t}{σ ɛ}) \frac{\partial k}{\partial x j}] + \frac{C | ɛ^{ɛ}}{k} G k - C 2 ɛ ρ \frac{ɛ^{2}}{k}

(10)

These default values are determined from the experiments for fundamental turbulent shear flows, including homogeneous shear flows and decaying isotropic grid turbulence.^3,10 Table 3 shows the values of the constants.

Table 3.

Values of empirical constants

$C μ$	$C \| ɛ$	$C 2 ɛ$	$σ k$	$σ ɛ$
$0.09$	$1.44$	1.92	$1.0$	1.2

The standard k-ɛ two-equation turbulence model only applies where the flow is fully turbulent, and the effects of molecular viscosity are negligible; hence, making it valid for fully turbulent flows with high Reynolds numbers.^3,20

Grid generation

In our study, the 3D diagram to represent this model was designed in solid works 2012^© and then exported to ICEM CFD of ANSYS 14.5©, where it was meshed (see Figure 3) in preparation for simulation. A program-controlled Triangular surface mesher was employed. Unstructured tetrahedral cells were applied in the computational domain as well as in the application of refinement meshing on grids in the near-wall regions that were denser than the other areas. To ensure better mesh quality, cell size and optimum cell size have to be controlled to suit standard requirements.^3,10

Figure 3.

Mesh topology and name selections of the three-dimensional model.

Boundary conditions

In our set up, unlike Lin et al.’s⁴ research, we had two velocity inlets, one at each independent transfer channel inlets. The velocity was varied, as shown in Table 1. The location of boundary conditions in the model is illustrated in Figure 3: the wall, the two velocity inlets at adjacent transfer channels, the pressure inlet, and the outlet.

Numerical method and solution strategy

To solve the governing equations, we employed the finite volume scheme based on FLUENT. We used a commercial CFD package ANSYS 14.5 to run the simulation. The conservation equations were solved based on the second-order upwind scheme to ensure a high level of accuracy. The second-order upwind scheme performs a simultaneous solution of momentum, conservation of mass, energy, turbulent kinetic energy, and its dissipation rate equation within the defined physical domain. Our solution was initialized by the hybrid method. We also set surface and volume monitors for velocity magnitude, static pressure, and turbulent kinetic energy k. The convergence criterion for both surface and volume monitors was set to 1 × 10^–4.

To focus on our main objective, geometrical and rotor spinning parameters were kept constant as per Lin et al.⁴ and Wang et al.¹⁷ The rotor speed was 100,000 r.p.m, pressure inlet 1.01 × 10⁵Pa, and pressure outlet –7000 Pa.

Results and discussion

Validation of the numerical code and grid independence

Numerical simulation results are greatly affected by the scheme of mesh generation. The mesh independence test is an important stage in computational simulation. In our study, we generated three different grids with various elements and nodes. The grids produced were as follows: Grid 1 (494,226 elements); Grid 2 (331,683elements); and Grid 3 (267,445 elements). In the three cases, grid refinements were done to ensure a well-defined mesh. Figure 4 shows a plot of velocity distribution within the interior of the model; the velocity curve for different grids show the same paths with a slight variation, which was attributed to airflow through the rotor opening. The curve trends reveal that the mesh grid is independent of the solution, which is a significant criterion to evaluate the mesh quality of a CFD model; as per the mesh independence tests, Grid 2 was chosen for further simulation. We executed ten simulation runs by varying velocity at the channel inlets as shown in Table 4.

Figure 4.

Velocity distribution for three different grid schemes: (a) along the x-axis; (b) along the y-axis.

Pressure and velocity vector distribution inside a dual-feed rotor spinning unit

Visualization of pressure distribution in the interior of the model shows a domination of negative pressure inside the rotor between –3000 and –7000 Pa with small traces of –1000 Pa, especially near the transfer channel outlets (see Figure 5). The negative pressure inside the rotor is due to the suction process during rotor spinning. The two transfer channels have positive pressure, since the direction of airflow is from the channel inlet towards the rotor. We observed traces of positive pressure near the transfer channel outlets; this is due to the positive pressure overflow from the channels. This behavior is quite similar to previous studies about the conventional rotor spinning unit.^4,13 The dual-feed rotor spinning unit is different in the sense that there are no traces of positive pressure inside the rotor.

Figure 5.

Interior pressure distribution: (a) three-dimensional view; (b) two-dimensional view.

We extracted velocity vectors, demonstrated in three forms. Figure 6(a) shows the velocity distribution in the whole model. As can be observed, the velocity magnitude increases along the transfer channels, demonstrating similar behavior to previous studies.^4,9 However, inside the rotor, it behaves differently. The velocity from end-to-end channel outlets crashes, and suddenly reduces to 40 m/s in the rotor interior and higher on the rotor walls, as well as in the doffing tube. The reduction in velocity regulates the turbulence strength and hence allows proper conveying of fibers inside the rotor. Figure 6(c) clearly shows the distribution of velocity along the x-axis; as can be seen, the maximum velocity is at 90 m/s. Visualization of 3D streams showed a vortex region in the rotor, as illustrated in Figure 6(c), caused by the effect of the rotor rotation and collision of the adjacent airstream from the two transfer channels. The vortex region can be minimized by regulating the negative pressure and airflow velocity; however, to a greater extent, the rotor dimension characteristics, such as the rotor groove, slide wall angle, and tapered transfer channel, play a significant role as well.

Figure 6.

Internal velocity profiles: (a) velocity vectors inside the model; (b) velocity distribution along the x-axis; (c) three-dimensional air velocity streams.

Effect of variation of velocity at inlet 1 and inlet 2

We varied velocity at the two inlets to determine the effect this might have, especially inside the rotor. The novel design is intended to improve the fiber alignment and ease blending of different yarn types; a variation of feeding rate or the velocity at the inlets was analyzed.¹⁷ To visualize the velocity profiles in the rotor interior, a slice was extracted at rotor vertical position y = 1.2 mm along the rotor x–z plane, where the airstreams from the dual channels mix and the effect of velocity can be observed clearly. This zone represents the formed fiber strand, as the yarn forms on the rotor groove to the doffing tube. We extracted velocity profile slices for 10 different simulation runs, and the velocity profiles were analyzed as shown in Figure 7. The increase in inlet velocities was associated with higher velocity profiles inside the rotor, particularly in the transfer channel exit positions. We also noted that there is a speed difference at the transfer channel outlet position. The velocity is two times higher than the position away from the channel outlets. The variation can be explained by an increase in velocity in the channel length; it is at maximum towards the channel end before the airflow enters the rotor. The airstream velocity is then crashed, hence accounting for the reduced velocities at the positions away from the outlet. This phenomenon coincides with previous research on the conventional rotor spinning unit,^4,13,14,27 with slight variation due to the effect of the second transfer channel.

Figure 7.

Velocity profiles at y = 1.2 mm for different simulation runs.

In brief, from our observation; in a dual channel design, the air from different directions crashed due to the rotor speed and negative pressure zone effect. Away from the channel outlets, the velocity profiles will increase steadily. There is a need to balance turbulence along the rotor groove as this might have an effect on the fiber alignment inside the rotor and, hence, affect the quality of the resulting yarns.

Figure 8 shows plots of velocity along the x-axis for all simulation runs; simulation runs (#1–#5) showed a similar path; as the velocity is increased, at simulation run #8, the velocity profile curve demonstrated a rather different path that can be attributed to uncontrollable turbulence. High airstream velocities will result in fiber breakages and poor configurations and alignment, especially for considerably long fibers as opposed to short fibers.¹⁸

Figure 8.

Velocity variation along the x-axis at different initial air velocities at inlets 1 and 2.

The pressure–velocity relationship along the rotor wall

We analyzed the velocity profiles and pressure distribution along the rotor wall. The velocity profiles, pressure distribution, and plot of velocity versus pressure are demonstrated in Figures 9(a)–(c), respectively. We considered simulation run #2 to clearly observe the velocity profiles along the rotor wall, where the transfer channel exit positions showed a lower velocity compared to other positions along the rotor wall. In Figures 9(a) and (b), the contours of pressure and velocity reveal that the positions of low speed (channel outlets) will have positive pressure up to 1000 Pa and the position away from the channel outlets will have negative pressure. The pressure–velocity relationship curve (Figure 9(c)) shows that the velocity increases with an increase in negative pressure. Therefore to control the velocity of the rotor, the negative pressure should be studied comprehensively with consideration of its impact on the fiber alignment as well as the final quality of the yarn.

Figure 9.

(a) Velocity distribution. (b) Pressure distribution. (c) Velocity–pressure curve.

Relationship between turbulence and air velocity inside the rotor (viscosity ratio)

Turbulence is an important factor to characterize when dealing with internal flows. For turbulent flows, the flow is characterized by very high Reynolds numbers.^13,28 In our study, we analyzed the viscosity ratio inside the rotor at different inlet 1 and 2 initial velocities. We extracted 2D contours for simulation runs #1–#5 and then we also extracted contours for the simulation run with the highest air velocity at inlet 1 and inlet 2.

In this section, we study the turbulence viscosity ratio at different velocities. The viscosity ratio is defined as the ratio of the turbulent viscosity to laminar or molecular viscosity. For a given flow, the comparison between the viscosity ratio and approximate Re can predict the turbulent strength inside the dual-feed rotor. When the inlet velocity is 100 m/s for simulation #10, the turbulent viscosity ratio is above 1100 according to the contours. In CFD, the Reynolds number, which shows the degree of turbulence, may be scaled against the viscosity ratio, as shown in Table 1.²⁰ For internal flows with higher turbulence at Re of 100,000, $β$ is equal to 100. In our study, traces of $β of 100$ are seen at simulation runs #1 and #2, where the air velocity is 20–30 m/s. In simulation runs #3–#5, the turbulent viscosity ratio ranges from 400 to 100, in which simulation run #10 has the highest viscosity ratio of up to 1100, and hence the highest turbulence (see Figure 10). Clearly, high velocities would cause poor alignment of fibers and higher yarn irregularities. Compared with Kong and Platfoot²⁷ in their study on fiber transportation, better Re numbers can be achieved between initial velocities of 18 and 25 m/s. Turbulent viscosity ratio is illustrated in equation (11)

β = \frac{V t}{V}

(11)

Figure 10.

Viscosity ratio at different inlet conditions.

Table 4.

Velocity variation at inlets 1 and 2

Sim. run	Velocity inlet 1 (m/s)	Velocity inlet 2 (m/s)
#1	20	20
#2	20	30
#3	30	40
#4	40	50
#5	50	60
#6	60	70
#7	70	80
#8	80	90
#9	90	100
#10	100	100

Dual-feed and conventional rotor spinning unit comparison

Air turbulence, velocity, and negative pressure distribution in the two rotor spinning units were compared. The airflow characteristics in the transfer channel showed similar behavior in the dual-feed and conventional rotor spinning unit. The airflow field fluctuations in the two methods lie in the rotor interior, because the process from the opening roller through the fiber tube is similar in principle. We extracted the flow field at y = 2 mm, a zone inside the rotor, as shown in Figure 11. The dual-feed and conventional simulation data were obtained at similar boundary conditions, inlet velocity of 20 m/s, pressure inlet of 1.01 × 10⁵, and pressure outlet of –7000 Pa.

Figure 11.

Turbulent viscosity and streamlines: (a) dual-feed; (b) conventional rotor spinning unit.

Figure 11 shows turbulent viscosity streamlines and contours. The streamlines show the paths of airstreams and the vortex region. In the conventional rotor spinning unit, a vortex can be observed, while the dual-feed unit forms ellipse-like streamlines that can be seen inside the rotor. The shape of the streamlines in the dual-feed can be due to the air mixing from the two channel exits, which may result in an instant drop in airflow speed because the flow is from opposite directions. Large vortex regions inside the rotor are not appropriate during fiber processing; they minimize chances for better fiber alignment and blending, in addition to causing fiber tangling and breakages. The absence of large vortices in a dual-feed rotor spinning unit presents a better flow behavior compared to the conventional one, which facilitates blending, fiber alignment, and then yarn quality characteristics. Turbulent viscosity characterizes the resistance to flow. Therefore, high turbulence strength is associated with low turbulent viscosity. Both cases show low turbulence near the wall, which is due to the high-speed rotating rotor and the turbulence strength being reduced considerably in zones towards the center of the rotor.

Figure 12 shows velocity contours in the two spinning units and comparison of velocity profiles at the same position inside the rotor. In principle, the tapered transfer channel flow increases towards the rotor; in the rotor, many factors affect the speed, rotor diameter, slide wall angle, rotor groove, rotational speed, and the negative pressure. The proposed design with two feed points presents a new concept that changes the flow field inside the rotor, due to the collision of the airstream from the two channel exits. The air velocity in the dual and conventional rotor spinning units shows varying contours, with the dual channel showing considerably even distribution compared to the single channel. We noticed that in the dual-feed, the velocity is slightly higher, particularly in the center compared to the conventional rotor spinning unit.

Figure 12.

Velocity profiles: (a) dual-feed; (b) conventional rotor spinning unit.

During the rotor spinning process, airflow in the transfer channel should be fast enough to exceed the surface speed of the opening roller, usually at a ratio of 1:5. To maintain the desired flow rate, negative pressure inside the rotor is a significant factor. The rotor should be kept in a partial vacuum environment using a self-pumping effect or an external pump mechanism to create negative pressure. We compared the pressure in a dual-feed and conventional rotor spinning method; the negative pressure in the two cases is measured at the rotor vertical axis y = 2 mm in the x–z plane, where 100 evenly positioned points are extracted along the centerline at z = 0 and plotted as shown in Figure 13. The results show that the dual-feed unit maintains pressure higher than the conventional one. To retain airflow through the air ducts at a reasonable speed, balancing of negative pressure is important to avoid fiber entanglements and disorientation during transfer. Comparing with previous studies, Lin et al.⁴ reported that a negative pressure range between –3000 and –7000 Pa is ideal for yarn processing, which agrees quite well with our results, besides the conventional unit that exhibits lower pressure up to –9000 Pa.

Figure 13.

Comparison of negative pressure in the rotor interior.

Simulation verification

Simulation verification is a significant procedure to determine the correlation between simulated and experimental data. The airflow field and the negative pressure inside the rotor should be measured. A rotor is an enclosed unit with a yarn guiding mouth as the only access point to its interior. Furthermore, it rotates at high speeds, making it challenging to use a velometer to measure the velocity or measure airflow viscosity at high speeds. The experimental rig presented in Figure 14 had a limitation for the measurement of negative pressure only because air turbulence and velocity variation require the use of a hot-wire anemometer or a 3D acoustic Doppler velocimetry. According to the turbulence and velocity measurement procedures, the system could not integrate into our experimental arrangement. Nevertheless, negative pressure as a crucial parameter in the rotor spinning process was considered in the verification process. Using a digital pressure sensor, measurement of the negative pressure was achievable by inserting the tube through the yarn guiding mouth. For precision, we considered two verification cases; one instance at evenly spaced points around the yarn guiding mouth and the other along the rotor height.

Figure 14.

Simulation verification experiment set up.

Pressure around the yarn guiding mouth inside the rotor spinning unit

We set up an experiment, illustrated in Figure 14, to measure the pressure at evenly spaced points (a–h) around the yarn guiding mouth (see Figure 15(a)). A hollow metal tube was curved and inserted through the doffing tube; two pressure gauges, one connected to the air pressure machine (suction) and the other to the rotor interior through the doffing tube, were used. Eight evenly marked points were made around the yarn guiding mouth, as illustrated on a simulated results extract (Figure 15(a)). Points a and e represent the position of inlets 1 and 2, respectively. The air pressure machine was adjusted to –7000 Pa, similar to the simulation conditions, by using a digital pressure gauge model PG7 with accuracy of 0.25%. The negative pressure at all points was then measured in a clockwise direction, and the simulation data extracted at the same position for comparison.

Figure 15.

(a) Selected points around the doffing tube of the dual-feed rotor spinning unit. (b) Experimental and simulated pressure values for the dual-feed and conventional rotor spinning systems. (c) Selected points around the doffing tube of the conventional rotor spinning unit.

As illustrated Figure 15, the simulation and experimental results agree well quantitatively; much as the experimental data is lower than the simulation values, the difference is due to the loss of pressure in the outlet tube. The results reveal a similar trend to that Lin et al.⁴ found for a single transfer channel rotor spinning unit. The experimentally measured pressure value at points f, g, and h in Figure 15 for the dual-feed rotor spinning unit are much higher than the simulated values. The degree of accuracy in data extraction in the simulation cases is much greater than in the experimental set up. We attribute the deviations in simulated and experimental results to the pressure loss in the outlet tube. Nevertheless, the results obtained in the rotor vertical direction showed a much better correlation between the measured and simulated results. We also observed that, at the position near the transfer channel inlets, the pressure is nearly the same and the pressure measured away from the inlets tends to increase. The increment is due to the effect of positive pressure from the transfer channel as well as the doffing tube. Comparison of the negative pressure in the conventional and the proposed dual-feed rotor spinning unit reveals higher negative pressure in the conventional method. This experimental set up justifies our simulation findings to conform with experimental data quantitatively.

Pressure along the rotor vertical axis

The second verification experiment was designed to measure the pressure along the rotor vertical axis (y-axis) at a 1 mm range from coordinates (0,0,0) to (0,11,0) (see Table 5). In this experimental set up, we used a 15 mm long tube, and this was inserted through the yarn guiding mouth and then connected to the pressure gauge by a flexible plastic tube. The tube was inserted through the yarn guiding mouth, and using a micrometer screw gauge, points were adjusted along the rotor height so as to ensure that the positions did not change during in the process of measuring pressure. The distance along the rotor height, as illustrated (AB) in Figure 16, was varied at a 1 mm range using a micrometer screw gauge. We performed this with extra care to make sure that the tube length did not change during measurement, which could compromise the measured results. We also closed the sliver feeding positions and the trash removal position to reduce pressure fluctuation due to the influence of the surrounding airflow. The pressure was also maintained at –7000 Pa, and the internal pressure values were measured at different positions in the rotor and recorded. For the simulation data, using Tecplot 360EX 2015 R2^© software, a probing tool was used to determine the pressure at desired points on the slice. The experimental and simulation data are compared under the same conditions. The results are presented in Figure 16. Even though the experimentally calculated values are slightly higher than the simulation data, the results quite agree well qualitatively.

Figure 16.

Pressure along rotor height (y-axis): (a) slice at (z = 0); (b) experimental versus simulated data along AB.

Table 5.

Experimental and simulated pressure inside the rotor at z = 0 mm

Position (mm)	Exp. (Pa)	Sim. (Pa)
0,0,0	5676	4217
0,1,0	5319	4229
0,2,0	5192	3441
0,3,0	4550	3584
0,4,0	4460	2826
0,5,0	4250	2704
0,6,0	4322	2669
0,7,0	4674	2878
0,8,0	4150	3852
0,9,0	3836	3062
0,10,0	3570	2870
0,11,0	3564	2364

Conclusion

In our study, we developed a 3D CFD model to simulate the airflow patterns inside the dual-feed rotor spinning unit. The main objective was to visualize the airflow zone, study the effect of air velocity variation to the entire flow regime, and compare results with previous research on the conventional rotor spinning unit. A standard k-ɛ model was adopted to simulate a 3D airflow in the dual channel rotor spinning box.The turbulence viscosity ratio β, velocity distribution, and pressure distribution inside the rotor were observed. A high turbulence viscosity ratio was witnessed, especiallyy on the positions near the transfer channel outlets and an increase in velocity at inlets 1 and 2 led to a high level of turbulence inside the rotor. Obervation of the streamlines and vortex regions in the dual-feed unit showed fewer vortices and the streamlines pattern was more orderly compared to the conventional rotor spinning unit.

Observations at the rotor wall showed increasing velocity at the positions away from the transfer channel outlets. At the same positions, there was positive pressure less that 1000 Pa; this is a position where the fibers are released to the rotor before they are deposited on the rotor groove. Comparisons between simulation results, experimental data, and previous researchs on the single channel rotor spinning unit reveal considerable agreeing results. To regulate turbulence strength inside the rotor, the channel inlet velocity should be minimized and the negative pressure balanced to improve fiber configurations, reduce fiber breakages, and hence improve yarn quality characteristics. Based on the airflow field in the dual-feed rotor spinning unit, it is shown that this method is capable of producing yarns with better configurations and fiber orientation compared to the conventional one. Further experimental comparisons should be made on the two methods to compare the yarn quality characteristics.

Footnotes

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Key Grant Project of Ministry of Education of the People’s Republic of China (grant number 113027A).

Appendix 1

References

Chen

Feng

Z-h

Dong

T-z

et al.

Numerical simulation of the internal flow field of a new main nozzle in an air-jet loom based on Fluent. Text Res J 2015; 85: 1590–1601.

Guo

. A 3D numerical model for a flexible fiber motion in compressible swirling airflow. Comput Model Eng Sci 2010; 61: 201–222.

Jungbecker

Veit

. Computational fluid dynamics (CFD) and its application to textile technology. Simulation in textile technology, Aachen, NRW, Germany: Woodhead Publishing, 2012, pp. 142–178e.

Lin

Zeng

Wang

. Computational simulation of air flow in the rotor spinning unit. Text Res J 2016; 86: 115–126.

Xiao

MDH

Chen

. Numerical simulations of air flow behavior in spinning rotor. J Text Res 2014; 35: 136–141.

Zeng

. Numerical simulation of fiber motion in the nozzle of an air-jet spinning machine. Text Res J 2004; 74: 117–122.

Khodadadi Yazdi

Ramazani

SAA

Kamyabi

et al.

Simulation of flow of short fiber suspensions through a planar contraction. Sci Iran 2012; 19: 579–584.

Wang

Zhang

Lian

. A CFD simulation of 3D air flow and temperature variation in refrigeration cabinet. Proc Eng 2015; 102: 1599–1611.

Xiao

Dou

et al.

Numerical simulations of air flow behavior in spinning rotor. J Text Res 2014; 35: 136–141.

10.

Jinjing Chen

. A numerical analysis of heat transfer in an evacuated flexible multilayer insulation material. J Therm Anal Calorimetr 2010; 101: 1183–1188.

11.

Lawrence

Chen

. Rotor-spinning. Text Progr 1984; 13: 1–73.

12.

Lawrence

Chen

. A study of the fibre-transfer-channel design in rotor-spinning. Part II: optimization of the transfer-channel design. J Text Inst 1988; 79: 393–408.

13.

Kong

Platfoot

. Two-dimensional simulation of air flow in the transfer channel of open-end rotor spinning machines. Text Res J 1996; 66: 641–650.

14.

Kong

Platfoot

. Computational two-phase air/fiber flow within transfer channels of rotor spinning machines. Text Res J 1997; 67: 269–278.

15.

Hajilari

Eskandarnejad

Behzadan

et al.

Effect of two separate fibre feed systems in rotor spinning on yarn properties. Fibers Polym 2007; 8: 543–549.

16.

Peyravi

Eskandarnejad

Moghadam

. Dual-feed rotor spinning of cotton fiber: trash separation and yarn properties. J Text Inst 2014; 105: 377–382.

17.

Wang Jun

JXY

Yushu

Zhang

et al.

Dual feed carding device. Patent D01H4/32, D01H4/38 ed. China, 2014.

18.

Lawrence

Chen

. A study of the fibre-transfer-channel design in rotor-spinning. Part I: the fibre trajectory. J Text Inst 1988; 79: 367–392.

19.

Lawrence

Overview of developments in yarn spinning technology. In: Lawrence

(ed). Advances in yarn spinning technology, Chichester, UK: Woodhead Publishing, 2010, pp. 3–41.

20.

Rhodes

. Introduction to particle technology, Wiley, 2013.

21.

Cai

. Computer modeling of fiber movements in high-speed airflow. J Text Apparel 2005; 9: 77–85.

22.

Das

Alagirusamy

Fundamental principles of open end yarn spinning A2. In: Lawrence

(ed). Advances in yarn spinning technology, Lawrence, CA: Woodhead Publishing, 2010, pp. 79–101.

23.

Elhawary

Fibre to yarn: staple-yarn spinning A2. In: Sinclair

(ed). Textiles and fashion, Alexandria, Egypt: Woodhead Publishing, 2015, pp. 191–212.

24.

Ting

DSK

. Turbulence scales. In: Elsevier Science (ed.) Basics of engineering turbulence, Windsor, Ontario, Canada: Academic Press, 2016, pp. 69–98.

25.

Kundu

Turbulence A2. In: Cohen

Dowling

(eds). Fluid mechanics, 5th ed. Boston, MA: Academic Press, 2012, pp. 541–620.

26.

Liang

Zhi-hua

Teng-zhong

et al.

Numerical simulation of the internal flow field of a new main nozzle in an air-jet loom based on Fluent. Text Res J 2015; 85: 1590–1601.

27.

Kong

Platfoot

. Fibre transportation in confined channel with recirculations. Comput Struct 2000; 78: 237–245.

28.

Guo

H-F

B-G

. A 3D numerical model for a flexible fiber motion in compressible swirling airflow. Comput Model Eng Sci 2010; 61: 201–222.

Sim. run	Velocity inlet 1 (m/s)	Velocity inlet 2 (m/s)
#1	20	20
#2	20	30
#3	30	40
#4	40	50
#5	50	60
#6	60	70
#7	70	80
#8	80	90
#9	90	100
#10	100	100

Sim. run	Velocity inlet 1 (m/s)	Velocity inlet 2 (m/s)
#1	20	20
#2	20	30
#3	30	40
#4	40	50
#5	50	60
#6	60	70
#7	70	80
#8	80	90
#9	90	100
#10	100	100

Sim. run	Velocity inlet 1 (m/s)	Velocity inlet 2 (m/s)
#1	20	20
#2	20	30
#3	30	40
#4	40	50
#5	50	60
#6	60	70
#7	70	80
#8	80	90
#9	90	100
#10	100	100