WELDING RESEARCH

-S283WELDING JOURNAL

ABSTRACT. In order to optimize lanceperformance, a nonreactive, nonmeltinglayer to cover the wire-core thermal cut-ting lance has been tried. The layer stabi-lized the lance combustion (aluminum-hybrid iron wire-core lances). Lances cov-ered by this layer produced a highly con-vergent stream of cutting oxygen and didnot suffer from major radial leakage ofcutting oxygen typical of commercial ther-mal lances. The layer also improved thecutting continuity by avoiding the solder-ing effect and lance extinction. A novellance, Sharp-Fire (O)TM has been devel-oped. Experimental data indicate that thenew lance cuts steel or iron two timesfaster than conventional commercial ther-mal lances. The cost of cutting with thenew lance is also economical. The cuttingdata reveal that the savings can be in therange of $1–$10/ft of material cut.

Introduction

The thermal lance process representsthe oldest commercial use of oxygen cut-ting or piercing of massive objects (metal-lic materials and concrete) (Ref. 1). Theinvention of the oxygen lance was basedon an exothermic reaction of iron pro-vided by the lance and a gaseous oxygensupply (Ref. 2). There are plenty of oxygenlance designs; however, the wire-corelance is the most common lance design inthe industry, because it provides more ironfor combustion and is easy to fabricate.

In the thermal lance cutting process,the lance rod, which is made of low-carbonsteel tube with several iron wires (Fig. 1),reacts with a supplied flow of oxygen. Theoxidation reaction generates a largeamount of heat, melting the target. Excessoxygen is used to oxidize and blow out themolten target (the slag). The process hasalways been regarded as a crude methodof cutting and severing, since it relied on

the evolution of a large amount of heat atthe point of cutting/piercing (Ref. 1).

The oxygen cutting jet plays an impor-tant role in the design of a thermal lancebecause most of the heat is provided by theburning of the cutting plate. This oxygenalso helps to blow out the unreacted ma-terials to make a cut. In the process ofheavy cutting, the pressure of the cuttingoxygen entering the cutting tip is of para-mount importance, because of the dis-tance the cutting jet must traverse throughthe kerf (Ref. 1). As opposed to oxyfuelcutting technology, the cutting tip of thethermal lance can always touch the cuttingspot to avoid this problem. The efficiencyof oxygen is therefore increased in thethermal lance cutting process.

In the past, research and inventionswere focused on questions like A) how tochange the lance design in order to get abetter lance performance; B) is it possibleto replace the iron fuel by other metals toachieve a higher lance burning tempera-ture (Refs. 1, 3, 4); or C) how to applychemical flux processes for increased per-formance (Refs. 5, 6).

In this paper, we focused our attentionon a nonmelting, nonreactive layer modi-fication of the wire-core thermal lance.This layer is applied around the thermallance and serves as a flame-focusing ele-ment and consequently improves thelance cutting efficiency. Although it is wellknown that plastic and ceramic layers areapplied for the insulation purpose of elec-tric-aided cutting rods (Refs. 4, 7), we didnot find any previous report on the flame-focusing effect of an inert layer.

Experiment Setup

The setup of a commercial thermallance system is depicted in Fig. 1.

The principle of the experimental ther-mal lance system is described as follows:

oxygen comes from the cylinder, throughthe oxygen hose and the torch, and finallyreaches the lance rod. Oxygen flow rate iscontrolled by both the cylinder regulatorand the torch trigger. The iron wire-corelance rod, commercially produced, hasbeen made of low-carbon steel tube withseveral low-carbon steel wires along theinside surface of the tube (some steel is re-placed by aluminum, in aluminum-hybridiron wire-core lances). There are hollowareas inside to allow the oxygen to passthrough — Fig. 1, upper-right corner. Inthe modified lance design, this assemblywas surrounded by a nonmelting, nonre-active inert jacket.

By scrubbing the lance tip on thestriker plate, sparks are generated on thelance tip. The sparks initiate the lanceburning. By adjusting the oxygen flow, thelance burning propagates by itself. Theburning lance targets the cutting plate tostart the cutting. In our experimental test-ing, we have measured the flow rate ofoxygen, time of cutting, consumption ofthe lance, and length of the cut. This in-formation made it possible to calculate thecutting speed. The experiments have beenrepeated several times and the values re-ported here represent a reliable averagevalue. Four types of lances were tested:commercially available iron wire-corelances (CAL), aluminum-hybrid lances,and inert-layer modified lances of bothCAL and aluminum-hybrid types.

Flame Focusing, Flame StabilityTheory, and Lance Modification

Combustion of metals in oxygen is astrongly exothermic process. It is wellknown that strongly exothermic processesfeature all types of instabilities occurringboth in space and in time. Typical time in-stability is represented by a longitudinaloscillation of the flame, and the tempera-ture in the flame can oscillate by severalhundred °C. We have observed this type ofbehavior for the flame produced by alu-minum or titanium lances. The angular os-cillation of the flame results in the rimburning at different speeds. We observedthat after lance extinction the lance rimlooked jagged.

In addition to flame instability, there isanother problem brought by the wire-core

Flame-Focusing Modification of a Wire-Core Thermal LanceImprovement shown to double cutting speed

BY H. WANG, P. PRANDA, AND V. HLAVACEK

H. WANG is a graduate student, Depart-ment of Chemical Engineering,The StateUniversity of New York at Buffalo. P.PRANDA is at Department of Chemicaland Biomolecular Engineering, Universityof Notre Dame, Ind., and V. HLAVACEKis at The State University of New York atBuffalo.

KEY WORDS

Thermal Cutting LanceExothermic ProcessesSoldering EffectFlame Stability

wang qwkcorr 8/27/04 8:40 AM Page 283

WELDING RESEARCH

OCTOBER 2004-S284

lance design. The lance is made of thinmetal tubing and several metal wires. Thedimension of the tube wall is usually muchless than the diameter of the wires (e.g.,for industrial standard iron wire-corelances, the wire diameter is three timeslarger than the tube wall thickness). Hi-rano (Ref. 8) and Sato (Ref. 9) indicatedthat the metal flame-spreading speed in-creases as the metal piece dimension de-creases. Therefore, it is no surprise thatthe lance rim burns faster than the insidewires. We observed this phenomenon inthe burning CAL lances.

Because of a jagged rim, together withdifferent flame-spreading velocities of thelance rim and body, a certain portion ofthe oxygen is ejected in the radial directionand the flame loses the focus. The burninglance looks like a spear — Fig. 2, left. Partsof the flame and oxygen are not focused onthe cutting target and they are wasted.

In order to keep the flame focused, it ispossible to wrap up the lance body with athin layer of an inert jacket. The jacketburns slower than the lance. A modifiedlance with the trademark Sharp-Fire(O)TM was developed by Ceramic and Ma-terials Processing, Inc., by this method.This provision allows the iron to burn, butthe jacket prohibits radial oxygen leaking.The effects of oxygen leaking and its elim-ination can be clearly observed in Fig. 2.The material for the focusing sleeve,shown in the figure, is of great importancefor the success of this method. The sleeveis made of a thin, flexible sheet of graphitewith the thickness of about 0.01 in. (0.25mm). The graphite sleeve is an inert ma-terial. By adjusting the sleeve thickness,the consumption speed of the sleeve is justslower than that of the lance, offering a fo-cusing effect without disturbing the cut-

ting. The reaction rate of the sleeve withoxygen is very slow, and for practical rea-sons we can assume nonreactivity. Sincethe sleeve is not consumed by oxygen in asignificant way, the metallic part of thelance burns inward. The sleeve is weakand, if not supported by a rigid metallicbody of the lance, it easily breaks off dur-ing the combustion process. Conse-quently, a self-regulation occurs withoutany interaction with the operator, and thecombustion process produces a combus-tion zone configuration that sits on the rimof the burning lance, which is surroundedby a “wall” of the sleeve that extends 1–3mm above the combustion front. Once thewall of the sleeve grows more, it is me-chanically broken off by the interaction ofthe flowing gas and slag. Small scales ofthe unreacted sleeve are part of the slag.Details are described elsewhere (Ref. 10).

Experimental Results

Cutting with CAL and Sharp-Fire (O)™Lances

In this section, the cutting performanceof the nonmelting, nonreactive jacket-modified iron wire-core lance is comparedwith original commercially available ironwire-core lances. A commercially avail-able iron wire-core lance (BROCO lance,produced by BROCO Inc.) has beentested against the modified lance (CALwith a thin, inert jacket).

Figure 3 displays the cutting speed as afunction of the plate thickness for 3⁄8-in.outside diameter (OD) CAL and modi-fied lances (0.375 in./0.952 cm OD, 0.028in./0.071 cm wall thickness, with sevenlow-carbon steel wires [0.09 in./0.229 cmOD]) under different oxygen flow rates:

282 L/min at oxygen pressure of 50lb/in.2/345 kPa) and 411 L/min (at oxygenpressure of 80 lb/in.2/345 kPa), respec-tively. Cutting plates used were mild steelslabs with thicknesses from 2.5 cm (0.984in.) to 12 cm (4.72 in.). The figure showsthat the modified lance cuts faster than theCAL lance at both oxygen flow rates.

In order to compare the performancedifference between CAL and modifiedlances, we introduced a parameter: rela-tive cutting speed index.

The definition of the relative cuttingspeed index (RCSI) is as follows:

The higher the RCSI number, the moreefficient is the use of modified lancesrather than CAL lances under the samecutting conditions.

Figure 4 shows that both curves have amaximum cutting plate thickness around 7cm (2.76 in.). As the cutting plate thick-ness increases from 0 to 7 cm (2.76 in.), thecutting performance difference increases.As the cutting plate thickness increasesfurther, the performance difference de-creases. No experiments were done at athickness above 12 cm (4.72 in.).

The lance burning speeds under all cut-ting conditions were measured. As wefound out, there is no obvious trend be-tween lance burning speed and types oflances (CAL and modified), cutting platethickness, or oxygen flow rate. The lanceburning speed oscillates in the range of0.65–0.95 cm/s (0.256–0.374 in./s).

Experiments with 1⁄4-in. OD lances (1⁄4-in./6.35-mm OD, 0.028-in./0.071-cm wallthickness, with six low-carbon steel wires

RCSI

Cutting Speed Cutting Speed

Cutting SpeedModified CAL

CAL

= −

Fig. 1 — Commercial thermal cutting setup.

Fig. 2 — Flames produced by a CAL, left, and a modified iron-wire corelance, right.

wang qwkcorr 8/27/04 8:40 AM Page 284

WELDING RESEARCH

-S285WELDING JOURNAL

[0.065 in./0.163 cm OD]), cutting on 1⁄2-in.-(12.7-mm-) thick plate at different oxygenflow rates, were performed to investigatethe influence of oxygen flow rate. The re-sults are shown in Fig. 5. Figure 5 showsthat the RCSI decreases abruptly as theoxygen flow rate increases.

The lance burning/cutting observa-tions (Fig. 2) show the inert jacket placedon the surface of a CAL modifies the oxy-gen flow, and a well-focused stream of thecutting oxygen is achieved. The fluid flowproperties are illustrated in Fig. 2. Thetiny droplets of iron oxides that are re-leased from the body of the burning lancefollow the streamlines. Figure 2 clearly in-dicates that for the modified lance, thestreamlines are not dispersed and, conse-quently, the cutting oxygen is transportedto the cut without any radial leakage. It is

well known that oxygen quality is impor-tant for steel cutting (Ref. 11); therefore,the oxygen focusing effect provided by thejacket improved the cutting speed sub-stantially (as shown in Fig. 4). Other ad-vantages of the modification from our cut-ting experiences are as follows:

1) The cover decreases the heat dissi-pation between the lance and the cuttingplate. As a result, the flame does not ex-tinguish during random perturbations,and high stability of the cutting regime re-sults. Thus the annoying reignition opera-tion is safely removed; this is important es-pecially for underwater operations.

2) The cover also eliminates the “sol-dering effect,” by avoiding direct touchingof the lance and the plate. The solderingeffect, which happens with iron thermallances, is induced by the fast cooling of the

cutting area during cutting. The fast cool-ing process solidifies the melted metalsand the lance sticks to the plate. The sol-dering effect (lance sticking to the cuttingplate) often happens in the case of a lowoxygen supply flow. The modification de-creases real cutting time and improves theconvenience.

Aluminum-Iron Lance and Flame Stability Control

Aluminum metal behaves differentlyfrom iron; it reacts with both oxygen and ni-trogen. Unlike iron, the melting point ofaluminum is low and the adiabatic temper-ature is high. The adiabatic temperature ofaluminum combustion in oxygen is approx-imately 3700°C (6700°F); the experimen-tally observed value is around 2500°C

Fig. 3 — Cutting speeds for different cutting plate thicknesses and oxygen pres-sures using 3⁄8-in. OD lances (1.2 cm/s = 0.472 in./s).

Fig. 4 — Relative cutting speed index at 50 lb/in.2 (345 kPa) and 80 lb/in.2 (552kPa).

Table 1 — Comparison of Aluminum-based Lance and CAL

Cutting # Type of Lance Time Length of cut Cutting Speed Observations in. (cm) in./s (cm/s)

Steel plate (thickness = 1.02 in. [2.6 cm]) at oxygen flow rate 80 L/min

1 CAL 89.59 8.86 (22.5) 0.099 (0.251)2 Fe tubing +7 Al wires — — — No ignition,3 Al (3003) tubing +7 Fe wires — — — Extinguished4 Al (6061) tubing +7 Fe wires — — — Al burns faster than Fe wires5 Al (6061) tubing + 7 Fe wires and inert jacket 38.29 6.1 (15.5) 0.159 (0.405) Combustion6 Al (3003) tubing + 7 Fe wires and inert jacket 61.45 8.86 (22.5) 0.144 (0.366) Combustion

Steel plate (thickness = 1.38 in. [3.5 cm]) at oxygen flow rate 80 L/min

7 Al (3003) tubing + 7 Fe wires and inert jacket 73.94 5.71 (14.5) 0.077 (0.196) Combustion8 Al (6061) tubing + 7 Fe wires and inert jacket 53.68 5.71 (14.5) 0.106 (0.270) Combustion9 CAL 79.99 5.71 (14.5) 0.071 (0.181) —

CAL, 50PSI

CAL, 50PSIModified lance,50PSI

Modified lance, 80PSI

1.2

1.0

0.8

0.6

0.4

0.2

0.00 2 4 6 10 12 121086420

0.0

0.2

0.4

0.6

0.8

1.0

8

Cut

ting

Spee

d (c

m/s

)

Rel

ativ

e C

uttin

g Sp

eed

Inde

x, R

CSI

wang qwkcorr 8/27/04 8:40 AM Page 285

WELDING RESEARCH

OCTOBER 2004-S286

(4500°F) (Ref. 12). There are studies thatreport that aluminum burns as vapor orboth as liquid and vapor (Refs. 13–15). Be-cause aluminum has a higher adiabatic tem-perature, it is widely used to replace iron toincrease the flame temperature. Due to in-stability of aluminum combustion, a stableself-propagating lance flame is possible onlyif a small part of iron is substituted by alu-minum (Refs. 4, 16).

We have tested the performance ofaluminum metal in the design of a wire-core lance, and the results have been com-pared with the performance of the com-mercially available lance (3⁄8-in. OD CAL)with the same dimension and design. Inthe experiments reported below, we usedFe wires of 0.09 in. (2.29 mm) and alu-minum wires of 0.08 in.(2.03 mm). The re-sults of the test are reported in Table 1.

Aluminum wires or aluminum tubing,if used in proper relation to the iron ma-terial of the lance, can substantially in-crease the cutting speed compared toCAL performance. However, this increaseof the cutting speed is at the expense of thecombustion stability of the lance. Alu-minum metal melts in the flame, the smalldroplets of liquid aluminum are carriedaway by the oxygen stream, and after-burning of aluminum takes place. Theself-propagating combustion regime is ex-tinguished as the oxygen flow rate in-creases above 80 L/min. An application ofan inert jacket on the aluminum lance isvery important for a stable combustion.Experimental observations indicated thatthe absence of the nonreactive, nonmelt-ing jacket would result in highly unstablecombustion, resulting in the flame extin-guishing. The implementation of the inertjacket improved the flame stability of thehybrid Fe-Al lance.

Discussion

Application ofother metals to aregular iron lancecan have positiveresults if the rela-tion between ironand the additionalmetals is kept in acertain range. Ex-periments de-scribed above re-vealed that a hybridFe-Al lance doesnot burn in a stablemode, and a stablecombustion can beachieved only if theiron material is inexcess and the lanceis protected by aninert jacket at thelow oxygen flowconditions. In particular, a combination ofiron wires and aluminum 6061 tubingturned out to be a very effective lance witha high cutting speed if an inert jacket wasapplied. The inert jacket has the effect offlame stability improvement. The follow-ing qualitative explanation of the im-proved flame stability can be presented. Itis known in the theory of strongly exother-mic, strongly nonadiabatic flames that thebehavior of the flame shows major longi-tudinal oscillation that eventually resultsin flame extinction. The theory also indi-cates that the decrease of heat loss fromthe flame improves stability behavior. Ob-viously, the unreacted piece of the sleeve,before being broken off, represents a ra-diation mirror, and the radiation flux inthe radial direction is smaller. Conse-

quently the flame is hotter and more alu-minum evaporates. A homogeneous flameusually exhibits higher stability comparedto a flame-consuming fine spray of drops.The thermal conductivity of graphite isalso much less than that of aluminum.These facts reduced the energy loss andheat dissipation to the environment. Theflame was therefore stabilized.

As the experimental results of 3⁄8-in. ODwire-core iron lances show, the inertjacket modified lance cuts faster than thecommercially available lances under thesame oxygen flow rates. With increasingcutting plate thickness, the difference be-tween CAL and modified lances increases.In cutting procedure, the lance flameheats up the plate surface first. After theplate is heated up to the ignition temper-

Fig. 5 — Relative cutting speed index as a function of oxygen flow rate (1.2cm/s = 0.472 in./s).

Fig. 6 — Overall cutting-cost comparison of CAL and modified lances at 50lb/in.2 (345 kPa).

Fig. 7 — Overall cutting-cost comparison of CAL and modified lances at 80lb/in.2 (552 kPa).

Rel

ativ

e C

uttin

g Sp

eed

Inde

x R

CSI

0.8

0.6

0.6

1.0

0.8

0.4

0.4

0.20.2

0.0 0.00 40 80 120 160 200

CAL

RCRI

1/4 inch OD lance (CAL, Modified)cutting on 1/2 inch thick steel Cost/ft Cut Cost ($) at 50 PSI

CAL

CAL

Modified

Modified

Cost/ft cut ($) at 80 PSI

Modified

wang qwkcorr 8/27/04 8:40 AM Page 286

WELDING RESEARCH

-S287WELDING JOURNAL

ature, the iron material from the plate re-acts with the excess oxygen provided bythe lance. When the plate is very thin, thepreheating time and focusing of the oxy-gen/flame stream play a minor role, andthe plate is cut immediately by the lanceflame. That is the reason RCSI is small forthin plates. When the cutting plate is thick,the cutting kerf actually played a role offocusing element; therefore, RCRI de-creases — Fig. 4. The performance differ-ence between modified and CAL lances incutting iron or steel plates is most obviousin the range of 6–8 cm (2–3 in.) of platethickness.

At a given cutting plate thickness, theperformance difference between CALand modified lances decreases with the in-creasing oxygen flow rate — Fig. 5, 1⁄4-in.OD lances. At very low oxygen flow, thefocusing effect substantially increases theefficiency of the oxygen. The modifiedlance cuts almost two times faster than theCAL lance. However, after the oxygenflow is increased, that difference disap-pears. It is also a proof that the inert wrapnot only focuses the flame, but oxygen aswell.

It is instructive to calculate the cost ofcutting both for the modified lance and forthe commercially available lance. Data oncutting speed, lance burning speed, andoxygen consumption must be considered.The following cost data have been used inour calculation: 3⁄8-in. OD lances, CALprice $4.06/3 ft, modified lance price$4.57/3 ft, oxygen price/L $0.00419.

The cost comparison is shown in Figs.6 and 7.

Figures 6 and 7 show the calculatedcutting cost based on the data of cuttingspeed, lance burning speed, oxygen flowrate, and the unit price of a lance and oxy-gen gas. The figures reveal that the over-all cutting costs are lower for the modifiedlance compared to commercial lances.

Oxygen consumption cost is about 35% ofthe overall cost for both the CAL and themodified lances at 50 lb/in.2 (345 kPa).Oxygen consumption cost is about 45% ofthe overall cost for the CAL lance and40% for the modified lance at 80 lb/in.2(552 kPa). The reduced cost of using themodified lance is due to the facts that cutspeed is increased and the material andfabrication cost of adding an inert jacket isalso low. Adding an inert jacket increasesthe lance cost less than 50 cents.

Conclusions

The nonmelting, nonreactive jacketimproves the wire-core lance performancethrough the following ways:

The jacket stabilizes the lance flame.The hybrid iron-aluminum lance does notproduce a stable self-propagating com-bustion regime without the nonmelting,nonreactive jacket at the oxygen flow rateof about 80 L/min. In the cutting process,the jacket reduces the flame extinctionand soldering effect by reducing the heatloss and by direct contact between thelance and the cutting plate.

Through the focusing effects of bothoxygen and the lance flame provided bythe jacket, the cutting time of the inertjacket modified lance is substantially re-duced. The cutting speed is increased by afactor of two for 6–8-cm-thick iron plates.

The cutting cost of the modified lanceis also much lower than that of the CALlance because the modified lance has ahigher cutting speed. The cost of the non-melting, nonreactive jacket is only mar-ginal.

Acknowledgment

The authors express their thanks toanonymous referees for their quantitativeand constructive comments.

References

1. Slottman, G. V., and Roper, E. H. 1951.Oxygen Cutting. New York: McGraw-Hill.

2. Brandenburger, E. 1975. New oxygenwire core lances. Metal Construction. 617.

3. Davies, I. L., Fleischer, C. C., and Paton,A. A. 1987. Developing techniques for remov-ing activated concrete. Nuclear Engineering In-ternational 32 (392): 45.

4. Brower, J. S. 1978. Underwater cuttingrod. U.S. patent no. 4,069,407.

5. Sweetman, W. G. 1958. Chemical cuttingmethod and apparatus. U.S. patent no.2,918,125.

6. Paaso, C., Cariello, B., and Palumbo, A.1991. Cutting electrode for underwater andland use. U.S. patent no. 5,043,552.

7. Moore, P. E., Strohl, R. L., and Soisson,L. R. 1985. Coated exothermic cutting. U.S.patent no. 4,544,139.

8. Hirano, T., Sato, K., and Sato, Y. 1983.Prediction of metal fire spread in high pressureoxygen. Combustion Science and Technology32:137–159.

9. Sato, J. I. 1983. Fire spread mechanismsalong steel cylinders in high pressure oxygen.Combustion and Flame 51(3): 279-287.

10. Hlavacek, V., and Pranda, P. 2001. High-speed chemical drill. U.S. patent filed on June1, 2001.

11. Kemp, M. D. N. 1986. The effect oflower quality oxygen purity in oxy-thermal cut-ting. Welding in the World 24(3/4): 28–34.

12. Grosse, A. V., and Conway, J. B. 1958.Combustion of metals in oxygen. Ind. Eng.Chem. 50:4663–4669.

13. Sircar, S., Gable, H., Stoltzfus, J., andBenz, F. 1991. Symposium on Flammability andSensitivity of Materials in Oxygen-Enriched At-mospheres: Fifth Volume. Eds. J. M. Stoltzfusand K. McCilroy. pp. 123–131. Philadelphia,Pa.: ASTM.

14. Sato, J., and Hirano, T. 1986. Symposiumon Flammability and Sensitivity of Materials inOxygen-Enriched Atmospheres: Second Volume.Ed. M. A. Benning. pp. 118–134. Philadelphia.Pa.: ASTM.

15. Mellor, A. M., and Glassman, I. 1965.Pyrodynamics 3:43–64.

16. Dean, J. L. 1985. Thermal burning rod.U.S. patent no. 4,541,616.

wang qwkcorr 8/27/04 8:40 AM Page 287