Strengthening and Closing Cracks for Existing Reinforced ...1).pdf · Strengthening and Closing Cracks for Existing Reinforced Concrete Girders Using Externally Post-Tensioned Tendons

Journal of Engineering Volume 19 November 2013 Number 11

1354

Strengthening and Closing Cracks for Existing Reinforced Concrete Girders Using Externally Post-Tensioned Tendons

Nazar K. Ali Oukaili (1) and Ihab Nabeel Issa Al-Shawi (2)

[email protected]. Dr., College of Engineering, University of Baghdad, Iraq. email: 1 2 M.Sc. Student, College of Engineering, University of Baghdad, Iraq. email: [email protected]

ABSTRACT

This research is devoted to study the strengthening technique for the existing reinforced concrete beams us-ing external post-tensioning. An analytical methodology is proposed to predict the value of the effective pre-stress force for the external tendons required to close cracks in existing beams. The external prestressing force required to close cracks in existing members is only a part from the total strengthening force.

A computer program created by Oukaili (1997) and developed by Alhawwassi (2008) to evaluate curvature and deflection for reinforced concrete beams or internally prestressed concrete beams is modified to evaluate the deflection and the stress of the external tendons for the externally strengthened beams using Matlab 7.0.

The analytical investigation is implemented on three ideal reinforced concrete beam models, each model is considered to be strengthened using three types of external tendon profile (straight, draped and double draped), where each type of tendon profile is analyzed separately. No comparisons were made with analyti-cal or experimental investigations, because no publications for this kind of studies were found.

KEYWORDS: Strengthening, Post-Tensioning, Draped Tendons, Deflection, Curvature and Rein-forced Concrete.

االجهاد المسبق بأستعمال المتواجدة الروافد الخرسانية المسلحةتقوية و غلق الشقوق في الخارجي

)2(و ايهاب نبيل عيسى الشاوي ) 1(كامل علي العقيلي نزار

طالب ماجستير، كلية الهندسة، جامعة بغداد، العراق. ) 2(أستاذ دكتور، كلية الهندسة، جامعة بغداد، العراق. )1(

الخالصةالمتواجدة باستعمال االجهاد المسبق الخارجي. حيث تم اقتراح ان الغرض من هذا البحث هودراسة تقنية تقوية الروافد الخرسانية المسلحة

للحديد طريقة تحليلية لحساب قوة االجهاد المسبق للحديد الخارجي الكافية لغلق الشقوق في الروافد الخرسانية المنشأة. ان قوة االجهاد المسبق ية.الخارجي المطلوبة لغلق الشقوق هي جزء من القوة الكلية المطلوبة للتقو

(الهواسي) والذي يقوم بحساب الهطول و التقوس في الروافد الخرسانية المسلحة طورته برنامج حسابي اقترحه (العقيلي) وعديل تم تة ية المسبقاالعتيادية او الروافد الخرسانية المسبقة االجهاد داخليا ليقوم بحساب هطول الروافد الخرسانية المقواة باألوتاد الفوالذية الخارج

االجهاد و حساب اجهاد الحديد الخارجي.

حديد تم تطبيق الطريقة الرياضية المقترحة على ثالثة نماذج تحليلية لثالثة روافد خرسانية مسلحة ليتم تقويتها باستعمال ثالثة اشكال من اللى حدى. حيث لم يتم اجراء مقارنة المسبق االجهاد خارجيا (مستقيم ، منحرف بموقع واحد، منحرف بموقعين). حيث تم تحليل كل شكل ع

نتائج الطريقة المقترحة مع البيانات المختبرية و ذلك لعدم توفر ابحاث منشوره بهذا الخصوص.

ةذات الشكل المنحرف، الهطول، التقوس، الخرسانة المسلح التقوية، االجهاد المسبق، االوتارالكلمات الرئيسية:

mailto:[email protected]

Nazar K. Ali Oukaili Strengthening and Closing Cracks for Existing Reinforced Ihab Nabeel Issa Al-Shawi Concrete Girders Using Externally Post-Tensioned Tendons

1355

INTRODUCTION

During the service life, concrete members may severe different types of deteriorated conditions in addition to the progressive structural aging lead to extreme cracking and deflection that may affect their performance for the rest of their lifetime. Serviceability requirements might be changed and required higher loading capacity than designed due to many possible reasons, like: increasing traffic volume in bridges and culverts or changing the type of floor occupancy in a building.

Cracking is a usual behavior in concrete structures in service, due to low tensile strength of concrete, therefore the internal steel reinforcement have the full responsibility to resist the tensile forces. Since, the cracking behavior is unavoidable; the steel reinforcement will be exposed to the exterior environment within the serviceability lifetime.

Accordingly, two alternatives can be considered: either to demolish and replace the existing mem-bers or to rehabilitate and restore the strength of the structures. The latter alternative, which is pre-ferred economically, involves either strengthening or repairing the member.

The external post-tensioning is an attractive tech-nique for strengthening existing structures. In which it introduce many advantages, like: in-creasing the load carrying capacity, improving serviceability performance and ease of installation and maintenance.

CRACKING BEHAVIOUR OF REIN-FORCED CONCRETE BEAMS Cracking is a disadvantageous phenomenon in concrete structures. Cracks formed in the concrete tension zone when the tensile stress exceeds the low tensile concrete strength. Cracking causes reduction in stiffness of the member that leads to larger curvature value at crack locations (with re-spect to uncracked location within the constant moment region), and exposing the steel rein-forcement to the exterior environment. Cracks in concrete beams can be classified into three types:

1- Normal-Flexural cracks: formed due to the effect of the flexural tensile stresses, and most-ly located within the middle third of the span. Flexural cracks are formed in the tensile zone and have a wedge shape, with a maximum width at the extreme bottom fiber and zero width at the tip of the crack.

2- Inclined shear-flexural cracks: occurs when critical combination of flexural and shear stresses develops near the top of a flexural crack.

3- Web-Shear cracks: formed when no flexural cracks are formed and occurred due to shear stresses being higher than flexural stresses in the web portion of the member at region near support. Thinner web encourage this type of cracking.

The formation of each type of cracking depends on the relative stiffness of the member and type of loading. The type of the crack that under consid-eration in this study is the flexural crack and the other types are not discussed. Crack Width Several formulas for prediction of crack width were developed by various investigators. These formulas contains miscellaneous set of variables, where no general agreement among these investi-gations on the significant variables affecting the crack width. The collected expressions are:

1. Clark (1956) :

( )

+−

−= 8156.1

*1029.1 6max

ss

ss

s fddh

ρρφω (1)

Where h : Total depth of the member; sd : Depth of the tensile non-prestressed steel reinforcement measured from extreme top fiber; sρ : Reinforce-ment ratio for the tension steel; sf : Steel stress for the tensile reinforcement; φ :Bar diameter

2. Chi and Kirestein (1958) :

−=τφ

τφω 4385max s

s

fE

(2)

Whereτ : Coefficient related to the assumed effec-tive concrete area in tension to the area of a single bar; sE : Modulus of elasticity of steel

3. Kaar and Mattock (1963) :

45max 10*57.1 ts Af−=ω (3)

Where tA : Effective concrete area in tension


1356

4. Gergely and Lutz (1968) :

36max .10*02.11 ecs Adfβω −= (4)

Where β : Ratio of distances from the extreme bottom fiber and from the steel centroid to the center gravity of the section; cd : Thickness of concrete cover measured from the extreme bottom fiber to the center of bar closest to that fiber; eA : Concrete area surrounding one bar, equal to total effective tension area of concrete surrounding re-inforcement and having same centriod divided by number of bars

5. Venkateswarlu and Gesund (1972) :

( )( )( )ss

s

fnf

−+−

=−

6621146210*4.2 5

max ρφω (5)

Where n : Modular ratio

6. Frotch (1999) :

2

22

max 22

+=

sdEf

cs

s βω (6)

Where s : Maximum bar spacing

7. Beeby (1979) :

( )( )

−−

+=

chda

a

clcr

mycr

21

3max

εω (7)

Where myε : Mean strain of concrete at the se-

lected level; c : Depth of the neutral axis; cld : Clear concrete cover; cra : Distance from the bar surface to the point where the crack width is cal-culated.

A simple comparison is made for the collected expressions of crack width and plotted with re-spect to the experimental data taken form (Hong el a. (2008)) for beam (F30-2D19-10), as shown in Fig. (1). the expression which was proposed by Gergely and Lutz (1968) shows good agreement with the experimental data.

STRESS-STRAIN RELATIONSHIP The model of Korpenko (1986) for concrete and steel is adopted in this study; the model takes the following form:

mmmm E νεσ = (8) In which ( mmE ν ) represent the secant modulus of elasticity at the nonlinear portion of the stress-strain curve, while ( mν ) equals to (1) in the linear portion and less than (1) in the nonlinear portion of the stress-strain curve. Korpenko derived the following expression for ( )mν :

( ) ( )[ ]

( ) 0ˆ

ˆ~ˆ2

ˆˆˆ

~ˆ1

2

1

2222

22

=

−−−

−−+

⋅−+

m

mommmm

momm

mmomm

e

e

νννε

νν

νννν

εννν

(9)

In which

m

peakm Eoε

σν =ˆ ;

oεεε m

m =~ ; mm ee 12 1−=

Where subscript (m) = refers to material.

peakσ : Ultimate strength of material.

oε : Material strain corresponding to ( )peakσ .

mE : Initial modulus of elasticity.

mε~ : Material strain level.

mm ee 21 , : Factors depend on material type.

oν : Factor depends on the stress level.

mν̂ : It is the value of ( )mν which correspond to the stress ( )peakσ . Fig. (2) shows the stress strain diagram for con-crete as suggested by Korpenko. MOMENT-CURVATURE MODEL An iteration method for analysis which adopted by Oukaili (1997) is programmed using Matlab 7.0. This method requires section meshing. The Cartesian coordinates for a cross-section is shown in Fig. (3), in which positive signs convention are shown for each force. The method is based on the following assumptions:

1. Strain of the concrete and reinforcement is proportional to the distance from the neutral axis in accordance to Bernoulli’s hypothesis “Cross-section shall remain plane after bending”.


1357

2. Shear and torsion stresses are ignored.

3. The behavior of steel and concrete is con-sidered to follow Korpenko's model, where all stresses in concrete and steel are related to se-cant modulus of elasticity.

4. Perfect bond exists between concrete and the internal reinforcement, the strain of the ordi-nary reinforcement due to external load is com-patible with the strain of the concrete fiber exists at the center gravity of that reinforcement. Also, the strain increment of the bonded prestressed steel is equal to the concrete fiber strain which exists at its center gravity.

5. Concrete is divided into a group of small cells having sizes related to the required accura-cy conditions. The individual steel reinforcement will not be meshed. Thus the reinforcement ele-ment acts as a system of linear elements exposed to axial compression or tension.

6. External tendons are not incorporated in this analysis.

the general relation between forces vector |F| F strain vector ε and stiffness matrix [ ]C can

be expressed as follows:

[ ] ε*CF = (10)

This expression can be detailed as follows:

y

x

o

y

x

KK

CCCCCCCCC

MMN ε

=

)3,3()2,3()1,3(

)3,2()2,2()1,2(

)3,1()2,1()1,1(

(11)

The elements of the stiffness matrix can be ex-pressed as follows:

( ) ∑∑==

+=p

isisici

r

ici AEAEC

111,1

(12)

( ) ∑∑==

+=p

isisisicici

r

ici yAEyAEC

112,1

(13)

( ) ∑∑==

+=p

isisisicici

r

ici xAExAEC

113,1

(14)

( ) ∑∑==

+=p

isisisicici

r

ici yAEyAEC

1

22

12,2

(15)

( ) ∑∑==

+=p

isisisicici

r

ici xAExAEC

1

22

13,3

(16)

( ) ∑∑==

+=p

isisisisicicici

r

ici yxAEyxAEC

113,2

(17)

The direct iteration method proposed by Cooke (1981) is adopted to solve a non-linear problem for determination of the strain vector for a cross-section subjected to known forces.

The evaluation of each element of the stiffness matrix [ ]C is dependent on secant modulus of elasticity in which it depends on the unknown value of strain in each material. The strain value depends on the strain vector as shown in the fol-lowing equation:

mymxam xKyK ++= εε (18) Where aε : Axial strain; xK : Curvature of the member longitudinal axis in OYZ plane; yK : Curvature of the member longitudinal axis in OYX plane. In other word, the matrix [ ]C is func-tion of strain vector. Accordingly eq. (10) can be rewritten in the following form:

( ) FC *1−

= εε (19)

In the first iteration all materials in section shall be assumed to be linear and the value of strain vector equal to zero. So the stiffness matrix can be calculated easily. Eq. (18) is used then to evaluate the strain vector resulted from the first iteration. For further iterations, the stiffness matrix will be updated according to the strain vector calculated from the previous iteration as shown below:

( ) FCii

*1

1

−

−= εε (20)

Where subscript ( )i : iteration number The procedure is repeated until the convergence of the load vector satisfies the following condition:

( )( ) FFFii

λ=−−1

(21)

Where Fλ : Convergence limit for the force vector which is considered a very small value LOAD-DEFLECTION MODEL The method of Newmark (1943) is adopted to determine deflection at each node from curvature


1358

values at these nodes. This method is based on the conjugate beam method, in which the beam is subjected to a fictitious load equal to (M/EI) which represents the curvature distribution along the beam. Hence the moment and shear values at a location in the conjugate beam represent the slope and deflection; respectively; values at that loca-tion in the actual beam. The model based on the following assumptions:

1. The beam is considered prismatic in which the cross-section geometry is same along the beam length. But the change of eccentricity of the internal prestressed tendon will be permit-ted, because the change in coordinate for the tendons between different locations has a mar-ginal effect on member stiffness.

2. The applied load on the beam is consid-ered symmetric.

3. The end supports of the beam are assumed to be simply supports only.

4. Sign convention for deflection is positive for downward deflection and negative for up-ward camber.

5. The beam shall be divided into segments of equal length, as shown on Fig. (3).

The (M/EI) curve between two nodes can be rep-resented by second order polynomials (parabola). Accordingly, the following equations can be used for evaluating deflection depending on curvature values determined at each node.

( ) ( )( ) xi

j ji ∆=∆ ∑ =*

2θ (22)

( ) ( ) ( ) 2/1

1l

c

ijlj KK +

= ∑

−

+=

θ (23)

( ) ( ) ( ) ( )( )11 1012 +− ++∆

= llll KKKxK (24)

Where subscript ( )i referes to node location

( )i∆ : deflection; ( )jθ : rotation; K : curvature PROPOSED ANALYTICAL METHODOLO-GY FOR CLOSING CRACKS IN EXISTING REINFORCED CONCRETE BEAMS The main goal is to predict the value of the effec-tive prestress force for the external tendons re-

quired to close all cracks in an existing beam. The major given data are the maximum crack width, crack spacing and number of cracks, while the corresponding output is the external prestressing force.

Assumptions 1. The beam cross section and the external tendons

are considered to be symmetric about a principle axis of the member that is parallel to member’s depth, creating no transverse curvature about that axis.

2. The external tendons are not incorporated in the strain compatibility conditions, and they are not compatible with surrounding concrete.

3. The stress increase in the external tendons be-yond the effective prestress is member depend-ent rather than section dependence.

4. The stress and strain in the external tendons are uniform along tendons’ length.

5. An idealized beam model is adopted in which. the deflection equals to ( )jcr∆ when the beam is

subjected to an externally uniform distributed load ( )jWcr .

6. At the moment when the deflection of the ideal-ized beam attains ( )jcr∆ , the tensile strain at the

extreme bottom fiber attains ( )rε , where ( )rε the strain is corresponds to the modulus of rupture of concrete ( )rf .

Method of Analysis Based on the assumptions mentioned above, the analysis will be performed for four separate mod-els:

1- Existing Beam Model: Based on the measured crack width ( )maxω and eq. (4), the stress of the

steel at the extreme bottom layer ( )sf can be de-termined. This model deals with the calculated value of ( )sf and by using the analytical mo-ment-curvature model, the analytical uniformly distributed load ( )exW can be evaluated. Accord-

ingly, the load ( )exW is that load which produces a maximum crack width at mid span equal to ( )maxω .


1359

Also, the depth of the cracks ( )jYcr can be esti-

mated from the analysis. In which subscript ( )j refers to crack number as shown in Fig. (7).

Fig. (9) shows the flow chart for the programming procedure to determine ( )exW .

2- Idealized Beam Model: Based the analytical moment-curvature model, the beam is subjected to external incremental load to reach ( )jWcr for

each crack. The external uniformly distributed load ( )jWcr produces curvature ( )iK and de-

flection ( )jcr∆ at location ( )j at this moment,

the tensile strain at the extreme bottom fiber at-tains ( )rε .

This model deals with determining ( )jcr∆ at loca-

tion ( )jXcr from the support as shown in Fig. (8). Fig. (10) shows the flow chart for the program-ming procedure to determine ( )jcr∆ .

3- Strengthening to close cracks: the beam is subjected to combination of uniformly distribut-ed load ( )exW and the incremental external pre-

stressing force ( )jFcr∆ . The external prestress-

ing force will be increased to attain ( )jFcr . The

value of ( )jFcr is the one that close crack num-

ber ( )j when the following condition is achieved:

( ) crcrs jj λ≤∆−∆

Where js∆ : Deflection value resulted from a

combination effect of the load ( )exW and the pre-stressing force ( )jFcr for crack ( )j .

crλ : Convergence limit for the deflection value (taken as 0.005) It is worth to mention that, the depth of the cracks ( )jYcr calculated from the existing beam model were used in this model to express the real stiff-ness of the beam by excluding concrete cells along the depth of cracks. Fig. (11) shows the flow chart for the programming procedure to de-termine ( )jFcr .

4- Optimum Strengthening: This model is simi-lar to model (3), in which the beam is subjected to combination of uniformly distributed load ( )exW and the incremental external prestressing

force ( )stF∆ . The external prestressing force will

be increased to attain ( )stF . The value of ( )stF is reach when the following condition is achieved: ( ) rrjct λεε ≤−=1,

Where 1, =jctε : Strain of the extreme top fiber at mid span. rλ : Convergence limit for the strain value (taken as 0.005). It is worth to mention that, the external prestress-ing force ( )stF used for optimum strengthening is larger than the calculated external prestressing forces ( )jFcr to close cracks. Numerical Applications on the Proposed Meth-odology The proposed methodology is implemented on three ideal beam models reinforced with nonpre-stressed reinforcement under service load, to be strengthened by external prestressed strands to close existing cracks.Three types of external pro-file for strengthening (straight, draped (one de-vaitor at mid span) and draped (two deviators at one-third span distance from each support)). The analytical study includes determining the incre-ment percentage in load carrying capacity for the models after strengthening for each chosen pro-file. Models are:

1. Model-01: the geometric and material proper-ties for this beam are shown in Fig. (4) and Table (1), respectively. The cracking is de-scribed in Table (3).The properties of the ex-ternal prestressing reinforcement used for strengthening is shown in Table (2).The out-put results for the beam analysis before and after strengthening is shown in Fig. (12).the increment in load carrying capacity after strengthening is shown in Table (4).

2. Model-02: the geometric and material proper-ties for this beam are shown in Fig. (5) and Table (5), respectively. The cracking is de-scribed in Table (7).The properties of the ex-ternal prestressing reinforcement used for strengthening is shown in Table (6).The out-put results for the beam analysis before and


1360

after strengthening is shown in Fig. (13).the increment in load carrying capacity after strengthening is shown in Table (8).

3. Model-03: the geometric and material proper-ties for this beam are shown in Fig. (6) and Table (9), respectively. The cracking is de-scribed in Table (11).The properties of the ex-ternal prestressing reinforcement used for strengthening is shown in Table (10).The out-put results for the beam analysis before and after strengthening is shown in Fig. (14).the increment in load carrying capacity after strengthening is shown in Table (12).

CONCLUSION 1. External prestressing is a very effective tech-

nique for strengthening existing concrete mem-bers, in which it allows to increase the load car-rying capacity of the member to (111%) for straight tendon profile, (104%) for draped ten-don profile (one deviator at mid span) and (103%) for draped tendon profile (two deviators at one third distance from the support).

2. The calculated external prestressing force that is required to close cracks for the existing con-crete members is found to be ((0.41) for straight tendon profile, (0.48) for draped tendon profile (one deviator at mid span) and (0.46) for draped tendon profile (two deviators at one third dis-tance from the support)) of the calculated ex-ternal prestressing force required for optimum strengthening.

3. Strengthening using the straight tendon profile requires higher prestressing force than the draped tendon profile by about (62%).

4. The empirical formula of Gergely and lutz (1968) for calculating the maximum crack width shows a very good agreement with the experimental data than others.

REFERENCE • ACI Committee 318, “Building Code require-

ments for Reinforced Concrete (ACI 318-2008),” American Concrete Institute.

• Alhawwassi I. F., “Short term deflection of Ordinary, partially prestressed and GFRB Bars

reinforced Concrete Beams” M.Sc thesis, Baghdad university, Iraq, 2008.

• Beeby, A. W., “The Prediction of Crack width in Hardened Concrete Cracking” Journal of Structural Engineering, Vol. 57, No. 1, pp 9-17, 1979.

• Chi, M. and Kirestein, A. F. , “Flexural Cracks in reinforced Concrete Beams” ACI Structural journal, Proceeding, Vol. 54, No. 10, pp 865-878, 1958.

• Clark, A. P., “Cracking in reinforced Concrete Flexural members” ACI Structural journal, Proceeding, Vol. 52, No. 8, pp 851-862, 1956.

• Frotch R., “Another Look at Cracking Crack Control in Reinforced Conrete” ACI Structural journal, Vol.96, No. 3, pp437-442, 1999.

• Gergely, P. and lutz, L. A., “Maximum crack width in Reinforced Concrete, Causes, Mecha-nism, and Control of Cracking Concrete”, SP-20, ACI Structural journal, pp87-177, 1968.

• Hong Sung Nam, Han Kyoung Bong, Kim Tae Wan, Beak Kyeong Seok , Park Sun Kye, and Ko Won Jun, “ Estimation of Flexural Crack Width in Reinforced Concrete Mem-bers” ,The 3rd ACF international conference – ACF/VCA 2008.

• Kaar , P. H., and Mattock, A. H. ” High Strength Bar as Concrete Reinforcement ’’, Part 4, Control of Cracking, Journal of Port-land Cement Association Research and De-velopment Laboratories , Vol. 7, No.1, pp 42-5.

• Korpenko, N. I., Mukhamediev, T.A. and Pe-trov, A. N., “The Initial and Transformed Stress - Strain Diagrams of Steel and Concrete. “ Special Publication, Stress-Strain Condition for Reinforced Concrete Construction, Rein-forced Concrete Research Center, Moscow, 7 – 25, 1986.

• Newmark, N. M., "Numerical Procedure for Computing Deflections, Moments, and Buck-ling Loads", Transactions, ASCE, 1-8, 1943.

• Oukaili, Nazar K. Ali, “Strength of Partially Prestressed Concrete Elements with Mixed Re-


1361

inforcement by Highly Strength Strands and Steel Bars”, PH.D. Thesis, Moscow Civil En-gineering University, Moscow, 1991.

• Oukaili, Nazar K. Ali, “Moment Capacity and Strength of Reinforced Concrete Members Us-ing Stress-Strain Diagrams of Concrete and Steel”, Journal of King Saud University, Vol. 10, pp. 23-44, 1997.

• Venkateswarlu, B. and Gesund, H., “Crack-ing and Bond Slip in Concrete Beams” Journal of Structural Engineering, ASCE, Vol. 98, No. ST11, pp 2663-2885, 1972.

• Cook R.D., "Concept and Application of Finite Element Analysis", Second Edition, John Wiley and Sons, New York, 1981.

NOTATION

ciA : Distance from the center of gravity to the concrete cell (i).

eA : Concrete area surrounding one bar, equal to total effective tension area of concrete surround-ing reinforcement and having same centriod di-vided by number of bars.

siA : Area of the internal reinforcement for bar (i).

tA : Effective concrete area in tension.

cra : Distance from the bar surface to the point where the crack width is calculated. c : Depth of the neutral axis.

sd : Depth of the tensile non-prestressed steel re-inforcement measured from extreme top fiber.

cd : Thickness of concrete cover measured from the extreme bottom fiber to the center of bar clos-est to that fiber.

cld : Clear concrete cover.

mE : Modulus of elasticity of material (m).

sE : Modulus of elasticity of steel.

E : Secant modulus of elasticity.

mm ee 21 , : Factors depend on material type.

jFcr : External prestressing force for closing

crack ( )j .

stF : External prestressing force for optimum strengthening.

sf : Steel stress for the tensile reinforcement.

rf : Modulus of rupture of the concrete. h : Total depth of the member.

xK : Curvature of the member longitudinal axis in OYZ plane.

yK : Curvature of the member longitudinal axis in

OYX plane.

xM : Bending moment about x-axis.

yM : Bending moment about y-axis.

N : Normal force. n : Modular ratio. s : Maximum bar spacing.

cici yx , : Distance to the center of gravity of con-

crete cell ( )i .

sisi yx , : Distance to the center of gravity of non-

prestrssed steel for bar ( )i .

exW : Applied uniform load for beam in service.

jWcr : Idealized cracking load.

jYcr : Crack depth for crack ( )i .

β : Ratio of distances from the extreme bottom fiber and from the steel centroid to the center gravity of the section. ∆ : Deflection. θ : Rotation.

jcr∆ : Deflection at cracking for crack ( )i .

aε : Axial strain.

mε : Strain in material.

sε : Strain of the tensile reinforcement.

myε : mean strain of concrete at the selected level.

rε : Strain correspond to ( )rf .

ctε : Strain in the concrete top fiber at ultimate

oε : Material strain corresponding to ( )peakσ .

mε~ : Material strain level.

mν : Material elastic modulus factor that expresses the ratio of elastic strains to the total strains.

oν : Factor depends on the stress level.


1362

mν̂ : It is the value of ( )mν which correspond to

the stress ( )peakσ .

sρ : Reinforcement ratio for the tension steel.

mσ : Stress in material.

peakσ : Ultimate strength of material.

φ : Bar diameter

maxω : Maximum crack width at extreme bottom fiber. τ : Coefficient related to the assumed effective concrete area in tension to the area of a single bar. F : Force vector.

[ ]C : Stiffness matrix.

ε : Strain vector.


1363

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

10 20 30 40 50

Clark

Chi and Kirestien

Kaar and mattock

Gergly and Lutz

Experimental

Frotch

Venkateswarlu andGesund

Beeby

Moment (KN.M)

Crac

k w

idth

(mm

)

Figure (1) Mid span moment - Maximum crack width (Hong el a. (2008) - beam (F30-2D19-10))

Figure (2) Stress-Strain diagram for concrete

(Korpeko (1986))

cε

peakσ

cσ cE

ccνE


1364

Figure (3) Sign Convention for internal forces

Figure (4) Geometric properties [Model-01]



Y

X X

Section Meshing

Non-Prestressed steel

Prestressed steel

Member division

Z

Z N

My

Mx O

Y Y


1365

Xcrj=4

Figure (7) Existing beam model

Figure (8) Idealized beam model

∆crj=4

∆crj=3 Xcrj=3

Xcrj=2 ∆crj=2

Xcrj=1 ∆crj=1

Xcrj=1

Xcrj=2

Xcrj=3

Xcrj=4

j=1 j=2 j=3 j=4


1366

Start iterations with j=1

Inputs Initial value for applied load (W), load increment (∆W), section geometry, material properties, profile and effective prestress for internal prestressed tendons, tolerances in strain vector(λ|ε|) value and applied load(λW), and maximum number of iterations for strain vector loop (Nt |ε|)

and for load loop (NtW) and 00=ε ; 1,, , =psisici ννν

Crack width at mid span is given (wcr(l/2))

Calculate steel stress at mid span of the extreme bottom layer (fsb

*)

( ) ελεε <−−1jj

Calculate strain for each element

)(1−j

εε As per eq. (18)

In which steel strain at the extreme bottom layer (εs

*) is determined

Calculate ,, sici νν for each element as per korpenko model

Calculate matrix ( )(EC j ) As per eq. (12 through 17)

FC jj

1−=ε

jj

εε =−1

WWWW λ<∆−− )( K=1

K=1

||Nt ε

(Wex) is determined

Calculate force vector F at mid span

NO

NO

YES

Calculate sν for steel at the extreme bottom layer

NO

YES

NO

YES

Calculate steel stress at the extreme bottom layer (fsb)

fsb*> fsb

Figure (9) Flow chart for the proposed existing beam model

|W|=

|W|-∆W

|∆W|=|∆W|

|∆W|=|∆W|/2

|W

|=|W

|-∆W


1367

Inputs Initial value for applied load (W), load increment (∆W), section geometry, material properties,

profile and effective prestress for internal prestressed tendons, tolerances in strain vector(λ|ε|) val-ue and applied load(λW), and maximum number of iterations for strain vector loop (Nt |ε|) and for

load loop (NtW) , Number of cracks (Ncr) , crack spacing (Scr) 00=ε ; 1,, , =psisici ννν

( ) ελεε <−−1jj

Calculate strain for each element

)(1−j

εε As per eq. (18)

Calculate ,, sici νν for each element as per korpenko model

Calculate matrix ( )(EC j ) As per equations (12 through 17)

FC jj

1−=ε

jj

εε =−1

WWWW λ<∆−− )( K=1

K=1

Calculate force vector F at mid span

NO

NO

YES

YES NO

YES

NO

YES

εcbot(i)> εr

Determine concrete strain at the extreme bottom fiber at location (i); (εcbot(i))

START ITERATION WITH j = 1

Determine crack locations and start numbering from mid span with i = 1

Calculate deflection ∆cr(i) at crack location (i)

Ncr

NtW

Nt|ϵ|

LOOP FOR EACH CRACK LOCATION START LOOP WITH i = 1

Figure (10) Flow chart for the proposed idealized beam model

|W|=

|W|-∆W

|∆W|=|∆W|

|∆W|=|∆W|/2

|W

|=|W

|-∆W


1368

`

Inputs Initial value for external prestressing force (P), load increment (∆P), section geometry, material properties, profile and effective prestress for internal prestressed tendons, tolerances in strain vector(λ|ε|) value and external prestressing force (λP), and maximum number of iterations for

strain vector loop (Nt |ε|) and for external prestressing force loop (NtP) , Number of cracks (Ncr) , crack spacing (Scr) and 0

0=ε ; 1,, , =psisici ννν

Calculate (Wex) as per Fig. (9)

Choose a profile for the external prestressing tendon

Calculate deflection ∆cr (i) for each crack location as per Fig. (10)

Calculate force vector F at crack location (i) As per (Wex) and (Pi)

LOOP FOR EACH CRACK LOCATION (i) START LOOP WITH i = 1

Calculate deflection values at each node

Determine deflection at crack location (i) (∆cr(i)`)

∆cr(i)` > ∆cr(i)

| ∆cr(i)` - ∆cr(i)| < λcr

K=1

NO

K=1

NO

∆Pi = |∆Pi|/2

Pi = P

i - |∆Pi |

P i =

Pi +

|∆P i

| ∆P

i = |∆

P i|/2

NO

YES

YES

YES The effective prestress forces are determined (Fcr = P ) Figure (11) Flow chart for the strengthened beam model


1369

(a) Existing beam model

(c) Ultimate state (strengthening – Type-01)

(b) Ultimate state (No strengthening)

(d) Ultimate state (strengthening – Type-02)

(e) Ultimate state (strengthening – Type-03)

Figure (12) Output results for beam model- Note : STRct stand for the strain of concrete at extreme top fiber


1370






Figure (13) Output results for beam model-02 Note : STRct stand for the strain of concrete at extreme top fiber


1371






Figure (14) Output results for beam model- Note : STRct stand for the strain of concrete at extreme top fiber


1372

Table (1) Material properties [Model-01]

Table (2) External tendon properties used for strengthening [Model-01]

Table (3) Cracking description in existing beam [Model-01]

Crack width at mid span (thousandth mm)

Crack Spacing (mm)

No of cracks

431 250 9

Table (4) Prestressing forces for different external prestressing profiles [Model-01]

Tendon profile

Prestressing force to close all

cracks (kN)

Prestressing force for optimum

strengthening (kN)

Ultimate load for the strengthened

beam (kN/m)

Increment percentage

in load carrying Capacity

(%)

Type

es (mm)

ed (mm)

No.

of d

evia

tors

xd (mm)

1 98.2 - - - 122.656 296.25 84.07 110.9

2 98.2 197.3 1 1500 85.937 165.87 77.1 93.43

3 98.2 197.3 2 1000 - 2000 81.250 165.51 77.09 93.4

Concrete Properties Ultimate Compres-

sive Strength

MPa

Modulus of Elasticity

MPa

Strain corresponds to

Ultimate Compres-sive Strain

Modulus of Rupture

MPa

30.00 25923.7 0.0020 0.003 3.41

Internal Steel Properties (Nonprstressed) Ultimate Ten-sile Strength

MPa

Yield Strength

MPa

Modulus of

Elasticity

MPa

Ultimate Ten-sile

Strain

Area

[ mm2 ]

Effective Depth

[ mm ] 420 280 200000 0.20 100 45 620 420 200000 0.125 339 255

External Prestressed Strand Properties Ultimate Tensile Strength MPa

Yield Strength

MPa

Modulus of

Elasticity

MPa

Ultimate Tensile Strain

Area

[ mm2 ]

Effective Prestress

MPa

Effective Depth

[ mm ]

1860 1625 180000 0.05 231 As per

prestressing force

As per profile


1373





Crack Spacing (mm)

No of cracks

475 214.3 11


Concrete Properties Ultimate

Compressive Strength

MPa


MPa


Ultimate Compressive

Strain

Modulus of Rupture

MPa

30.00 25923.7 0.0020 0.003 3.41

Internal Steel Properties (Nonprstressed) Ultimate Tensile Strength MPa

Yield Strength

MPa

Modulus of

Elasticity

MPa


Area

[ mm2 ]

Effective Depth

[ mm ] 420 280 200000 0.20 226 45 620 420 200000 0.125 565 355


Yield Strength

MPa

Modulus of

Elasticity

MPa


Area

[ mm2 ]

Effective Prestress

MPa

Effective Depth

[ mm ]

1860 1625 180000 0.05 462 As per

prestressing force

As per profile

Tendon profile

Prestressing force to close all

cracks (kN)

Prestressing force for optimum

strengthen-ing (kN)


beam (kN/m)



(%) Type

es (mm)

ed (mm)

No.

of d

evia

tors

xd (mm)

1 129.8 - - - 232.031 541.96 200.3 116.92 2 129.8 244.7 1 1500 170.703 343.33 192.91 108.93

3 129.8 244.7 2 1000 - 2000 164.453 342.3 192.58 108.56


1374





Crack Spacing (mm)

No of cracks

400 166.67 15


Tendon profile

Prestressing force to close all cracks (KN)

Prestressing force for op-

timum strengthening

(KN)


beam (KN/m)



(%) Type

es (mm)

ed (mm)

No.

of d

evia

tors

xd (mm)

1 148.94 - - - 293.750 732.75 276.21 107.72

2 148.94 273.4 1 1500 211.718 467.82 269.21 102.46

3 148.94 273.4 2 1000 - 2000 201.562 466.22 269.22 102.47

Concrete Properties Ultimate

Compressive Strength

MPa


MPa


Ultimate Compressive

Strain

Modulus of Rupture

MPa

30.00 25923.7 0.0020 0.003 3.41

Internal Steel Properties (Nonprstressed) Ultimate Tensile Strength MPa

Yield Strength

MPa

Modulus of

Elasticity

MPa


Area

[ mm2 ]

Effective Depth

[ mm ] 420 280 200000 0.20 339 45 620 420 200000 0.125 800 355


Yield Strength

MPa

Modulus of

Elasticity

MPa


Area

[ mm2 ]

Effective Prestress

[ Mpa ]

Effective Depth

[ mm ]

1860 1625 180000 0.05 616 As per

prestressing force

As per profile


1375

Effect of Velocity on Dissolved Oxygen Cathodic Polarization using a Rotating Cylinder Electrode

Prof. Dr. Qasim Jasim Muhammed Slaiman

Hala Muhammed Hussain Al – Nahrain University

[email protected] [email protected]

ABSTRACT The aim of the present work to study the effect of changing velocity (Reynold's number) on oxygen cathodic polarization using brass rotating cylinder electrode in 0.1, 0.3 and 0.5N NaCl solutions (PH = 7) at temperatures 40, 50 and 600C. Cathodic polarization experiments were conducted as a function of electrode rotational speed and concentration. KEY WORDS: Brass, oxygen, polarization, rotating cylinder electrode

حمد سليمانأ.د. قاسم جاسم م م.م. هالة محمد حسين

تأثير السرعة على األستقطاب الكاثودي لألوكسجين بأستخدام القطب الدوار

المقدمهدواره فـي محاليـل بـراص بأسـتخدام أسـطوانه الكاثودي لألوكسـجين ستقطابتهدف الدراسه الحاليه الى متابعة تأثير تغير سرعة الدوران على أأل

NaCl أجريـــت درجـــه مئويـــه . 60و 50و 40الحـــراره درجـــات ) عنـــد7مـــوالري (عنـــدما يكـــون األس الحامضـــي = 0.5و 0.3و 0.1بتراكيـــز . NaClسرعة القطب الدوار و تركيز الملح بداللة التجارب

مفتاح الكلمات: براص، أوكسجين ، األستقطاب، القطب الدوار



Qasim Jasim Muhammed Effect of Velocity on Dissolved Oxygen Cathodic Hala Muhammed Hussain Polarization using a Rotating Cylinder Electrode

1376

1. INTRODUCTION A brass alloy, 39%- Zn-Cu, has been employed to investigate the cathodic polarization of dissolved oxygen [3]. Polarization methods involve changing the potential of the working electrode and monitoring the current which is produced as a function of time or potential. For anodic polarization, the potential is changed in the anodic (or more positive) direction, causing the working electrode to become the anode and causing electrons to be withdrawn from it. For cathodic polarization, the working electrode becomes more negative and electrons are added to the surface, in some cases causing electrodeposition.[6] The limiting current is defined as the maximum current that can be generated by a given electrochemical reaction, at a given reactant concentration, under well-established hydrodynamic conditions, in the steady state. This definition implies that the limiting rate is determined by the composition and transport properties of electrolytic solution and by the hydrodynamic conditions at the electrode surface[4]. In general, for a rotating cylinder, when Reynolds number is greater than 200 the flow is turbulent [2]. The aim of the present work is to study the effect of changing the velocity (Reynolds number) on oxygen cathodic polarization curve using on surface of brass rotating cylinder electrode at different concentrations of NaCl: 0.1, 0.3 and 0.5N at 40, 50 and 600C. 2. EXPERIMENTAL DETAILS The chemical composition (in wt %) of the brass alloy used analyzed in Ebn Siena labortary was (Cu = 60.24, Zn = 39.22, Sn = 0.52). The dimensions of the cylindrically- shaped metal specimen was 3 cm long and 3 cm in diameter. The specimen was connected to a rotating shaft driven by a motor. The brass specimen was ground sequentially with Sic papers to 600, 400, 300, 250, 200, 150 and 100 grit and immersed in alcohol 1 min. and acetone 1 min and dried by paper tissues and placed in desiccator over night before the electrochemical tests. The polarization runs were conducted in 0.1, 0.3 and 0.5N NaCl solutions (PH = 7). A saturated Calomel electrode (SCE) was used as the reference electrode and

graphite was used as the counter electrode. The rotation speed was varied from 0 to 2000 rpm. 3. RESULTS AND DISCUSSION Figures 2 to 10 show experimental results conducted to demonstrate the effect of rotational speed on O2 cathodic polarization curves. On increasing the velocity, the limiting current will be increased at constant temperature and concentration this appear in table (1 to 9). These results are in agreement with Stern and Uhlig [5,7]. Velocity primarily affects electrochemical reaction rate through its influence on diffusion phenomena. It has no effect on activation-controlled processes. The manner in which velocity affects the limiting diffusion current is a marked function of the physical geometry of the system. In addition the diffusion process is affected differently by velocity when the flow conditions are laminar as compared to a situation where turbulence exists. For most conditions the limiting diffusion current can be expressed by the equation:

nUKi ×= (1) Where (K) is a constant, (U) is the velocity of the environment relative to the surface and (n) is a constant for a particular system. The value of n varies from 0.2 to 1 [5,7]. Figures 11 and 12 show experiments conducted to find the effect of concentration on O2 cathodic polarization curves. When the concentration of NaCl is increased the limiting current will be increase at constant Reynolds number (rotational velocity) and temperature show in table (10) which is in agreement with Fontana and Greene [1]. The effect of oxidizer additions or the presence of oxygen on electrochemical rate depends on both the medium and the metals involved. The rate of (limiting current) may be increased by the addition of oxidizers, oxidizers may have no effect on the corrosion rate, or a very complex behavior may be observed. By knowing the basic characteristics of a metal or alloy and the environment to which it is exposed, it is possible to predict in many instances the effect of oxidizer additions [1]. For diffusion-controlled process, an increase in concentration of the diffusing species in the bulk of the environment increases the concentration


1377

gradient at the metal interface. The concentration gradient provides the driving force for the diffusion process. Thus the maximum rate at which oxygen can diffuse to the surface (the limiting diffusion current) would be essentially directly proportional to the concentration in solution. Figs.11 and 12 are examples of the cathodic polarization diagrams which are operative for this system [5]. 4. CONCLUSIONS On brass alloy when increasing rotational speed limiting current will be increased. For diffusion controlled process, an increase in concentration of the diffusing species in the bulk of the environment increases the concentration gradient at the metal/solution interface. The concentration gradient provides the driving force for the diffusion process. Thus the maximum rate at which oxygen can diffuse to the interface surface (the limiting diffusion current) would be essentially directly proportional to the concentration in solution and temperature in presence of increasing NaCl concentration.

5. REFERENCE [1] Fontana and Greene, "Corrosion Engineering", McGraw Hill, 1984. [2] Gabe" Technical Note" 2006 [3] H. H. Uhlig "Corrosion and Corrosion Control", John Wily and Son, New Jersey, 2008. [4] J. R. Selman, "AICHE", Vol.77, No.204, P.88, 1981. [5] M. Stern, "Corrosion-NACE", Vol. 13, P.97, 1957. [6] P. R. Roberge, "Handbook of Corrosion Engineering", McGraw Hill, America, 2000. [7] H. H. Uhlig, "The Corrosion Handbook", John Wiley and Sons, New York, 1976 [8] Ming GENG, Zhenhao DUAN, Geochimica et. Cosmochimica Acta 74(2010) 5631-5640.

Fig.1. Rotating cylinder system.


1378

Fig. 2 Dissolved oxygen cathodic polarization curves on brass in 0.1N NaCl solution at 400C.







1379

ω (rpm)

Re i (mA/cm2) i (μA/cm2)

0 static 0.011323425 11.323425 500 8823.155 1.0049540 1004.9540 1000 17646.31 1.3164510 1316.451 1500 26469.46 1.569.3560 1569.356 2000 35292.62 1.8683650 1868.365

Table.1. Experimental limiting current results of dissolved oxygen cathodic polarization on brass in 0.1 N NaCl solutions at T= 40 oC.




Fig. 11 effect concentration change on polarization curve on brass at 400C at constant rotational speed (ω = 1000 rpm)

Fig. 12 effect concentration change on polarization curve on brass at 500C at constant rotational speed (ω = 1000 rpm)


1380

ω (rpm)


0 static 0.01433121 14.33121 500 14785.87 1.1004955 1100.4955

1000 29571.74 1.3694508 1369.4508 1500 44357.6 1.6719745 1671.9745 2000 59143.47 1.9125974 1912.5974

ω (rpm)


0 static 0.016100495 16.100495 500 14907.89 1.1150000 1115.0000

1000 29815.77 1.4081755 1408.1755 1500 44723.66 1.7055910 1705.5910 2000 59631.54 2.0099080 2009.9080

ω (rpm)


0 static 0.036447275 36.447275 500 17840.91 1.2853855 1285.3855

1000 35681.82 1.6277425 1627.7425 1500 53522.73 1.8966735 1896.6735 2000 71363.64 1.9975230 1997.5230

ω (rpm)


0 static 0.01167700 11.67700 500 17987.2 1.3145760 1314.576

1000 35974.41 1.6808210 1680.821 1500 53961.61 1.9709835 1970.9835 2000 71948.82 2.0169850 2016.985

ω (rpm)


0 static 0.057356705 57.356705 500 12602.43 1.1431325 1143.1325 1000 25204.86 1.6135880 1613.5880 1500 37807.3 2.1443735 2144.3735 2000 50409.73 2.2381455 2238.1455

ω (rpm)


0 static 0.042285915 42.285915 500 10651.02 1.2652160 1265.2160 1000 21302.05 1.6047415 1604.7415 1500 31953.07 1.8825195 1882.5195 2000 42604.09 2.1939135 2193.9135








1381

ω (rpm)


0 static 0.01127728 11.27728 500 21290.47 1.0744870 1074.487 1000 42580.95 1.3959660 1395.966 1500 63871.42 1.6737440 1673.744 2000 85161.89 1.7374380 1737.438

ω (rpm)


0 static 0.03556263 35.56263 500 21117.77 1.3782730 1378.2730 1000 42235.55 1.6808210 1680.8210 1500 63353.32 1.9904460 1990.4460 2000 84471.09 1.9986765 1998.6765

C(N)

T(0C)

Oxygen solubility in (mg/l)

0.1 0.3 0.5 40 6.389 6.1945 6

50 5.399 5.0495 4.7 60 4.490 3.795 3.1

Table.10. Oxygen solubilities at atmospheric pressure[8]




1382

An Investigation to the Abrasive Wear in Pipes Used for Oil Industry

Asst.Prof. Dr.Ahmed Abdul-Hussein Ali Eng. Mohanad kassim Abdul-Razzaq Zalzala

University of Baghdad University of Baghdad College of Engineering College of Engineering

Mechanical Engineering Department Mechanical Engineering Department [email protected] (M.Sc) [email protected]

ABSTRACT

The work reported in this study focusing on the abrasive wear behavior for three types of pipes used in oil industries (Carbone steel, Alloy steel and Stainless steel) using a wear apparatus for dry and wet tests, manufactured according to ASTM G65. Silica sand with hardness (1000-1100) HV was used as abrasive material. The abrasive wear of these pipes has been measured experimentally by measuring the wear rate for each case under different sliding speeds, applied loads, and sand conditions (dry or wet). All tests have been conducted using sand of particle size (200-425) µm, ambient temperature of 34.5 °C and humidity 22% (Lab conditions).

The results show that the material loss due to abrasive wear increased monotonically with the applied load at constant sliding speed and constant grit size due to increasing depth of penetration in both dry and wet sand which agrees with Archard´s equation. Sliding speed show insignificant effect on the wear loss of metals at constant load and constant grit size in both dry and wet sand. Wet sand results show higher wear losses than dry sand (20-70) % due to micro abrasion – corrosion wear and high slurry concentration.

Keywords: Abrasive wear ,Wear loss; Dry abrasion; Wet abrasion; Stainless steel; Carbon steel; Alloy steel;ASTMG65; Wear modeling; Finite element method.

دراسة التأكل بألحك على أالنابيب ألمستخدمه في المنشات النفطيه

الخالصةسلوك التاكل او البليان بالحك لثالثة انواع من االنابيب المستخدمه في على الدراسهالعمل الذي تم انجازه في هذه يركز

مقياس التاكل الذي جهازباستخدام لصدأل المقاوموالفوالذ والفوالذ السبائكي الفوالذ الكاربوني المصنعه منالصناعات النفطيه ر تاكل المعادن في الظروف التي تستخدم لقياس مقدا ASTM G65تم تصنيعه باالسواق المحليه طبقا للمواصفه القياسيه

)1100-1000(مسبب للتاكل هو رمل زجاجي (سيليكا ) وبصالده حوالي ان الرمل الذي تم استخدامه كعاملالجافه والرطبه .مم\( كغ

2).بتغـييـر كل من الحمل والسرعه ووسط التاكل من جاف الى رطب يمكن قياس التاكل بالحك لهذه االنابيب عن

) 425-200(جم حبيبات طريق معرفة مقدار الخساره بالحجم لكل من هذه االنابيب عمليا.ان جميع االختبارات تمت تحت ح ( ظروف المختبر).ان %22درجه مئويه ونسبة رطوبه حوالي 34.5مايكرو متر وخشونة سطح ثابته ودرجة حراره حوالي


Ahmed Abdul-Hussein Ali An Investigation to the Abrasive Wear in Pipes Mohanad kassim Abdul-Razzaq Zalzala Used for Oil Industry

1383

النتائج اظهرت ان مقدار الخسارة بالحجم لهذه المعادن التي تم استخدامها تزداد طرديا مع زيادة الحمل المسلط عند سرعه ثابته وحجم حبيبات ثابت نتيجة زيادة عمق التأكل او التغلغل في حالتي الرمل الجاف والرطب وهذا يتفق مع

Archard´s equation)(ها لم تظهر تاثيـر مقنع على مقادر الخساره بالحجم لهذه المعادن في ، اما فيما يخص السرعه فاننتيجة للتداخل %)70-20(نتائج الرمل الرطب تظهرخسائربالحجم أعلى من نتائج الرمل الجافحالتي الـرمل الجاف والرطب.

بة الـرمل العاليه في وكذلك نتيجة لنس (micro abrasion – corrosion wear) في اكثر من نوع واحد من التاكل .(slurry concentration)الماء

هالعناصر المحددطريقة ; G65المواصفه ;السبائكي;الكاربوني ;فوالذ عديم الصدأ ;الكلمات الرئيسيه: التاكل بالحك (جاف ورطب)

1. INTRODUCTION

The dry sand/rubber-wheel abrasion test is widely used to evaluate low-stress abrasive wear of materials, particularly for evaluating wear-resistant materials used in the mining, oil sand, oil pipe lines and agricultural machinery industries. During such a test, a specimen is loaded against the rim of a rotating rubber wheel, a sand flow is directed to the gap between the wheel and specimen, abrading the specimen under an applied normal load at a certain sliding speed. Abrasion resistance of a material is evaluated by measuring its volume loss [C. Hilerio, 2004].

Abrasive wear defines by ASTM (American Society for Testing and Materials) as the loss of material due to hard particles or hard protuberances that are forced against and move along a solid surface, [D. Hewitt, 2009]. 50% of all wear problems in industry are due to abrasion, and as such, much laboratory work has examined and sought to rationalize the abrasive wear behavior of a wide range of materials [S. Wirojanupatump, 2000]. Based on the analysis of parameters responsible for the wear of mechanical parts, about 50% (of the parts) works in abrasive wear, 15% - adhesive wear, 8% - erosion, 8% - fretting, 5% - wear is due to corrosion and about 14% is just a combination of abrasive, erosive and corrosive wear[M. Adamiak, 2009].

Abrasive wear in pipes results when solids make up a large percentage of the fluid being transported [D. Hewitt, 2009].

The sand particles interact with each other, the fluid media and the pipe wall. This

can be described by measuring the lost mass of the pipe [D. Hewitt, 2009].

For many industrial applications, the rubber-wheel test is performed often under a fixed load and at a fixed sliding speed to have all tested materials evaluated under the same condition. ASTM G65 has specified such

abrasion test with fixed loads and fixed sliding speeds for ranking materials in different classes. However, ranking materials using this method may not be always accurate and misleading information might be generated [X. Ma, 2000].

Wear resistances of a Be–Cu alloy, 17-4 PH steel and D2 tool steel using a rubber-wheel tester have been studied. Different loads and sliding speeds were chosen for the abrasion test and SiO2 sand was used as the abrasive [X. Ma, 2000].

In this work the abrasive wear behavior of three types of pipes used in oil industries (Carbone steel, Alloy steel and Stainless steel) has been studied using a wear apparatus according to ASTM G65 for dry and wet tests.

Different sliding speeds (1.7954-3.5908) m/s, different applied loads (50-150) N and different test condition (dry and wet) had been set for this work and the abrasive wear behavior were compared with other behaviors obtained by [C. Hilerio, 2004] , [D. Hewitt, 2009],[M.Adamiak, 2009], [S. Wirojanupatump, 2000] and [X. Ma, 2000],etc.


1384

2. EXPERIMENTAL DETAILS The dry and wet sand/rubber wheel

apparatus used in this work is schematically illustrated in Fig. 1. It was built based on ASTM G65 standard [ASTM G65, 2001]. The rubber wheel is in contact with a specimen under an applied load. A flow of sand particles is directed to the gap between a rotating rubber wheel and the specimen. The sand particles scratch the surface of the specimen under the applied load at a sliding speed of Rω, where ω is the angular speed of the rubber wheel and R is its radius.

The rubber wheel was made of chlorabutyl and its hardness was Durometer A-62. In the study, standard silica sand (Sio2) with hardness (1000 -1100) HV, (200-425)µm of rounded silica sand abrasive were obtained by sieving, as recommended by ASTM G65 for the dry and wet sand/rubber wheel abrasion test. Specimens for the wear test had a rectangular shape with dimension of 75mm long, 26 mm wide and 6 mm thickness. Specimens of different material, namely, Carbon steel A106 grade C, Alloy steel A213 grade T9, Stainless steel A312 grade S20400 with hardness of 153 Hv, 174 Hv, and 253 Hv respectively.

Chemical composition of the tested materials can be shown in table (1).

Mechanical properties of these materials were determined by tensile test according to ASTM E8M using a tensile machine (Instron8516). The mechanical properties of the tested materials can be shown in table (2).

Specimens for the wear test were polished

using emery paper of different grit size to obtain surface roughness of Ra less than (0.8) µm (between 0.146-0.236) µm according to ASTM G65. Since the sand flow rate affects the wear rate, a constant sand flow of 320 g/min was used for both dry and wet tests.

Feed water added to enable the apparatus to measure the wear losses in wet conditions according to modified ASTM G65 apparatus.

Wet test used a mixture of 1.5kg of abrasive and 0.940 Kg of water according to modified ASTM G65 for wet test. Wear loss of

a steel specimens was evaluated by measuring the volume losses of the specimens. Volume losses were measured by specimen mass loss using an electronic balance (KERN) with an accuracy of 0.001 g. The volume losses is the mass loss divided by the density of each type of steel.

Specimens were well cleaned before a new tests are carried out.

In industry, a constant force of 130 N is usually used for the abrasion test. Such a load is recommended by ASTM G65 for testing most metallic materials in a wide range of abrasion resistance. In the present study, range of loads from (50 to 150) N were used while the sliding speed was also changed from 1.7954 to 3.5908 m/s. We did not exactly follow the ASTM G 65 standard and, instead, used various loads and speeds to investigate effects of the load and sliding speed on the wear loss. The difference in wear between different materials could vary markedly when the applied load or the sliding speed is changed. Some materials have excellent wear resistance under low loads or speeds but may perform poorly under higher loads or speeds, while the other materials may show opposite behavior. Therefore, using one fixed load and sliding speed to rank industrial materials may not be sufficient to obtain accurate information. The test duration was (5) min according to ASTM G65 procedure E for low and medium abrasion materials resistance with a linear sliding distance between (538.62 -1077.24) m according to the rotational speed (150-300) r.p.m. The specific wear coefficient of the dry and wet tests was determined by the standard Archard´s equation, [T.A. Rodil, 2006], [R.L. Norton, 2011]. KS = Vloss

L.W eq. (1)

Where KS is the specific wear coefficient in units of mm3/N.m, Vloss is the volume removed


1385

from materials in mm3 , L is per unit sliding distance (m) and W is the applied load (N). All dry tests were done under apparent temperature range between (34-75) °C using the infrared scan temperature model (410), the scan temperature apparatus was calibrated and reported with an error of 8% between the actual and apparent temperature. To determine the actual temperature the following equation could be used. Tact.= Tr × (1 + 0.08

0.92 ) eq. (2)

Finally it should be mentioned that for dry tests, an apparatus according to ASTM G65 was used, and for wet tests the same apparatus was used with a feed water added according to modified ASTM G65 as shown in Fig. 2. 3. EXPERIMENTAL RESULTS 3.1. Wear Loss with Respect to the Applied Load Volume losses of the Carbon steel (A106), Alloy steel (A213) and Stainless steel (A312) were measured with respect to the applied load and results of the measurement are illustrated in Fig. 3 for dry tests and Fig.4 for wet tests.

The data was an averaged result of a two measurements with its error range less than 5%.

It was demonstrated that from all experimental and numerical results obtained the wear losses of Carbon steel (A106), Alloy steel (A213) and Stainless steel (A312) increased monotonically with applied load due to increasing depth of penetration in both dry and wet tests which agrees with Archard´s equation.

3.2. Effect of Sliding Speed The sliding speed is another parameter that influences the wear loss of a material. In order to determine the effect of sliding speed on wear loss, four sliding speeds were used for the abrasion test in dry condition and three sliding speeds were used for abrasion test in wet condition. Volume losses of C.S, A.S and S.S specimens at different sliding speeds under different loads were measured.

However, carbon steel and alloy steel showed different response to the sliding speeds. 3.2.1 Effect of sliding speed on the wear rate of carbon steel and alloy steel in dry test.

The volume losses of these metals in dry tests are illustrated in Fig.5.

It seem that the interaction between the wear mechanisms during the dry tests is a reason for this behavior, for carbon steel and alloy steel at sliding speeds less than 2.3938m/s, wear loss decreased rapidly in oxidational wear condition due to the protective layers of oxidized debris. Sliding speeds higher than 2.3938m/s abrasive wear was predominant resulting in the highest wear loss [K. Elalem, 2001] and [N. N .Aung, 2008].

3.2.2 Effect of sliding speed on the wear rate of carbon steel and alloy steel in wet test. The result of abrasive wear behavior of these metals in wet sand are illustrated in Fig. 6.

It was demonstrated that the increasing of sliding speeds increase the volume losses of these metals which agrees with Archard´s equation.

3.2.3 Effect of sliding speed on the wear rate of stainless steel in dry and wet tests.

The result of abrasive wear behavior illustrated in Fig. 7. It was demonstrated that the increase in sliding speeds cause a decrease in the volume loss of stainless steel in both dry and wet conditions, this is can explained by referring to the following: 1. The heat generated due to friction within test duration (300s) is negligible. Therefore it seems the effect of work hardening in dry conditions is predominant factor here, because ASTM A312 grade S20400 has a very high work by increasing the sliding speeds, [A.K Steel, 2007]. 2. The SiO2 sand was not strong enough to significantly damage the surface of stainless steel under loads. Instead, it seems that the surface of SiO2 sand was damaged to a


1386

considerable degree, thus diminishing its role in abrading the tested material, [X. Ma, 2000]. 3.3. Wear Loss with Respect to Dry and Wet Tests It is obvious from all dry and wet results, the volume losses in wet sand conditions is higher than dry sand conditions (20-70)% for all metallic materials used in experiment tests as shown in Fig. 8. This can be attributed to: 1. Wear-corrosion interaction can lead to either increase in the overall mass loss or a decrease in the overall mass loss. The change in the mass loss due to the synergistic effects of coupling wear and corrosion is often referred to as synergy (S). Positive synergy results in accelerated material loss due to the combined action of wear and corrosion and is an undesirable material property,[J.O. Bello, 2007]. On the other hand, negative synergy results in a decrease in the overall loss of material due to improvement in either wear or corrosion resistance and is a desirable material property According to the ASTM G119 standard guide for determining synergism between wear and corrosion, the total wear during the process of abrasive wear-corrosion is defined by the following equations, [J.O. Bello, 2007] Total Wear (AC) = Pure Abrasion (PA) + Pure Corrosion (PC) + Synergy (S) Synergy (S) = ΔPCA + ΔPAC Wear may be accelerating by corrosion, increased removal of the protecting oxide layer from the surface during the friction of the corrosion process. Friction provides continuous removal of the oxide layer.

The transition from mild to severe wear is linked to the level of oxidation of the metal in contact.

During corrosion, the oxide layer decrease due to interaction between water and surface layer of metals and therefore the protective oxide layer removed by wear faster than it is regenerate.

The influence of water on the wear of metals, however, is much more significant than on friction. The greatest effects occur during

wear in the corrosive regime where chemical reactions with the environment lead to either increase the removal of the protecting oxide layer from the surface and then increase the mass losses, [j.k. Lancaster, 1990]. 2. It is obvious that the volume loss exhibited relatively greater dependence on slurry concentration than normal load which increase the severity of wear, in other word, the slurry concentration has more effect than the normal load which was agree with [S.G. Sapate, 2010].

It should be mention that the slurry concentration used in this study is (150%) which considered as a high concentration.

3.4. Temperature Measurement The induced temperature due to applying a certain load and sliding speed have been measured and plotted against time as shown in Fig. 9.

The temperature rises during the test depend on sliding speed more than normal load, in other word the temperature increase with sliding speed more than normal load. 3.5. Correlation between Volume Loss and Hardness of Materials

According to the experimental results of hardness for S.S (253Hv), A.S (174Hv) and C.S (153Hv).

It seem that from mass losses of the three types of metals (C.S A.S, S.S) that the mass losses of stainless steel in both dry and wet conditions have wear losses more than Alloy steel at 150 R.P.M and 200 R.P.M and less than carbon steel. After increasing the sliding speed to 250 R.P.M and 300 R.P.M the wear losses of stainless steel became less than Alloy steel and also less than carbon steel. The reason for that is that the hardness of the material should not always mean that higher hardness means lower wear losses, in fact in some cases higher hardness produce higher wear losses, [J.G.C. Nava, 2010].


1387

Fig. 10. show the relationship between the hardness and the materials type used in this work and does not represent the volume loss of these materials.

4. FINITE ELEMENT METHOD AND WEAR SIMULATION Predicting wear and scuffing risk in metallic contacts is an important task. Influential factors such as temperature, elastic–plastic deformations, wear, surface topography, material properties and chemical composition all contribute to the complex contact conditions.

Three-dimensional components. Thus, it is very important to identify when and how much material should be removed from the models. The basic approach is to: (1) identify the important parameters affecting the material removal rates, (2) determine appropriate wear rates from specimen-level tests, and (3) perform iterative finite element analyses to progressively remove materials during simulation. In this study, the wear loss that is obtained from the ASTM G65 apparatus is used to perform a series of finite element analyses and to estimate the profile of the worn surface [N.H. Kim, 2005].

For finite element modeling and analysis a commercial program, ANSYS V13, is used to solve the contact problem and wear strain.

The most frequently model used is based on the Archard’s abrasive wear law. It is assumed here that wear can be evaluated by applying modified Archard’s equation to local contact conditions along a differential width of the contact interface.

Solid95 is used to represent the steel specimen with 20 node and plane183 is used to represent the sand particles with 8 node.

The finite element mesh used was map and hexahedral.

Each line of steel specimen and sand was divided into eight equal lines before meshing using the command (esize).

In the present study, it has been observed that maximum number of a bout (6400) iteration is generally sufficient to predict

the solution divergence or failure for the abrasive wear problems.

This maximum number of iterations depends on the type of the problem, extent of nonlinearities, and on the specified tolerance.

Maximum number of elements used is (512) elements and max number of nodes (2673) for steel specimens and same number used for sand at esize equal to (8).

It should be mention that at esize (6) and esize (4) the error between the experimental and numerical results was greater than at esize (8),also using lesize cause divergence during the test and long time running , therefor all results were obtained at esize (8) which could be represent the optimum number for meshing.

4.1 Calculation of Contact Pressure A surface-to-surface contact modeling technique that prevents contact elements (CONTA174) and target elements (TARGE170) from penetrating each other is used. In this contact–target strategy the contact pressure is only calculated for nodes on the contact elements. In order to calculate the contact pressure on steel specimen surface, symmetric contact is used.

To find the contact point locations and pressure augmented lagrange approach was used in conjunction with standard behavior of contact, symmetric stiffness matrix and automatic time increment was used during all tests. Modified Newton Raphson iterations are required to find the converged configuration for each time step. 4.2 Calculation of Wear Strain Some researchers used the modified Archard’s model to determine the wear strain as follow [J.M. Thompson, 2006]: Vloss = KD. 𝑆C2.R eq. (3) Where: KD is the dimension abrasive wear coefficient C2= 1


1388

R is the number of repetition of load If (Vloss) represents the change in

volume of the specimen due to wear, then we can define wear strain as the change in volume divided by the original volume of the specimen and rewrite the wear equation as ,[ J.M. Thompson, 2006]:

𝑒𝑤= C1. 𝑆C2.R eq. (4) Where: 𝑒𝑤 is the wear strain C1 = (KD/V)

Wear strain as proposed here is different from wear as proposed by Archard. The Archard equation is a systems approach where the applied load is assumed to be distributed over the entire loading area. Wear would be expected to occur uniformly over the entire surface. The wear strain proposed here is a function of stress and load repetitions. This implies that where load is applied to the surface, wear will occur and that parts of the surface which are (currently) unloaded will not experience change due to wear. Explicit creep is used since the plan is to calculate the wear strain based upon the final configuration of the surface at the end of the load step. In ANSYS, the explicit creep calculation is performed as in [J.M. Thompson, 2006]. The strain hardening creep equation used is given by [J.M. Thompson, 2006]:

ecr = ddt

C1.SC2 . ecrC3 . e(−C4 T� ) eq. (5) 4.3 Finite Element Results The volume losses of all materials used in this study are determined through wear strain equation as shown in Fig. 11. and Fig. 12. These figures represent the maximum wear strain at the end of the tests (5) min, it is obvious that the maximum wear strain form a volume at the middle of the specimen which agrees with the experimental results.

From wear strain we calculated the volume loss by multiply the maximum wear strain by the original volume of each specimen at specific load and sliding speed. Wear strain was calculated using eq. (5) and set C3 and C4 to zero. The contact pressure of all materials used in this study are determined through surface to surface contact as shown in Fig. 13. and Fig.14 These figures represent the maximum contact pressure at the end of the tests (5) min, it is obvious that the maximum contact pressure form an area at the middle of the specimen which agrees with the experimental results. The contact pressure was calculated by using the contact wizard. The contact pressure was calculated experimentally be divided the applied load on the apparent contact area of the specimen and these results compared with the contact pressure obtained by ANSYS. 5. DISCUSSION

It was demonstrated that from all experimental and numerical results obtained the wear loss of stainless steel, alloy steel and carbon steel increased monotonically (linearly) with applied load due to increasing depth of penetration which agrees with Archard’s equation.

The Relationship between the load and volume loss may be linearly described in Equation of the following form using linear curve fitting:

Y(x) =𝑎1(x) +𝑎𝑜 eq. (6)

Sliding speed didn’t show significant effect on the volume loss of materials used in this study, many of these relationships may be described linearly, some of these relationship as shown in Fig. 5. Show that linear curve fitting is not appropriate for these relationships and nonlinear curve will be better.


1389

The present results did not show a defined relationship between materials hardness and volume loss of materials. The specific wear coefficient (KS) show marked dependency on both applied load, sliding speed and test condition (dry or wet). The error percentage between experimental and numerical results was calculated as follows:

Error = �𝐸𝑥𝑝−𝑁𝑢𝑚𝑒𝑟𝑖𝑐𝑎𝑙𝑁𝑢𝑚𝑒𝑟𝑖𝑐𝑎𝑙

� ×100 eq. (7) The error was calculated between the experimental and numerical contact pressure results and between the experimental and numerical volume losses results for each test of the specimens (dry and wet). From dry test five values of load and four values of sliding speed were used which mean that , 20 test for each steel type has been done , and in total 60 test for dry condition. From wet test five values of load and three values of sliding speed were used which mean that , 15 test for each steel type has been done , and in total 45 test for wet condition.

Therefore the results show that the standard deviation (σ) between the experimental and numerical results is:

(12.85034) volume loss for dry condition (60) test and (7.93811) % for wet condition (45) test.

(0.72753)% contact pressure for dry condition (60) test and (0.2160) %for wet condition (60) test.

The standard deviation was calculated for the (60) dry tests and for the (45) wet tests.

6. CONCLUSION 1. The abrasive wear loss of metals increased monotonically with applied load according to Archard´s equation at constant sliding speed and constant grit size. 2. Sliding speed show insignificant effect on the wear loss of metals at constant load and constant grit size for both dry and wet sand. 3. Wet sand results show higher wear losses than dry sand results (20-70)% due to micro abrasion – corrosion wear and slurry concentration.

4. The wear losses of stainless steel decrease with increasing sliding speed in both dry and wet conditions due to work hardening. 5. The temperature of metals increase with increasing both applied load and sliding speed, but it show relative depends on sliding speed than applied load. 6. Results shows that a linear relationship could be used which agrees with Archard´s equation. 7. Some figures show that linear curve fitting is not the best choice and it seems that nonlinear curve will be more appropriate for these figures. 8. It should mention here that in order to study the effect of any other parameters like particle size, surface roughness, temperature and humidity, the effect of these parameters could be included by the specific wear coefficient(KS). 9. Increase the hardness should not always mean that the higher hardness means the lower wear losses.

REFERENCES

• A.K Steel Corporation UNS S20400, West Chester, OH 45069, [2007].

• ASTM G65: Standard Test Method for Measuring Abrasion Using the Dry Sand/Rubber Wheel Apparatus, [2001].

• C. Hilerio, M.A. Barron, A. Altamirano, "Wet and Dry Abrasion Behavior of AISI 8620 Steel Boriding”, University of Metropolitana Unidad Azcapotzalco, Department of Materials, Mexico,D.F,[2004].

• D. Hewitt ,S. Allard , P. Radziszewski , "Pipe Lining Abrasion Testing for Paste Backfill Operations", Minerals Engineering 22,(1088–1090), [2009].

• J.G.C. Nava , A.M. Villafan , F.A. Calderon , J.A. Cabral , M.M. Stack , "Some Remarks on Particle Size Effects on the Abrasion of a Range of Fe Based Alloys”, Science Direct, Tribology International Vol 43, p (1307–1317), [2010].


1390

• J.k. Lancaster, "A Review of the Influence of Environmental Humidity and Water on Friction, Lubrication and Wear", Tribology International, [1990].

• J.M. Thompson, M.K. Thompson, "Proposal for the Calculation of Wear", Mechanical Engineering Dept., MIT, [2006].

• J.O. Bello, R.J.K. Wood, "Synergistic

Effects of Micro-Abrasion–Corrosion of UNS S30403, S31603 and S32760 Stainless Steels”, Wear 263, p (149–159), Science Direct, [2007].

• K. Elalem , D.Y. Li , "Variations in Wear Loss with Respect to Load and Sliding Speed under Dry Sand/Rubber - wheel Abrasion Condition" , Wear 250, (59–65),[2001].

• M. Adamiak, J. Górka, T. Kik, "Comparison of Abrasion Resistance of Selected Constructional Materials", Journal of Achievement in Materials and Manufacturing Engineering, Volume37, Issue2, [2009].

• N.H. Kim, D. Won, D. Burris, P.

Swanson, "Finite Element Analysis and Experiments of Metal/Metal Wear in Oscillatory Contacts", Science Direct, Wear258, (1787-1793), [2005].

• N. N .Aung, W. Zhou, L. E.N. Lim,

“Wear Behavior of AZ91D Alloy at Low Sliding Speeds”, Wear 265, p (780-786), Science Direct, [2008].

• R.L. Norton "Machine Design", Worcester, Massachusetts, Fourth Edition, [2011].

• S.G. Sapate, A. Selokar, N. Garg,

"Experimental Investigation of Hard-faced Martensitic Steel under Slurry Abrasion Conditions", Materials and

Design, Volume31, Pages (4001-4006), Science Direct, [2010].

• S. Wirojanupatump, P.H. Shipway, "Abrasion of Mild Steel in Wet and Dry Conditions with the Rubber and Steel Wheel Abrasion Apparatus", Wear 239, 91–101, [2000].

• T.A. Rodil, "Edge Effect on Abrasive Wear Mechanisms and Wear Resistance in WC-6wt. %Co Hard Metals", Karlstad University, [2006].

• X. Ma , R. Liu , D.Y. Li , "Abrasive

Wear Behavior of D2 Tool Steel with Respect to Load and Sliding Speed under Dry Sand/ Rubber Wheel Abrasion Condition" , Wear 241 , (79–85), [2000].


1391

Fig. 1. Schematic illustration of Fig. 2. Abrasive wear apparatus a rubber wheel abrasion testing apparatus.

Table 1 Compositions of the materials under study Table (2) Mechanical properties of the tested materials

Metals 𝜎𝑦

(Mpa) 𝜎𝑢

(Mpa) Elongation

% (ϵ)

Young's Modulus E(Gpa)

C.S 379.5 490.96 26 200 S.S 445.99 696.86 58 200 A.S 373.2 500.98 30 200

Type C% Si% Mn% Cr% Ni% Mo% Cu%

C.S 0.17 0.41 0.52 0.1 - - -

S.S 0.06 0.37 9.3 15.2 1.41 - 1.7

A.S 0.14 0.44 0.47 9.75 - 1.09 -


1392

Fig. 3. Relationship between load and Fig.4. Relationship between load and Volume loss for dry steel volume loss for wet steel

Fig.5. Relationship between volume loss Fig.6. Relationship between volume loss and and speed sliding speed

0

10

20

30

40

0 50 100 150 200Vol

ume

loss

(mm

^3)

Load (N)

200 R.P.M and particle size (200-425μm)

Dry A.S Dry C.S Dry S.S

0

50

100

150

200

0 50 100 150 200Vol

ume

loss

(mm

^3)

Load (N)


Wet A.S Wet C.S Wet S.S

0

20

40

60

80

0 1 2 3 4Volu

me

loss

(mm

^3)

Sliding speed (m/s)

load 75N and particle size (200-425µm)

Wet A.S Wet C.S

05

1015202530

0 1 2 3 4volu

me

loss

(mm

^3)

sliding speed (m/s)

load 100N and particle size (200-425μm)

Dry A.S Dry C.S


1393

Fig. 7. Relationship between volume loss and Fig.8. Relationship between volume loss and sliding for dry and wet stainless steel load for dry and wet test

Fig. 9. Relationship between temperature Fig. 10. Relationship between Hardness and sliding speed and materials type

0

50

100

0 50 100 150 200

Volu

me

loss

(mm

3)

Load (N)


Dry S.S Wet S.S

0

20

40

60

80

0 2 4 6

Tem

pera

ture

(c )

Time (min)

300 R.P.M and 125 N

Exp

0

50

100

150

0 1 2 3 4

Volu

me

loss

(mm

^3)

Sliding speed (m/s)

load 125N and part ic le s ize 425µm

Dry S.S Wet S.S

0

50

100

150

200

250

300

S.S A.S C.S

Hard

ness

Hv

Materials type


1394

Fig. 11. Abrasive wear on dry stainless steel Fig. 12. Abrasive wear on wet stainless at time (5) min steel at time (5)min

Fig. 13. Contact pressure on dry stainless Fig. 14. Contact pressure on wet stainless Steel at time (5) min steel at time (5) min


1395

Adsorption of Chromium (Vi) from Aqueous Solutions using Low Cost Adsorbent: Equilibrium and Regeneration Studies

Ihsan Habib Dakhil Chemical Engineering Department, College of Engineering, Al-Muthanna University

E-mail: [email protected]

ABSTRACT

The adsorption of Cr (VI) from aqueous solution by spent tea leaves (STL) was studied at different initial Cr (VI) concentrations, adsorbent dose, pH and contact time under batch isotherm experiments The adsorption experiments were carried out at 30°C and the effects of the four parameters on chromium uptake to establish a mathematical model description percentage removal of Cr (VI). The analysis results showed that the experimental data were adequately fitted to second order polynomial model with correlation coefficients for this model was (R2 = 0.9891). The optimum operating parameters of initial Cr (VI) concentrations, adsorbent dose, pH and contact time were 50 mg/l, 0.7625 g, 3 and 100 min, respectively. At these conditions, the maximum percentage removal of Cr (VI) was 92.88%. The amounts of Cr (VI) adsorbed onto STL were highly affected by the solution pH value. Equilibrium data was modeled with Langmuir and Freundlich models isotherms. Langmuir model is found very well represent the equilibrium data with correlation factor is close to unity than the Freundlich model. The maximum monolayer adsorption capacity was found to be 47.98 mg/g at optimum conditions. The saturated adsorbent was regenerated by base treatment and found to be reuse efficiently after fourth cycle at optimum conditions as well as for safe disposal of base that contains high concentration of Cr (VI) is precipitated as barium chromate.

KEYWORDS: Adsorption, Removal, Low cost adsorbent, Chromium (VI), Regeneration.

زةامحاليل المائية بأستخدام مواد مأمتزاز الكروم الخماسي من ال واطئة الكلفة : دراسات التوازن وٕاعادة التنشيط

الخالصة:اسـي كيز األبتدائيـة للكـروم الخمأمتزاز الكروم الخماسي من المحاليل المائية بواسطة ورق الشاي المستهلك بتغيير التراتم دراسة

تجـارب االمتــزاز . بثبـوت درجـة الحـرارة بأسـتخدام نمـط تجـارب الوجبــات مــن الـتالمس، كميـة المـادة المـازة ، الدالـة الحامضـية وز كفـاءةف للحصـول علـى موديـل رياضـي يصـعلى امتزاز الكروم لمتغيرات االربعة م لدراسة تأثير ا ْ 30حدثت تحت درجة حرارة

معامـل تصـحيح تحليل وضحت ان النتائج العملية كانت متطابقـة مـع معادلـة مـن الدرجـة الثانيـة بنتائج اللكروم الخماسي. اازالة لتراكيــز االبتدائيــة للكــروم الخماســي، كميــة المــادة المــازة، الدالــة الحامضــية ا مــن ). العوامــل التشــغيلية المثلــى0.9891( مســاوياً

أعلى نسبة لالزالة كانـت عند تلك الظروف على الترتيب. دقيقة، 100و 3غم، 0.7625ملغم/غم، 50كانت وزمن التالمسمع التجارب العملية. كمية الكروم الخماسي الممتز على ورق الشاي ي% اثبتت بتطابق المحسوب من الموديل الرياض92.88

) و Langmuir( بيانـات االتـزان مثلـت مـع مـوديالت متسـاوية الحـرارة الحامضـية للمحلـول. المسـتهلك تـأثر بشـكل كبيـر بالدالـة)Freundlich) موديــل .(Langmuir مثــل بصــورة جيـــدة بيانــات االتــزان مـــع معامــل تصــحيح اقتــرب مـــن الواحــد مقارنــًة مـــع (

ملغم/غــم عنــد الظــروف المثلــى. تــم اعــادة تنشــيط المــادة المــازة 47.98) . اعلــى نســبة لالمتــزاز كانــت Freundlichموديــل (


Ihsan Habib Dakhil Adsorption of Chromium (Vi) from Aqueous Solutions using Low Cost Adsorbent: Equilibrium and Regeneration Studies

1396

عند الظروف المثلى. ولتصريف آمن للقاعدة المحتوية على بكفاءة ألربع مراحلة استخدامها المشبعة بمعاملتها مع قاعدة واعادتراكيــــــــــــــــــــــــــــــز عاليـــــــــــــــــــــــــــــــة للكـــــــــــــــــــــــــــــــروم الخماســــــــــــــــــــــــــــــي رســـــــــــــــــــــــــــــــب علـــــــــــــــــــــــــــــــى شــــــــــــــــــــــــــــــكل كرومـــــــــــــــــــــــــــــــات البـــــــــــــــــــــــــــــــاريوم.

INTRODUCTION:

The pollution of water resources due to the disposal of heavy metals has been an increasing worldwide concern for the last few decades. Chromium is one of the most toxic heavy metals discharged into the environment through various industrial wastewaters, constituting one of the major causes of environmental pollution. The main industrial sources of chromium pollution are leather tanning, electroplating, metal processing, wood preservatives, paints and pigments, textile, dyeing, steel fabrication and canning industry (Singh et al., 2009). The most common forms of chromium are trivalent chromium [Cr (III)] and hexavalent chromium [Cr (VI)]. Hexavalent chromium forms chromate (CrO4

-2) or hydrogen chromate (HCrO4

−) that is more toxic and more soluble. The exposure of Cr (VI) to human causes nausea, diarrhea, liver and kidney damage, dermatitis, internal hemorrhage and respiratory problems. The maximum concentration limit for chromium (VI) for drinking water is 0.05 -0.1 mg/L (Surendra and Dharmendra, 2012).

There are various treatment technologies available to remove Cr (VI) from wastewater such as chemical precipitation (Uysal and Irfan, 2007), ion-exchange (Jianlong and Xinmin, 2000), membrane separation (Kozlowski and Walkowiak, 2002), electrocoagulation (Roundhill and Koch, 2002), solvent extraction (Li et al., 2004), and adsorption (Baral et al., 2007) and (Mohan et al., 2005). These techniques are economically expensive for the removal of Cr (VI) from wastewater. The above mentioned removal techniques have many disadvantages such as incomplete metal removal, high reagent and energy requirements, and generation of toxic sludge or waste products which require proper disposal without creating any problem to the environment (Aliabadi et al., 2006) and (Mohan and Pittman, 2006). Therefore, there is a dire need of a treatment method for Cr

(VI) removal from wastewater which is simple, effective and inexpensive (Babu and Gupta, 2008).

Because of high performance and ease of use, adsorption is introduced as one of the most applied methods (Bailey et al., 1999). In this method heavy metals are adsorbed in the pore surface of adsorbent which is insoluble in water. One of the most common adsorbent for heavy metals is activated carbon, which because of its high cost of activating processes is very expensive. In recent years cheap adsorbents have been attractive to many of researchers. Cheap adsorbents are widely and easily in reach and their preparation cost is low. These adsorbents are mainly a waste result of industrial and agricultural activity and have cellulose base (Hassan et al., 2012). Adsorption technology is easy to use, no need for processing and complex reforming processes, efficient and selective for heavy metals.

The objective of the present study is to investigate the possible use of STL as a cheap adsorbent material for the removal of Cr (VI) from wastewater. Batch experiments are carried out for studies the removal of Cr (VI) from aqueous solution. The influence of various important parameters such as initial Cr (VI) concentration, adsorbent amount, pH, and contact time are investigated. The Langmuir and Freundlich models are used to fit the experimental equilibrium isotherm data obtained in this study. MATERIALS AND METHODS:

Materials

The adsorbent STL was collected from the local sources. The tea leaves was washed several times in distillate water to remove any adhering dirt and repeatedly boiled with water until the filtered water was cleared. Then it was oven dried at 80 oC for 24 h. Finally, the


1397

dried sample was ground and sieved to obtain a particle size range of 3–6 mm and stored in plastic bottle for further use. Chemicals

All the chemicals used are of analytical grade. A stock solution of 1000 mg/l of Cr (VI) is prepared by dissolving 2.8287 g of 99.9% potassium dichromate (K2Cr2O7) in 1000 ml of solution. This solution is diluted as required to obtain the standard solutions containing (50-500) mg/l of Cr (VI). The concentration of Cr (VI) ions in the effluent is determined spectrophotometrically by developing a purple violet color with 1,5-diphenyl carbazide in acidic medium by following APHA, AWWA standard methods for examination of water and wastewater (APHA, AWWA, 1998). The absorbance of the purple-violet colored solution is read at 540 nm after 20 min. pH adjustment is carried out by using 0.5N HCl and 0.5N NaOH solutions.

Batch Experiments

The adsorption experiments were carried out in isothermal batch process at 30°C

± 1°C and the effect of different parameters such as initial Cr (VI) concentration, adsorbent dose, pH and contact time were studied. The experiments were conducted by adding an amount of adsorbent varied between (0.05-1) g with 100 ml of Cr (VI) solution of different initial concentration (50-500) mg/l in 250 ml stopper conical flasks. These flasks were placed on a rotating shaker with constant shaking at 150 rpm to maintain the equilibrium condition. The range of experimental variables could be represented in the Table 1.

Table 1 Range of Experimental Variables

At different intervals of time, samples were drawn out of the adsorber using syringe (2ml). Cr (VI) concentration was measured using a spectrophotometer in the visible range at maximum wave length. The percentage of removal efficiency (% R) for Cr (VI) was calculated according to the following equation:

100×=−

o

CC

CR% eo (1)

Also, the amount of adsorption at equilibrium, qe (mg/g), was calculated by:

Wq )CC(V

eeo −= (2)

where Co and Ce (mg/l) are the

concentrations of Cr (VI) at initial and equilibrium, respectively. V (L) is the volume of the solution and W (g) is the mass of dry sorbent used.

Statistical Design of Experiments:

Box-Wilson experimental design method, commonly called “Central Composite Design”, was used to establish a mathematical model relating the removal efficiency (Y) of Cr (VI) with various operating variables such as initial Cr(VI) concentration (X1), adsorbent dosage (X2), pH (X3) and contact time (X4) (Box and Hunter, 1957). The relationship between coded level and corresponding real variables were tabulated in Table 2.

The needed number of experiments (N) depends on number of variable (q) and was estimated according to the following equation (Montogomery, 1984):

122 ++= qN q (3)

In accordance with eq.(3), twenty-nine experiment were carried out with four operating conditions. Also, the form of a quadratic (second order model) polynomial is

illustrated by the following equation (Montogomery, 1984):

Contact Time (min)

pH

Adsorbent Dose (g)

Initial Conc. of Cr(VI) (mg/l)

20 - 180 1 - 9 0.05 - 1 50 - 500

Y = Ao+ A1X1+ A2X2 + A3X3 + A4X4+ A5X12

+A6X22 + A7X3

2 +A8X42 +A9X1X2 + A10X1X3

+ A11X1X4 + A12X2X3+A13X2X4 + A14X3X4 (4)


1398

where Y is the predicted response, X is the coded value of the independent variable and A is a coefficient. At the present study, the central composite design was used for optimization of the adsorption process and to evaluate the effects and interactions of the process variables. The experimental data was analyzed by using Statistics–Software program version 7. It was made under “non-linear estimation”. The obtained equation was verified by applying the F–test and analyzing the correlation coefficients (R) and variance explained (S) by comparison between the calculated and experimental values. To evaluate the optimum conditions that give the maximum response (highest removal efficiency) were determined using analysis by a central composite experimental to get desirability operating variables (Zivorad, 2004).

Table 2 The Relationship Between Coded

Level and Corresponding Real Variables

Analysis of Experimental Results:

The response of experimental work conducted according to Box-Wilson design is represented by the mathematical model that described the removal efficiency (Y) of Cr (VI) from aqueous solution on STL which gave:

The results were analyzed to compute F-value for each term in polynomial equation then compared with the critical F-value. In this work (α) was chosen to be 0.05 (95% confidence) with 14 term to give the critical F-value = 4.60. It was seen the term of interaction between the variable X1X2 was insignificant. Thus, the best form of the equation representing the removal efficiency can be written as follows: Correlation coefficient (R2) = 0.9891 Variance Explained (S) = 97.839 % The optimum conditions for the four factors that give maximum adsorption capacity are: X1

*: optimum initial Cr (VI) concentration = 50 mg/l X2

*: optimum amount of adsorbent dose = 0.7625 g X3

*: optimum pH solution value = 3 X4

*: optimum time contact = 100 min Ymax.: max. removal percentage = 92.88 % RESULTS AND DISCUSSION:

Batch experiment were carried out to investigate the effects of initial Cr (VI) concentration, adsorbent dose, pH and contact time on percent removal of Cr (VI) from aqueous solution.

Effect of Initial Cr (VI) Concentration

Fig.1 shows the effect of Initial Cr (VI) concentration on the removal of Cr (VI) for different amount of adsorbent dose (0.05-1) g and at constant optimum value of pH and contact time, 3 and 100 min, respectively. The result indicated that the percentage of Cr (VI) removal increases with increasing the adsorbent dose. The high sorption at the initial concentration may be due to an increased number of vacant sites on the adsorbent

Coded

Level

Initial

Conc.

of Cr

(VI)

(mg/l)

Adsorb-

ent Dose

(g)

pH

value

Contact

Time

(min)

-2 50 0.05 1 20

-1 162.5 0.2875 3 60

0 275 0.525 5 100

1 387.5 0.7625 7 140

2 500 1 9 180

Y = 45.3513 + 0.005348 X1 + 86.35611 X2 –

0.3463 X3 + 0.27418 X4 – 0.00009 X12 –

40.1108 X22– 0.159375 X3

2 – 0.000789 X42

+0.018713X1X2+0.005556X1X3– 0.000306

X1X4–3.42105X2X3–0.105263X2X4+ 0.0125

X3X4 (5)

Y = 45.3513 + 0.005348 X1 + 86.35611 X2 –

0.3463 X3 + 0.27418 X4 – 0.00009 X12 –

40.1108 X22 – 0.159375 X3

2 – 0.000789 X42 +

0.005556 X1X3 – 0.000306 X1X4–3.42105

X2X3–0.105263X2X4+0.0125X3X4 (6)


1399

available at the initial stage. As concentration of initial Cr (VI) is increased there is a decrease in percentage removal of Cr (VI). This can be attributed to the accumulation of Cr (VI) particles on the surface of adsorbent. This observation is in a good agreement with the findings of Gupta and Babu (2009) on sawdust and Singh et al. (2009) on wheat bran.

Fig.2 shows the same effect but for different pH value and optimum conditions of adsorbent dose (0.7625 g), optimum contact time (100 min). Also, it can be seen that the percent of removal is favored at pH value near 3.

Fig.3 shows the same effect but for different contact time and optimum conditions of adsorbent dose (0.7625 g) and pH value (3). The results showed that more than 90% of the Cr (VI) was adsorbed within a period of 80 min. The maximum uptake of Cr (VI) was (47.98 mg/g) observed within 100 min.

Effect of Adsorbent Dosage

Fig.4 shows the effect of adsorbent dose on the removal of Cr (VI) for different initial Cr (VI) concentrations (50-500) mg/l and at constant optimum pH value of 3 and constant optimum contact time of 100 min. The result showed that the percentage removal of Cr (VI) increases with the increase in adsorbent dosage till optimal amount of adsorbent (0.7625 g), after this, the removal of Cr (VI) curves are smooth and continues leading to saturation. This can be attributed to increased adsorbent surface area and availability of more adsorption sites resulting from the increasing adsorbent dosage. This result agrees with the finding of most researchers such as Mohammad et al. (2011) and Hassan et al. (2012).

Fig.5 shows the same effect but for different pH value and optimum conditions of initial of Cr (VI) concentration (50) mg/l and optimum contact time (100 min). The result showed that the percentage of Cr (VI) removal increases with increasing the adsorbent dose.

Fig.6 shows the same effect but for different contact times and optimum conditions of initial Cr (VI) concentration (50) mg/l and optimum pH value (3). It can be, also, seen that increasing the contact time shall

increase the percentage removal of Cr (VI). Similar results were obtained by Cimino et al. (2000) and Ali (2010). The results also clearly indicated that the removal efficiency increases up to the optimum dosage beyond which the removal efficiency is negligible.

Effect of pH Value

The pH of the solution is an important parameter in the adsorption process because it affects the solubility of the metal ions concentration of the counter ions on the functional groups of the adsorbent (Surendra and Dharmendra, 2012). The effect of pH value on the percentage removal of Cr (VI) at optimum values of the others factor is shown in Figs. 7,8 and 9. From these figures it can be shown that the percentage removal of Cr (VI) decrease with increasing the pH value from 3 to 9. Chromium exists mostly in two oxidation states which are Cr (VI) and Cr (III) and the stability of these forms is dependent on the pH of the system (Cimino et al., 2000). It is well known that the dominant form of Cr (VI) at aforesaid pH is HCrO4

− which arises from the hydrolysis reaction of the dichromate ion (Cr2O7 -2) according to the equation:

2HCrO4− + H+ 2H2CrO4 (7)

2H2CrO4 Cr2O7-2 + 2H2O (8)

Increasing the pH will shift the concentration of HCrO4

− to Cr2O7-2. Maximum

adsorption at pH range 1 to 3 indicates that it was the HCrO4

− form of Cr (VI), which was the predominant species at this pH range and adsorbed preferentially on the adsorbents. Better adsorption capacity observed at low pH values with desired range of pH between 1 and 3 may be attributed to the large number of H+ ions present at these pH values, which in turn neutralize the negatively charged hydroxyl group (−OH) on adsorbed surface thereby reducing the hindrance to the diffusion of dichromate ions. At higher pH values, the reduction in adsorption may be possible due to abundance of OH− ions causing increased hindrance to diffusion of dichromate ions. Similar results were obtained by Bhattacharya et al. (2008) and Surendra and Dharmendra (2012).


1400

Effect of Contact Time:

The effect of contact time on the percentage removal of Cr (VI) at optimum conditions of the other factors is the presented in Figs. 10,11 and 12. These figures show that the removal of Cr (VI) increases with increasing mixing contact time attains equilibrium in 100 min that equilibrium time was independent of initial Cr (VI) concentration. . After this period the removal curves are single smooth and continuous, suggesting the formation of monolayer of adsorbate on the surface of the adsorbent. These results indicated that the sorption process can be considered very fast because of the large amount of Cr (VI) attached to the sorbent within the first 60 min of adsorption. The higher sorption rate at initial period can be attributed to the increase of number of vacant site on the adsorbent available at the initial stage. This result of equilibrium time is in agreement with those obtained by Surendra and Dharmendra (2012) on Spirogyra algae and Singh et al. (2009) on wheat bran were attained equilibrium time at 100 and 110 min, respectively.

ADSORPTION ISOTHERMS:

The experimental data are analyzed by the Langmuir and Freundlich equilibrium adsorption isotherm. The Langmuir isotherm was represented by the following equation (Davis et al., 2003):

eC.

lKoqoqeq1111

+= (9)

The linear plots of 1/qe versus 1/Ce suggest the applicability of the Langmuir isotherms Fig.13. The values of qo and Kl were determined from slope and intercepts of the plots and are presented in Table 3. The Langmuir constants qo and Kl are related to the adsorption capacity (amount of adsorb ate adsorbed per unit mass of the adsorbent to complete monolayer coverage) and energy of adsorption, respectively. The highest value of adsorption capacity qo (maximum uptake) was (47.98 mg/g). To confirm the favorability of the adsorption process, the essential

characteristics of the Langmuir isotherm may be expressed in terms of a dimensionless constant separation factor or equilibrium parameter (Weber and Chakraborti, 1974), RL, which is defined as:

olL CK

R+

=1

1 (10)

where Co is the initial concentration (mg/l) and KL is the Langmuir constant related to the energy of adsorption (l/mg). The value of RL indicates the shape of the isotherms to be either unfavorable (RL > 1), linear (RL = 1), favorable (0 < RL < 1) or irreversible (RL = 0) (McKay et al., 1982). The calculated RL values at different initial Cr (VI) concentration which is found in the range of 0.3945 and 0.06118 (0 < RL < 1) which confirms the favorable adsorption process for Cr (VI) removal using STL. The calculated RL values are shown in Fig.14. Also, higher RL values at lower ion concentrations showed that adsorption was more favorable at lower concentration.

The Freundlich equation is an empirical equation employed to describe heterogeneous systems, in which it is characterized by the heterogeneity factor 1/n and is as the following (Freundlich, 1906):

nefe C.Kq1

= (11)

where Kf and n are the Freundlich constants that indicate the adsorption capacity and intensity, respectively. The linear form of Freundlich model can be written as:

efe Clnn

Klnqln 1+= (12)

The values of Kf and n are evaluated from both intercept and slope, respectively, of the linear plot of the experimental data of ln qe versus ln Ce as illustrated in Fig. 15.The values of Kf and n given in the Table 3 show that the increase in negative charges on the adsorbent surface makes electrostatic forces between the adsorbent surface and Cr (VI) ion.


1401

Table 3 Parameters of Langmuir and Freundlich isotherm models

In general, R2 values, which are a

measure and describe of goodness fit adsorption data with Langmuir and Freundlich isotherm models. From Table 3, it can be seen that the values of correlation factor R2 is close to unity for Langmuir model which indicates that the experimental results on adsorption were fitted very well to the Langmuir adsorption isotherms than the Freundlich model.

Fig.16 shows the maximum adsorption capacity at different Cr (IV) concentration and optimum other factors.

REGENERATION STUDIES:

The saturated adsorbent which contains Cr (VI) is not safe for the disposal due to the stringent environmental constraints. It is important and appropriate to propose a method for the regeneration and reuse of adsorbent so as to reduce the load on environment in terms of disposal of polluted adsorbent as well as reduction cost of adsorption process. In the present study, after the optimized conditions for the removal of Cr (VI) were determined, STL was recovered by filtration and regenerated using 2N NaOH then washed by deionized water until the pH of the wash effluent stabilized near 7. Finally, it was dried in oven at 80 oC for 24 hr for reuse it to the removal of Cr (VI) at optimum conditions for initial Cr (VI) concentration, adsorbent dose, pH and contact time were 50 mg/l, 0.7625 g, 3 and 100 min respectively. Fig.17 shows the percentage removal of Cr (VI) using regenerated STL. The percentage removal of Cr (VI) obtained found decrease from 88.3 %

to 71.8 % after fourth regeneration cycle. The regenerated STL can be recycled effectively for the adsorption of Cr (VI) which makes the process cost effective.

The major problem of adsorption process is the disposal of the base solution obtained which contains high concentration of Cr (VI). One of the methods to tackle this problem is precipitation of Cr (VI) from the aqueous solution using barium chloride. Addition of barium chloride solution to Cr (VI) solution precipitates a bright yellow barium chromate, as given by the following reaction:

Ba2+

(aq) + CrO4 -2(aq) BaCrO4(s) (13)

The precipitated solid volume is very less as compared to the volume of the solution. Also, the chromium present in the complex solid can be recovered and reused by the industries. So this way the problem of disposal which is a major disadvantage of adsorption operation can be solved effectively and efficiently.

CONCLUSION:

The present work shows that STL is an efficient and more costly adsorbent for the removal of Cr (VI) from aqueous solution. The response surface methodology based on Box-Wilson design was used to develop mathematical model for predicting Cr (VI) removal by STL and employed to determine the optimal process parameters such as initial Cr (VI) concentration, STL dosage, pH and mixing contact time on the adsorption of Cr (VI) were obtain 50 mg/l, 0.7625 g, 3 and 100 min, respectively, the maximum percentage removal of Cr (VI) was 92.88% at optimum conditions. Analysis of variance showed a high correlation coefficient of determination value (R2= 0.9891), thus ensuring a satisfactory adjustment of the second order regression model with the experimental data. The results show that the adsorption of Cr (VI) onto the STL is highly pH dependent. Hence, adsorption of Cr (VI) is accomplished by increasing the pH value. The equilibrium adsorption data are tested with Langmuir and

Langmuir Constants

qo (mg/g) Kl (l/mg) R2 62.1041 0.03068 0.99978

Freundlich Constants

Kf (mg/g)/(mg/l)1/n 1/n R2

3.5629 0.5702 0.97473


1402

Freundlich isotherm models. The equilibrium data are best fitted with Langmuir than Freundlich isotherm model. The maximum adsorption capacity was 47.98 mg/g. The essential factor RL revealed the favorability of STL on Cr (VI) adsorption. The saturated adsorbent was regenerated by base treatment and can be reuse after four times with high efficient. The disposal of the base solution obtained that contains high concentration of Cr (VI) is the precipitation of Chromium as barium chromate.

REFERENCES:

- Aliabadi M., Morshedzadeh K. and Soheyli H., (2006), “Removal of Hexavalent Chromium from Aqueous Solution by Lignocellulosic Solid Wastes”, Int. J. Environ. Sci. Technol. 3, pp. 321–325.

- APHA, (1998), Standard Methods for the Examination of Water and Wastewater, 20th ed., APHA, AWWA, WPCF, Washington D.C., New York.

- Bailey. S., Olin T., Bricka R. and Adrian D., (1999), “A review of Potentially Low Cost Sorbents for Heavy Metals”, Water Research, 33, pp. 2469-2479.

- Baral S., Das S., Rath P. and Chaudhury R., (2007),“Chromium (VI) Removal by Calcined Bauxite”, Biochem. Eng. J., 34, pp. 69–75.

- Babu B. and Gupta S., (2008), “Adsorption of Cr (VI) Using Activated Neem Leaves as an Adsorbent: Kinetic Studies”, Adsorption, 14, pp. 85–92.

- Box, G. E. and Hunter, J. C., (1957), “Multifactor Experimental Design for Exploring Response Surface”, Ann. Math, vol.28, p.195.

- Bhattacharya A., Naiya T., Mandal S. and Das S., (2008), “Adsorption, Kinetics and Equilibrium Studies on Removal of Cr(VI) from Aqueous Solutions Using Different Low Cost Adsorbents”,

Chemical Engineering Journal, 137, pp. 529–541.

- Cimino G., Passerini A. and Toscano G., (2000), “Removal of Toxic Cations and Cr(VI) from Aqueous Solution by Hazelnut Shell”, Water Res. 34, pp. 2955–2962.

- Davis, T.A., Volesky, B., Mucci, A., (2003), “ A review of the Biochemistry of Heavy Metal Biosorption by Brown Algae”, Wat. Res. 37, pp. 4311-4330.

- Freundlich H., (1906), “Adsorption in Solution”, Phys.Chem. Soc., 40, pp. 1361–1368.

- Gupta S. and Babu B., (2009), “Removal of Toxic Metal Cr(VI) from Aqueous Solutions Using Sawdust as Adsorbent”, Chemical Engineering Journal, 150, pp. 352-365.

- Hassan M., Samaneh A. and Morteza B., (2012), “Removal of Chromium (VI) and Cadmium from Aqueous Solution Using Orange Peel as A Cheap Sorbent”, Journal of Applied Sciences in Environmental Sanitation Vol. 7, 4, pp. 269-273.

- Jianlong W., Xinmin Z. and Yi Q.,

(2002), “Removal of Cr (VI) from Aqueous Solution by Macroporous Resin Adsorption”, J. Environ. Sci. Health A 35, 1211–1230, 2000.

- Kozlowski C. and Walkowiak W., (2002),“Removal of Chromium (VI) from Aqueous Solutions by Polymer Inclusion Membranes”, Water Res., 36, pp. 4870–4876.

- Li C., Chen H. and Li Z., (2004),“Adsorptive Removal of Cr(VI) by Fe-Modified Steam Exploded Wheat Straw”, Proc. Biochem., 39, pp. 541–545.

- Mohan D., Singh K. and Singh V., (2005), “Removal of Hexavalent


1403

Chromium from Aqueous Solution Using Low Cost Activated Carbons Derived from Agricultural Waste Materials and Activated Carbon Fabric Cloth”, Ind. Eng. Chem. Res., 44, pp. 1027–1042.

- Mohan D. and Pittman C., (2006), “Activated Carbons and Low Cost Adsorbents for Remediation of Trivalent and Hexavalent Chromium from Water”, J. Hazard. Mater, B137, pp. 762–811.

- Montogomery, D.G., (1984), “Design and Analysis of Experiments”, 2nd ed., John Wiley.

- Mohammad R., Mina S. and Pourya B., (2011), “Removal of Cr (VI) from Aqueous solution Using Pine Needles Powder as Adsorbent”, Journal of Applied Sciences in Environmental Sanitation, vol. 6, No. 1, pp. 1-13.

- McKay, G., Blair, H. and Gardener J., (1982), “Adsorption of Dyes on Chitin. I. Equilibrium Studies”, J. Appl. Polym. Sci., 27, pp. 3043 – 3057.

- Roundhill D. and Koch H., (2002), “Methods and Techniques for the Selective Extraction and Recovery of

Oxo-Anions”, Chem. Soc. Rev., 31, pp. 60–67.

- Singh K., Hasan S., Talat M, Singh V. and Gangwar S., (2009), “Removal of Cr (VI) from aqueous solutions using wheat bran”, Chemical Engineering Journal, 151,113–121.

- Surendra B. and Dharmendra S., (2012), “Adsorption of Cr (VI) Metal Ion from Aqueous Solutions on Low Cost Adsorbent”, Journal Biointerface Research in Applied Chemistry, vol. 2, 2, pp. 284-290.

- Uysal M. and Irfan A., (2007), “Removal of Cr (VI) from Industrial Wastewaters by Adsorption”, Part I: Determination of optimum Condition, J. Hazard Mater, 149, pp. 482–491.

- Weber T. and Chakraborti R., (1974), “Pore and Solid Diffusion Models for Fixed Bed Adsorbents”, AIChE J., 20, pp. 228–238.

- Zivorad R., (2004), “Design of Experiments in Chemical Engineering”, Wiley-VCH, ISBN: 3.527.31142.4.


1404

Fig. 3 Effect of initial Cr (VI) concentration on

removal efficiency at different contact time Fig. 4 Effect of adsorbent dose on removal efficiency

at different initial Cr (VI) concentration

Fig. 5 Effect of adsorbent dose on removal efficiency at different pH

Fig. 6 Effect of adsorbent dose on removal efficiency at different contact time

Fig. 1 Effect of initial Cr (VI) concentration on removal efficiency at different adsorbent doses

Fig. 2 Effect of initial Cr (VI) concentration on removal efficiency at different pH


1405

Fig. 7 Effect of pH on removal efficiency at different initial Cr (VI) concentration

Fig. 8 Effect of pH on removal efficiency at different adsorbent dose

Fig. 9 Effect of pH on removal efficiency at different contact time

Fig. 10 Effect of contact time on removal efficiency at different initial Cr (VI) concentration

Fig. 11 Effect of contact time on removal efficiency at different adsorbent dose

Fig. 12 Effect of contact time on removal efficiency at different pH


1406

Fig. 13 Langmuir isotherm for Cr (VI) adsorption on STL

Fig. 15 Freundlich isotherm for Cr (VI) adsorption on STL

Fig 14 The Separation Factor for Cr (VI)

Fig. 16 Maximum uptake for Cr (VI) adsorption on STL at optimum conditions

Fig. 17 Regeneration STL adsorbent at optimum conditions


1407

Image Compression Using 3-D Two-Level Techniques

Zainab Ibraheem Abood Email: [email protected]

Academic Status: Asist. Teacher ABSTRUCT In this paper three techniques for image compression are implemented. The proposed techniques consist of three dimension (3-D) two level discrete wavelet transform (DWT), 3-D two level discrete multi-wavelet transform (DMWT) and 3-D two level hybrid (wavelet-multi-wavelet transform) technique. Daubechies and Haar are used in discrete wavelet transform and Critically Sampled preprocessing is used in discrete multi-wavelet transform. The aim is to maintain to increase the compression ratio (CR) with respect to increase the level of the transformation in case of 3-D transformation, so, the compression ratio is measured for each level. To get a good compression, the image data properties, were measured, such as, image entropy (He), percent root-mean-square difference (PRD %), energy retained (Er) and Peak Signal to Noise Ratio (PSNR). Based on testing results, a comparison between the three techniques is presented. CR in the three techniques is the same and has the largest value in the 2nd level of 3-D. The hybrid technique has the highest PSNR values in the 1st and 2nd level of 3-D and has the lowest values of (PRD %). so, the 3-D 2-level hybrid is the best technique for image compression.

KEYWORD: Image compression, 3-D two level wavelet transform, 3-D two level multi-wavelet transform, 3-D two level hybrid technique, Image data properties.

لثالثي االبعاد المستوى ثنائيةتقنيات ضغط الصورة باستخدام زينب ابراهيم عبود

الخالصةالثالثي –اسلوب التحويل الثنائي المستوى المقترحة تتضمن اتلضغط الصورة. التقني ثالث تقنياتتم في هذا البحث بناء

الثالثي االبعاد –واسلوب التحويل الثنائي المستوى Haar )( و (Daubechies) جة وقد تم هذا باستعمال ياالبعاد للموى استخدام االسلوب خروالتقنية اال )Critically Sampled preprocessingجة المتعددة وتم هذا باستعمال نوع ( يللمو

ظ على زيادة نسبة الضغط الهدف هنا هو الحفا الثالثي االبعاد. –الثنائي المستوى جة المتعددة)يالمو –جة ي(تحويل الموالهجينلكل مستوى من ي االبعاد لذلك تم هنا قياس نسبة الضغط بالصورةثبالصورة مع زيادة مستوى التحويل في حالة التحويل الثال

مستويات التحويل. ايضا تم قياس خواص بيانات الصورة وذلك للحصول على ضغط جيد للصورة ومن هذه الخواص ) ، النسبة المئوية للجذر التربيعي لمتوسط مربع الفرق بين الصورة االصلية والصورة entropy ( حةمقياس الطاقة المتا

ذروة) ، percent root-mean-square differenceنسبة الى متوسط مربع الصورة االصلية ( (قيمة الخطأ)المرجعة .(energy retained)والحفاظ على الطاقة ، )Peak Signal to Noise Ratio( فيها ه و التشويشالتشوي شارة نسبة الىاال

تقنية االسلوب الهجين للمستويين االول والثاني الثالثي تم مقارنة التقنيات الثالث وتبين أن اعتمادا على نتائج االختبار للجذر التربيعي لمتوسط مربع ئوية للنسبة المواقل قيم فيها ه و التشويشالتشوي شارة نسبة الىاال روةلذاالبعاد اعطى اعلى قيم

نسبة الضغط فانها لالفرق بين الصورة االصلية والصورة المرجعة نسبة الى متوسط مربع الصورة االصلية. أما بالنسبة

Zainab Ibraheem Abood Image Compression Using 3-D Two-Level Techniques

1408

الثالثي –االسلوب الهجين الثنائي المستوى ذلك فان لمتساوية للتقنيات الثالث واعلى قيمة لها في المستوى الثاني لثالثي االبعاد. االبعاد هو افضل تقنية لضغط الصورة.

الثالثي –جة، اسلوب التحويل الثنائي المستوى يالثالثي االبعاد للمو–ضغط الصورة، اسلوب التحويل الثنائي المستوى كلمات رئيسية:

.ورةخواص بيانات الص ،الثالثي االبعاد –اسلوب الهجين الثنائي المستوى جة المتعددة، ياالبعاد للمو

INTRODUCTION Image compression algorithms aim is to remove redundancy in data in a way which makes image reconstruction possible. This basically means that image compression algorithms try to exploit redundancies in the data; they recognize which data needs to be kept in order to reconstruct the original image and therefore which data can be ’thrown away’. By removing the redundant data, the image can be represented in a smaller number of bits, and hence can be compressed [Karen Lees, 2002], Related Works Talib M. Jawad Abbas presented two techniques for comparison. The first technique was the hybrid technique, which used Multi-walidlet. This technique is a combination of 2-dimentional Discrete Multi-wavelet Transform (DMWT) and Walidlet Transform, which converts the speech signal from (1-D) into two dimensional (2-D) forms. Next, the 2-D Multi-walidlet transform is applied to each 2-D signal. The second technique used 3D-(DMWT) on multi-walidelet coefficients matrices using GHM four multi-filters and using an over-sampled schema of preprocessing [Talib M. Jawad Abbas, 2008]. A method for image compression is described, in the wavelet transform technique the coefficients below a certain threshold are removed so a global threshold is used to improve the wavelet compression technique. The aim is to maintain the retained energy and to increase the compression ratio with respect to other global thresholds commonly used [Macarena Boix, 2010]. In order to develop an efficient compression scheme and to obtain better quality and higher

compression ratio using multi-wavelet transform and embedded coding of multi-wavelet coefficients through set partitioning in hierarchical trees algorithm (SPIHT) is used [ Muna F. Al-Sammaraie, 2011]. Different wavelets are used to carry out the transform of test image and the results analyzed according to the values of peak signal to noise ratio obtained and the computation time taken for decomposition and reconstruction [P.M.K. Prasad, 2012]. COMPRESSION USING WAVELET TRANSFORM: Wavelet analysis can be used to divide the information of an image into approximation and detail sub signals. The approximation sub signal shows the general trend of pixel values, and three detail sub signals show the vertical, horizontal and diagonal details or fast changes in the image [P.M.K. Prasad, 2012]. Discrete wavelet transform employs two sets of functions, called scaling functions and wavelet functions, which are associated with low pass and high filters, respectively. The first level decomposition mathematical expressions are [Tara Othman, 2006]:

(1)

(2)

In the decomposition level one, the image will be divided into 4 sub-bands, called LL, LH, HL, and HH. The LL sub-band is a low-resolution residue that has low frequency components, which are often referred to as the average image, LH provides vertical detailed images, HL provides detailed images in the


1409

horizontal direction, finally, the HH sub-band image gives details on the diagonal, while the LL sub-band is divided again at the time of decomposition at a higher level i.e. LL sub-band can be further decomposed into four sub-bands labeled as LL2, LH2, HL2, and HH2 as shown in Fig. (1). The process is repeated in accordance with the desired level. In this research, a 2- level decomposition is considered. In the discrete wavelet transform (DWT), there are properties for precise reconstruction. This nature gives a sense that in fact no information is lost after the transformed image is set to its original form. But there are missing information on image data compression with wavelet transform that occurs during quantization. Information loss due to compression should be minimized to keep the quality of the compression. A good quality compression is generally achieved in the process of memory consolidation, which generates a small reduction, and vice versa. The quality of an image is subjective and relative, depending on the observation of the user [P.M.K. Prasad, 2012]. COMPRESSION USING MULTI-WAVELET TRANSFORMS: Multi-wavelet possess more than one scaling function offer the possibility of superior performance and high degree of freedom for image processing applications, compared with scalar wavelets. Multi-wavelets can achieve better level of performance than scalar wavelets with similar computational complexity. In the case of nonlinear approximation with multi-wavelet basis, the multi-wavelet coefficients are effectively “reordered” according to how significant they are in reducing the approximation error [S. Esakkirajan, 2008]. The multi scaling function and the multi-wavelet function will satisfy matrix dilation as in the following equations [Muna F. Al-Sammaraie, 2011]:

(3)

(4)

The multi-wavelet used here have two channels, so there will be two sets of scaling coefficients and two sets of wavelet coefficients. Since, multiple iteration over the low - pass data is desired, the scaling coefficients for the two channels are stored together. Likewise, the wavelet coefficients for the two channels are also stored together. The multi-wavelet decomposition sub-bands are shown in Fig. (2). for multi-wavelets the L and H have subscripts denoting the channel which the data corresponds. For example, the sub-band labeled L1H2 corresponds to data from the second channel high pass filter in the horizontal direction and the first channel low pass filter in the vertical direction. This shows how a single level of decomposition is done. In practice, more than one decomposition performed on the image. Successive iterations are performed on the low pass coefficients from the previous stage to further reduce the number of low pass coefficients. Since the low pass coefficients contain most of the original signal energy, this iteration process yields better image compression [P.V.N.Reddy, 2011]. IMAGE DATA PROPERTIES: In order to make meaningful comparison of different image compression techniques, it is necessary to know the properties of the image. One property is the image entropy; a less details picture will have low entropy. For example a very low frequency, highly correlated image will be compressed well by many different techniques; one way of calculating entropy is suggested by:

(5)


1410

where G is the image grey-levels and the probability of grey-level k is p(k). The entropy also can be calculated using:

= Image Entropy (6)

where is the original image [Karen Lees, 2002]. The most common criterion to measure reconstructed image quality is the Percent Root-mean-square Difference (PRD % ). Let and be the original and reconstructed signals, respectively, N is the length of the window over which the metrics are calculated, and the coding noise [Carlos Bazán-Prieto, 2012]. PRD parameter as quality measurement, can mask the real performance of an algorithm since it depends on the mean value of the original signal [Manuel Blanco, 2005], it is given by:

×100 (7)

The PRD describes the error in terms of percentage of image energy which is useful to assess the impact of the error on the image. [Carlos Bazán-Prieto, 2012]: There are two things that can be used as benchmarks of compression quality, the Peak Signal to Noise Ratio (PSNR) and compression ratio (CR). PSNR is one of the parameters that can be used to quantify image quality. PSNR parameter is often used as a benchmark level of similarity between reconstructed images with the original image. A larger PSNR produces better image quality. PSNR equation is illustrated below [P.M.K. Prasad, 2012]:

(8)

where

(9)

where is the reconstructed image, m and n are the dimensions of the image. Compression ratio is the ratio of number of bits required to represent the data before compression to the number of bits required to represent data after compression. Increase of compression ratio causes more effective compression technique employed and vice versa [Nagamani .K, 2012]. To reach this goal, compression methods introduce certain, sometimes undesirable, effects such as the increase of computational complexity, processing delays, and coding noise or distortion. In order to quantify the effect of distortion, two objective metrics are used: PRD and Root Mean Square Error (RMSE) [Carlos Bazán-Prieto, 2012]. To analyze the efficiency of the compressor, one can use as a parameter, the energy retained:

(10)

where represents the compressed image [Macarena Boix, 2010]. THE PROPOSED TECHNIQUES BLOCK DIAGRAM: The proposed technique consists of three techniques applied to the image after image preprocessing step, these techniques are:

1. Three dimension (3-D) two-level discrete wavelet transform.

2. 3-D two-level discrete multi-wavelet transform.

3. 3-D two-level hybrid (wavelet- multi-wavelet transform) technique.

The transformation algorithm is applied in x, y direction then applied in the z-direction [Talib M. Jawad Abbas, 2008], i.e., a 2-D transform (1st and 2nd level) is applied in x, y direction then 1-D transform (1st and 2nd level) is applied in the z- direction. Fig. (3) shows the block diagram of 3-D two-level wavelet image compression, Fig. (4) shows the block diagram of 3-D two-level multi-wavelet image compression and Fig. (5) shows


1411

the block diagram of proposed 3-D two-level hybrid technique In this research, the parameters, PRD, PSNR and CR are calculated for all proposed techniques, i.e., for each first and second level in 3-D wavelet, multi-wavelet and hybrid techniques. IMAGE PREPROCESSING: The first step is to deal with the image using some of image processing techniques in order to prepare it to the next step. So the following steps must be followed:

1. Input the color or grey image of any size or format.

2. Convert the image to a grey-scale form (if it is color). By using matlab functions, one can reconstruct the color image.

3. Resize the image into a nearest square and to power of two in order to apply DWT or DMWT, i.e., the conditions of DWT or DMWT.

4. Some of images are resized to be of size (256*256) which is the nearest square and power of two to their original sizes, some of them are resized to be of size (512*512) while the other are resized to be of size (1024*1024).

COMPUTATION OF THE PROPOSED TECHNIQUE ALGORITHM: A: 3-D Two-Level Wavelet Transform: The following steps illustrate the computation of 3-D two-level wavelet transform: 1. For a general NxNxM, 3-D array, where N*N is the dimension of the image and M=4=no. of matrices, i.e.4 input images, apply a single level discrete 2-D wavelet transform using Daubechies wavelet transform for all matrices. 2. Apply a 2nd level 2-D DWT using Daubechies wavelet transform for each low –low sub-band of each matrix. 3. Apply a single level 1-D Haar DWT to each low-low sub band (those produced from step 2) in the z-direction. This can be done as follows: a. Construct a vector (v) containing four

elements (this number as the number of the matrices), v (1, 1) = [ ], where

are the first elements in each matrix. b. Construct a second vector containing four elements, v (1, 2) = [ ], where

are the elements in the position first raw and second column in each matrix. c. The same procedure continue till reach to the vector numbered (N/4*N/4), where (N/4*N/4) is equal to the size of the low-low sub-band that produced from step 2. d. Apply a single level 1-D Haar DWT to each vector. 4. Apply a 2nd level 1-D Haar DWT to the approximation coefficients vector of each vector that produced from step 3. 5. Finally, Take only the approximation coefficients vector that produced from step 4. B: 3-D Two - Level Multi-wavelet Transform: The following steps illustrate the computation of 3-D two-level multi-wavelet transform: 1. For a general NxNxM, 3-D array, where N*N is the dimension of the image and M=4=no. of matrices, i.e.4 input images, apply a single level critically sampled preprocessing 2-D DMWT for all matrices. 2. Apply a 2nd level critically sampled preprocessing 2-D DMWT for each low–low sub-band of each matrix. 3. Apply a single level 1-D DMWT to each low-low sub-band (those produced from step 2) in the z-direction. This can be done as follows: a. Construct a vector (u) containing four elements (this number as the number of the matrices), u(1,1) = [ ], where

are the first elements in each matrix. b. Construct a second vector which contain four elements, u (1, 2) = [ ], where

are the elements in the


1412

position first raw and second column in each matrix. c. The same procedure continue till reach to the vector numbered (N/4*N/4), where (N/4*N/4) is equal to the size of the low-low sub-band that produced from step 2. d. Apply a single level 1-D critically sampled preprocessing DMWT to each vector. 4. Apply a second level 1-D critically sampled preprocessing DMWT to the low sub-band that produced from step 3. 5. Finally, Take only the low sub-band that produced from step 4. C: 3-D Two - Level Hybrid Technique: The following steps illustrate the computation of 3-D two-level wavelet-multi-wavelet transform: 1. For a general NxNxM, 3-D array, where N*N is the dimension of the image and M=4=no. of matrices, i.e.4 input images, apply a single level discrete 2-D wavelet transform using Daubechies wavelet transform for all matrices. 2. Apply a 2nd level 2-D DWT using Daubechies wavelet transform for each low –low sub-band of each matrix. 3. Apply a single level 1-D DMWT to each low-low sub-band (those produced from step 2) in the z-direction. This can be done as follows: a. Construct a vector (w) containing four elements (this number as the number of the matrices), w(1,1)=[ ], where

are the first elements in each matrix. b. Construct a second vector which contain four elements, w (1, 2) = [ ], where

are the elements in the position first raw and second column in each matrix. c. The same procedure continue till reach to the vector numbered (N/4*N/4), where (N/4*N/4) is equal to the size of the low-low sub-band that produced from step 2. d. Apply a single level 1-D critically sampled preprocessing DMWT to each vector. 4. Apply a second level 1-D critically sampled preprocessing DMWT to the low sub-band that produced from step 3.

5. Finally, Take only the low sub-band that produced from step 4. An example test is applied to a standard image, “Lena”, of size 1024*1024 pixels. Figures 6, 7 and 8 illustrate the results of applying the proposed techniques on “Lena” image. TESTING AND EVALUATION OF RESULTS: There are five tables show the results measured for the proposed system when applied on the data base images. The bold values are the best results. For a comparison between 3-D one and two level wavelet transform, Haar, Db3, Db5, and Coiflet, the energy retained (Er), entropy (He), percent root-mean-square difference (PRD%) and peak signal to noise ratio (PSNR) from eq.’s 10, 6, 7 and 8 are measured and the results are shown as tabulated in table 1. Energy retained in the 2nd level of Db5 and Coiflet1 is better than that in the 1st level, but the entropy is good in the 2nd level for Db3, Db5, Coiflet1 and Haar (which is the best one) than that in the 1st level. Db5 is the best one in measuring PRD% (PRD% = 38.2121) and PSNR (PSNR=31.5256) which are the lowest and highest values, respectively. It can be conclude that, Haar and Db5 are compressed better than Db3 and Coifle1. In table 2, measurements of PRD% and PSNR for 2-D and 3-D one and two level wavelet transform for samples of (14) images, The PRD% in 3-D is half that of 2-D, that’s means that, when using 3-D, the error in terms of percentage of image energy is reduced to half of its value of 2-D. In the other side, the PSNR in 3-D is greater than that in 2-D. In 3-D, the second level PRD% and PSNR is better than that in first level, it can be conclude that, the image is compressed better when using 3-D tow level. Table 3 and 4, show 2-D and 3-D critical DMWT and hybrid results, in 3-D, the PRD% is quarter that of 2-D and the PSNR is greater than that of 2-D, also whenever the decomposition Levels increases the PSNR values are also increasing .


1413

Table 5 shows a comparison between the proposed 3-D two-level techniques, 3-D and 2-D one and two level techniques according to the values of PRD%, PSNR and CR. As shown from the results, in 2-D the results of wavelet and hybrid are the same for PRD%, PSNR and CR because the hybrid is 2-D wavelet in x and y direction and 1-D multi-wavelet in z-direction. in 3-D, CR values are 32:1 and 64:1 in the 1st and 2nd level respectively for wavelet, multi-wavelet and hybrid, while in the 2-D CR values are 4:1 and 16:1, also, PRD% in 3-D is less half the value in 2-D in 1st and 2nd level in wavelet and quarter that of 2-D in multi-wavelet and hybrid, hybrid technique has a largest values of PSNR (35.2721) in the 1st Level and (40.6906) in the 2ndLevel in 3-D. CONCLUSIONS: The techniques presented in this paper produce some of the best-reported results to date for a 3-D two-level discrete wavelet transform (DWT), discrete multi-wavelet transform (DMWT) and hybrid technique-based image compression and comparison between them. It is obvious from the results that any wavelet giving good results for decomposition will produce good results for advanced techniques being used for image compression. Based on testing results, it can be concluded that the hybrid technique has the highest PSNR value in the 1st and 2nd level of 3-D and has the lowest values of (PRD %). CR in the three techniques is the same and has the largest value in the 2nd level of 3-D. So, when the image is compressed to a high CR, then it is increasing the speed of computation and decreasing the wasting time, so, one can use it in authentication, recognition, using as a feature, etc...Therefore, the 3-D 2-level hybrid is the best technique for image compression. REFERENCES: “Arabic Speech Recognition Using Two Techniques Hybrid & 3D-Multiwavelet”, By Talib M. Jawad Abbas, Journal Al-Rafidain University College For Sciences ISSN:

16816870 Year: 2008 Issue: 22 Pages: 116-131 Publisher: Rafidain University College.

“Authentication using Wavelet and Multi-wavelet with Neural Network”, By Tara Othman ALsaraf, thesis, university of Sulayimani, 2006.

Evaluation of SPIHT Compression Scheme for Satellite Imageries Based on Statistical Parameters”, By Nagamani .K and A G Ananth, International Journal of Soft Computing and Engineering (IJSCE) ISSN: 2231-2307, Volume-2, Issue-2, May 2012, www.ivsl.org .

“Image Compression Using Multiwavelet and Multi-stage Vector Quantization”, By S. Esakkirajan, T. Veerakumar, V. Senthil Murugan and P. Navaneethan, International Journal of Information and Communication Engineering 4:4 2008. “Image Compression using Orthogonal Wavelets Viewed from Peak Signal to Noise Ratio and Computation Time”, By P.M.K. Prasad ,Prabhakar Telagarapu and G. Uma Madhuri, International Journal of Computer Applications (0975 – 888)Volume 47– No.4, June 2012, www.ivsl.org

“Image Compression using Wavelets”,By Karen Lees, thesis, May 2002, www.ivsl.org.

“Medical Images Compression using Modified SPIHT Algorithm and Multiwavelets Transformation”, By Muna F. Al-Sammaraie, Computer and Information Science Vol. 4, No. 6, Nov. 2011, www.ccsenet.org/cis , www.ivsl.org .

“Multiwavelet Based Texture Features for Content based Image Retrieval”, By P.V.N.Reddy and K.Satya Prasad, International Journal of Computer Applications (0975 – 8887), Volume 17– No.1, March 2011, www.ivsl.org. “On the use of PRD and CR parameters for ECG Compression”, By Manuel Blanco-Velasco, Fernando Cruz-Rold´an, J. Ignacio Godino-Llorente, Joaqu´ın Blanco-Velasco, Carlos

http://www.iasj.net/iasj?func=issues&jId=242&uiLanguage=en

http://www.iasj.net/iasj?func=issues&jId=242&uiLanguage=en

http://www.ivsl.org/



http://www.ccsenet.org/cis




1414

Armiens-Aparicio and Francisco L´opez-Ferreras, 2005, [email protected], www.elsevier.com/locate/medengphy. “Retained Energy-Based Coding for EEG Signals”, By Carlos Bazán-Prieto, Manuel Blanco-Velasco, Julián Cárdenas-Barrera and Fernando Cruz-Roldán, Medical Engineering and Physics 34 (2012),892–899, www.elsevier.com/locate/medengphy , www.ivsl.org . “Speech Recognition by Wavelet Analysis”, By Nitin Trivedi, Vikesh Kumar, Saurabh Singh, Sachin Ahuja and Raman Chadha, International Journal of Computer Applications, Vol. 15– No.8, February 2011

“Wavelet Transform Application to the Compression of Images”, By Macarena Boix, Begoña Cantó, “Mathematical and Computer Modeling, www.elsevier.com/locate/mcm ..

LIST OF ABBREVIATIONS: 1-D: One dimension 2-D: Two dimension 3-D: Three dimension DMWT: Discrete multi-wavelet transform DWT: Discrete wavelet transform CR: Compression ratio He: Entropy PRD: Percent root-mean-square difference MSE: mean square error Er : Energy retained SPIHT: Set Partitioning In Hierarchical Trees GHM Geronimo, Hardian, Masopute

LL

LH

HL

HH

LL2

LH2

LH HL2 HH2

HL

HH

Fig. 1: One and two level wavelet decomposition


http://www.elsevier.com/locate/medengphy

http://www.elsevier.com/locate/medengphy


http://www.elsevier.com/locate/mcm


1415

L1L1

L1L2

L1H1

L1H2

L2L1

L2L2

L2H1

L2H2

H1L1

H1L2

H1H1

H1H2

H2L1

H2L2

H2H1

H2H2

L1L1 L1L2

L2L1 L2L2

L1H1 L1H2

L2H1 L2H2

L1H1

L1H2

H1L1 H1L2

H2L1 H2L2

H1H1 H1H2

H2H1 H2H2

L2H1

L2H2

H1L1

H1L2

H1H1

H1H2

H2L1

H2L2

H2H1

H2H2

Fig. 2: One and two level multi-wavelet decomposition

Input 4 image Image

Preprocessing

Compressed image

1st level

2-D wavelet transform

Take low-low sub-band and apply

2nd level 2-D wavelet

transform

Convert to a 3-D array

2nd level


Take low-low sub-band and apply 1st

level 1-D wavelet transform in

z- direction

Fig. 3: Block diagram of 3-D two-level wavelet image compression


1416

Image Preprocessing

Compressed image

Input 4 image

1st level

2-D multi-wavelet transform


2nd level 2-D multi-wavelet

transform

Convert to a

3-D array


transform

Take low-low sub-band and apply 1st level 1-D multi-

wavelet transform in z-direction

Fig. 4: Block diagram of 3-D two-level multi-wavelet image compression

Image Preprocessing

Compressed image

Input 4 image

1st level



2nd level 2-D wavelet

transform

Convert to a

3-D array


transform

Take low-low sub-band and apply 1st level 1-D multi-

wavelet transform in z-direction

Fig. 5: Block diagram of the proposed 3-D two-level hybrid technique


1417

Original image

Image after 1st level 2-D wavelet transform

Image after 2nd level 2-D wavelet transform

Image after 3-D 2-level wavelet transform

Fig. 6: 3-D 2- level wavelet transform


1418

Original image

Image after 1st level 2-D multi- wavelet transform

Image after 2nd level 2-D multi-wavelet transform

Image after 1st level 1-D multi-wavelet transform


1419

Image after 2nd level 1-D multi-wavelet transform

Image after 3-D 2- level multi-wavelet transform

Fig.7: 3-D 2- level multi-wavelet transform

Original image

Image after 1st level 2-D wavelet transform


1420

Image after 2nd level 2-D wavelet transform

Image after 1st level 1-D multi-wavelet transform

Image after 2nd level 1-D multi- wavelet transform

Image after 3-D 2- level hybrid technique

Fig. 8: 3-D 2- level hybrid technique


1421

Table 1: 3-D Wavelet transform

Table 2: PRD% and PSNR for wavelet transform


Haar Db3 Db5 Coiflet1

1st Level 2nd Level 1st Level 2nd Level 1st Level 2nd Level 1st Level 2nd Level

Er 99.9578 99.8707 100.002 99.9897 100.0564 100.1867 99.9753 99.9874

En 1.0879 0.9080 1.1331 1.0471 1.1633 1.1443 1.0603 1.0036 PRD% 41.4125 41.4030 39.8878 40.4483 39.3609 38.2121 40.0527 40.8571 PSNR 30.5364 30.5714 30.8640 30.8321 30.981 31.5256 30.8291 30.7300

Wavelet

transform





1st Level 2nd Level 1st Level 2nd Level 1st Level 2ndLevel 1st Level 2nd Level

PRD% PRD% PRD% PRD% PSNR PSNR PSNR PSNR Image1 99.9906 99.9729 41.3898 41.3086 29.3574 26.3478 30.9957 31.0102 Image2 100.0001 99.9982 41.4170 41.4211 27.2455 24.2354 28.8816 28.8814 Image3 99.9997 100.0001 41.4246 41.4167 22.7343 19.7240 24.3685 24.3683 Image4 100.0010 100.0050 41.4219 41.4319 25.9887 22.9781 27.6240 27.6224 Image5 100.0098 100.0191 41.4330 41.4316 26.8338 23.8223 28.4661 28.4663 Image6 99.9996 99.9985 41.4215 41.4087 30.7860 27.7758 32.4207 32.4227 Image7 100.0464 100.0121 41.4425 41.3157 34.8994 31.8972 36.5336 36.5346 Image8 100.0072 100.0151 41.3917 41.3423 27.8306 24.8201 29.4711 29.4767 Image9 100.0683 100.1911 41.4273 41.5572 31.1759 28.1606 32.8273 32.8062

Image10 99.9242 99.7350 41.3840 41.5758 36.8591 33.8816 38.5224 38.4658 Image11 99.9889 99.9975 41.3996 41.4156 31.7132 28.7029 33.3540 33.3486 Image12 100.0033 100.0043 41.4159 41.4158 27.1797 24.1692 28.8158 28.8176 Image13 99.9905 99.9278 41.4550 41.3583 38.0153 35.0095 39.6446 39.6701 Image14 100.0000 100.0006 41.4091 41.3621 35.4521 32.4418 37.0899 37.1009


1422

Table 3: PRD% and PSNR for critical multi-wavelet transform

Table 4: Hybrid technique

Hybrid

technique

2-D multi-wavelet

3-D multi-wavelet

2-D multi-wavelet

3-D multi-wavelet

1st Level 2nd Level 1st Level 2nd Level 1st Level 2ndLevel 1stLevel 2ndLevel PRD% PRD% PRD% PRD% PSNR PSNR PSNR PSNR

Image1 99.9906 99.9729 28.7947 30.7630 29.3574 26.3478 34.1473 39.5286 Image2 100.0001 99.9982 28.5648 27.2455 30.2824 24.2354 32.1085 37.5328 Image3 99.9997 100.0001 28.6128 30.4313 22.7343 19.7240 27.5824 32.9832 Image4 100.0010 100.0050 28.5853 30.2768 25.9887 22.9781 30.8457 36.2836 Image5 100.0098 100.0191 28.5985 30.3223 26.8338 23.8223 31.6861 37.1110 Image6 99.9996 99.9985 28.5035 30.1603 30.7860 27.7758 35.6672 41.0985 Image7 100.0464 100.0121 27.7841 29.5763 34.8994 31.8972 40.0066 45.2617 Image8 100.0072 100.0151 28.7231 30.6488 27.8306 24.8201 32.6447 38.0319 Image9 100.0683 100.1911 28.6583 30.3581 31.1759 28.1606 36.0280 41.4989 Image10 99.9242 99.7350 29.5543 31.3428 36.8591 33.8816 41.4466 46.9716 Image11 99.9889 99.9975 28.9327 30.7174 31.7132 28.7029 36.4661 41.9254 Image12 100.0033 100.0043 28.5789 30.2839 27.1797 24.1692 32.0383 37.4706 Image13 99.9905 99.9278 28.8425 30.7758 35.0095 38.0153 42.7955 48.2112 Image14 100.0000 100.0006 28.4585 30.1674 35.4521 32.4418 40.3475 45.7599

Critical multi-

wavelet transform

2-D multi-wavelet

3-D multi-wavelet

2-D multi-wavelet

3-D multi-wavelet

1st Level 2nd Level 1st Level 2nd Level 1st Level 2ndLevel 1stLevel 2ndLevel PRD% PRD% PRD% PRD% PSNR PSNR PSNR PSNR

Image1 100.0161 100.0712 28.7822 30.7654 29.3552 26.3381 34.1453 39.5178 Image2 99.9873 99.9752 28.5843 30.2937 27.2466 24.2381 32.1046 37.5340 Image3 99.9932 100.0426 28.6053 30.4233 23.7349 29.7205 27.5830 32.9834 Image4 100.0619 100.0865 28.6006 30.2948 25.9834 22.9683 30.8344 36.2728 Image5 100.0029 100.0138 28.5816 30.3118 26.8344 23.8230 31.6916 37.1135 Image6 99.9990 99.9544 28.5092 30.1489 30.7861 27.7797 35.6675 41.1047 Image7 99.9302 99.8310 28.0404 29.9174 34.9095 31.9157 39.9389 45.2079 Image8 99.0109 99.9537 28.7686 30.6766 27.8303 24.8251 32.6330 38.0277 Image9 99.1084 100.0712 28.7854 30.4758 31.1724 28.1668 35.9802 41.4620 Image10 100.5874 99.3157 29.6639 31.2546 36.8016 33.7125 41.3269 46.9136 Image11 99.9985 99.2686 28.9320 30.7105 31.7124 28.5929 36.4080 41.8685 Image12 100.8408 99.9715 28.5796 30.3084 27.1073 24.1357 32.0035 37.4289 Image13 99.5162 99.5109 28.5047 31.0393 38.0566 35.0655 42.9431 48.1322 Image14 100.0046 100.0237 28.4475 30.1683 35.4517 32.4396 40.3497 45.7566


1423

Table 5: Comparison between three proposed techniques

Samples of Images from Data-Base:

1 2 3 4

5 6 7

Algor- ithm

2-D

3-D

1st Level 2ndLevel 1st Level

2ndLevel

PRD% PSNR CR PRD% PSNR CR PRD% PSNR CR PRD% PSNR CR

Wavelet

100.0021

30.4360

4:1

99.9912

27.6408

16:1

41.4166

32.0725

32:1

41.4115

32.0708

64:1

Multi-wavelet

99.9326

30.4987

4:1

99.8634

28.1228

16:1

28.6703

35.2578

32:1

30.4847

40.6659

64:1

Hybrid

100.0021

30.4360

4:1

99.9912

27.6408

16:1

28.6565

35.2721

32:1

30.2192

40.6906

64:1


1424

8 9 10

11 12 13 14


1425

Construction of a General-Purpose Infrastructure for Rfid – Based Applications

Assist. Lect. Basheera Mohammed Ridha Mahmood

Department of Computer Engineering College of Engineering

Baghdad University [email protected]

Lect. Mohammed Isam. Younis

Department of Computer Engineering College of Engineering

Baghdad University [email protected]

Assist. Prof. Hamid Mohammed Ali Department of Computer Engineering

College of Engineering Baghdad University

[email protected]

ABSTRACT:

The aim of advancements in technologies is to increase scientific development and get the overall human satisfaction and comfortability. One of the active research area in recent years that addresses the above mentioned issues, is the integration of radio frequency identification (RFID) technology into network-based systems. Even though, RFID is considered as a promising technology, it has some bleeding points. This paper identifies seven intertwined deficiencies, namely: remote setting, scalability, power saving, remote and concurrent tracking, reusability, automation, and continuity in work. This paper proposes the construction of a general purpose infrastructure for RFID-based applications (IRFID) to tackle these deficiencies. Finally, the proposed IRFID is compared against eight existing systems. As a result, IRFID can be considered as a prototype for the futuristic with flexibility and generality in a wide-range of automation and development areas. Keywords: RFID. Remotely Setting. Scalability. Automation. Power Saving. Continuity. Backup.

تشييد البنية التحتية للتطبيقات العامة لنظام التعريف باستخذام الموجات الراديوية حامد محمد عليأ.م.

قسم هندسة الحاسبات كلية الهندسة –جامعة بغداد

د. محمد عصام يونسم. قسم هندسة الحاسبات

كلية الهندسة –جامعة بغداد

محمود بشيرة محمدرضام.م قسم هندسة الحاسبات

كلية الهندسة –جامعة بغداد

لخالصة:ا

ع ظهور عصر االتمتة اصبح الهدف من التحديث في التقنيات هو زيادة التطور العلمي وبالتالي تحقيق الرضا العام و توفير مااللية التي لها القدرة على التحكم بمتطلبات الراحة للبشرية مع عصرنة الحياة. ونتيجة لذلك ،اصبح لزاماً ايجاد و تطوير االنظمة

ان من اهم مجاالت البحوث النشطة في السنوات .العمل االلي والتي تقوم بجمع المعلومات تلقائيا لتتفاعل مع العالم بشكل فعلي) في انظمة الشبكات. RFIDاالخيرة لمعالجة القضايا المذكورة اعاله، هو دمج تقنية تحديد الهوية باستعمال الترددات الراديوية (

هذا البحث يسلط الضوء على أن وعلى الرغم من ان هذه التقنية تعتبر من التقنيات الواعدة، اال انه يوجد بعض نقاط الضعف فيها. ذلك، وللقيام بوما تم التوصل اليه في االنظمة الموجودة حاليا و التي تم ادراجها في البحث. مستويات التطور احدث المستجدات في

فان هذا البحث يعرف سبعة اوجه من اشكال القصور وهي: الضبط عن بعد، قابلية التوسع، توفير الطاقة، التتبع عن بعد بشكل وبالتصويب لالعمال السابقة وتطويرها، يقترح هذا البحث متزامن، إعادة االستخدام، التشغيل اآللي، واالستمرارية في العمل.

المقترح IRFIDان . ) IRFID( عامة للتطبيقات المعتمدة على التعريف االلي باستعمال الترددات الراديويةتشييد البنية التحتية الالمقترح من IRFIDتمت مقارنته مع ثمانية انظمة قائمة. ومن خالل تقييم مزايا و عيوب هذه االنظمة، فلقد تم عرض مزايا الـ

نموذجا مستقبليا النظمة مستقبلية تملك المرونة الالزمة للعمل ضمن يعتبر IRFID . وهكذا، فان الـخالل قائمة التدقيق المجدولة مجاالت االتمتة وتطويرها و ضمن نطاق واسع.

تحديد الهوية باستعمال الموجات الراديوية، الضبط عن بعد، قابلية التوسع، التشغيل االلي، توفير الطاقة، االستمرارية : الكلمات الرئيسية الدعم.في العمل،


Basheera Mohammed Ridha Construction of a General-Purpose Infrastructure Mohammed Isam. Younis for Rfid – Based Applications Hamid Mohammed Ali

1426

1. INTRODUCTION As a result of the increasing complexity, and modernization of life, it is necessary to develop and enhance the life styles in all of its branches. Many branches appear to fulfill the demands of modernizing and digitizing like the Information and Communication Technology (ICT), distributed system, networking, automated system and automatic identification technology (Ali M. F. M. et al, 2010). In the new world, computer networks gave businesses the ability to cleverly handle vast amounts of information and the Internet gave them the resources to share them. However, computers remain unable to interact with the real world and gather information automatically, without human intervention which is subject to human error and waste of human's resources. Many technologies in the Auto-ID field have appeared and many researchers nowadays focus on RFID technology combined with computer networks to make one of the core technologies for realizing ubiquitous computing identification (David S. and Cobain C, 2002, Byun Y.C. et al, 2009). 1.1 RFID Technology

RFID is an electronic and wireless sensor technology which is based on the detection of electromagnetic signals (Mishra D. et al, 2012). The use of RFID tag which identifies an object and thing will be expanded with the growth of ubiquitous society; therefore, it is essential to construct an RFID network system as a social infrastructure (Kuribayashi S. and Osana Y., 2010). Recently, the hierarchical organization in any network (e.g. RFID network) has been highly recommended to make a centralized control possible to the entire system to improve network management (Wang Y., 2010). 1.2 Components of RFID System

RFID System components can be divided into four parts: first part is the RFID tag that is a small device which can store and transmit data in a contactless manner (wirelessly) using radio waves. Most common tags today consist firstly of an Integrated Circuit (IC) with memory and secondly of an antenna to communicate with the reader through this antenna [Ali M. F. M. et al, 2010, Al-Tameemi Z. F. A., 2011, Manish

B. and Shahram M., 2005), the second part is the RFID reader (interrogator) that is a device gathers information from tags and sends data to the host computers which have the software application (Ali M. F. M. et al, 2010). The third part is the antenna which is the conduits for data communication between the reader and the tag. It sends wireless signals from the reader to the tags, and receives wireless information from the tags in the coverage zone of the reader. The last part is the host computer, the data acquired from the tags must be put for practical use, and this is done by using a host computer to process the data. The arrangement of all components of an RFID system is shown in Figure (1).

2- DEVELOPING RFID BASED

APPLICATIONS ISSUES Due to the complexity and diversity of RFID network's resources, the design and implementation of recourses manageability and usability of RFID network system is a complex and a challenging process (Abed al Hussain S.H., 2010, Su X. et al, 2008). However, many intertwined deficiencies appeared in the implementation of RFID systems. These deficiencies are scheduled as follows: 1. Remotely setting: It is the ability to remotely make the configuration setting of RFID network system parameters by a central monitoring and management computer. If RFID networks are set up in remote geographical areas, central monitoring and management will be important (Roberti M., 2007). 2. Scalability: It is a desirable attribute of a network, system, or process that gives a rich system performance and robot capability without the need to duplicate or reengineer the system. This accommodates an increasing of objects and elements number to handle the growing volume of work gracefully (Bondi A. B., 2000). These attribute features are stated as follows: a- Networking and management: It is a set of technologies that enable immediate, automatic identification, monitoring, devices configuration, deployment, initialization, control of receivers and transponders and sharing of information on items in the RFID application. By this way, the RFID network will make organizations more effective by enabling true visibility of information about items


1427

in the application (Ham Y. H., 2005, Ismail M. N. and Zin A. M., 2010).

b- Structural extendibility: The system is said to be structurally scalable if its standards or implementation do not impede the growth of object numbers [(Bondi A. B., 2000).

3- Remote and Concurrent Tracking: Remote tracking is the ability of tracking the object situation remotely (Al-Tameemi Z. F. A., 2011); whilst the concurrent tracking of any system is the ability of concurrently tracking the activities of a number of individuals in a specified unit of time and position (Aguzzi J. et al, 2011).

4- Reusability: It is the ability to use all or a greater part of the same programming code or system design in another application, thus avoiding time wastage and reducing the cost (Al-Tameemi Z. F. A., 2011).

5- Automation: This feature includes supporting a fully automated start and a termination of every executed acquisition process (Al-Tameemi Z. F. A., 2011).

6- Remote power saving: Power save uses intelligent energy management to ensure that the system is available when system resources are required, and conserves power during productivity downtimes. Sometimes it needs a centralized system power status control (Faronics Corporation, 2009).

7- Continuity in work: This feature includes supporting the continuity of work in case of failure.

It must be mentioned that RFID network needs some other feature such as reliability of reaching data in the network, economical, information availability, universal timing and green system. 2. RELATED WORK

This section gives the art-of-the-practice of the up-to-date RFID-based systems. Qaiser A. and Khan S.A. 2006 (Qaiser A. and Khan S.A., 2006) proposed an improvement in the university based on RFID technology. A system is implemented for the Automation of Time and Attendance using RFID Systems (ATAS). Lim T.S. et al. 2009 (Lim T.S., et al, 2009) proposed a system called RFID Based Attendance System (BAS). BAS takes the attendance of students in school, college, and university. Hornback G. et al. 2010 (Hornback G., 2010) proposed an Automatic Attendance System using RFID network called (AAS). AAS takes the attendance of students in a classroom by using passive RFID reader. Abdul hussain S. H. 2010

(Abed al Hussain S.H., 2010) proposed a simulation protocol for Vehicle Tracking System (VTS) using a passive RFID technology. Ali M. F.M. et al. 2010 (Ali M. F. M. et al, 2010) proposed a reusable RFID Tracking & Monitoring Application Programmable Interface (RFIDTMAPI)

for heterogeneous RFID environment. Al-Tameemi Z. F. A. 2011 (Al-Tameemi Z. F. A., 2011) proposed a Scalable and Automated RFID-based Attendance System (SAAS). Khor J.H. et al. 2012 (Khor J. et al, 2012) presented the potential of using Electronic Product Code (EPC) Class-1 Generation-2 RFID-based Malaysian University

Communities (RFIDBMUC). RFIDBMUS is used in an automate data management system for different applications. Younis M.I. 2012 (Younis M.I., 2012) proposed a Smart Library Management System (SLMS). SLMS integrates the passive RFID technology into a library management system.

4. HIERARCHICAL ORGANIZATION OF THE IRFID The IRFID has a hierarchical organization as a tree topology with four levels as shown in Figure (2). The four levels are defined briefly as follows: Level 1: It is the top level of hierarchy. This level constitutes the Main Management Administrator (MMA), which in turn consists of the following devices: main management controller server, backup server, database server and web server. Level 2: This level represents Edge Management Controller (s) (EMCs), which is a computer or a number of computers that represents client(s). Level 3: This level represents RFID readers. They are distributed to cover a special region as desired. Level 4: This level is considered as a leaf of the tree (i.e. the low level of the topology); it represents all tag's ID used in the infrastructure. The hierarchical relationships of controlling and management from top to bottom are shown in Figure (2). 5. ARCHITECTURAL DESIGN OF THE IRFID This section focuses on the components of the infrastructure and unifies them as follows:

5.1 Main Management Administration (MMA) MMA is the top level of the proposed architecture topology. It consists of several equipments as follows:


1428

a. Main Management Control Server (MMC) It is the main management control server in the system that categorizes the managerial functions to the level beneath it. Management functions include organizing processes of planning, directing and controlling operations to meet the goal of the infrastructure. MMC handles the management functions by using specific commands. These commands are listed as follows: System initialization, Clock synchronization (using universal time clock), Remote setting of EMCs and readers parameters, Acquisition of EMCs parameters and readers parameters, Acquisition of the number of tag IDs and send poll command.MMC controls the above actions automatically and sends commands to EMC(s) and RFID reader(s). The Schematic Diagram of IRFID is shown in Figure (3).

b- Database Server The main database contains tables. Each table contains fields; therefore, the role of this server is recording data that comes from MMC, EMCs, and RFID readers in tables. For any event, data will be recorded with its time. Also, it makes updating on the stored information and any authorized user can read from it. The stored tables of database can be classified into five tables as follows: System events table: this table records any event that occurred in the system with its time, EMC parameters information table: this table records the parameters of EMCs, reader parameters information table: this table records the parameters of RFID readers, tag information table: this table contains the information regarding the identification objects, tag ID table: this table contains EMC name, reader name, and tag ID associates to that tag, c- MMC Backup Server

The backup server is a copy of MMC server. It will work instead of MMC when the MMC is fault. It has all facilities of the MMC server; thereby, it gives continuous working to the system.

d- Web Server

To give remote information access and reporting of the data that exists in the database as web pages, a Web server is made as one of the elements of the proposed IRFID. Any authorized

person has the right to login into the system and access to the data. Only the administrator has ability to update or remove information from the tables. The Web Server helps to transport accessed content through the internet.

5.2 Edge Management Controller (EMC) The EMC job is briefly defined as follows:

1- Response to the initialization command that comes from MMC, gathers its own initial parameters and the initial parameters of readers and then sends them to the MMC. 2- Response to the clock synchronization command comes from MMC. So, synchronization will be done between EMC and MMC for purpose of clock synchronization depending on NTP (Network Time Protocol) in addition to using universal time server. 3- Taking the RFID readers information, which is gathered automatically from the tags and sending this information to the main database to store it with its time. 4- Sending acquired parameters to MMC in an acquisition process. 5- Setting its parameters and the readers parameters when a command arrives from the MMC. The parameters of EMC and RFID reader are listed as follows:

EMC parameters are listed as: EMC name, EMC ID, EMC password, EMC Time Interval for Auto Acquisition (TIAA, Number of Tag ID for Auto Acquisition (QTYAA), EMC mode: this parameter defines the ways of sending Tag IDs from EMC to the main database, which are automatically continuous, automatically with a certain time interval, automatically with a certain number of tags, and poll mode. The poll mode means that the collected tag IDs can not be sent from EMC to the main database unless MMC sends a poll command to that EMC to sending IDs. RFID reader parameters are listed as: Reader name, Reader protocol, Reader mode: this parameter determines the modes of reading tag ID which are; continuous mode and poll mode, Reader frequency, Reader power, Number of slots, Reader type, Reader kill password. 5.3 RFID Readers (Interrogators)

It is the third level of the proposed infrastructure. Each group of readers is distributed in a special manner to cover a dedicated area; these


1429

interrogators receive embedded information from tags and send this information to the responsible

EMC. On the other hand, they take their setting from the EMC.

5.4 RFID Tag (Transponder) In the proposed infrastructure, every physical object will be identified by an RFID tag attached to it. The tag ID is embedded inside the tag, and all other information related to that object is stored in the main database (tag information table).

6. DESIGN THE CORE MODULES OF IRFID The proposed IRFID can be described as a set of interconnected structural modules that provide a framework supporting the entire structure of an RFID network based applications. These modules are: Main module, Main Database module, Web server module, MMC Backup server module, Communication modules, Utility module and Application module as shown in Figure (4). Each of these modules has its own functionalities in supporting the IRFID.

The key module is the Main module that has its own sub modules inside it. These sub modules are shown in Figure (5). The modules of EMC module are similar to the sub modules of main module but there are additional modules which are: read tag ID, EMC database and RFID reader communication modules as shown in Figure (6). Since backup server is considered as a copy of the MMC server, the modules of the whole IRFID that work inside the backup server are similar to the modules in the MMC server. But it has two additional modules to support its work. The additional modules are: MMC Monitor Module and Database MMC Copier Module as shown in Figure (7).

7. MAPPING IRFID REQUIREMENTS TO CORE MODULES Summing up, the entire modules of the IRFID can be classified according to their interrelated work with the server and their functionalities into three types:

• MMC server modules (Figure 4). • Backup server modules (Figure 7). • EMC modules (Figure 6).

7.1 MMC server modules

The entire modules of IRFID that work with MMC server are explained as follows:

• Main module

The Main Module is responsible for making remote setting, initialization, synchronization, acquisition, logging and communication with other modules. It has the following sub modules:

1- GUI Module: This module is responsible for the interfaces between the user and the IRFID. Usually, any command is excuted via GUI module.

2- Job Assignment Module: This module receives the commands and data that arrive from the GUI or the communication modules and then classifies this information to the intended modules to perform the required job.

3- Initialization Module: This module is responsible for starting the communication process between the MMC and EMC.

4- Clock Synchronization Module: This module is responsible for clock synchronization between the MMC and EMC(s).

5- EMC Setting /Acquire Parameters Module: - this module is responsible for performing two actions. The first action is the remote setting of the EMC's parameters. The second action is acquiring the EMC data automatically.

6- Log Module: This module provides the flexibility of displaying any event that happened in the system with its time.

7- Reader Setting/Acquire Parameters Module: This module is responsible for two actions. The first action is the remote setting of the parameters of RFID reader. The second action is acquiring the RFID reader data automatically.

8- Acquisition Module: This module does two actions. The first action is to send an acquire command to send tag IDs to main database (in the case of EMC is in Poll mode, by using the Poll command from the MMC). The second action is asking about the quantity of tag IDs in case a specific application needs this information.

9- MMC Database module: This module is resposible for transferring data from MMC to main database.

• Web Server Module This module is responsible for delivering pages

content that can be accessed through network by an authorized user; therefore, any authorized user


1430

has the ability to make a remote tracking and a concurrent tracking of objects.


1431

• Main Database Module This module represents the main database used in IRFID and contains five tables. It is connected to MMC database module, EMC database module, Web server module, MMC Backup communication module and application module. This module is responsible for storing tables of data, displaying and providing a copy of these tables by opening a connection with other modules. In this module, the updating of tables occurs automatically.

• Utility Module

This module acts as a tool to support specific functions in the proposed IRFID. These functions involve universal time synchronization, finding available com port and checking the availability of network devices in the IRFID.

• Application Module This module represents the application software. The

IRFID is supposed to handle multiple applications. Hence, the reusability could be achieved.

• MMC Backup Module This module works on the basis of a message called

the heartbeat messages. This message is sent to the backup server to inform it that the main server is still working. Therefore, if anything happened to the MMC server the backup server is working instead of it.

• Communication Module

This module is responsible for the communication among the modules.

7.2 Backup server modules

The entire modules of IRFID that work with the backup server are similar to the modules of IRFID that work with the MMC server; in addition, there are two other modules:

• MMC Monitor Module

This module operates on the basis of heartbeat messages which are sent from MMC server to the backup server. Receiving these messages by this module to ensure that the MMC server is still working; therefore, the continuity of work will be accomplished due to the fact that the Backup server will operate instead of the MMC.

• Database MMC Copier Module

This module operates periodically and automatically to get an updated copy of database from the main database and store it in the backup server. Therefore, there is a continuous data backup.

7.3 EMC Modules EMC modules are similar to the sub modules of the main modules; in addition there are other modules:

• 0BRead Tag ID Module The functionality of this module is to deliver the

information arrived from RFID readers to the job assignment module. This module works automatically.

• RFID Communication Module This module is responsible for the communication

between EMC(s) and RFID readers.

• EMC Database Module This module is responsible for transferring Tag IDs to

the main database directly through a dedicated communication module.

8. THE IMPLEMENTATION OF IRFID

These sections give the implementation of the proposed IRFID. In doing so, these sections explain the implementation of RFID tags (Level 4), RFID Reader (Level 3), and finally the GUI activities for both EMC (Level 2) and MMA (Level 1).

8.1 RFID Equipment The RFID equipments needed by the IRFID are as follows:

• RFID Tags (Level 4)

The tags that are used in the proposed system are passive RFID tags which are programmed to store the identities of the system related objects. Every tag represents single object, which is programmed by the RFID reader/writer device to contain one unique identity. The shapes of tags used in the implementation are shown in Figure (8).

• RFID Reader (Level 3) The type of RFID reader used in the implementation is a passive tagsense Nano-UHF RFID Reader; it is a very smal, low-power and low-cost RFID reader. Figure (8) shows tagSense Nano-UHF RFID Reader. For hardware installation, tagsense reader is

connected with the other components. The power


1432

connection is connected to an adapter with a voltage of 3.3 V (regulated). The RFID reader is connected to the PC (EMC) using a simple 3-wire serial technique; the 3 data lines are: Ground, Data Transmit (TX) and Data receive (RX). The RS-232 serial adapter includes a DB-9 connector which is attached to a standard serial cable (Tag Sense, Inc, 2013) as shown in Figure (8).

• Antenna

Tag sense half patch antenna are used in the implementation and the reader connected via Sub Miniature version A (SMA) connector.

8.2 The Implementation of IRFID The MMC which is the main management controller server must have the ability to control and monitor almost all the stations beneath its level, so the capability of MMC is implemented by using GUI which is the user-visible interface. The main program window of the MMC is shown in Figure (9). EMC is the edge management controller that is making control on the levels beneath it. The GUI of the EMC is shown in the Figure (10). When backup server starts working, the window shown in Figure (11) will appear. The IP address of MMC must be set in its filed in the first time, and when the connection is established between MMC server and backup server the heartbeat message begins transmitting from MMC to the backup and the counter shown in Figure (11) begins counting every 10 seconds. All data is organized and displayed faster by using MySQL databases. MySQL is used because it is a relational database system, fast, supporting large databases, and customizable (Glass M. et al, 2004). In order to create a new record in a specific table, login as an administrator user account with the appropriate privileges is needed. The tables that are created for the proposed system are shown in Tables 1, 2, 3, 4 and 5. Apache Web Server is responsible for accepting HTTP requests from clients and serving them as HTTP responses, usually in the form of web pages containing table's content of the main database. The language used for designing webpage is PHP that used to create web pages. The types of interface used between EMC and RFID reader are serial (RS232).

9. APPLYING SYSTEM FEATURES The applying of each property will be demonstrated in the following sections. Applying of Remotely Setting is accomplished through using a remote main management control server (MMC), which controls every device setting (EMCs, RFID Readers) from remote distance, by installing the

proposed software in MMC. Applying of Scalability is implemented by using multiple PCs (MMC, multiple number of EMCs) and RFID readers also the system expand in a chosen dimension without major modifications to its software. Applying of Automation is efficiently improved, throughout controls the automatic initialization and terminates acquisition processes also the automated dealing with RFID equipment and MySQL commands, updating the database accordingly and website updating. Applying of Remotely Power Saving is done by making MMC controls the mode of reading tag ID. If the RFID reader is in polling mode (i.e. power saving mode or standby mode), it will automatically turn off the power to all its radio circuitry when it is not transmitting. Applying of Reusability is achieved by implementing two applications practically as case studies for small scale applications .The case study can be implemented by installing only small external software in the user's PC, and when this software runs, the GUI in Figure (12) will appear. As seen in Figure (12), two applications can be selected, and each one represents a case study: Case Study one (Tracking System): In the College of Engineering / the University of Baghdad the proposed system can be implemented for tracking the employees. The MMA can be installed in the administrative building of the College of Engineering and 32 EMCs can be distributed in the College of Engineering departments to cover the area of the college. Therefore, the RFID readers can be distributed in every (entry, exit) gate in all partitions as shown in Figure (13). every employee has a tag ID which contains the ID of this person. All other information related to this person is stored earlier in the main database. The illustration of tracking will be displayed on the map and on the GUI fields as shown in Figure (14). As a proof of system reusability, a second case study (Attendance System) was implemented on a small scale application to take the attendance of employees, and by the same manner. The application can be implemented in the college's department and the same EMC used in the previous application can be used here, but the readers are putting in the staff member room. The GUI of this application is shown in Figure (15). Applying of Remotely and Concurrent Tracking is done by using website dedicated for this purpose; the generated tables will be browsed. Applying of Continuity in work is achieved through the using of a backup server. Also, there is a continuous backup of data by taking a copy of data from the main database and storing this data in the backup server.


1433

10. COMPARISON BETWEEN IRFID AND RELATED WORK

As a summary of the foregoing features evaluation, Table 6 demonstrates the main functionalities of the IRFID compared with related work discussed before.

11. CONCLUSIONS AND FUTURE WORK This paper presented a fully scalable, automated, reusable, remote setting, power saving RFID infrastructure system called IRFID. Existing RFID systems are embedded in this research, in which the investigations involved evaluating the performance of each one of these systems in terms of chosen characteristics. There are some points that might be taken into consideration in further work such as using other types of connections such as, Ethernet, and/or Wireless connection using towers, and/or VSAT (very-small-aperture terminal) the extendibility feature can be improved, on the other hand, it is highly desired to integrate other modules into the IRFID like LF, HF, and active RFID. Also the security feature and mobile module must add in the infrastructure to serve mobile devices facilities and increase the ability of IRFID to use in many applications. 12. REFERENCES Abed al Hussain S.H., "Design and Simulation of RFID Reader Protocol for Vehicle Tracking System for University of Baghdad-Aljaderia Campus", Proceeding of 2nd Information Technology Conference, University of Technology, Baghdad, Iraq, pp. 142-158, April 2010. Aguzzi J., Sbragaglia V., Sarria D., Garcia J.A., Costa C., Río J. d., Manuel A., Menesatti P. and Sarda F.,"A New Laboratory Radio Frequency Identification (RFID) System for Behavioural Tracking of Marine Organisms", Sensors, Vol. 11, No. 10, pp. 9532-9548, 2011. Ali M. F. M., Younis M. I., Zamli, K. Z. and Ismail, W., "Development of Java based RFID Application Programmable Interface for Heterogeneous RFID System", The Journal of Systems and Software, Vol. 83, No. 11, pp. 2322-2331, 2010. Al-Tameemi Z. F. A., "Design and Implementation of a Scalable Automated RFID based Attendance System with Scheduling Technique", M.Sc. Thesis,

School of Electrical and Electronic Engineering, University of Science Malaysia, 2011. Bondi A. B. , "Characteristics of Scalability and Their Impact on Performance", Proceedings of the 2nd international workshop on Software and performance (WOSP '00), Ottawa, Canada, pp. 195-203, September 2000. Byun Y.C. , Byun J.W., Byun S.Y. and Gerardo B. D.," ALE-compliant RFID Middleware for Mobile Environment", Proceedings of the 10th IEEE International Conference on Software Engineering, Artificial Intelligences, Networking and Parallel/Distributed Computing (SNPD '09), Daegu ,South Korea, pp. 606-611, May 2009. David S. and Cobain C., "The New Network", Auto-ID Center, White paper, Available at: http://www.ifm.eng.cam.ac.uk/automation/documents/centerguide.pdf. Glass M., Scouarnec Y. L., Naramore E., Mailer G., Stolz J. and Gerner J., "Beginning PHP, Apache, MySQL Web Development", First Edition, Wiley Publishing, Inc, 2004. Ham Y. H., Kim N. , Pyo C. and Chung J. W., "A Study on Establishment of Secure RFID Network Using DNS Security Extension", Proceeding of IEEE Conference on Communications, Perth, Western Australia, pp. 525-529, October, 2005. Hornback G., Babu A., Martin B., Zoghi B., Pappu M. and Singhal R., "Automatic Attendance System", Technical Report, RFID Sens Net Lab, Smart Distributed System Group, 2010. Ismail M. N. and Zin A. M., "Network Traffic and Utilization: Reliability Of Network Analyzer Development With Independent Data, OPNET Simulation Tool and Real Network", Annals Computer Science Series, Vol. 8, No. 1, pp. 109- 136, 2010. Khor J., Ismail W., Kamarulazizi K. and Rahman M. G., "EPC Class-1 Generation-2 Radio Frequency Identification (RFID) -based Malaysian University Communities", Scientific Research and Essays, Vol. 7, No.8, pp. 852- 864, 2012. Kuribayashi S. and Osana Y., "System Virtualization and Efficient ID Transmission Method for RFID Tag Infrastructure Network", International Journal of Computer Networks & Communications (IJCNC), Vol. 2, No. 6, pp. 183-194, 2010.

http://www.ifm.eng.cam.ac.uk/automation/documents/centerguide.pdf

http://www.ifm.eng.cam.ac.uk/automation/documents/centerguide.pdf


1434

Lim T.S., Sim S.C. and Mansor M.M., "RFID Based Attendance System", IEEE Symposium on Industrial Electronics and Applications (ISIEA), Kuala Lumpur, Malaysia, pp.778–782, October 2009. Manish B. and Shahram M., "RFID Field Guide: Deploying Radio Frequency Identification Systems", First Edition, Prentice Hall PTR, 2005. Mishra D., Vasal A. and Tandon P., "A Novel and Cost Effective Approach to Public Vehicle Tracking System", International Journal of UbiComp (IJU), Vol. 3, No.1, pp. 33-44, 2012. "Power Save", User Guide, Faronics Corporation, 2009.

Qaiser A. and Khan S.A., "Automation of Time and Attendance using RFID Systems", Proceeding of the 2nd IEEE International Conference on Emerging Technologies, Peshawar, Pakistan, pp. 60-63, November 2006. Roberti M., "Plan an RFID Rollout That Stays On Track", RFID Journal, Article No. 3083, pp. 1-10, 2007.

Su X., Chu. C. C., Prabhu B.S. and Gadh R., "Service Organization and Discovery for Facilitating RFID Network Manageability and Usability via Win RFID Middleware", Proceeding of IEEE Symposium on Wireless Telecommunications (WTS 2008), California, pp. 392-398, April 2008. "Tag Sense Nano-UHF RFID Reader", Data Sheet v2.2, Tag Sense, Inc.,available at http://www.tagsense.com/, last accessed on 22 January 2013. Wang Y., Liu J. B. and Yao G., "Hierarchical Architecture for Better Network Management", Proceedings of the third meeting on Mobile and Wireless, Seoul National University, Seoul, Korea, pp. 1-4, February, 2010. Younis M.I., "SLMS: A Smart Library Management System Based on an RFID Technology", International Journal of Reasoning-based Intelligent Systems (IJRIS), Vol. 4, No. 4, pp. 186-191, 2012.

Figure (1) Data Communication between Parts of an RFID System

http://www.tagsense.com/


1435

Figure (2) The Hierarchical Tree Organization of the IRFID

Figure (3) The Schematic Diagram of IRFID


1436

Figure (4) The Entire Modules of the IRFID Dedicated to MMC Server

Figure (5) The Entire Sub Modules of the Main Module


1437

Figure (6) The Entire Modules of EMC

Figure (7) The Entire Modules of the IRFID with Back Up Server


1438

Figure (8) RFID Reader, Tags and the Antenna

Figure (9) The Main Program Window of the MMC

The types of tags used in the implementation of IRFID

The type of RFID reader used in the implementation of IRFID

The combination of RFID reader with the antenna used in the implementation of IRFID


1439

Figure (10) EMC GUI Window

Figure (11) Backup Server Startup Window

Table 1 The Construction of the RFID Readers Parameters' Table

Table 2 The Construkction of the EMC Parameters' Table


1440

Table 3 The Construction of the 'System Events' Table

Table 4 The Construction of the 'Tag ID' Table

Table 5 The Construction of the 'Tag Information' Table

Figure (12) The Applications' GUI


1441

Figure (13) The Distributing of RFID Reader on the College's Map

Figure (14) GUI of Remote Tracking and Monitoring Application


1442

Table 6 Comparison between Proposed IRFID and Related Systems

Implemented feature (√) Not supported feature (X)

Limited supported feature (L)

RFID-based Systems

Functionalities

AT

AS

BA

S

AA

S

VT

S

RFID

TM

API

SAA

S

RFID

BM

UC

SLM

S

Proposed IRFID

1- Remotely setting X X X L L X X X accepted

2- Scalability

Networking and management X X X X L √ √ √ accepted

System extendibility X X X X X √ √ √ accepted

3- Tracking

Remotely tracking √ X √ X √ √ √ √ accepted

Concurrently tracking X √ √ √ √ √ √ √ accepted

4- Reusability X X X X √ √ √ X accepted

5- Automation L L L L L √ √ √ accepted

6- Power saving X X X X X X X X accepted

7- Continuity in work X X X X X X X X accepted

Figure (15) The Attendance of the Selected Employee

Documents

Strengthening and Closing Cracks for Existing Reinforced ...1).pdf · Strengthening and Closing Cracks for Existing Reinforced Concrete Girders Using Externally Post-Tensioned Tendons