Volume 20 Issue 3 Article 2

2021

Flow Characterization In Mine Ventilation Fan Blade Design Using Flow Characterization In Mine Ventilation Fan Blade Design Using

CFD CFD

Follow this and additional works at: https://jsm.gig.eu/journal-of-sustainable-mining

Part of the Computer-Aided Engineering and Design Commons, Mining Engineering Commons, and

the Sustainability Commons

Recommended Citation Recommended Citation Hassen, Anwar Endris (2021) "Flow Characterization In Mine Ventilation Fan Blade Design Using CFD," Journal of Sustainable Mining: Vol. 20 : Iss. 3 , Article 2. Available at: https://doi.org/10.46873/2300-3960.1063

This Research Article is brought to you for free and open access by Journal of Sustainable Mining. It has been accepted for inclusion in Journal of Sustainable Mining by an authorized editor of Journal of Sustainable Mining.

Flow Characterization in Mine Ventilation Fan BladeDesign Using CFD

Anwar Endris Hassen

FDRE Ministry of Mines and Petroleum, Research and Development Directorate, Addis Ababa, Ethiopia

Abstract

In axial ventilation fans, the generation of a uniform flow velocity is desirable for better efficiency. To that end,different fan blade types have been developed to achieve better flow uniformity. This article aimed to characterizethe flow distribution and its uniformity in four blade designs, namely constant chord, tapered blade, skewed blade,and tapered skewed blade, using Computational Fluid Dynamics (CFD). The study employs an iterative study wherekey study decisions are made as the study progresses. The study began with the selection of a blade profile for thestudy. A comparative study between the NACA seven-digit and four-digit series was conducted and for its higherflow throughput, the four-digit airfoil profile was selected. Next, with 30 and 40� Angle of Attack (AoA), the constantchord blade flow pattern is characterized. At 40� AoA flow disturbance and high-velocity spots were observedestablishing the problem statement. Following that, three optimization strategies (tapering, skewing, and taperskewing) were applied in the design, and the flow pattern of each design was studied. Using a dispersion studya flow uniformity comparison between the models conducted. The property trade-off between three key performanceindicators: efficiency, flow rate, and flow uniformity studied. The result shows an axial fan having a higher efficiencydoesn't necessarily mean it has higher throughput whereas lower flow dispersion relates to the system's higherefficiency. Therefore, it can be concluded that seeking higher efficiency and flow uniformity in the design anddevelopment of axial fans comes with system throughput trade-off.

Keywords: mine ventilation, axial flow fan, tapered blade, skewed blade, CFD

1. Introduction

T he mining industry highly benefits from theuse of mechanical ventilation. The ventila-

tion system aims to serve a critical role in un-derground mining operations through thedilution and removal of gas and dust, heatextraction, and oxygen supply for the safety ofproduction and staff. The system has such crucialsignificance; it is expected to be operational24 hours a day. Within the operation hours, theventilation system estimated to account for one-third of the mining operation's electrical con-sumption [1]. With this understanding and anoutlook for sustainable mining [2], the quest forhigher energy efficiency in the system continues.It is plain fact that the efficiency of the ventilation

systems and the system power consumption has an

inverse relation [3]. Even if in the end, the totalsystem integration efficiency determines the powerconsumption of the ventilation system, every smallefficiency improvement that can be achieved at theindividual component level is significant. Theconsideration of fans as a primary component of thesystem is challenged by Papar et al. [4], where theyrecommend system-based approach instead ofcomponent-based efficiency improvement. Sucharguments shape the research focus in the fieldwhere most studies conducted in mine ventilationlook at energy efficiency from the system integrationlevel. To validate the move Demirel [5] highlightedventilation system suffers a higher efficiencyreduction due to the outlet dumper's low efficiencyon her review paper. However, with the rising use ofmore advanced control systems such as ventilationon demand, nonlinear system control, and novelintermittent approaches, the use of damper as theonly flow regulator is expected to decline. This fact

Received 16 March 2021; revised 22 June 2021; accepted 30 June 2021.Available online 28 July 2021.E-mail address: eanwre@gmail.com.

https://doi.org/10.46873/2300-3960.10632211-8020/© Central Mining Institute, Katowice, Poland. This is an open-access article under the CC-BY 4.0 license(https://creativecommons.org/licenses/by/4.0/).

RESEARCH

ARTIC

LE

leaves fans as the lowest efficient component in thesystem. In addition, most of the components in thesystem can be found on-shelf whereas fans need tobe tailored with site-specific data and its design in-fluences the selection of other components in theventilation system. This makes fans arguably theprimary component of the system.What makes fan design more challenging than

other site-specific components of the system is thatthe fan itself contains several parts as shown in Fig. 1and each part has their own design control param-eters. Controlled parameters in axial fan bladedesign include but are not limited to, blade airfoilprofile, blade span, blade breadth, angle of attack,the hub to tip ratio, tip clearance, and blade number.This parameter determines the aerodynamic char-acteristics of the axial fan and ultimately its effi-ciency. Poor study of each parameter in the designprocess will result in various energy loss factorsindicated in [6]. Hence, axial ventilation fan bladedesign presents a significant challenge and goodresearch area for energy conservation. To that end,the effect of design-controlled parameters on theenergy efficiency of axial fans has been scrutinized inother applications such as in electronics cooling. Yetthe literature from the mine ventilation perspectiveis limited. [7] Present an optimization strategy bymanipulating the airfoil profile to achieve betterefficiency using blade element theory. Also recentlyPanigrahi and Mishra [8] conducted a comparativestudy to select an efficient airfoil profile for mineventilation application by employing CFD. In theirresearch Petrove and Popov [9] entertained the useof variable pitch fan blades for higher efficiency. Theeffect AoA has on the power consumption of theventilation discussed by Kazakov et al. [10].Flow distribution study in axial fans is highly

important as it provides a comprehensive representation of the effect each control parameters haveon the system. Different configurations of thecontrolled parameters in the design process due

result in different flow patterns. The manipulationof these parameters will allow the designer to seekmore uniformity in the flow that is challenged by thefan wheel velocity governing equation. Equation (1)relates the fan circumferential speed (u) with theradius of the blade (r) and its angular velocity (u).The governing equation states that a point particleat the mid radius of the fan wheel is only 25% of thevelocity at the tip. Differences in speed affect themagnitude of exerted forces such as the centrifugaland shear force on different parts of the fan bladesurface. Hence, the resolved forces affect the ve-locity triangle components resulting from deflectionof discharge.

u¼ ru; ð1ÞThe velocity triangle component in turboma-

chinery is highly influenced by the friction coeffi-cient of the contact surface, fluid properties, andangular velocity (u). The circumferential speedmarks its influence by creating a higher fluid shearforce on the blade surface near the center of rotationdue to its low circumferential speed. This leads todelayed fluid particle escape time from the workingsurface causing it to be highly pressurized. Thepressure difference between certain portions of theblade added with centrifugal force can make thefluid particles susceptible to the pull effect the low-pressure side has at the tip.Any fan wheel working to discharge air in the

axial direction through a rotational motion cannotachieve an absolute axial discharge velocity due tothe rotation speed. There must also be a rotationalcomponent creating the air deflection angle be-tween the actual wheel velocity and axial direction.In general, when the velocity of an axial fan isstudied, an axial component of the velocity triangleis referred to and it is defined as the average velocityof air moving in the axial direction through the netarea of the fan in Equation (2).

Flow Velocity¼Actual Flow Rate / (Throat Area eHub Area), (2)

Fig. 1. The schematic of axial flow fan [11].

Abbreviation

u circumferential velocityr radiusD diameteru angular velocityq velocity pressurer densityv velocitys standard deviations modified standard deviation

JOURNAL OF SUSTAINABLE MINING 2021;XX:144e156 145

RESEARCH

ARTIC

LE

The fan wheel circumferential speed effect onsystem efficiency can also be observed in the ve-locity pressure (q) calculation of the system usingEquation (3). Velocity pressure, as pertains to fans,refers to the impact of a moving air stream relativeto a fixed object and is proportional to the kineticenergy. Equation (3) shows that the velocity pres-sure increases as a square of fan speed (y) where thedensity of the working fluid (r) assumed constant inincompressible state. A different discharge velocityalong the working surface of the fan blade whenconverted into velocity pressure, the lower flowvelocity side will average the higher velocity side.When a system has a higher velocity pressure re-quires higher power to perform work, falling shortwhen efficiency is considered. Thus it is necessary tomaintain uniform discharge velocity for lower ve-locity pressure to avoid energy wastage and achievehigher efficiency in the system.

q¼12

�ry2

�; ð3Þ

Therefore, the use of constant chord blades inaxial fans will result in highly dispersed flow andlower efficiency as defined in Equations (1) and (3).To counter the problem, different types of fan bladedesigns emerge as a result of design control pa-rameters manipulation. Tapered blades, curvedblades (forward and backward), and skewed bladesare the most common types of blades in the in-dustry. As individual blade types with uniquedesign parameter configuration, it is no brainer toexpect a unique aerodynamic characteristic in eachdesign model. To that end, the current study aims toinform on flow characteristics of four-blade typeswhich are constant chord, tapered blade, skewedblade, and tapered skewed blade using CFD. Theflow characterization will be conducted by employ-ing 16 test points in the flow domain and flow ve-locity will be recorded at the measuring points. Therecorded results will be analyzed for dispersion rateusing a specially modified standard deviation for-mula. The effect design control parameters, in bladedesign, have on flow characteristics will be dis-cussed. The study will be the first to study propertytrade-offs between three key performance in-dicators (KPI) of axial fans which are efficiency, flowrate, and flow uniformity.

2. Approach and Methods

This study aims to characterize the flow distri-bution in four blade designs: constant chord un-twisted blade, tapered bladed, skewed blade, andtapered skewed blade. An iterative research

approach was employed where key study decisionsare made as the study progresses. And each bladetype flow characterization was conducted and dis-cussed by case section. The first case section mainlyseeks to (a) select an airfoil profile to use in thisstudy, and (b) establish the problem that ariseswhen using a constant chord blade as discussedtheoretically in the introduction section.For airfoil profile selection the study relied on

existing literature suggestions and a comparativestudy between the NACA seven and four digits se-ries performed. Based on higher throughput gen-eration, the airfoil profile was selected for geometrypreparation in the remaining part of the study.Secondly, the first section aims to validate thestatement made in the introduction section. Theintroduction section theoretically highlights that asthe AoA increases, the bottom working surface ofthe blade will be exposed to large amounts of theworking fluid. This will increase the drag force thatwill amplify the centrifugal force allowing morefluid particles to be drawn to the blade tip andescaping the surface at higher velocity. Thus, section3.1.2a. higher blade AoA will be used to observe thephenomena graphically.Following the selection of the airfoil profile and

establishing the problem, the second case sectionwill focus on flow characterization of the remainingblade types designed to eliminate the observedproblem. Therefore, the section will use flowdispersion study as flow characterization parameters.In the flow dispersion and uniformity study, flowvelocity values in the flow passage remain a vitalvariable. To help capture the velocity value, sixteen

Fig. 2. Proposed test points.

146 JOURNAL OF SUSTAINABLE MINING 2021;XX:144e156

RESEARCH

ARTIC

LE

probe points were applied at 0.05m distance in theexit section of the fan blade in all models as shown inFig. 2. The placement of the monitor points aims toget a sample representation of the blade passageflow velocity. Test point 1 is placed on the principleindicated in Equation (2) (flow velocity should bemeasured in the net area). The position placement ofpoint 1 is used as a reference to set the rest of thetesting points. Thus, the number of the probe pointsbecomes the function of the net area of the flowdomain, the location of the reference point, and thepoint separation distance. For this study the authorused 0.05m point separation, resulting in a total of 16probe points. Since the fan wheel is expected toportray a similar flow pattern in between twoconsequent fan blades, the placement of the probepoints in one region was found to be adequate forflow representation of the system. The probe pointsrecord the velocity of the point particle in all threedimensions independently, pressure, turbulence ki-netic energy (k), and the specific rate of dissipation.Nevertheless, the probe points were able to capturedifferent parameters, only the velocity in the axialdirection will be used to perform a comparative flowdispersion study. Besides the probe points, twosample plane segments of the flow result will beused to make graphical inspection and flow ratemeasurement.In general, the study contains four fan model sets

with different fan geometries, where model set onecontains four geometries and the remaining modelset will contain two geometries each, a total of 10geometries. Model set one will only contain aconstant chord blade, the second set containstapered blade design geometry, the third set con-tains the skewed blade, and the last set contains atapered skewed blade. Fig. 3 shows a general rep-resentation of each model set blade type. A detaileddesign of each blade and the principle of theintended design alternation will be discussed ineach case section. Care was taken to maintaingeneral similarities with all fan designs except theparameters that remain under investigation. All fangeometries are designed with the same hub-to-tipratio, tip clearance, and effective blade span. To

maintain low hub influence on the aerodynamicsperformance of the system, a tapered blade neckwas introduced to all geometries between the huband the effective blade. To achieve full flowdevelopment and far-field condition, both the inletand outlet boundaries are extended by a factor of2D and 4D respectively, where D is the diameter ofthe fan wheel.The study contains a total number of 10

geometrical models for CFD analysis and con-ducting a grid sensitivity study for each analysisposes to be a computationally expensive task. Toobtain the acceptable mesh size, the author pro-posed to conduct the mesh dependency study onone randomly selected model from the geometryset. The test was conducted on tapered blade pro-file fan geometry with three different coarser,moderate, and finer mesh sizes containing 6.8million, 11.8 million, and 22.1 million cells,respectively. The velocity result of the simulation atthe proposed probe points compared to investi-gating the error margin occurs due to the differencein grid size. An average of 0.22% relative diver-gence deviation was recorded between the coarserand moderate mesh sizes. However, the differencemargin fell to 0.09% between the last two grid sizesshowing better result divergence. The same meshdevelopment setup is used in all cases includingregion refinement near the rotating region to bettercapture the turbulence development and a localrefinement of the fan blade geometry. The meshcreator algorithm's ability to automate grid sizingfollowing the applied physics and geometry makesit difficult to obtain the same size of cell number ineach case. Despite that, the remaining mesh gen-eration operation on the rest of the model isdesigned to size in between the moderate and finermesh cell count.The same simulation setup is also used in all case

sections. The problem is set on the assumption thatthe pressure difference created due to the rotation ofthe axial fanwill create airmovement. Therefore, theair inlet and outlet boundary condition applied witha pressure gauge value of zero pascal. The fan bladeand the fan casing have a no-slip wall boundary

Fig. 3. General representation of designed fan blade types: (a) constant chord, (b) tapered, (c) skewed blade, (d) tapered/twisted.

JOURNAL OF SUSTAINABLE MINING 2021;XX:144e156 147

RESEARCH

ARTIC

LE

condition.With itsmerit of goodbehavior in adversepressure gradients and capturing flow separation,the SST k-omega turbulence model with highlyrefined mesh should be adequate for this study.For its relatively simple yet robustness, the use ofMultiple Reference Frames (MRF) was employed tomodel the rotating part of the problem. A study byZadravec et al. [12] shows that the size of the rotatingdomain used in the MRF model influences theresult. However, their study did not propose anexclusive suggestion on how to scale the domain.Thus, the author used a ratio of 1.4D and 1.1D toscale the domain height and diameter, respectively,where D is the diameter of the fan wheel. To startthe problem, a fixed rotational hub speed of 50 rad/sec was introduced to all models. The problem wassolved using OpenFOAM solver, where the flow isconsidered incompressible and steady-state timedependent on the SimScale platform. Both thenumerical results and graphical representation ofthe results were inspected to characterize the flowvelocity distribution and to make a comparisonbetween models. To assist the comparison, tworesult segments were prepared using clip filters at0.01 m away from the fan wheel from the entry sec-tion and 0.3 m towards the exit section perpendic-ular to the axial flow direction. To calculate theefficiency of the fan, in the simulation a force andmoment result control parameter is used. The resultcontrol parameters were able to record the usefulpower that is coming out of the fan which is a mea-sure of power that the fan is producing and themeasure of force imparted by the fluid on the fanblade which is equivalent to the torque required torotate the fan wheel.

3. Results and Discussion

This chapter is structured in two major sectionscontaining sub-sections within. The first section in-cludes the selection of airfoil profiles that will beused for blade geometry design while flow distri-bution in constant chord blade type is characterized.The section also establishes a linkage between thetheoretical principles highlighted in the introduc-tion section and the CFD analysis result. Followingthat the second section focuses on flow character-ization in blade types optimized for better flowdistribution.

3.1. Constant chord blade

3.1.1. Airfoil profile selectionThe selection of the CFD analysis model was the

first step to begin the process. For its low

computational cost requirement and simplicity,most literature employs the use of a 2D analysismodel. However, even in free stream applicationsstudies conducted by Talib et al. [13] conclude thatthe 2D analysis model failed to capture the fullflow structure and unsteadiness of the case prob-lem. In the current case where the ventilation fanis constructed and operates in the casing,controlled parameters such as tip clearance andboundary layer development due to the wallfunction of the fan duct attribute to the flowpattern. Hence, a more complex and unsteady flowstructure is expected. Therefore, to best studythese phenomena, a 3D analysis type is selected.For the selection of a comparative candidate airfoil

profile, the study relies on existing knowledge. Tothat end, after conducting a comparative study be-tween six airfoil profiles Panigrahi and Mishra [8]concluded that a NACA 747A315 airfoil providedbetter efficiency. On the other hand, a summarymade on the NACA airfoil series by Pier Marzocca,a professor from Clarkson University, highlightssome properties that could make the seven-digitseries unfavorable for ventilation applications.1

Properties such as reduction of maximum lift coef-ficient and poor stall behavior as the AoA increasesare mentioned as disadvantages of the serieswhereas the four-digit airfoil shows "good stallcharacteristics" that favor it for ventilation fan pur-poses. Therefore, NACA 747A315 and NACA 6412were selected for comparison and the selection ofthe NACA 6412 airfoil profile as a candidate isarbitrary.At the same time, the current study aims to

characterize the effect AoA has on flow distributionand throughput generation. Panigrahi and Mishra's[8] study concludes the NACA 747A315 airfoil pro-file could perform better at 15� AoA while studiessuch as [14], [15] conclude a ventilation fan systemcould perform better at AoA range between 40 to45�. The methods and approaches used in thestudies are different and the conclusions made onboth sides have good merit. Based on the suggestionan AoA of 15, 30, and 40� was selected for investi-gation in this study. In most comparative studies ofsuch systems, efficiency is used as a key perfor-mance indicator. However, this study uses flowthroughput as a comparison criterion because mineventilation systems are designed based on therequired flow rate. The analysis was first conductedon NACA 747A315 at 15 and 30� AoA and the 15�

1“ The NACA airfoil series” https://people.clarkson.edu/~pmarzocc/

AE429/The%20NACA%20airfoil%20series.pdf. Accessed 20 Apr. 2020.

148 JOURNAL OF SUSTAINABLE MINING 2021;XX:144e156

RESEARCH

ARTIC

LE

AoA generated the lower throughput, hence thedesign dropped from further study. Following thatthe analysis result of 747A315 at 30� AoA comparedwith NACA 6412 at 30� AoA. From the comparisonthe NACA 6412 at 30� AoA generates a largerthroughput. Therefore, the whole NACA 747A315design set dropped and the NACA 6412 airfoilprofile was selected for further investigation at AoAof 40�.To further discuss the result obtained from each

analysis, two segment planes in the inlet and exitsection of the fan wheel from each analysis werestudied, and the result is presented in Table 1. Theuse of arbitrary sampled planes to investigate re-sults in CFD analysis has also been adopted in thedifferent scientific literature. The bulk result at theselected section plane is calculated using theaverage of all cell values projected on the sampleddata plane. Properties of flow rate and velocity atsectioned sample planes are used to make a nu-merical comparison between the two NACA airfoilfamilies. The numerical simulation result showsa higher degree of attack was able to draw a largerthroughput. The same NACA 747A315 profiled fanblade generates a 114% higher flow rate when theAoA33 increases from 15 to 30�. Moreover, the four-digit family has recorded a flow rate increment of22% when the AoA increased to 40 from 30�. Thecomparison between the two families at the sameAoA of 30� shows the four-digit airfoil family wasable to increase the volume flow rate of the systemby 10%, making it a superior choice for suchapplication.The fourth design in the set has presented

a higher average flow velocity and a higher velocityrecord in the system. The increments of themaximum velocity as per the increment of the AoAin the other designs were moderate. However, themaximum velocity in design four reachesa magnitude of 28 m/s. Due to the higher averagevelocity, the design was able to generate a higherflow rate from the model set. However, the flowrate increment rate and system efficiency fall whencompared to the first model set. This result

indicates the system flow distribution lacks uni-formity and baleful characteristics such as noiseand higher vibration may unveil in the system. Thefirst two designs in the model set with seven-digitairfoil series have a flow rate increment of 0.03 m3/sfor each 1� AoA increment from 15 to 30� and theefficiency of the system increases significantly. Thelast two models with the NACA four-digit record a0.026 m3/s flow rate increment for each 1� AoAincrement from 30 to 40� and its efficiency declinedby almost 1%.

3.1.2. Flow distribution study in constant chord bladeTable 1 can give a clear numerical result to help

make a selection decision based on the comparisonof the needed volumetric flow rate. However, thenumerical result alone can fall short to give thenecessary information related to the main objectiveof this work. The introduction section highlights theeffect of AoA variation on the total flow generation.To recall, it is stated that as the AoA increases, thelower surface of the blade is exposed to a higheramount of working fluid. The large exposure meansa higher drag and more flow generation. The higherdrag makes the rotating fluid particles more sus-ceptible to centrifugal force where more fluid willescape to the blade tip at higher velocity. In addi-tion, the rise of AoA more than optimum value willresult in flow separation affecting performance.

Table 1. CFD analysis results calculated at the sampled data segment of Model Set 1.

Angleof Attack

Working areaper blade (cm2)

Studied section (m) NACA 747A315 NACA 6412

Maxvelocity (m/s)

Flowrate (m3/s)

Efficiency (%) Maxvelocity (m/s)

Flowrate (m3/s)

Efficiency (%)

15� 369.36 0.3 inlet section 3.821 4.023e-1 51.7 e e e

0.1 outlet section 3.345 e

30� 369.36 0.3 inlet section 6.875 8.595e-1 83.59 7.901 9.435e-1 83.140.1 outlet section 6.711 7.626

40� 369.36 0.3 inlet section e e e 21.6 1.209 82.410.1 outlet section e 17.45

Fig. 4. Model Set 1 Axial flow velocity magnitude at probe points.

JOURNAL OF SUSTAINABLE MINING 2021;XX:144e156 149

RESEARCH

ARTIC

LE

Therefore this section will characterize the dis-cussed phenomena.To help us study flow distribution, the velocity

values obtained at the monitoring points aresuperimposed on the same graph as shown in Fig. 4.The plot shows, all designs in the model set adheredto the relation defined in the introduction sectionbetween the circumferential velocity of the fanwheel and the velocity distribution along the fanblade. The plot shows there is a lower flow velocitynear the hub and increases along the blade span asthe circumferential speed of the wheel increasestowards the tip of the blade from probe points 1 to 4.The obtained CFD analysis result is valid as it sat-isfies the relation defined by the governing Equation(1). From the velocity plot, it can also be observedthat a lower AoA near the hub can lead to significantflow velocity loss. As the AoA increases from 15� indesign 1 to 30� in designs 2 and 3, the starting ve-locity values at point 1 shows a significant incre-ment. Which indicates higher AoA near the hubcould help compensate velocity loss resulted fromlow circumferential speed near the hub. Whereasthe flow velocity at point 1 shows a slight incrementwhen the AoA increase from 30� in design 3 to 40� indesign 4, indicating the AoA reaching its optimumvalue and the flow starts to separate. The overallflow distribution in the first three cases shows thesame flow pattern where the velocity deviation fromone probe point to the next remains moderatewhereas design 4 recorded undesired sudden up-take in the last four probe points indicating high-velocity spots.In general, in both airfoil cases, a constant chord

blade develops a similar flow distribution along the

work surface of the blade where the flow velocityincreases gradually towards the tip. However, themaximum flow velocity achieved by 40� AoA isthree times higher than that of 30� AoA. The highervelocity development in the system lacks uniformityas shown in the color mapped result in Fig. 5b. Thecolor mapped result scaled to the zone flow velocityat 2m/s difference. The first model has four velocityzones while the second model has eight velocityzones. Since Equation (2) recommends the calcula-tion of flow velocity only in the net area of the fanwheel, the velocity zoning at the hub and the bladecan be neglected. Therefore, the model will havetwo velocity zones from Fig. 5a and Fig. 5b. Fig. 5ahas the same flow patterns in between the workingblades, yet Fig. 5b shows a different flow pattern. Inthe model also the flow velocity reaches up to 28m/sin a small part of the casing corner. Such unevenflow distribution and high-velocity concentrationcan induce noise and vibration and result in an ef-ficiency decrease.

3.2. Design optimization methods for optimal flowdistribution

The previous section has defined the total flowdistribution characteristic in a constant chord bladewith different airfoil profiles and AoA. The sectionalso highlights the larger AoA in the model setintroducing undesired flow phenomena that couldundercut the system efficiency. This section focuseson characterizing properties trade-off betweencontrolled parameters and performance indicators.The flow characterization will be conducted on themost common modification strategies designed to

Fig. 5. Clip filter with velocity magnitude in the fluid flow direction: (a) AoA 30�; (b) AoA 40� (Velocity color mapping scale: 0e10 m/s for a and0e20 m/s for b).

150 JOURNAL OF SUSTAINABLE MINING 2021;XX:144e156

RESEARCH

ARTIC

LE

avoid the shortcomings observed in the previoussection. Efficiency, flow throughput, and flowdispersion rate will be used as KPI to characterizethe system. The numerically calculated result of thedefining KPI is presented in Table 3.Equation (1) states in cases where the angular ve-

locity remains constant, radius remains the onlyvariable that the circumferential velocity outcomedepends on. The variable radius has a uniformdevelopment from zero at the hub to full span lengthreaching the tip. Thus, the design of geometriescontaining the optimization strategies (tapering andtwisting) will be dependent on the same principlewhere the tapper percentage and the twist anglevariation from the hub to the tip will be a function ofblade radius as shown in Table 2. The modificationmade to the geometry and the principle behind it willbe discussed in detail in case sections.The proposed probe points will be employed to

perform a comparative study within the design andacross the designs. The comparison across designsis straightforward, where the same positionedprobe point in different designs can be compareddirectly and the location placement of the pointswill serve as a relation. However, the comparisonbetween probe points within the same model couldpose a challenge since the probe points do not havedefined relations. For an optimal flow distribution,it is ideal to achieve an equal flow velocity resultbetween the proposed probe points. In otherwords, the velocity value obtained at 16 test pointsneeds to converge at the same result. However, anabsolute value of zero difference between probepoints cannot be achieved due to the constraints aducted fan design comes with. The development ofthe boundary layer in ducted fun is the majorreason. Therefore, such optimization strategies aimto achieve a minimum value of dispersion withinthe collected data set. For dispersion study, stan-dard deviation statistics presented in Equation (4)are widely used.

s¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiPni¼1

�xi � x

�2

N� 1

vuut; ð4Þ

s ¼Pn

i¼1ðxi � x1Þn� 1

; ð5ÞThe measurement of dispersion in the Equa-

tion (4) is relative to the mean value (x ) of the dataset. However, in this study optimization strategiesaim to maintain similar velocity with the velocitynearest the hub. In our case, the probe point valueat position 1 is the reference for the rest of thevalues in the data set for the dispersion study.Therefore, the mean value in Equation (4) wasreplaced with probe point 1 value. The squarepower operation of the bracket value in the stan-dard deviation study aims to eliminate the possi-bility of a negative value from the subtractionoperation. Since the reference value of the currentdata set is the smallest in the set, the operation inthe bracket would not result in a negative value.Thus, the standard deviation equation was modi-fied in the form of Equation (5) to align it with thecurrent study objective. The operation in the divi-dend, the subtraction of one from the data set, willbe held in the assumption of excluding the refer-ence value (x1) from the dispersion study because itserves as the reference value. Equation (5) createsthe much-needed mathematically defined relationfor velocity values obtained in the same design andserves as a comparison medium in the study.

3.2.1. Tapered bladeThe surface area and velocity of the effective blade

are two major factors that come into play for flowoptimization in tapered blade designs. To compen-sate for the slower speed of the fan wheel near thehub, the tapered blade design aims to increase thechord breadth, increasing the working surface of theblade. In reverse, where the wheel velocity is highertowards the tip of the blade, the method aims to

Table 2. Geometry description.

Model sets Working area per blade (cm2) AoA in degree chord length (mm)

hub midspan tip difference (%) hub midspan tip difference (%)

Tapered blade Design 1 308.82 40 40 40 0 90 75 60 33Design 2 328.73 40 40 40 0 90 80 70 22

Skewed blade Design 1 366.42 40 35 30 25 90 90 90 0Design 2 364.23 40 30 20 50 90 90 90 0

Taper/twist blade Design 1 308 40 35 30 25 90 75 60 33Design 2 328 40 35 30 25 90 80 70 22

JOURNAL OF SUSTAINABLE MINING 2021;XX:144e156 151

RESEARCH

ARTIC

LE

decrease the surface area of the blade by decreasingthe breadth of the blade profile. This design wouldresult in surface area reduction as a variable ofblade length. Thus, each blade will have a narrowerblade airfoil profile at the tip where the fan whilehas higher speed and the airfoil profile near the hubwill have a longer chord length. The optimizationmethods used at the National Renewable EnergyLaboratory (NREL) combined experimental rotordesign [16]. To investigate the optimization strat-egy's approach to overcoming the defined problem,design 4, from section 3.1.2., used as a benchmark.To establish a relation with flow pattern and

tapered percentage, a model set was prepared withtwo different geometries. The geometries aredesigned with NACA 6412 airfoil profile at bothends with different chord lengths. Designs 1 and 2have chord length differences of 33 and 22%,respectively, between the larger and smaller airfoilprofile as shown in Table 2. The chord length in-between remains the function of the two airfoilprofiles chord lengths at both ends of the blade andthe radius. The modification led to a significantreduction in the surface area of the effective blade.The selected benchmark study has a working sur-face area of 369.36 cm2 whereas the new geometrieshave a surface area of 308.82 cm2 and 328.73 cm2 perblade in designs 1 and 2 respectively. All othergeometrical features remain the same as thebenchmark study with the same AoA of 40�. Thesame CFD simulation setup as the previous sectionwas used to conduct the flow analysis.For graphical inspection, the velocity color map

results at 0.03 m sampled segment plane in the inletdirection presented in Fig. 5. The first design wasable to generate a total of 1.096 m3/s flow rate andachieved a maximum flow of 8.988 m/s. The secondmodel generates a 1.136 m3/s flow rate anda maximum flow velocity of 9.639 m/s recorded.The optimization method was able to cut down thesudden velocity uptake recorded at probe points14, 15, and 16 in the benchmark model. The velocitydispersion rate within the design at recognized

probe points was calculated as 4.6 in design 1 and 5in design 2. Hence, the applied optimization strat-egy was able to reduce the velocity dispersion ofthe system by 20.6% in design 1 and 13.7% indesign 2 from 5.8 m/s velocity dispersion recordedin the benchmark design.Looking at Fig. 6a, a significant flow velocity

decrease at probe points 8 and 12 recorded andvelocity uptake at point 13 observed in all modelssuperimposed in the plot. In Fig. 4 this phenomenaare observed only in design 4 and the authorassumed the observed flow distribution is due to themodel's high AoA characteristic. However, as pre-sented in Fig. 6a, the same phenomena can beobserved in the second model set in total alignmentwith the benchmark design. Through a preliminaryreview of the placement of the probe points and thecolor mapped result of the velocity distribution inFig. 7, it can be understood that the sudden declineof velocity at probe points 8 and 12 mainly occursdue to the boundary layer development near the fancasing. Whereas the velocity uptake at probe point13 occurs because of the placement of the probepoint near the tail of the airfoil, exposing the point toa higher flow escape region. The relation governedunder Equation (1) does not take into account thevelocity drop and uptake that occurs due to suchfactors. Thus, the author proposed to exclude probepoints influenced by those factors in the rest of thestudy. The graph is re-plotted in Fig. 6b excludingthe three points and the velocity dispersionrecalculated.With the remaining 13 probe points, flow velocity

difference was recalculated using Equation (5) andthe benchmark design recorded a flow dispersionrate of 6.2. Designs 1 and 2 of the tapered model setrecorded a dispersion rate of 4.8 and 5.1 respec-tively. The dispersion rate difference between thebenchmark model and optimized designs calcu-lated to be 22.5 and 17% in designs 1 and 2respectively. Both designs improve the flow uni-formity of the benchmark study. Design 1, witha higher tapering percentage, shows better flow

Table 3. CFD analysis result (Efficiency, flow rate, dispersion rate).

Model sets Averagevelocity (m/s)

Efficiency(%)

Flow rate(m3/s)

Dispersionrate

Constant chord/untwisted blade Design 1 NACA 7 15� 2.544840769 51.74825175 0.402 2.2122175Design 2 NACA 7 30� 5.598164615 83.59447005 0.860 4.1779625Design 3 NACA 4 30� 6.169536154 83.14285714 0.944 4.640959167Design 4 NACA 4 40� 8.087791538 82.41758242 1.209 6.216591667

Tapered blade Design 1 Tap 33 6.91232 86.18705036 1.099 4.895103333Design 2 Tap 22 7.139298462 83.33333333 1.138 5.191985833

Skewed blade Design 1 Skewed 25 6.577762308 84.71698113 1.016 4.587668333Design 2 Skewed 50 5.420409231 87.5 0.834 3.760093333

Tapered/twisted blade Design 1 Tap/twist 22 6.19675 86.02150538 0.973 4.491631667Design 2 Tap/twist 33 6.41811 85.55555556 0.999 4.621030833

152 JOURNAL OF SUSTAINABLE MINING 2021;XX:144e156

RESEARCH

ARTIC

LE

uniformity in the model set. This establishes aninverse relationship between tapering percentageand flow uniformity. As expected when the flowuniformity improved, the system efficiency alsoimproved showing direct relation. The benchmarksystem has an efficiency of 82.4% design 1 and 2recorded a system efficiency of 86 and 83%,respectively. However, system throughput gener-ation shows an inverse relation with flow unifor-mity. Design 1 generates a 1.096 m3/s volumetricflow rate whereas design 2 generates a 1.138 m3/svolumetric flow rate. Therefore, it can be concludedthat as the tapper percentage increases, systemflow uniformity will improve and system efficiencyalso improved. However, system flow uniformityand efficiency improvement come at the cost ofsystem throughput.

3.2.2. Skewed bladeThe introduction of skewed angles is highly

practiced in axial fan design and has been studied in

detail in other sectors. Better performance in thereduction of noise [17e20] and total pressure loss[20] have been documented in published literature.Efficiency improvement is also highlighted in pa-pers like [19, 21, 22]. In the flow velocity distributionoptimization, the skewed blade principle works ondefining the effective blade exposure to the air. Withlow AoA, the wheel blade parts the air through theblade leading edge. The aerodynamic shape of theleading edge of an airfoil is designed to have a lowersurface area that results in lower exposure of theblade to the working fluid. Thus, as the fan wheelrotates at a lower angle of attack, the air deflectionangle in the desired axial direction will be lower. Insuch cases, since the contact between the workingfluid and the blade is minimal, a loss occurring dueto drag will remain low. Whereas as the AoA of theworking blade increases, the bottom surface of theblade that is relatively flat will be exposed to moreworking fluid increasing the deflection angleallowing the system to generate a higher flow rate.

Fig. 6. Model Set 2 Axial flow velocity magnitude at the probe point: (a) 16 probe points (b) 13 probe points.

Fig. 7. Tapered blade clip filter with velocity magnitude: (a) 33% chord length difference; (b) 22% chord length difference.

JOURNAL OF SUSTAINABLE MINING 2021;XX:144e156 153

RESEARCH

ARTIC

LE

The higher exposure will contribute to energy lossdue to higher drag and the AoA should not beincreased more than optimal to control the devel-opment of stall characteristics. Therefore, the ge-ometry proposed for the study will have a higherAoA near the hub and decrease towards the tip ofthe blade and the design parameters are presentedin Table 2.Following the method used in the previous sec-

tion, a model set was prepared to investigate thetwist blade effect on the flow pattern. The model setcontains two designs with different twist angles.The first geometry has a twist-angle blade thatgradually varies from 40 to 30� AoA from the hub tothe tip respectively. The second geometry has 40 to20� AoA variation without any tapered modifica-tion. In this section, the result obtained for NACA6412 at 40� of AoA from amodel set one will be usedas a benchmark. The first geometry has a surfacearea of 366.42 cm2 per blade whereas the secondgeometry has a surface area of 364.23 cm2 perblade. The introduction of twist angle strategy hasshown a slight effective blade surface area losswhen compared to the benchmark design having369.36 cm2 working area. The total number of threeblades used in the design and all other geometricalfeatures remain to be the same.At the sampled segment, design 1 records

1.016 m3/s, just over a 16% flow rate shrink from thevalue of the benchmark with a maximum velocityof 8.33 m/s. Design 2 generates a 31% lower flowrate from the benchmark and the maximum flowvelocity recorded at probe point 11 to be 6.4 m/s. Inboth cases, it can be concluded that the designmodification plays a significant role in the redac-tion of a high-pressure area near the center ofrotation indicated in the benchmark study. Theresult exhibits the larger twist angle can causea significant flow velocity fall affecting the averageflow rate of the system. Since Fig. 8 show relatively

flat trend in flow velocity, a flow uniformityimprovement expected. Using Equation (5), thevelocity dispersion in the system calculated anddesign 1 managed to reduce the total velocity dif-ference recorded in the benchmark design by 33%while design 2 cut it by 45% achieving better flowuniformity in the model sets. Improved flow uni-formity in the system has also led design 2 to ach-ieve a higher system efficiency of 87.5%. Inconclusion, as the twist angle percentage increasedthe system achieved lower flow dispersion andimproved system efficiency. However, the optimi-zation strategy comes with a flow rate reduction,same as tapered optimization.

3.2.3. Tapered/Twisted BladeA tapered/twisted blade design has been used on

the NREL Combined Experiment Rotors [23]. Thecombined design between both methods aims totrade off the characteristics that both models fallshort of meeting individually. This includes thetapered blade model set achieved a higherthroughput than the skewed model. And theskewed model achieved better flow uniformity thanthe tapered model. Now in this section, both opti-mization strategies will be applied on a single ge-ometry to investigate the aerodynamiccharacteristics trade-off between the models.In the skewed blade model set, design 1 has

recorded higher flow velocity dispersion. Byapplying a tapered feature to the design, this sectionwill investigate the change in the flow pattern. Thus,for this section, the result obtained in design 1 froma skewed model set will serve as a benchmark of thestudy. Following the same geometry preparationconcepts as the previous section, two geometricalmodels were prepared for CFD analysis in thissection. The first design has a twist angle that variesfrom 40 to 30� and a taper percentage or chord dif-ference of 22%. The second geometry has a similartwist angle and the chord difference changed to33%. The new geometry has not shown any signifi-cant working surface area difference for the bench-mark design. Using the same CFD analysis setup,the obtained velocity in the axial direction at themeasurement probe points plotted and presented inFig. 9.At the sampled segment of 0.03m towards the exit

direction, the design records a slight reduction inaverage velocity and dispersion rate from thebenchmark study. Design 1 records a 6.19m/saverage velocity at the probe points showing a 5.7%reduction from the benchmark value. The disper-sion rate also shows a 2.7% decrease from the 4.5value of the benchmark. The system generate

Fig. 8. Model Set 3 Axial flow velocity magnitude at probe point withbenchmark.

154 JOURNAL OF SUSTAINABLE MINING 2021;XX:144e156

RESEARCH

ARTIC

LE

0.973m3/s flow rate, showing 0.043m3/s flow ratedecline than the benchmark study. Regarding theefficiency of the system, the design improved thebenchmark study efficiency of 84.7 to 86% showingmoderate improvement.Whereas, design 2 developed an approximate

value of 1m3/s throughput with a maximum flowvelocity of 8.154m/s recorded at the 12th probepoint. The flow rate difference occurred betweendesign 2 and the benchmark value is very small witha different percentage of 1.5%. The average flowvelocity of the benchmark study within the probepoint accounts 6.5 m/s while design 1 records6.41m/s, a 0.09m/s difference. Even though thedesign shows a decrease in average velocity, thedesign recorded a slightly higher flow dispersionrate than the benchmark, indicating relativelydisturbed flow as shown in Table 3. Design 2recorded a velocity value that surpasses the bench-marked value at probe points 3, 5, and 10 as shownin Fig. 9 attributed to its higher flow dispersion. Thesystem recorded efficiency of 85.5%, a 0.8%improvement from the benchmark study. However,design 2 records a lower efficiency from design 1 inthe same model set. In conclusion, a tapered modelset record flow uniformity and efficiency improve-ment as tapering percentage increase. However, inthe twisted/taper model set the result was observedto be reversed where a higher tapering percentageled to flow dispersion increment and efficiencyreduction.

4. Conclusion

In conclusion, through the comparison study, thefour-digit NACA airfoil series was found to deliver aneeded higher throughput when compared to theseven-digit family. In constant chord blade design,as the AoA increases, the flow rate of the system isalso observed to increase. However, as the AoA

increases more than the optimal amount, the systemexhibits an undesired flow pattern affecting systemefficiency. To overcome the problem, three differentoptimization strategies were applied and eachoptimization method was studied in case sections.The applied modification methods were designedwith two scales in a model set. The scale differencehelps to investigate improvement trends onapplying optimization strategies. All optimizationmethods were able to eliminate the undesired flowbehavior observed in the constant chord blade type.Yet the level of efficiency, flow uniformity, andthroughput within each optimization method wasfound to be substantially different. The taperedblade type was able to cut down the sudden velocityuptake recorded at the last 4 probe points in thebenchmarked study introducing normalcy in thesystem. However, the design comes within surfacereduction, in a way it could be good for reducing fanwheel load, but also result in throughput decline. Byobserving the flow pattern of both designs in thetapered model set, it can be forecasted that as thetapering percentage increases, better flow distribu-tion and higher efficiency could be achieved withthe flow rate trade-off. The same conclusion can beextruded for a skewed blade model set where ahigher twist angle could lead to better flow distri-bution and efficiency improvement, yet the methoddoes not result in significant surface area reduction.The skewed model set was able to achieve thelowest flow velocity dispersion and higher efficiencyin all designs investigated for this project. In thetapered/twisted model set, both tapered and twistedstrategies were applied and the surface areareduction was the same as the tapered blade set.The strategy was found to work in reverse from atapered model set where when the tapering per-centage increased the system efficiency drops.

Conflicts of interest

None declared.

Ethical statement

The authors state that the research was conductedaccording to ethical standards.

Funding body

None.

Acknowledgments

The author would like to thank SimScale GmbHfor providing the tool used in the research. The

Fig. 9. Model set 4 axial flow velocity magnitudes at probe point withbenchmark.

JOURNAL OF SUSTAINABLE MINING 2021;XX:144e156 155

RESEARCH

ARTIC

LE

author has also received editorial assistance fromCarolina Giavedoni and Jousef Murad fromSimScale.

Reference

[1] De la Vergne J. Hard rock miner's handbook. Edmonton:Stantec Consulting; 2008 [Online]. Available: http://www.stantec.com/content/dam/stantec/files/PDFAssets/2014/Hard%20Rock%20Miner%27s%20Handbook%20Edition%205_3.pdf. [Accessed 12 March 2021].

[2] Sterling D, Johnson G. Identifying opportunities to reducethe consumption of energy across mining and processingplants. 2010.

[3] Sharma RN. Economics of mine ventilation. Minetech 2002;23.

[4] Papar R, Szady A, Huffer WD, Martin V, McKane A.Increasing energy efficiency of mine ventilation systems.1999.

[5] Demirel N. Energy-Efficient Mine Ventilation Practices. En-ergy Effic Miner Ind 2018:287e99.

[6] Eck B. Design and operation of centrifugal, axial-flow andcross-flow fans. Oxford: Pergamon Press; 1973.

[7] Dugao Z, Jiang Z. Optimization design of an axial-flow fanused for mining local-ventilation. Comput Ind Eng 1996;31(3e4):691e6.

[8] Panigrahi DC, Mishra DP. CFD simulations for the selectionof an appropriate blade profile for improving energy effi-ciency in axial flow mine ventilation fans. J Sustain Min 2014;13(1):15e21.

[9] Petrov NN, Popov NA. Ways of improving economy andreliability of mine ventilation. J Min Sci 2004;40(5):531e6.

[10] Kazakov BP, Shalimov AV, Kiryakov AS. Energy-savingmine ventilation. J Min Sci 2013;49(3):475e81.

[11] Vutukuri VS, Lama RD. Environmental Engineering inMines. Cambridge University Press; 1986.

[12] Zadravec M, Basic S, Hribersek M. The influence of rotatingdomain size in a rotating frame of reference approach for

simulation of rotating impeller in a mixing vessel. J Eng SciTechnol 2007;2(2):126e38.

[13] Tabib M, Rasheed A, Siddiqui MS, Kvamsdal T. A full-scale3D Vs 2.5 D Vs 2D analysis of flow pattern and forces for anindustrial-scale 5MW NREL reference wind-turbine. EnergyProcedia 2017;137:477e86.

[14] Izadi MJ, Falahat A. Effect of blade angle of attack and hub totip ratio on mass flow rate in an axial fan at a fixed rotationalspeed. In: Fluids Engineering Division Summer Meeting.48401; 2008. p. 903e13.

[15] Falahat A. Numerical and experimental optimization of flowcoefficient in tubeaxial fan. Int J Multidiscip Sci Eng. 2011;2(5):24e9.

[16] Butterfield CP, Musial WP, Simms DA. Combined experi-ment phase 1. Final report. Golden, CO (United States):National Renewable Energy Lab; 1992.

[17] Xingshuang ZSWJW. Study on aerodynamic performance ofaxial flow fan with bowed blade based on Bezier function. JHuazhong Univ Sci Technol Nat Sci Ed. 2013;3.

[18] Kr€omer F, Müller J, Becker S. Investigation of aeroacousticproperties of low-pressure axial fans with different bladestacking. AIAA J 2018;56(4):1507e18.

[19] Ouyang H, Li Y, Du Z-H, Zhong F-Y. Experimental study onaerodynamic and aero-acoustic performance of low pressureaxial flow fan with circumferential skewed blades. J AerospPower 2006;21(4):668e74.

[20] Cai N, Li D, Zhong F. Optimum design and experiment onskewed-swept rotating blades. J Shanghai Jiaotong Univ1997;vol. 31:81e5.

[21] Yang L, Ouyang H, Du Z. Experimental research on aero-dynamic performance and exit flow field of low pressureaxial flow fan with circumferential skewed blades. J Hydro-dyn Ser B 2007;19(5):579e86.

[22] Vad J, Kwedikha ARA, Horv�ath C, Balcz�o M, Loh�asz MM,R�egert T. Aerodynamic effects of forward blade skew in axialflow rotors of controlled vortex design. Proc Inst Mech EngPart J Power Energy 2007;221(7):1011e23.

[23] Giguere P, Selig MS. Design of a tapered and twisted bladefor the NREL combined experiment rotor. Golden, CO (US):” National Renewable Energy Lab.; 1999.

156 JOURNAL OF SUSTAINABLE MINING 2021;XX:144e156

RESEARCH

ARTIC

LE