,

“)

SYNTHESIS R13PORT

FOR PUBLICATION”

CONTRACT N“ : BREU-CT91 -0548 (RZJE)

PROJECT No : SE -4538

TITLE :Solution of Electrostatic ProbleIns in

Powder Handling and Processing

PROJECTCO-ORDINATOR : Chilworth Technology Ltd.

Beta House - Chilw-orth Rese ~ch CentreSOUTHAMPTON S016 I’N:,England.

PARTNERS : L.O.M.Alenza 1 y228003 Madrid -

ESPAGNE

L. C.I.E.33 avenue de General Lecler:92260 FONTENAY AUX RQSESFRANCE-.

CEMENTOS REZOLAAvenida de Anorga36/2009 SAN SEBASTIANESPAGNE

INERISPare Technologique ALATAB.P. N“.260550 VERNEUIL - EN ~~ HALATTEFRANCE

S T A R T I N G D A T E : 1 9 9 2 - 0 1 - 0 1 DURATION :36 MONTHS——.-.-.—-—..————— -— -—-—-.-,— —

PROJECT FUNDED BY THE EUROPEANC o m m i s s i o n ” U N D E R T H EBRITE/EURAld P R O G R A M M E

SOLUTION OF ELECTROSTATIC PROBLEMS

IN POWDER HANDLING

.4ND P r o c e s s i n g~

A B S T R A C T

‘,

)

Electrostatic tribocharging ofpowder causes it to be at !racted to, or repelledji-orn, a nearby surface or szu-rounding par[icles. This con lead to variations inbulk density, choking of machines and a range of other u~zpredictable industrialproblems. Research work described by the report ~imed to improve theeconomic eflciency of milling and mixing in the test-zase cement plant, byreducing electrostatic powder adhesion. The jh-st stage of Ihe project was tofind the best powder categorisation methods. In the se(:ond stage, laboratorytest rigs lo model three relevant cement processing ma( hines were developed;a pneumatic conveyor, a cement mill and a jluidised bed homogeniser. Thereport deals principally with the obsenations andpredi( tions made using theserigs. In the third stage, laborato~ charge control tec)miques were repeatedusing the full-scale industrial machines, giving gener, d agreement with ~heexperimental resuits. Two important conclusions are that powders cannotreadily be classljied in terms of laboratory charge( lbili~ tests, and thatlaboratory equipment which models a machine closely may allow predictionsof powder behaviour. Experience gained jl-om this p ~oject will be used toimprove similar industrial situations via ongoing constdtancy work.

--

—

SOLUTION OF ELECTROSTATIC PROBLEMS

IN POWDER HANDLING

AND PROCESSING

Mr. Ian D. Pavey

CHILWORTH TECHNOLOGY LtdBeta house - ChiIworth Research CentreSOUTHAMPTON S016 7NSENGLAND

~ Dr. Angel Vega Remesal

L.O.M.Alenza 1 y228003 MADRIDESPAGNE

Dr. Claude Menguy

L. C.I.E.33avenue de Gerkral Leclerc ~92260 FONTENAY AUX ROSESFRANCE

) Snr. Juan Luis Querejeta

CEMENTOS REZOLAAvenida de Anorga36/2009 SAN SEBASTIANESPAGNE

M. Dominique Guionnet

INERISPare Technologique A L A T A

B.P. No.260550 Verneuil-en-HalatteFRANCE

1.

2.

3.

4.

5.

-.‘)

6.

7 .

8.

9.

CONTENTS

1NTROI)UCTION

TECHNICAL DESCRIPTION

2.1 Setting Up the Equipment and Procedures2.2 ‘ Chargeability Method for Powders and Fibres2.3 Design of the Simulation Apparatus

i) P n e u m a t i c C o n v e y o r ‘i~ Milling Test Apparatus

iii) Fluidised Bed for Mixing Homogeneity2.4 Available Charge Control Techniques

PNEUMATIC CONVEYING TRIALS

3.1 Electrostatic Effects at Laboratory Scale3.2 Single Point D.C. Neutraliser3.3 On-Site Measurements. Surface Contamination of Pipes

MILLING TRIALS

4.1 Electrostatic Effects in Laboratory Scale Mill4.2 On-Site Measurements. Milling Plant of the Inclustrial Partner4.3 Research at Other Manufacturing Plants

IMPROVEMENTS IN MIXING AND MILLINGAT THE CEMENT PLANT

5.1 Particle Size Determination5.2 Mixing (Raw M’ix Homogenisation)5.3 Cement Milling (Substitution and Improvement of Additives)

CONCLUSIONS

ACKNOWLEDGEMENTS

REFERENCES

FIGURES

l’age No.

1

2

3

4

6

‘7

8

9

10

11

12

14

14

15

1

)

INTRODUCTION

Undesirable electrostatic charging eff+cts are invariably caused by the complexphenomenon of contact electrification, that is, tribocharging. Even the most simpleprocess using a fairly res,istive powder may rapidly cre[:te very large fields and highpowder charges. Typical troublesome situations of this type exist where a powderhas been mixed, poured, ground, sieved, or micronised. One should be particulflrlywary when the po~vder has been entrained within a mo-{ing air stream, for exampleas occurs during pneumatic conveying.

A good example of an electrostatic side-effect concernt the variation in bulkdensity of powder according to ambient humidity. In t} (O practical examples, bagsfilled to the brim \vith food-stuff powder were sometin Ies found to be underweight;a polymer powder transported by tanker filled to a particular level often had a bulkdensity outside the specification. The ‘damp’ appearam :e of the powder in thesecond case was a tell-tale sign of charging.

Unpredictability in the results of tribocharging originat ss from the fact that thepowder charge at any time is very susceptible to small changes in the features of aplant, the humidity and the surface conditions. This being so, the best remedy foran electrostatic powder handling problem will always I equire specific analysis ofeach individual case, in terms of the above parameters Even so, there is often asingle, quite simple method which will remedy a wide array of different, unwantedcharging scenarios.

Test rigs of Laboratory scale were designed and built tl } reproduce the circumstancesleading to three different powder handling problems ir a cement manufacturingplant. Electrostatic charging in the experimental machines (a fluidised bed, apneumatic conveyor and a grinding mill) was monitor(:d in order to gauge thesuccess of a series of neutralizing techniques.

The moiect had three aims: firstly, to enable industry to recognise an electrostatic.-cause in powder handling problems; secondly, to quantify the problem as bestpossible; thirdly, to enable an appropriate form of soh Ition to be chosen.

‘Experience gained from the work wil[ be continually ind widely disseminatedseveral partners, leading to a more widespread awareness in industry of thephenomenon of tribocharging.

as

by

1 “

I

L.

.

)

TECHNICAL DESCRIPTION

2 . 1 Setting Up the Equipment and Procedures

In the first phases of the project, which occupie ~ the five partners for over ayear, the best available techniques of measurem mt, in terms of bothprinciples and equipment, were reviewed. The i litial plan was to categorisepo~vders according to their properties and therel )y create a database.Resistivity’ and charge decay time, pernlittivity, cl~argeabi~ity, Particle $izeand shape, adhesion and cohesion were the imp xtant parameters looked at.

2.2 Chargeability Method for Powders and Fibrus

The term ‘chargeability’, in this case refers to 1 he degree of charge on apowder as a result of a simple, standard tribock arging process. No standardtest for chargeability y currently exists, and form i of equipment representativeof real conditions were evolved empirically by two of the partners. It wasfotind quite difficult to create a reasonably con’:rolled charge using finepowders. Intuitively, one may expect that case: ding powder down aninclined stainless steel plane, connected to a se ~sitive ammeter recording thestreaming current, would be a useable system. [n practice, the powdercharging was much too aggressive, rapid, and ] esults unrepeatable.

Ultimately, chargeability of larger particles waj quantified by stirring themin an earthed stainless steel blender. Powder 01’ finer granule size wassprinkled fi-om a vibratory feeder to impact up m a metal plate. In bothcases, charging was followed with direct trader of the test material to aFaraday Pail, connected to an electrometer set to the charge range.

These methods permit the measurement of cha rgeability of powders underclimatic controlled conditions, and are;

- reproducible- valid for substances with high charge ~bility- useable with small qu~tities of fine (lusts, fibres and pellets- completed in a short time.

In addition, information from the tests;

- highlights the effect of ambient cond tions- enables study of side effects such as adhesion and agglomeration.

The effect of solid and liquid additives on the specific charge of powderswas most usefully anal yzed by the blender m( :thod.

2

II

I

I

),/

2.3 Design of the Simulation Apparatus

The three experimental rigs will be introduced sc paratel y.

i) Pneumuiic Conveyor (See Figw-e 1.)

Charging of a material being transported in a stfi iight pipe depends partlyupon the orientation of the pipe with respect to the ,,force of gravity. As aconsequence, horizontal, vertical up and vertical down flow sections weredesigned into the rig. Fittings, particularly those giving rise to a change inflow direction, were anticipated to have a significant effect on powdercharge. To limit the number of possibilities, fitti ~gs used were restricted to90° elbow bends. Taking all these points into consideration led to the basicsquare circuit in the vertical plane, all pipework being 30mm internaldiameter stainless steel tubing.

I

The system was run below atmospheric pressure, by using suction on thecyclone air outlet. Separate pipe sections were j( )ined with simple insulatingpush fittings. Sensors and other items of equipment could pass through thepipe walls without risk of powder escaping into the lab. The upstream endof the circuit was open to atmosphere, and powt ler fed directly into the pipefrom a vibratory feeder. The rig was constructed i in such a way that any partcould be electrically isolated, allowing streamirq~ currents, and consequentlythe local powder charging effect, to be monitort d. To ensure a reproducible,quantified air flow, an orifice plate and water rr anometer was fittedimmediately downstream of the cyclone.

iiy ~ Milling Test Apparatus (See Figures 2 & 3)

Figure 2 shows a schematic diagram of the exp a-imental mill in relation tothe electrostatic instrumentation. The mill itself is depicted from the side byFigure 3..

A small cement mill, for evaluating electrostatic: charge behaviour anddeveloping the best control techniques, was bui .t to laboratory scale.Instrumentation comprised specially designed field meters, electrometers anda device to measure and record the power cons~mption. The machine hasbeen installed in a climatically controlled enclosure having a metaliic screenlike a Faraday Cage. Relative humidity can be ~djusted to any pointbetween 30’%0 and 60?L0, and the temperature to between 18°C and 30°C.

Principally, the mill used a revolving steel cylinder, 60cm in length anddiameter 30cm, closed by two end caps and driven through gearing by anelectric motor. One cap supported a robust tra,i in the entrance for theremoval of processed material. Blasting by san{i,jet removed every trace ofoxide or paint from all surfaces. Inside the mil , an additional tray waslocated, the purpose of this being to remove th t output of grinding balls(a lso termed gr inder heads).

3

I

I

),,

Above a critical velocity, all of the grinding ball:, will remain attached tothe inner wall by a centrifugal effect. In normal :]se, the mill turns at 65-70% of critical velocity, nominal speed being 42 r.p.nl. ,

The number of grinder heads and their diametral size is shown below:

100 of 20mm; 100 of 25nm~; 100 of 30mm; 25 of 35nml; 18 of 40 nm.

All mechanical connections to, and supports of, 1 he mill were made fromPTFE (Teflon) to obtain high insulation with res ~ect to surroundinggrounded metallic parts. Insulation resistance bet ween the mill and themotor was greater than 100 GigaOhm.

ii~ FIuidised Bed for Mixing Homogeneip (See Figure 4)

Mixture quality is a major concern of many ind~ stries, such as chemicalproduct manufacture (gas-solid reaction) and ph: rmaceuticals. The major.,‘difficulties in sohd-solid mixing operations are s segregation and the inabilityof mixers to break down agglomerates. A segregation or de-mixing processalways occurs in competition with homogenizing. The mixing qualitydepends upon the dynamic equilibrium between mixing and segregation. Amixture is homogeneous if any sample of the mixture has the samecomposition and properties as any other.

The fundamental aspects governing the mechani jm ‘of gas-flow within thebed are not yet fully understood. The fluidisatio~ of divided solid particlesoccurs in the form of bubbles, analogous to the upward flow of gas througha column of liquid. Figure 4 shows the schemat c diagram of theequipment; the actual set-up was designed more to simulate the phenomenonat the Cementos Rezola cement milling plant.

The fluidised bed was divided into three stacket. parts. The lowest andmiddle parts, separated by a porous membraqe, were again divided into fourcompartments in which mean and sequential pressure could be applied. Themean pressure just fluidises all particles; sequential pulses of higher pressure

,,

perform the mixing. In the uppermost section, the smallest particles wereremoved. The air feed pressure was measured with a digital manometer, andthe aeration and fluidisation flow with two independent flow meters.

2.4 Available Charge Control Techniques

Quite a large proportion of the project time wa:: spent reviewing methods of “electrostatic charge control, both in the literature and privatecommunications from experienced cement mini ~g operatives. The idea wasto use the laboratory scale equipment to rank charge control techniques inorder of effectiveness, expose any side effects, md select the optimum for.trials in full scale equipment.

4

,

There were two important methods by which po’vder charge generation inthe three test rigs could possibly be subdued:

a) Powdered and liquid additivesb) High voltage neutralisers

A, small proportion of powdered additive can be mixed with the primarymaterial (i.e. cement powder) to reduce the volu ne resistivity. Liquidadditives comprised }vat er, in the form of ambie It humidity, and a well-known cement milling additive, SAT. Both worl. by speeding the chargedecay time, allowing more rapid relaxation of the powder charge wheneverthe particles contact a grounded surface.

The degree of success of each treatment strong]: r depends upon the primarypowder; several kinds of material were tested to =xpose these variations.

Commercially available high-voltage neutraliser: ~ usually take the form of aset of points supplied with high voltage of alter] Lating polarity, or with eachalternate point positive and negative respective]). Corona discharge provides

I a stream of ions from each point, the dispersal i If ions often assisted withair jets. Comparatively large charges of equal a~ Ld opposite polarity sprayedat the powder overcome existing tribocharge le~ els.

Preliminary Trials

Prior to any experiment with the test rigs, the e !fect of additives uponprimary powder properties was investigated. Two powdered additives,graphite and alumina, were used to treat three pimary powders, LDPEfines, cement and sepiolite. An addition of 5% graphite to LDPE fines wasable to reduce the charge decay time from mor( ~ than an hour to less thanone second. Similar effects were observed in th a other two powders, and asa consequence the chargeability much reduced, and improved, by graphite.The chargeability of cement only could be reduced by addition of alumina.

)

For liquid additive testing, the concentrations u ;ed were 35% & 50% forrelative humidity, and 150 & 300 ppm for SAI”. For cement and sepiolite,but not LDPE, an increase in relative humidity produced an importantreduction in resistivity. All powders were less [johesive/adhesive, in a higherhumidity atmosphere.

A liquid additive currently widely used to cont:ol charge in cement millingwas found to have no significant effect upon tl e electrical properties of anyprimary powder, cement included.

It was hoped that the information given by these standard laboratory testswould correlate with, and couldmore complex tests in industrial

~,

therefore have predicted, the outcome ofsituations.

5

3. PNEUMATIC CONVEYING TRIALS

3.1 Electrostatic Effects at Laboratory Scale

Using untreated sepiolite as a primary powder, tile space charge developedwithin this equipment soon became appreciable :md internal walls powdercoated. The resultant cyclone blockages are thou ght to occur because of anintensified space charge density close to the inte ma] wal 1s. That is, powderentrained in the internal vortex of air in the cyclone is centrifugal] ycompressed, giving particles an increased electrc static adhesion force.

Powdered Additive: GraphiteConcentrations of O. 5% and 5% graphite/primar:r powder were pre-mixedvigorous shaking. Three types of primary povvd~ :r were modified; cement,sepiolite and LDPE fines.

I Liquid Additive: SAT

by

Mixing was achieved by electrostatically sprayir ~g the “SAT (using a solutionof 5°/0 volume of SAT in isopropyl alcohol) on the powder. Two types ofprimary powder were modified; cement and sep iolite.

CommentsThe most effective additive, either liquid or po~ {dered was graphite,combined with the powder in a concentration o~” 5°/0. This managed toreduce tribocharging levels on all the primary powders and stopped thesepiolite cyclone blockages.

For sepiolite and cement, SAT seems unable to reduce, the charge level ofthe powder. A comparison between the initial c ~argeability test results andthe charge developed during pneumatic transpo]t ruled out any accuratepredictions. Tribocharging occurring due to panicles impacting a plate didnot correlate with tribocharging from powder b own through pipeworkentrained in a volume of air. Put another way, ndividual powder treatmentswhich appeared successful in the simple experi] nental model were notreflected in “a different kind of charging scenario.

3.2 Single Point D.C. Neutraliser

The basic system comprised a needle connected to a high voltage powersupply protected by a plastic tube. Of the three powders tested, sepiolite,cement and LDPE fines, a low local leveI of c~ mrge was measured. Thedevice reduced the electrostatic space charge o~’ sepiolite at the inlet to thecyclone, hence preventing any blockages. Theeffective than the additives, but a single pointcharge throughout the entire transport system.

leutraliser was certainly more~:ould ‘not control powder -

3.3 On-Site ,Measurements. Surface Contamination of Pipes

A 40m run of 80mm diameter stainless steel piping was used to convey fine ‘metal oxide powder of 2 microns mean particle diameter by pneumatics.The powder was fed into the system at the rate of 1 tonne/hour andtransported in an air flow of approximately 3601 n3/hour. After the operation,‘up to 10kg could be, found adhering to the walk of the pipes.

Using the laboratory technique, streaming curre] Its from’ eight sections ofpipework, insulated from each other and also fn }m ground, were monitoredto record specific charge exchanged per unit length of pipe. Charge-drivencurrent from the receiving hopper, and the chaq ;e on the powder in thehopper at the end of conveying, could also be n )easured with goodreproducibility.

IAs expected, under low humidity conditions the powder was much morelikely to acquire charge by triboelectrification, the degree being typical ofpneumatic transport.

)

An increase in the mass flow rate of powder le: ids to a decrease in thespecific charge on the powder, but this did not appear to have any influenceon the contamination problem. The most impor ant aspect of the results wasthe hypothesis that the space ch~ge density of powder [pC/m3] is thegoverning factor for powder charging. Space c1 arge density is a directfunction of streaming current and relatively ind tpendent of mass flow rate.

,Where right-angled bends were included in the pipework, we found that the

polarity of powder charging was opposite to thtit in straight pipe sections.This finding confirmed identical phenomena pr wiously observed in thelaboratory scale rig. Powder traveling around the inside of a bend willbecome compressed against the inside wall, du~: to a centrifugal effectsimilar to the operation of a cyclone. The spacl> charge density close to theinternal pipe surface will be much increased, zld as a consequence, thedirection of the driving force for charge exchmlge could reverse,

We concluded that the charge levels [1 to 3 p{;/m3] measured on thepowder were enough to initiate powder adhesitm, and recommended that thepowder be conveyed under more humid conditions. Electrostatic chargingwould then be implicated as the cause of the c mtarnination problem.

7’I

4. MILLING TRIALS

,.

4.1 Electrostatic Effects in Laboratory Scale Mill

Typically, the problem associated with electrosta: ic charge generation duringmilling is one of efficiency loss due to agglomer ~tion and adhesion of thepowder to milling elements. Additives and static eliminators neutralise thepowder charge and prevent sticking, saving ener~ :y and increasingproductivity.

A laboratory scale mill was the starting point for the study of the problemsassociated with milling; also, to initiate solutions to be applied in the realplant. The test rig was used for measurements of some effects and variablesdescribed as follows:

- Electrostatic charge of the powder subs~ antes during milling- Electrical power consumption- Effects of additives and neutralizers on power consumption- Effects of ambient conditions (Relative Humidity, Temperature)

The well controlled conditions and systematic m tthods of operation of thelab-scale mill have ciearly established that electrostatic effects areresponsible for loss of milling efficiency. It was useful to apply the samesolutions in the real plants. Additives, and the el actrostatic neutraliser probein the scale mill both decreased the electrostatic charge. Power consumptionwas significantly lowered as a consequence.

4.3 On-Site Measurements. Milling Plant of the Industrial Partner

Measurements of electrostatic charge of powder: during a real process were

)made using similar principles to those used to measure charge in laboratorytests. Simple equipment based on a Faraday cag~~ received powder collectedin the plant. This permitted immediate “on-line” knowledge of the specificcharge of the powder taken from a real stage of the process. Themeasurements were carried out with or without ]owder additives and/orneutralizers.

The following items summarise the schedule of the electrostatic audit:

- Determination of the critical measuring points- Location of accessible points to collect the sample powder- Careful collection of sample powder to avoid superimposed charges

Measurements of electrostatic charge o‘ the sampleWeighing the sample to ,calculate the specific chargeComplementary measurement of electri c field

8

Electrostatic measurements in the plant, accord in,; to this systematic method,illuminated the location of the problem and dem{lnstrated the results ofadditives, neutralizers and other types of devices used to, improveperformance and reduce the power consumption.

4.4 Research at Other Manufacturing Plants /

Two industrial plants were anaiyzed according tf the systematic techniquesdeveloped during the work with our industrial p: rtner. These processes havein common the use of fine powders with reco@i zed electrostatic problems,as described below.

a) Sulphur plant with a great probler 1 of productivity loss due topowder adhesion.

I

.)

b) Large wheat flour plant with operational problems such asshock discharges to personnel ‘and Ioss of productivity duringsieving.

In both cases, an important problem not related [o the objective of theprogramme was revealed; namely the explosive ~ehaviour of the powders.

A laboratory study of sulphur powder shows that a small proportion ofadditive to powder drastically changes the electr sstatic behaviour. Thisprinciple could be applied to plant to improve tile productivity, also to limitstoppage of the process due to powder aggiome~’ation.

In the case of flour plants no electrostatic additives are permitted. Theresearch for these food powders was focused on the improvement of theprocess based upon modifications of the plant. IJactors such as geometry,materials in contact with the powder, climatic cmditions were researchedand changed. Also, the analysis of powders and material at laboratory scaIehelped to solve problems in the plant.

1.,

\

I

1.

9

5-. IMPROVEMENTS

5.1 Particle Size

IN MIXING AND MILLING AT ~,HE CEMENT PLANT

Determination

Several interesting and useful comparisons betwc en different methods todetermine particle size \vere compiled by the inii ial research. As a result ofdevelopment in this area milled Quartz has been adopted as an internalstandard to verify periodically the laser granulon letry equipment. Thestatistical study of this data confirms the repeata’’}ility and reproducibility ofthe equipment. In a valuable outcome, particle .4 zebetter adjustment of granulometry set points on {he

5.2 Mixing (Raw Mix Homogenisation)

verification enablesmilling equipment.

One of the characteristics most appreciated in th t cement industry is thehomogeneity, as much in the feedstock as in the final product. The mainfeedstock in a cement factory is the raw mix which is fed to the kiln toobtain the portland cement clinker. As a conseq~ Ience, all improvements inraw mix homogeneity improve kiln performance, economy and final producthomogeneity.

In mixing by fluidisation, a technique has been >erfected to quantify the ‘efficiency of the factory’s raw mix homogenisation silo in real work.conditions. By the application of this technique, it has been shown that theactual operating conditions of this silo allows a reduction of 4 to 5 times thewiriation of the chemical composition between the entrance and exit of thesilo.

.)-In the following table the results of two tests based on chemicalcharacterisation (characteristic studied: Lime S: ~turation Factor) are shown.Samples of the m-ix were taken from the entram:e and exit of thehomogeneisation silo.

,,

Tests n“l Test n02

L.S.F. Inlet L.S.F. Outlet L.S.F. inlet L.S.F. Outlet

M a x . 1 3 0 , 3 7 109,55 123,45 108,21

Min- 9 8 , 7 3 102,04 101,32 101,79

St. Deviation 9 , 1 6 2,14 7,22 1,54

Table 1. Lime Saturation Factor at entrance and exit to the silo.

10

I

5.3

From these results, the homogenisation silo is clew] y not responsible ior ourproblems in raw mix homogeneity. So, the other elements (such as transportinstallations, intermediate storage silos etc. ) mus. be checked as suspectmix segregation elements.

Cement Milting (Substitution and Improvement of Additives)

The specific electrical energy consumption in cer~ent milling is in the rangefrom 41 kW/ton to 67 kW/ton depending on the cement type. To reduce thespecific consumption, chemical milling aids are \lsually employed.

In this area, via experimental work carried out ir the plant, it has beendemonstrated (using a high voltage neutralizer) t!~at electrostatic phenomenamay be partly responsible for reduced efficiency in the process. This has ledto the partial replacement of chemical additive b{ an electric neutralizer inthe smaller cement production mill. -

On other hand, as a consequence of the project, :he use of the millingadditives in the factory has been rationalised. By co-operation with anadditive supplier, new and more economical add tives have been developed.

We can evaluate in a general way the influence lf the project on the millingprocess. Energy consumption remains the same 1 Jut the costs of millingadditive in 1995 are more or less 10°/0 lower than in 1994, for the samequantity of cement production.

11,.

‘ 6. CONCLUSIONS

The work completed has a very wide application in an il]exhaustible number ofindustrial situations. There are two or three solutions which may solve the majorityof unwanted charging effects, but each ‘case is different md inevitably creates itsown problems. For example, variations in surface conditions, materials andhumidity, or the geometric features of a plant, wi~l all r~:quire specific ana~ysis inany case.

All powders cannot be categorised by chargeability, sim :e different chargingscenarios can cause quite unpredictable levels of charge Fine powders in particularmay give unrepeatable results for the same test. Even ii a detailed database of ~material charging behaviour could be prepared, it will always be difficult to foreseeelectrostatic problems using this knowledge alone. It is therefore important to seekguidance from expert consultants if the situation is unddy complex, poorlyunderstood or at all hazardous.

I “’1“

I

.)

Laboratory analysis of the powder was crucial to the u] ~derstanding of electrostaticproblems and should be the first step in any research ri :Iated to powder handling.Test-rigs to model each specific problem were used su~:cessfully to categorisepowder charge control techniques in order of effective~ less. In all cases the bestresults could be repeated at large-scale in similar math ines.

There are a limited number of options for dealing witt static charge. Most oftenused are additives, which lower the volume resistivity of the powder, or high-voltage corona neutralisers which overcome the powder tribocharge with muchhigher levels of positive and negative ions.

The degree of success of additives in increasing condt .ctivity depends very muchon the primary powder. But there may be additional physical reasons to preventtheir adoption. Looking at the application to cement o niy, although the resistivitycan be usefully lowered by the addition of 5°/0 graphi’ e, there were two undesirableside-effects. Firstly, the cement was unacceptably disc oloured; secondly, thelubricating effect on the grinder heads reduced their efficiency.

Of the liquid additives tested with cement, water in t} [e form of relative humiditywas found to be most successful at modifying electri( al properties, and SAT not atall. However, conditioning cement powder using humidity was not a practicalindustrial treatment. The whole picture should alway: be considered.

Pneumatic conveying highlights one of the important differences between additivesand neutralisers; that is, their, different spheres of inf uence. A neutraliser can ord ydischarge the powder in its locaIity, and charge will :apidly be re-generated further

-., along the pipework. Local discharging by neutraliser in a key area such as just‘before a cyclone has been shown to prevent cyclone blockages, so the device canbe an appropriate solution. Alternatively, modifying [he powder with an additivewill control charge build up through the whole syste .m. But we should bear in mindthat the powder will not necessarily be completely n mtral at any stage.

I

I 12

Greater certainty can be expected from predictions of clinging when the physicalmodel is close to the real situation. For example, powdt:r charge levels in lab-scalepneumatic conveying did not correlate with chargeabifit y recorded from simpleimpact plate experiments. However, consider the impori ant finding that powder

\ travel ling around a 90” bend, and along straight pipes, ~ vill charge in the oppositedirection. This phenomena was observed in the both the small scale and industrialpneumatic conveyers. In hindsight, it could have been predicted.

Space charge density, and not specific charge, correspol Ids directly to streamingcurrent in pneumatically conveyed powder; thought to Ee the govering factor fortribocharging. Where powder flow is centrifugally conll messed, the space chargedensity increases and the situation is likely to cause po~~der adhesion or choking.

In laboratory scale cement milling, electrostatic charge 1 las been isoiated as thecause. of efficiency 10SS, mostly due to powder sticking :0 milling elements.Neutralisers and additives were both successful at reduc ng power consumption atsmall scale. The ultimate aim was to substitute current c heroical additives with amore economical A.C. corona neutraliser at full scale. 1 unfortunately, the internal

., environment of a cement mill is very aggressive, and ar y ordinary neutraliserwould probably be destroyed by falling rocks, or erodec. A specially designedhigh-power system would be necessary to tackle these problems.

L.O.M. have drawn upon their Brite EuRam experience in two post-projectconsultancies. In both the sulphur plant and the wheat hour plant investigated, anexplosion risk has been identified, and in the second ca’ e, unsafe shocks have beenreceived by personnel. Using an accurate model of the relevant features of thesulphur plant, L.O.M. predict that low productivity COU1 ~ be markedly improved,simply by using specified additives. Where no chemical additives are permitted, asin the case of wheat flour, the options are reduced to changing materials, geometryand ambient conditions, backed up with laboratory anal>’sis of powder.

To Cementos Rezola, as an industrial partner, the most relevant aspects of the

)whole project we~e those related to the mixing and miIli ng processes. As a directresult of. the research into the best methods of particle size measurement, Rezolahave adopted milled quartz as an internal standard. As a consequence, on-lineverification of their granulometry equipment has improved powder quality control.

Analysis of the raw-mix homogenisation silo has shown that although there may beconsiderable static charge at the surface, this is probably insignificant compared tothe other forces, Because of the extremely large size of :he. silo (9.7m diameter and23m height) neutralisation of the surface charge to redui:e the overall static build-up would not be worthwhile.

Experience gained during the Iife of the” project will be continually disseminated toclients through consultancy work on individual industria. problems as they arise.Because of the permanent increases in productivity, or rt~ductions in powerconsumption which may be achieved using the results, tile cost of the project “willbe recovered after a small number of successful solutions have been applied.

13

7.

8.

[1]

1 [2]

[3]

[4]

[5]

[6]

.)[7]

ACKNOWLEDGEMENTS

The work described above could not have been carried i Jut without the support ofthe European Community Brite-EuRam Programrne.

( C o n t r a c t N“. B R E U - C T 9 1 - 0 5 4 8 (RZJE); Project N“. 4538 .

The research opportunities opened up by this funding h: .ve been gratefullyappreciated by al i of the five partners.

REFERENCES.

British Standard B.S. 5958 “Code of Practice for Con, rol of Undesirable StaticElectric@”. (1980)

GLOR, MARTIN. ““Electrostatic Hazards in Powder handling”, Research StudiesPress Limited, John Wiley and Sons inc. (1988)

BAUDET, M. “Mines et Carridres”, (June 1990)

COLE B.N., BAUM M. R., MOBBS F.R. “An investig~ lion of electrostaticcharging effects in high speed gas-solids pipe flow”, Prf )c. Inst. Mech. Eng., 184(3C) (1969-1970), pp 77-83.

EBADAT V., BAILEY A. G.” AND SINGH S. ‘The 1) !fluence of a Super-imposedElectric fieid on Triboelectrfrcation of Powder Particle r in a Pneumatic ConveyingSystem”, Journal of Electrostatics, Vol. 3 (1990), pp 25”’-270

CROSS J. A. “Electrostatics: Principles, Problems ana Applications”, AdamHilger (1987).

BOSHUNG P:, GLOR, M. “Methods for Investigating the Electrostatic Behaviourof Powders’: JournaI of Electrostatics; Vol. 8 (1980), p]) 205-219.

,

14

. ... ,,

/’”Manometer

Airout @ 7

I

Total Pipe Length

Nomina

Powder ---c~llectior

F PowderC y c l o n e In m

II

Y 1

1tt:h.--lf iv,,Vvw vu.”,,

Feeder

D iamete r 30mm.

Figure 1. Pneumatic Conveying Experimental Apparatus

InsulatingBushes

,

\,

EccLen.-:

[;:::,:. .

~ ------------- r - 1 - - - - - - - - - - - - - - - - *,,,,,, ~,,I,*,,I,1,

I

t

tt,1,I,,I

::*IUI

*:;:

● 151Lt

,tI,r)1 - - - - -

- 1- 1H

z

I,t:,,

,1

*,,IB,I

J --

1,

I

,I

.:,,0,,

1,

.1

/1 - - - - - - - - I

T

L

00

“jux

,.)

z+E zn

1 T

,,

,I,1,,aI

I,,

‘------IE==7!

I

I

llllb~llllllll,llllllc

- - - - - - - - - - - - - - - - - - # -- - - - - - - - - - - - - - - - - - - - - - - - - -

E

#l

Air

L

78 , - , /

d.

7,(i

Air Jet

‘l-

2-3-4 -

5 “6-7-8-9 -

2 3

L6

I

FiiterPressure regulatorRotameterSolenoid valve a) vent valve

b) inlet valveDistributor (Whatrnan no 4 filter paper+ support)Glass columnCycloneBag filterCopper wire

,.

around the column for earthing

I Figure 4, Experimental Fluiclised Bed for Mixing Homogeneity Testing