Analsis of Wire Drawing Machine

Highlights Drawing high carbon wires: a comparison...

DRAWING HIGH CARBON WIRESA comparison between theoretical analysis and actual experience on a multipass line with efficienthigh heat exchange capstans

By Eng. Angelo Zinutti: Research & Development Manager and Quality AssuranceManager of Wire Technologies S.p.A. andEng. Giancarlo Saro: Technical Director of Wire Technologies S.p.A.

Preface

This report intends to analyse, both theoretically and experimentally, the coolingmethods of multipass dry-drawing lines utilized for the production of high and lowcarbon steel wires. In the drawing of high carbon wires, it is essential for both theprocess and the product’s quality to be able to efficaciously cool the material beingprocessed: for this reason many drawing machine manufacturers have studied andproduced capstans in such a way as to improve the heat exchange capacity. Thisanalysis intends to evaluate the construction and environmental parameters whichaffect the capacity of absorbing heat from the wire being processed: this task is for themost part performed by the capstans as the environment and the die boxes have alimited heat absorbing capacity.Therefore the analysis has been integrated with tests and measurements performed inthe workshop, in order to verify how the study of a new machine, with the aid oftheoretical thermal parameters, has optimized the cooling of the wire being processed.

WIRE TECHNOLOGIES - Highlights http://www.wireworld.com/wiretech/drawing_wires.htm

1 of 15 4/12/2011 1:20 PM

Heat generation in the wire

Before proceeding with the determination of the parameters which define the heatexchange it is necessary to evaluate the quantity of heat generated during the drawingprocess: in fact a part of this must then be absorbed by the capstan.By equating the drawing power with the quantity of heat which the material beingprocessed is able to absorb in the time unit, it is possible to determine the maximumtheoretical temperature increase in the wire in adiabatic conditions. For this analysiswe can neglect the difference in temperature between the core and the surface of thewire since this usually is eliminated before it touches the capstan. The increase foreach pass can be determined as follows:

(1)

where:

T max. theoretical temperature increase (K);FPULL pull force necessary for drawing (Newton);A2 wire cross section at exit of the die (m²);

PULL pull stress that is applied to the wire (N/m²);CS specific heat of the drawn material (joule/(kg·K));

material density (kg/m3).

It can be noted that in the final formula the speed factor is not present, so theoreticallyit does not concur to increase the temperature. Actually, this factor will be consideredwhen analysing the cooling methods; in fact the time period during which the wire is incontact with the capstan is related to the number of wound wire spirals, to thecapstan’s diameter and to the wire speed.Formula (1) supplies the maximum theoretical temperature increase in adiabaticconditions; it does not consider the heat reduction either in the die boxes or in the freeatmosphere preceeding the contact point of the wire on the capstan: in this zone thetemperature, besides becoming uniform in the wire, decreases because of convectionand radiation. Moreover, as soon as the wire is wound on the capstan, it is struck by aflow of forced air which creates a further heat loss. It can be estimated that the actualincrease in temperature, evaluated on the first wire spirals wound on the capstan, is 10

WIRE TECHNOLOGIES - Highlights http://www.wireworld.com/wiretech/drawing_wires.htm

2 of 15 4/12/2011 1:20 PM

- 30 % lower than the calculated value using the above formula (1). The increase in temperature, as a function of the pull stress and the presumedpercentage of heat loss, is represented in fig.1.

The following empirical formula determines the water flow rate normally used forcooling the capstans for each pass of wire reduction:

(2)

PINST power installed per pass (kW);WH2O cooling water flow rate per pass (litres/min);f empirical coefficient ranging from 0.7 to 1.

Once defined the water flow rate, the capacity to cool the wire depends largely on theeffectiveness of the heat exchange obtained in the capstan. The aim is to notaccumulate the increases in temperature in the passes of the line because it maycompromise the process or even the final product.

Heat exchange with the capstan

WIRE TECHNOLOGIES - Highlights http://www.wireworld.com/wiretech/drawing_wires.htm

3 of 15 4/12/2011 1:20 PM

WIRE TECHNOLOGIES - Highlights http://www.wireworld.com/wiretech/drawing_wires.htm

4 of 15 4/12/2011 1:20 PM

After several wire spirals have been wound on the capstan (fig.2), a heat exchangeprocess initiates with the water flowing inside the capstan.The heat flow from the wire to the water can be evaluated through a heat exchangecoefficient (KT) which is dependent on both the physical and geometrical aspects ofthe elements present between the wire and the cooling fluid as well as on the physicaland dynamic features of the flowing water.By determining this coefficient one is able to define the quantity of heat transferredfrom the wire to the capstan and the efficiency of the cooling system. The quantity ofheat transferred can be determined by the following formula:

(3)

therefore:

(4)

QCOOL heat transferred to the capstan (Watt);KT global exchange coefficient (Watt/(m²·K));S conventional exchange surface (m²);TaWIRE average wire temperature (K);TaH20 average water temperature (K);WH20 cooling water flow rate per pass (kg/sec);v wire speed at exit of the die (m/sec);CH20 water specific heat (joule/(kg·K));

TH20 increase in the average water temperature (K);TWIRE decrease in the wire temperature (K).

It can be noticed that the exchange coefficient KT and the conventional exchangesurface S determine the quantity of heat exchanged. From an operational point of viewit is opportune that the diameter of the capstan, which with the height of the woundwire spirals determines such surface, should not increase excessively while it is

WIRE TECHNOLOGIES - Highlights http://www.wireworld.com/wiretech/drawing_wires.htm

5 of 15 4/12/2011 1:20 PM

extremely important that the water inside be fully in contact with the surface throughwhich there is the heat exchange. Formula (4) shows how the possibility of decreasingthe wire temperature diminishes when increasing its speed and therefore the lineproductivity. The exchange surface is conventional because the wire is in contact withthe capstan only on a limited part and not on its whole semicircle; this fact can beconsidered in the global exchange coefficient through a parameter which considers theheat resistance between wire and capstan (Rw). It is also important to keep undercontrol the water temperature at the entry point of the capstan, which generally shouldnot exceed 20-25°C. Formulas (3) and (4) show how the water flow rate is directlyproportional to the quantity of heat exchanged; therefore in order to reduce the wiretemperature, it is fundamental to control all the parameters which influence this value(feeding pressure, crossing sections, water speed, etc.). Since the capstan’s diameter isalways much greater than its thickness, the exchange coefficient can be determined asfollows:

(5)

Rw heat resistance between the wire and the outside surface of capstan (m²·K/Watt)

heat resistance of each layer of material between the wire and the water with thickness si (m) andconductivity (Watt/(m·K)); in this parameter the possible deposits inside the capstan should

also be taken into account;

convection coefficient of water with the inside surface (Watt/(m²·K)).

With all other conditions being equal (i.e. wire diameter, surface condition, lubricantused, etc.), the heat exchange between the wire and the capstan can be improved bylowering the outside temperature of the capstan. To do this, all the manufacturers’efforts have been focused particularly on the improvement of . In reality it is also

very important to pay close attention to the other parameters, in the denominator offormula (5), which can add resistance that may jeopardise the heat exchange. Due tothis special care should be taken in avoiding the formation of deposits and adoptingonly coatings with high conductivity. In order to theoretically determine the waterconvection coefficient on the inside surfaces ( ) one can use the dimensionless

WIRE TECHNOLOGIES - Highlights http://www.wireworld.com/wiretech/drawing_wires.htm

6 of 15 4/12/2011 1:20 PM

values normally indicated in the technical literature: Reynolds number, Prandl numberand Nusselt number. The Reynolds number defines also the type of exit flow: laminar,mixed or turbulent. In order to have an ideal heat exchange it is better to have athoroughly developed turbulent flow i.e. it is necessary that the speed by which thewater laps the surface exceeds a certain value in relation to the narrow gap sn.By joining the relevant equations it results that in case of turbulent flow:

(6)

That is, the convection coefficient with the surface: a) is proportional to a coefficient rrelated to the fluid characteristics;b) increases with the speed VH20 by which the water laps the exchange surface;c) increases on reducing the narrow-gap sn.

Degree of various parameters influence

Water velocity and narrow gap.The Reynolds number depends on the narrow-gap sn and on the water velocity incontact with the inside surface: therefore it is determined by the constructioncharacteristics of the capstan. In particular it is important to define whether theinternal chamber of the capstan is solid to it or is fastened. The first case is generallymarked by a laminar flow where the water velocity is mainly determined by the flowrate and the passage section. On the contrary it is important to try and make the fluidmotion turbulent by increasing the velocity: this can be achieved by conveying theflow of the water under pressure in a spiral direction through the narrow gap instead ofmaking it flow along the capstan’s axis. Thus the passage area is reduced therebyincreasing the velocity to such values as to obtain a good convection coefficient .In the second case the water is subject to a dragging shear effect by which its velocityis a fraction of the periferical speed of the capstan and therefore, in this operativecondition, it is extremely difficult to determine the type of fluid motion.As can be seen in formula (5), the global exchange coefficient is not only determinedby the convection coefficient of the water but also by the conductivity and thicknessof the capstan.

Capstan’s and coatings’ thicknessThe steel capstans on the market often appear to be too thick considering the pullstress they support. Although it is important to pay attention to vibrations and to a

WIRE TECHNOLOGIES - Highlights http://www.wireworld.com/wiretech/drawing_wires.htm

7 of 15 4/12/2011 1:20 PM

certain thermal inertia which are useful in transient times, in some cases thesethickness could be reduced. Equally, it is fundamental to pay attention to the coatingsto be used. It is useful to point out, in particular, that the painting of the inside capstanwall by resinous and polymeric products is to be avoided since a thickness of 100 isthe same as adding a cast iron thickness of 20 mm: it is much better to use metalliccoatings that are oxidation-resistant. Similarly on the outside wall, while the use oftungsten carbide does not endanger the heat conductivity, the use of ceramic should beavoided when producing high carbon wire, (the conduction coefficient is lower thanthat of the surface). Finally, an incorrect inner or outer coating can double theequivalent thickness of cast iron or steel wall and consequently the thermal resistanceto heat exchange.

DepositsWhat has been said for the capstan’s coatings is valid also for deposits: since theirconductivity is very low it is important to avoid their formation (use of softened water,coatings and outside surfaces such as to prevent rusting and calcareous deposits) andto periodically carry out checks and maintenance so as to eliminate the problem.Regarding this, the significant increase of the exchange surface, which theoreticallycan be obtained through internal threading of the surface, can be easily vanified by thehigher tendency to form deposits on the same threads. The fig. 3 shows how evensmall deposits can endanger the heat transfer in comparison with a newly installed, andtherefore clean, capstan: the graph shows how the water-capstan heat transfercoefficient can be reduced by up to 50%.

WIRE TECHNOLOGIES - Highlights http://www.wireworld.com/wiretech/drawing_wires.htm

8 of 15 4/12/2011 1:20 PM

Performed Tests

To verify the effectiveness of the heat exchange a device was manufactured tomeasure in real time, and with the machine in operation, the actual water flow rate in acapstan and the inlet and outlet temperature of the same. The device, as shown inpicture 1, is able to measure the flow rate with a precision of one per cent and thetemperatures with a precision of one tenth of a degree.

Picture 1 - Heat flow measuring device

In this manner the effective heat quantity removed by the water, flowing in thecapstan under examination, can be easily registered. A nine draft machine, withhorizontal capstans (see picture 2) which are cooled with water under pressure in aclosed circuit, was utilized for the tests.

Picture 2 - View of the multipass line

WIRE TECHNOLOGIES - Highlights http://www.wireworld.com/wiretech/drawing_wires.htm

9 of 15 4/12/2011 1:20 PM

The wire collection was performed by a drawing coiler equipped with capstan cooledby water under pressure. A water cooled die box was present in the drawing coiler.

Picture 3 - Drawing Coiler

The regulation of the cooling water flow rate in the various capstans was obtained byvarying the feeding pressure. The capstans were designed to obtain a spiral flow inclosed chamber so that, with different working conditions, the water velocity wasalways such to obtain a completely developed turbulent flow (Reynolds > > 10000).This permitted not only to obtain an excellent heat exchange coefficient betweenwater and inside surface but also the possibility to optimize the water quantity even inthe presence of low feeding pressures. The measuring device was connected to theeight draft of the drawing line and therefore the registered flow data refers to theoperative conditions on this pass. The tests were performed by drawing two types of

WIRE TECHNOLOGIES - Highlights http://www.wireworld.com/wiretech/drawing_wires.htm

10 of 15 4/12/2011 1:20 PM

entry material (see table A): 1. G7 low carbon used for production of welding wire(mechanically descaled 2. C63 high carbon typically used for spring production(pickled and phosphatized)

Tab.A Characteristics of the two entry materials

Rm N/ mm² C % Mn % Si % Cr % Ni % P % S % Ceq.

G7515.5 0.090 1.489 0.872 0.046 0.031 0.017 0.014 0.36

C63952.5 0.630 0.638 0.214 0.044 0.028 0.009 0.009 0.75

The die drawing sequence in the multipass line was maintained constant for all testswhile both the die and the cooling of the coiler were varied in accordance with thematerial being tested. The following table shows the working sequence:

Tab.B Drawing sequence - Material G7: Low Carbon Steel.

Draft

N°

Diam.

(mm)

Reduct.

(%)

Tot.Red.

(%)

Tot.Red.

(daN/mm²)

Pull

daN

Speed

(m/sec)

5.500

51.5

2.1

1 4.800 23.83 23.83 63.7 738 2.8

2 4.210 23.07 41.41 72.4 503 3.7

3 3.700 22.76 54.74 78.2 416 4.7

4 3.260 22.37 64.87 84.0 344 6.1

5 2.880 21.95 72.58 89.5 282 7.8

6 2.560 20.99 78.34 94.1 228 9.9

7 2.270 21.37 82.97 99.9 192 12.6

8* 2.020 20.81 86.51 105.2 158 15.9

9 1.800 20.60 89.29 109.4 130 20.0

101.650

15.97 91.00114.4

9823.8

* Capstan subject to heat flow measurement

WIRE TECHNOLOGIES - Highlights http://www.wireworld.com/wiretech/drawing_wires.htm

11 of 15 4/12/2011 1:20 PM

Tab.C Drawing sequence - Material C63 High Carbon Steel.

Draft

N°

Diam.

(mm)

Reduct.

(%)

Tot.Red.

(%)

Tot.Red.

(daN/mm²)

Pull

daN

Speed

(m/sec)

5.500

95.0

2.1

1 4.800 23.83 23.83 107.7 936 2.8

2 4.210 23.07 41.41 120.7 778 3.7

3 3.700 22.76 54.74 128.3 629 4.7

4 3.260 22.37 64.87 135.9 507 6.1

5 2.880 21.95 72.58 143.1 407 7.8

6 2.560 20.99 78.34 150.0 321 9.9

7 2.270 21.37 82.97 163.6 281 12.6

8* 2.020 20.81 86.51 177.4 234 15.9

9 1.800 20.60 89.29 188.2 195 20.0

101.700

10.80 90.45191.2

8822.4

* Capstan subject to heat flow measurement

To ensure the worst possible working conditions, a narrow band of wire spirals i.e. 100mm equal to 50 wire spirals, was intentionally maintained on the capstan underexamination. On the remaining capstans the average band width was equal to 110 mmfor the G7 and 120 mm for the C63 material. During line operation the wiretemperatures, in the pass under examination and in the final part of the coiler capstan,were registered using an infrared optic pyrometer.

Synthesis of the test measurements

During the tests was registered a great number of data which for the simplicity issynthesised in the following tables

Tab.D

Low Carbon G7 WireInitial Tensile Strength: 515.7 N/mm² - Final Tensile Strength: 1116.1 N/mm²

Initial diameter: 5.50 mm - final diameter: 1.65 mm

Without cooling on the coiler - Air temperature: 16.2 °C

WIRE TECHNOLOGIES - Highlights http://www.wireworld.com/wiretech/drawing_wires.htm

12 of 15 4/12/2011 1:20 PM

Multipassexit

speed

Coilerspeed

Speedon the

analizedcapstan

Waterflow

Waterinlet

temp.

Wateroutlettemp.

: Watertemp.

increase

Heattransferred

by the water

Wire temp.on the capstan

coiler

m/sec m/sec m/sec m3/h °C °C °C Watt °C

5,0 5,9 4,0 2,09 15,4 16,8 1,4 3403 45

10,0 11,9 8,0 2,10 15,7 18,7 3,0 7327 70

15,0 17,9 12,0 2,09 15,8 20,0 4,2 10209 89

20,0 23,8 16,0 2,09 15,3 20,3 5,0 12153 93

Tab.E

High Carbon C63 WireInitial Tensile Strength: 952.5 N/mm² - Final Tensile Strength: 1916.0 N/mm²

Initial diameter: 5.50 - final diameter: 1.70 mm

With cooling on the coiler - Air temperature: 17.1 °C

Multipassexit

speed

Coilerspeed

Speedon the analizedcapstan

Waterflow

Waterinlet

temp.

Wateroutlettemp.

: Watertemp.

increase

Heattransferred

by the water

Wiretemp. on

the capstancoiler

m/sec m/sec m/sec m3/h °C °C °C Watt °C

5,0 5,6 4,0 2,08 14,9 16,6 1,7 4112 54

10,0 11,2 8,0 2,10 15,1 19,2 4,1 10013 70

15,0 16,8 12,0 2,09 15,0 20,4 5,4 13126 105

20,0 22,4 16,0 2,10 15,2 22,2 7,1 17340 122

From other test measurements it was noticed that by increasing the flow rate, whilemaintaining all other conditions constant, the heat flow rises even if the watertemperature increment diminishes. Using formula (3), in combination with theregistered data, one is able to determine the speed at which there is an equilibriumbetween the heat flow absorbed by the water and the deformation work. Under the testenvironmental conditions and with the approximation in the calculation of the heat

WIRE TECHNOLOGIES - Highlights http://www.wireworld.com/wiretech/drawing_wires.htm

13 of 15 4/12/2011 1:20 PM

dissipated by the wire (fig. 1), this equilibrium is theoretically at about 19 m/sec for thelow carbon wire and at 16.5 m/sec for the high carbon wire (coiling speed).Furthermore it was noticed how the relation between the heat removed from either thehigh carbon or low carbon wire corresponds roughly to the behaviour of the wiretemperature increments as calculated with formula (1). According to this formula thetemperature increment is proportional to the pulling force in the pass underexamination, in our case 48%, while the measurements indicated an average flowincrement of approximately 35-40%.It was also possible to verify formula (5) and (6) by carefully analysing the influenceof the various heat resistance in the total exchange: in particular an excellent waterexchange coefficient with the inside surface was registered ( > 3500Watt/(m²·K)).Furthermore the elaboration of the test data illustrated the importance of a correctregulation of the water flow rate and of the speed with which it laps the internalcapstan surface. It is important that this speed be independent of the capstan rotationvelocity, which as well-known varies from the first to the last pass, since the coolingrequirements are almost constant along all the entire line (second part of formula (3)).

Conclusions

One must note that, even if the positive heat exchange was facilitated by the low inlettemperature of the cooling water, this was compensated by the intentionally severeworking conditions adopted in order to discover any eventual critical conditions of theequipment and of the cooling system.In fact for this reason it was decided to operate with the following severe workingconditions:

working band width of wire wound on the capstan limited to 100 mm with time ofpermanence of the wire on the 8th pass analized at only 5.5 seconds at the maximum coilingspeed;

1.

Partial capstan coating with a ceramic band (for a section of about 20 mm), with consequentlimitation of the heat exchange between capstan and wire;

2.

Limitation of the water flow rate to only 35 litre/minute (2.1 cubic metre hour). The line usedin the test is also able to work with a flow rate of 50 litre/minute;

3.

Maximum working speed sufficiently high (approximately 24 m/sec.) for the above mentionedoperating conditions;

4.

The tests were not performed under laboratory conditions but on a line operating regularly ina workshop.

5.

In consideration of these rigorous and severe working conditions, one can concludethat the adopted technical solutions, and in particular the original cooling system(exclusive Pittini Impianti Industriali technology), confirm the design indicationsfurnished by the theoretical parameters illustrated in the first part of this paper.

WIRE TECHNOLOGIES - Highlights http://www.wireworld.com/wiretech/drawing_wires.htm

14 of 15 4/12/2011 1:20 PM

The resulting excellent heat exchange values permit not only to save on waterconsumption, during low carbon wire production, but also to have large margins forincrements in the heat exchange, with high carbon wire, by increasing both the bandwidth of wire on the capstan and the water flow rate.

Therefore the analytical definition of the heat exchange, and the consequent designoptimization, has permitted the carrying out of reliable tests of verification which havewidely confirmed the validity of the technical solutions adopted.

Bibliography

1. Wire Association International Inc., Ferrous Wire Handbook, The Wire AssociationInternational Inc., USA, 1989.2. G.Perotti, Tecnologie siderurgiche.3. L. Mattarolo, Trasmissione del Calore, CLEUP, Padova, Italia, 1980.4. P. Bartolini, Scambiatori di calore, CLEUP, Padova, Italia, 1982.

WIRE TECHNOLOGIES S.p.A. Via Buttrio 57/B - 33050 Pozzuolo del Friuli (Udine) - ITALY Tel. +39 0432 525911 Fax +39 0432 525921

[email protected]

Web Projects MULTIWAYS© 1999-2000 -2001-2002

WIRE TECHNOLOGIES - Highlights http://www.wireworld.com/wiretech/drawing_wires.htm

15 of 15 4/12/2011 1:20 PM