REPORTS

HEAT PROCESSING · (3) · ISSUE 2 · 2005 1

Introduction

In many areas of technique, material andheating exchange processes play animportant role. To achieve homogenioustemperature division in chamberfurnace, the theoretic as well as theexperimental registration of decreasingprocesses is of big interest. In the metalworking industry, the chamber furnace isan important heat technical furnace forheat treatment of e.g. wrapped metalbelts, - wires, small parts, soldering etc.The chamber furnaces KS/S is a insertfurnace. This fact requires a periodicaloperation: loading, heating up, holdingand cooling down.

Increasing demands on chamberfurnaces concerning quality of productsand the obligation to minimize specificenergy consumption as well as the cycletime lead to the need to homogenize theheating of insert goods over the wholecross section and depth of insert goodsduring heating and heat treatment ofribbon or copper alloys. One importantoperation area is e.g. the recry-stallisation annealing of cold rolled belts.During this process, textures andstrengthening accelerated during rollingare removed by grain growth duringannealing. The heat treatment has toeffected over the recrystallisation, whichare important for the material.

The heat treatment of alumina andcopper alloys is done in a protective gasatmosphere in order to keep a largelyblank surface. A high convective heat

transmission can be achieved by bigvelocity of flow on the surface. Thesehigh velocities of flow can be achievedby flow machines, which revolute thegas in the treatment room. Therefore it isused a pseudo radial ventilator.Compact constructions limit the room,which is necessary for the fully mixingprocess of gases, so that the demandfor a possible homogenous as well aslow-loss mixture get more and moreimportant. Inflow technique ventilators inthe chamber furnace construction,pressure loss rich velocity of flowconcerning returning and distribution ofrevolution material, are also effectingagainst the mixture process.

The presented problem with thetemperature comparison in chamberfurnaces makes a detailed research verysuitable.

Therefore advices for optimal tem-perature distribution in indirect heatedchamber furnaces with forced revolutionand furnaces with similar velocity of floware prepared.

Positioning and dimensioning ofmeasuring grid

The auxiliary grid in Fig. 1 with thedimensions 300x325x500 mm is divided100x100x125 mm segments, that means36 subdivisions are possible. With this

Experimental researches of temperaturedistribution in an electrically heatedprotective gas tight industrial furnacewith gas revolution

The following report describes the extent to which temperature homogenizationcan be achieved in a Type KS 80 S single-zone chamber furnace by means ofcombinations of parameters and forced gas recirculation. Temperature profile wasmeasured at five measuring levels in the downstream direction in the rectangularcross-section muffle. In addition, velocity was measured at room temperature, inorder to permit subsequent quantitative interpretation of flow distribution. Finally,a numerical method was used for determination of temperature distribution in thefurnace. These results were then utilized for optimization of gas circulation in thefurnace system under examination and an attempt made to apply this to a wholeseries of inert-gas muffle furnaces of up to 500 l effective capacity.

Dipl.-Ing. Ahmet TürkmenLinn High Therm GmbH,Eschenfelden (Germany)Tel. +49 (0)9665 / 91400Email: info@linn.de

Fig. 1:Auxiliary grid withmeasuring points

HEAT PROCESSING · (3) · ISSUE 2 · 20052

REPORTS

auxiliary grid are spatial temperaturemeasurements possible. The measuringpoints are equipped with fixednumeration to be able to gauge thetemperature processes on everymeasuring point during and after theresearches. To give a detailed statementabout temperature division, the auxiliarygrid has been placed symmetrical in thechamber. The viewing direction is fromthe door insight the chamber (Fig. 2).

The arrangement of the thermocouplesin auxiliary grid (Fig. 1):

• The diagonal A-F (green marking) with6 measuring points from in front left tobehind right at the top.

• One diagonal G-H (turquoise marking)with 2 measuring points from in frontright to behind left beyond.

• All 6 remaining edges 1-6 (redmarking), which are not included inthe diagonal.

• In the middle of the longitudinal axisa-c (black marking), that means in themedium square in front, middle,behind.

• Lateral face 1’-8’ (blue marking) at thetop, beyond, left and right, in themiddle of the each 2 squares with

Structure and performance ofrevolution ventilator

In the chamber furnace construction,revolution ventilator are often used. Therevolution wheels are directly coupledwith the flow drive, which is installed inthe door in order to assure a fastexchange. The flow and heat technicalefficiency of the furnace depends on inand out flow conditions beforeinstallation in the flow circulation. Thepseudo radial ventilator has advantagescompared to the axial ventilator. Thetransforming of the flow off cross sectionto the rectangle cross section of the flowlead through, can be designed moreeasy with the radial ventilator than withthe axial ventilator. The deflection by 90 °C in the radial ventilator additionallyallows a more reasonable installation inthe furnace. The type of operation of therevolution ventilator can be described inFig. 3. The material is sucked on thesuction side and then flowed throughradial with high circumference speed.This principal is valid for the radialventilator. The radial flowed throughmaterial is deflected by 90 °C on thefurnace diffusion in longitudinal directionof the furnace. During this process,losses arise in the meaning of pressurelosses. The order of and the positioningof the screen shields can be recognized.These screen shields serves for radiationinsulating.

In order to minimize the losses of gasvolume flows, the protective gas is insertbeyond the first three tight adjoiningscreen shields, which pre-heat the gasesbefore inlet in the chamber. The tight

adjoining screen shields are mounted indistances of x = 10 mm from each other,so that the enough radiation energy canbe released to the gases and to achieveas less as possible energy loss.

Heat transmission

The temperature profile is a result ofbalance between the recourses andreleased heat power in stationarycondition. The mathematical replica ofthe process is ordinary complex as it is aplurality of coupled filed problems.Therefore results the need for limitationon the essential mechanism. The heattransmission mechanism is divided inheat conduction, radiation and con-vection.

The energy balancing for heattransmission mechanism per volumeunit is [1]:

(1)

Heat conduction

The result of this equation (1) withoutconsideration of the flow, under neglectof radiation rate and under the conditionhomogeny and stationary proportion forthe volume power density:

(2)

tridemensional in vectorial form

(3)

Radiation

A lot of gases emit in form electro-magnetically waves. The radiations areabsorbed or reflected if the abut againsta body [2].

(4)

Convection

The heat transmission in gases byconduction and convection areconnected with each other aredescribed therefore together [2]. Theconvective heat transmission containsthe heat exchange between a solid bodyand a gas. On the boundary surface

Fig. 2: Symetrical installation of auxilary gridin the chamber

Fig. 3:Structure ofrevolution ventilator

REPORTS

HEAT PROCESSING · (3) · ISSUE 2 · 2005 3

between both materials, the heatingpower is transferred by conductions,while the transport inside the gas ismainly effected by movement of thesematerials.

(5)

The flow process can be calculated bysolving of the equation [3], [4]:

• mass balance

• energy balance, see equation (1)

• impulse balance

Partial results

The heating up curve under Argonatmosphere (Fig. 4) makes clear theprocess of the single measuring pointsin the chamber, while a stationary valueis adjusted in the area of app. 10 min(marking 1). One presumable reason is,that there is a energy storage of theinsert in this area. During On closesexamination it can be seen that a smalloscillating process of the controlthermocouple (numbering x), whicheffects the rough behaviour of thefurnace. The stationary condition alwaysadjusts during turning point of controlthermocouple process. This oscillatingtemperature process depends on theheating-up rhythm. This can berecognized up to T = 300 °C. In thisstationary phase, it can be seen a veryhigh temperature difference between theuppermost and lowest curve.

As already mentioned it was selectedArgon atmosphere in this example. Onthe marking 2 and 3, see Fig. 4, it hasbeen compared to judge the influence ofthe gas during heating and stationarycondition, the range of temperature aswell as possible changes. The result of∆ϑ = 34 K shows no difference, thatmeans that is does not change anythingin both cases. With the estimation of themaximal deviation from measurement, itis possible to include the magnitude,which can appear under disad-vantageous circumstances. These are ± 3,7 K at T = 900 °C, that means in theworst case the range would be ∆ϑ = 34K ± 3,7 K. The graphs, which are out ofthe range, have not been consideredduring this judgment as the measuringpoints joints do not tower into thefurnace that means the experimentee donot tower until the measuring points.

The first investigation for theventilation wheel 1 at changing ofnumber of rotaries

It was determined how big the influenceof the suction cross section of themeasuring metals sheets on the mixtureof changed gas volume flow V

.Gas

depending of diameter d of the mea-suring metal sheet series 1. In Fig. 5 thecurve processes of three differentvolume current proportions. As sup-posed, the exchanged gas amountincreased with increasing cross sectiond2. But no better results have beenachieved, as a worse temperaturedistribution was determined with thebiggest cross section. The reasontherefore could have many aspects.Presumably small significant turbul-ences evaluated in the chamber, whichdid not help for the homogenisation ofthe temperature. The increasing numberof rotaries has the effect thattemperature difference decreases.

Graphical temperature course ofthe measuring values withdifferent parameters in thechamber furnace KS 80 S

With the help of the parametercombination and the forced circulationof the gases a homogeneity of thetemperature was achieved. In thechannel with the rectangular cross-section was the measuring of thetemperature profile in 5 measuring areasdownstream. Additionally it wasrecorded a speed-measuring at roomtemperature on the level 1 to renderafterwards a quantitative explication ofthe stream distribution.

The following parameters were atdisposal: Suction sheet metal diameter,speed changes, fan wheels, gases, gasinlet configuration. In Fig. 6 and 7 thetemperature history at T = 800 °C isshown depending on the parameter (thelegend with the figure 81 is the time from

Fig. 4:Heating-up curveunder Argonatmosphere

Fig. 5: dT/dt as function of suction metal sheet at various rotaries

HEAT PROCESSING · (3) · ISSUE 2 · 20054

REPORTS

the upper pic. 6, instantaneous expo-sure).

Temperature distribution

The temperature homogeneity isdetermined of the geometry, theradiation conditions, the convection andthe heat conduction of the gas. Theheating is made from two sides, bottomand rear. At the unheated door occurs anintense temperature decrease which canreach further into the furnace volume,depending on the size of the door. Theheat conduction in the gas (not helium)can be omitted. The main actors areconvection and radiation. With naturalconvection these is the defined factor upto approx. 300 °C. Afterwards thetemperature transmission by radiation

overweights. If a reinforced convectionis forced by gas circulation, it will besignificantly valid up to approx. 400 °C.Above this the heat transmission andtemperature will be dominated by theradiation. End result: if an exacttemperature distribution via hugevolumes are necessary, then it shall beused a gas circulation at temperaturesup to 400 °C. Above 800 °C a more-zoneheating is reasonable and necessary.Between 400 °C and 800 °C the best is

to use both possibilities. Gas circulationcan be used up to 950 °C. If it isnecessary because of the process touse it up to max. 1050 °C, e.g. debindingat 300 °C and pre-sintering at 1100 °Cwith a good temperature distribution, theventilator must be able to be loweredwith a special wheel to a minimumspeed, e.g. 2/s, in order not to bedestroyed. But a rotation is necessary asotherwise the axel drive will bedeformed.

Studies were made on a more-zoneheating with three control paths,distributed over the furnace length (doorarea, middle part, rear) by the Universityof Erlangen for FT (Fig. 8). With theactions can be achieved temperaturederivations of +/- 3 to +/- 7 K over thefurnace volume. With a higher effort, e.g.a six-sided heating with trimmingpossibility, are for special applicationsalso higher accuracies to achieve, but itis often easier to limit the usefulchamber (next higher furnace type) inproportion to the muffle dimension inorder to achieve a similar goodtemperature distribution with lowercosts.

Temperature compensation inmeasuring cube (100x100x100) ina VMK 250 S laboratory furnace

Additionally the heat transmission fromthe furnace chamber to differentmaterials, as to say the soaking of thesematerials, was studied. The heattransmission is determined by theconvection and radiation. The aut-horitative mechanism in the solid itself isthe heat conduction. The heat conuctiondescribes a heat stream dQ/dt through aarea A in a material, which is in atemperature gradient dT/dx.

Fig. 6: Argon atmosphere, T = 800 °C, fan impeller 1, diameter 140 mm

Fig. 7: Instantaneous image at t = 81 min

Fig. 8: Temperature distributionn in a three-zone chamber furnace Fig. 9: Temperature process of different materials

REPORTS

HEAT PROCESSING · (3) · ISSUE 2 · 2005 5

Test process

The temperature compensation isdetermined in this test by means of astationary measuring method. Cubeswith an edge length of 100 mm are usedand are placed in the middle of thebottom in the furnace type VMK 250(250x400x250 – wxdxh). In the top-sideof the samples was made a hole for theinput of a thermocouple. Then it will beheated up with an approximate constantheating rate to T = 1000 °C. Thetemperatures can be seen directly on ameasuring unit. Therefore now themeasuring results and the geometricdimensions of the samples can be takento calculate the thermal conductivityafter the conversion of equation 2.

Measuring results

The temperature history of differentmaterials in Fig. 9 gives informationabout the thermal conductivity ofindividual materials. Graphite achievesthe desired value in shortest time and istherefore a very good heat conductor.The radiation is noticeable in graphite

over 500 °C, as the graph will becomesteeper, therefore graphite absorbsnearly all radiation (similar to a blackbody). Normally copper has the bestthermal conductivity characteristic, butthe heat transmission through radiationis very dependable of the surfacecharacteristic and on plain metals verylow. In the opposite it is to see that theceramic materials as SiC/Al2O3 have abad thermal conductivity and aretherefore good isolators.

Theoretical determination of thetemperature distribution in thefurnace with the numericalmethod

Finally the numerical method with asimulation program for the temperaturedistribution in the furnace was applied.After the input of the furnace geometryand of all material data the softwareallows the calculation of the temperaturedistribution in stationary and non-stationary status. Flows are not con-sidered. The listed furnaces in Table 1with different rectangular cross-sectionswere studied. Here only the level 3 wasinspected, thus exactly the middlecross-section of the furnace. Linn HighTherm offers the KS-series in severalsizes. Now it should be shown, in whichway the known recognitions of thestudied KS 80 S can be applied in biggerunits.

Afterwards the calculation of the small -/laboratory furnace lines VMK-S typewith minor other heater configuration ismade. Fig. 8 shows the temperaturedistribution for the chamber furnace. Atemperature difference of some Kelvin in

vertical direction is to note. Also thestudied chamber furnace KS 80 S hassimilar temperature distributions, but ata desired value input of T = 800 °C in thechamber furnace there are significantlyhigher temperature differences as shownin Fig. 10. The temperature in level 3 onthe measuring point b is in the stationarystatus at the same condition T = 836 °C,that means ∆ϑ = 36 K higher. Theseincrease of the temperature canoriginated because the rear wall was notconsidered in this numerical process.

Also this plot in Fig. 11 refers to themedium cross section (level 3, centre).One characteristic has to be considered,as the heat insert is affected by foursides. It can be seen that thetemperature is almost homogeny overthe whole cross section. The furnacetype VMK 135 S differs from the desiredvalue T = 800 °C in the centre by app. 2-3 Kelvin. The small furnace types VMK-S reach better temperature homogeneityover the whole cross section comparedto KS/S.

Reference

[1] Dubbel, H.: paperback book for machineconstruction, Springer publishing houseBerlin 1997

[2] VDI Wärmeatlas, calculation sheets forthe heat transmission, 8the edition,Springer publishing house Berlin 1997

[3] ANSYS 7.1 User Manuel[4] Groth, C.; Müller, G.: FEM for practical

man – temperature fields, expertpublishing house Renningen- Malsms-heim

Table 1: Geometrical data for the furnaces

Fig. 10: Temperature distribution in the KS 80 S

Fig. 11: Temperature distribution in the VMK 135 S