Fuel, Water and Gas Analysis for Steam Users

Home , Bell jar, Retort stand

FUE L WAT ER A N D , GAS A NA LY S IS

M US ER S FOR STEA ,

BY H N B C K H W I E S F. . C . O . . R A $ ,

A u tl wr o S mok P r v n ti n e a tc f e e e o , t , e

W ith 50 Illustratio ns .

LONDO N.

ARC IB CONSTAB E C LTD D . H AL L 8: O .

1 0 9 7 .

1 4 1 0 2 5 AP R 1 5 1910 9 9 52 0 T H N 1 M A$$ $

P REFA C E .

TEAM-USERS have shown a tendency in the past to - - neglect the boiler house for the engine room , and have concentrated their efforts for the improvement of the

e fﬁ ciency of the plant almost exclusively upon the latter . A study of the losses incurred during the conversion of the thermal energy stored in coal into the thermal energy

o f - steam , will show that it is in the boiler house that the

greater preventable losses are occurring , and that the ratio ma y be expressed by the numbers 25 and 5.

It is however , now beginning to be recognized that a s cientiﬁcally managed boiler-house is a sine quanon for the

e i conom c generation of steam power , and considerable attention is being given by steam-engineers to this portion

of their power generating plant .

The chemical examination of the fuel , water, and of the waste gases has been found to be of great service in attain ing the highest efﬁciency from the boiler plant but no ﬁ work has hitherto been published , at once scienti c and - practical , covering the ground required by the boiler house en lneer g .

The author , in the following pages , has therefore attempted

to meet a real want , and has brought together descriptions of apparatus and information as to methods of analysis , which it is hoped may prove of value to all engaged in the m anagement of steam boilers .

as The descriptions and information are b ed , as far as

ri possible , upon practical expe ence with the apparatus and v PREFACE m E ethods Of work described . special attention may be directed to the three chapters in which the practical applica

Of t l tions the tes results are i lustrated ; since , often when

to tests have been made , there is failure make use of the

Of fﬁ Of information obtained , for improvement the e ciency

the plant . It is hoped that a book which is thus based upon science and practice may prove Of considerable service to all en gineers w ho are anxious to obtain the highest possible fﬁ t e ciency from their s eam generating plant . K O B . . $ HN . C ERSHAW E EST ANCAS IRE AB ORATOR TH W L H L Y , ATE R OO IVERPOO W L , L L , anuar 1 1 0 j y 3 , 9 7 .

vi E S CONT NT .

— PART I FUEL .

H P P G E C A . A ORIG IN OF NATURAL FUELS — PHYSICAL AND CHEMICAL C CT I TIC OF W O P T LIGNIT B IT HARA ER S S O D, EA , E , U I N C S I-ANT R CIT ANTH R AC1 TE N M OUS OAL, EM H A E, , A D CO K E— LIQUID FUELS — B RIQUETTE FUELS— FUEL SAMPLING II THE APPARATUS AND ME THODS USED FOR THE A P P R ox 1 MATE ANALYSIS OF FUELS— SPECIAL TESTS FOR SULPHUR 1 7 —2 9 III THE APPARATUS AND ME THODS USED FOR THE CALOR - I FI C VALUATION OF FUELS 30 49 IV THE PRACTICAL APPLICATIONS OF THE TEST RESULTS FO RMS OF FUEL CONTRACT B ASED UPON THE TEST

— PART II WATER .

7 ‘ SOUR CES OF FEED-W ATER SUPPLY— PHYSICAL AND C I C C CT I TIC OF W -W T HEM AL HARA ER S S ELL A ERS, RIVE W T S CE W T N RAI N W T R A ERS, URFA A ERS A D A ER - —COLLECTING THE SAM P LE S 53 7 0 VI TH E APP ARATUS AND ME THODS USED FOR THE AP — P Rox 1 MATE ANALYSIS OF FEED-WATERS SPECI AL - TES TS FOR OIL 7 1 8 5

— VII THE PRACTICAL APPLICATIONS OF TH E TEST RESULTS 8 6 90

V II I T US E S T NING A P T N R G NT HE OF OF E P ARA US A D EA E S , AND THE TESTS NECESS ARY FOR O B TAINING THE B EST - RESULTS 9 1 1 0 1 Vii CONTENTS

— P ART III WASTE GASES . C HAP. PAG E IX THE CHEMICAL AND PHYSICAL CHARACTERISTICS OF THE WASTE GASES FROM VARIOUS FUE LS é— COL - LE CTING GAS SAMPLES 1 0 5 1 1 5 THE APPARATUS AND METHODS USED FOR THE AP PROX IMATE ANALYS IS OF THE WASTE GASE S XI THE APPLICATIONS AND US E OF CONTINUOUS AND RECORDING GAS-TESTING APPARATUS 1 2 8—1 39 XII THE PRACTICAL APPLICATIONS OF THE GAS TEST — RESULTS 1 40 1 51

APPEND IX .

RULES FOR SAMPLING FUEL 1 55 II TYP ICAL TESTS OF ENG LISH AND SOUTH WALES COALS 1 56

’ III GOUTAL S FORMULA FOR CALCULATING THE CALO R IFIC VALUE FROM THE RE SULTS OF THE APPROXIMATE ANALYSIS IV THE US E OF FUELS OF Low CALORIFI C VALUE F O R STEAM-RAISING PURPOSES TES TS OF NATURAL AND FE ED WATERS FROM VARIOUS SOURCES VI TYP ICAL TESTS OF EX IT GASES

IND E x

viii LI ST OF I LLUSTRATI N O S .

FUEL ANALY SIS .

S ampling P late an d H ammer

$ $ G Mi M aand S e e rinding ll , ort r i v s

N alae an d We s E B nc ight - P Copper Air bath with The rmome ter and Gas Regulator

Copper Foil for Weighing P urposes

P orcelain Basins an d S upports for Ash Te sts

Arrangement of Apparatus for Ash Test

P latinum Crucible an d S upport for Coke Test

Lewis Thompson Calorimeter ( original form)

Lewis Thompson Calorime ter (modiﬁe d form) Darling Calorimeter (original form ) Darling Calorimete r (modiﬁe d form) Battery of Bichromate Cells for Fuel Ignition Mould for Making Fuel Briquettes

Wash Bottle for Regulating Oxygen S upply - Mahler Donkin Bomb Calorimeter

Water Vessel an d Mixing P addles

Bomb Cover with P latinum Dish an d Ignition Wire Bomb with Oxygen S upply Conne ctions an d P ressure gauge

Oxygen Cylinder with Reducing Valve an d P ressure Gauge Bomb Immersed and Ready for Ignition of Fuel 1X LIST OF ILLUSTRATIONS

PAG E

WATER ANALY SIS .

Automatic Wate r -sampling Apparatus Me asuring Flasks

Water-testing Apparatus

On e L e 1 . . Mea F a itr , OOO c c ) suring l sk

w -ﬁv P T enty e c . c . ipette

Meniscus Reading of the Burette

Filter-drying Tube

P latinum Basin an d Wate r-bath

S eparating Funnel

Diagram to S implify Water S ofteni ng Calculations

a ar E W e S e 3 2 . Gl ss $ for xamining at r ampl s

GAS ANALY SIS .

’ H onigman s Modiﬁed Gas -sampling Apparatus S ampling Apparatus for taking Average S amples over one hour S ampling Apparatus for taking Average S ample s ove r tw elve hours

’ Improve d Form Of Ors at s Gas -testing Apparatus

’ Improve d Form of Ors at s Gas -testing Apparatus S oot Filter

Moisture Absorption Apparatus

Water Aspirator for taking S ampl es

Ad os Recording Apparatus ( Old form) Ad os Recording Apparatus ( ne w form) Typical Records Of the Ad os Apparatus Uehling Recording Apparatus

’ 4 5 Arndt s Gas Balance

e e aa 46 . Kr ll R cording App r tus

el Re A aa 4 7 . Kr l cording pp r tus X LIST OF ILLUSTRATIONS

F I G G . PA E H - - 4 8 . aber Loewe Gas testing Instrument

4 9 . Diagram of H eat Losses in E xit Gase s

APPEND IX .

’ 0 Diara f use i Gou tal s F a 5 . g m or w th ormul

° P A RT I FUE L.

CHAPTER I .

Naturaand A rt ﬁ a Fue — The r Or n l i ci l ls i igi , om o on f C p siti and Me thods o S ampli ng . H E form in which fuel may be used under steam

boilers , in order to produce the heat energy which to i it is desired util ze in the form of steam , is exceedingly

d . are iaverse Natural fuels either saolid or liquaid . but both v rieties , by aid Of suitable appar tus and pl nt , may be A converted into gaseous fuel . natural gas which may be used as fuel is also found in certain localities in E urope i an d America. In the latter country it s largely used for A lighting an d heating purposes . t the present time there is a decided tendency to employ fuel-gas in place Of solid fuel for many purposes in the arts and industries . The c ontrol Of the combustion process is much more easily c arried out w hen employing fuel-gas than When using solid fuel and the loss Of heat which occurs in the gas-produc ing plant is to some extent balanced by the value Of the a i mmonia and tar products , Obta ned by suitable treatment l Of the producer gases . The su phur impurities of the coal - al b . may so be removed , and Obtained as valuable y products Although gaseous-fuel has not been applied on any con sid erable to ﬁr scale the ing of steam boilers in this country , this method of steam raising will receive increased atten ’ tion and application in the future ; and , in the writer s O use Of - as is pinion , the increased fuel g likely to prove one solution for the smoke problem which now confronts manufacturers using solid fuels for heating purposes in large cities . Carbon and hydrogen are the essential heat-producing elements in all natural or artiﬁcial fuels . The amount Of heat produced by the complete combustion of one pound of fuel depends upon the percentage of these tw o elements i . C s present in the free or combined state 5 ombustion in com reality an oxidation process . The products Of the lete m are C0 p co bustion Of fuel carbon dioxide gas ( 2) and aqueous vapour The incomplete combustion of AND GAS FUEL , WATER ANALYSIS

of fuel produces a number of secondary products , chief which are carbon monoxide gas (CO) and various hydro carbon gases (methane , ethylene and acetylene) ; these latter being either present as such in the fuel , or formed by com bination of the carbon and hydrogen of the fuel when it ﬁ an . O is rst heated n further heating , in presence of Of adequate supply of air , these gases burn with evolution heat and production of carbon dioxide gas an d aqueous s vapour . The further chemistry of the combustion proces w i C ill be dealt w th in hapter IX . It may , however , be stated here that the exact nature of the physical and chemical changes which occur when coal is heated is still the unknown . Chemists have not yet discovered whether hydrogen present in solid fuels exists in the free or combined B e i state . o rnste n has recently published an investigation M h I d . r on hem. n see o . C bea ing this subject ( S c , arc

1 1 0 6 . 5, 9 , p Natural solid fuels may be classified in the following - : m al m an order wood , peat , lignite , bitu inous co , se i - thracite and anthracite while petroleum and its by pro ducts stand as the representatives of liquid fuels ; an d - - coal gas , coke , peat coke and briquette fuels represent the fi arti cial derivatives of the above natural fuels . All naturally occurring solid fuels are held to be modifi Of fi an d cations wood and woody bre , absence of air , heat pressure being the agencies which bring about decay an d z fossili ation of the vegetable matter . Geologists assert that anthracite is the Oldest of the

al . C natur solid fuels ellulose , the chief constituent of and ﬁ wood woody bre , is represented by the formula C H 0 and n a 6 1 0 5 , contains 44 per cent . o ly Of c rbon while an 0 e Of n thracite contains 9 per cent . Th process conversio of wood into coal is thus accompanied by a gradual escape

Of . ortion of 6 p the oxygen and hydrogen , which make up 5

. Of of l and al per cent the weight ce lulose , a gradu increase the in density and hardness of the product Of this change . L or i or al ignite , soft bitum nous coal , and hard steam co , are the successive stages in the conversion Of wood into P l anthracite . eat is a similar product from the plant ife an d of which grows in bogs , Of less age therefore less fos silized character, than coal . In peat , the percentage of volatile matter which is given Off on heating is Often as 6 er . is n high as 5 p cent , and the solid carbon thus less tha 4 FUEL on e- third Of the total weight Of the peat . Time , heat , and pressure would no doubt convert peat into a fuel resembling al co . P etroleum and natural gas are fuels that have been pro d uced by an other series Of changes in the strata Of the ’ earth s crust but limits Of space will not permit Of any discussion Of the various theories relating to the formation O f these products .

WOOD .

Although wood is not used to any great extent as a fuel u a in co ntries with extensive and av ilable supplies Of coal , yet as the original material from which our coal supplies r have been de ived by natural processes , its composition and characteristics deserve some mention here . Wood al is of an l in its natur state composed thickened pl t ce ls , which on microscopic examination reveal three constituents - ﬁ or al the woody bre cellulose , the sap , and the miner s alts taken up from the soil . sa al The p consists chieﬂy Of water , with the organic s ts ,

ri Of or . required for nou shment the tree plant , in solution In freshly felled timber the water present in the sap may f l amount to 45 per cent . o the totaweight Of the wood ; by air drying this can be reduced to from 1 0 to 0 to 2 per cent . according the nature Of the wood -ﬁbre i and density Of its ﬁbres . The woody conta ns from i - 0 to . a s 4 44 per cent Of c rbon , and the real heat producing e l t ement Of wood . Owing o the large amount Of hydrogen u and oxygen present in cellulose (C6 H 1 00 5 ) in the molec lar l proportions Of aqueous vapour , the thermal va ue of wood

1 . as a fuel is not high . Direct tests have shown that lb Of - air dried wood on the average will only evaporate 4 to 5 lbs . '

low 1 . Of water . Owing to the percentage (under per cent ) Of al miner matter present in wood , however , after dry dis tillation has removed the water and other volatile constitu e Of w ood nts , it is one Of the purest forms carbon , and ckarcoal an , as the product is called , has evaporating efﬁciency more than double that Of the wood from which i use . o s it is derived The c st , however , prohibitive Of its as a fuel on a large scale . 5 AND GAS FUEL , WATER ANALYSIS

EAT P .

P eat is formed from plants and shrubs which ﬁnd their ri ro habitat in swampy dist cts , under conditions which p duce decay in the absence Of oxygen or air under water . Owing to the peculiar conditions under which peat is r an d ash is fo med , the amount Of water which it contains al tw o ns on e ways high , and these co tituents an averag amount to more than three-fourths Of its composition by l weight . The following are approximate analyses Of typ ica samples Of peat

English Peat.

Ah w k md .

Moisture As h Coke Volatile matter Fixe d carbon

Air- l 0 dried peat sti l retains from 20 per cent . to 3 per

. an d on its al as cent Of water , this account v ue a fuel is very low . B y mechanical and Chemical treatment this water one an d can be reduced by half , peat in this state then becomes a fuel comparable in appearan ce an d evaporative al v ue to bituminous fuel . - It is probable that as the coal ﬁeld s Of this and other Of countries become exhausted , the immense reserves peat

’ all i be d raw n and which exist in nearly countries w ll upon , ﬁ use that by modi cations Of processes now in , supplies Of peat-coke and peat-fuel will be made available forindustrial

and household use .

LIGNITE OR B ROWN COAL .

Lignite is a natural fuel intermediate between peat and

bituminous coal in age an d appearance . Lignite is formed by the gradual decay Of shrubs and trees under conditions which have excluded air an d oxida f ar al e tion , but have di fered m te i l y from those which hav

Obtained in the formation Of peat . The percentage of water present in freshly w on lignite is however high— vary

AND GAS FUEL , WATER ANALYSIS

B ritish thermal units . It is interesting to note e that it contained 3 89 per cent . Of hydrogen ov r and above to 8 Of that required combine with the 7 2 per cent . free a 1 6 oxygen present in the co l ; while 4 5 per cent . Of the carbon w as present in some form which yielded volatile - a A h hydro c rbons on heating . lthoug much has been Of written upon the subj ect coal , the exact constitution an d state Of the various Chemical elements before heating are still unknown and it is an undecided point as to whether the oxygen an d hydrogen and volatile portions Of the carbon are present in coal in the free or combined state . A l Of pproximate ana yses other bituminous fuels , with al ﬁ lb - ri . their c culated calo c values in centigrade units , are given belo w

Calculated Coke .

Caloriﬁc Value .

All the above tests were made on samples dried at ° sh in 2 0 F . a 3 , and the low percentages Of contained samples N e to Of OS . 3 , 4 and 5, wer due these being washed Slacks e xceptional cleanliness . l When the percentage Of volati e matter , yielded by a fuel on heating , rises above 35 per cent . , the fuel is exceed ingly diflicult to burn without smoke production . The percentage Of volatile matter given Off by a fuel is in reality ffi Of a measure Of the di culty burning it under the boilers , B with perfect combustion Of its component parts . itu minous coals containing over 36 per cent . Of volatile matter cannel an d as are known as coals , are used chiefly for g All making purposes . bituminous fuels require careful - management and large brick lined combustion chambers ,

if they are to be burned without smoke . The length Of ﬂame produced by bituminous coals de Of pends upon the percentage volatile matter , and this in its turn depends upon the amount Of hydrogen they con tain over and above that required to form water with the

oxygen present in the coal . 8 FUEL

This surplus hydrogen combines with the carbon to form CH C H C H methane ( 4 ) , ethylene ( 2 4 ) , and acetylene ( 2 2) three hydro-carbon gases which yield carbon dioxide and aqueous vapour when burned under perfect conditions of c i ombustion . When the air supply or the temperature s ffi t al o insu cient , par i oxidation nly occurs , and a portion Of the carbon Of these gases separates in the form Of soot . ae h The percent g Of surplus ydrogen is , therefore , another difli l measure of the cu ty of burning bituminous fuel . This d ifliculty increases with the ease and rapidity with which these hydro-carbon gases are evolved by the fuel on heating . Nearly all bituminous fuels contain sulphur in the form of hi Fe an ( S ) as ri . iron sulp de , ”impu ty The percentage of or a f iron sulphide , brasses , v ries greatly in di ferent

. O is classes Of this fuel n heating , half Of the sulphur liberated and burns with evolution Of heat to form sul phurous acid gas (SO which sooner or later condenses in . of ri H un the form sulphu c acid ( It is this compo d , and or not the carbon soot , which produces the larger part Of the destruction Of vegetation and of buildings in the n f eighbourhood Of industrial districts and o large cities . 1 Recent calculations have been made to Show that in London over tons Of sulphuric acid are produced annually by the sulphur impurity contained in the coal A burned in the metropolitan area . similar calculation for the coal burned annually in the whole of the United K Th ingdom gives to tons of acid . e use Of lime has been advised for fixation Of this sulphur in the r of fo m Of calcium sulphate in the ash the fuel , but a a far more certain and safer method would be to pass the products of combustion through scrubbing towers filled w i flints - as th . The sulphur contained in the exit g could of b - then be recovered in the form a valuable y product , - n of or . As amely hypo sulphite lime soda already stated , is h not it the sulp ur , and the more visible soot and smoke , f which is the chief cause of the damage e fected by smoke . r m The carbon blackens ; but the sulphur co rodes etal ,

d . An rac estroys brick and stone , and kills vegetation y p al tic method which would remove the sulphur from coal , o r its products from the waste gases , deserves , therefore ,

1 D d L S A e C fe e e Dec . 1 0 R a e ae . . i eat r l ondon mok b t m nt on r nc , 9 5 9 AND GAS FUEL , WATER ANALYSIS

- U l h most careful consideration by steam users . nti suc method has been generally adopted the sulphur contents al to as low of coal ought ways be kept as possible , by care ful picking Of the”fuel and removal of the veins Of pyrites al r . s Of or co b asses In large purcha es fuel by contract , a clause ought to be added w ith respect to the amount of al this impurity, and no co should be accepted containing over one and a quarter per cent . of sulphur .

- - SEMI ANTHRACITE OR STEAM COAL . 0 Fuels which when heated yield from 1 0 per cent . to 2 - es per cent . Of volatile matter are known as semi anthracit - or steam coals . They are very largely used for steam r R al an d r d raising pu poses in the oy Navy , by facto ies locate A in cities where smoke prosecutions are frequent . p proximate an alyses Oi typical semi-anthracites are given below

Coke .

Fuels of this type can be burned without smoke produc Of -fum ace n ot e tion in most types boiler , and they do requir large combustion chambers or very skilled management in s order to Obtain perfect combustion . A shown by the l approximate analyses , the greater portion Of the fue remains as an incandescent mas s of coke after the pre r ﬁre- of an d limina y heating upon the bars the furnace , radiation transfers the heat from this mass of glowing fuel to the water in the boiler .

NT RACITE A H . Anthracite is the nearest approach to coke in chemical f m constitution which is found naturally , but it di fers fro of coke in its physical properties , being very dense and A to shining appearance . nthracite gives Off only from 3 I O FUEL

1 0 f . O s per cent volatile matter when heated , and consist

ﬁ . chieﬂy Of xed carbon , with little oxygen or hydrogen The following is an approximate analysis of a typ ical South Wales anthracite fuel

Ash. Volatile matter. Coke . Caloriﬁc alue . - - v 1 0 e r e . 6 er e . 6 6 er e . 8 0 7 p c nt 3 4 p c nt 9 3 p c nt , 3 5

A s diﬂicult n to nthracite are to ig ite , owing their density and low percentage Of volatile matter ; but under proper Of conditions this type fuel can be burned completely , and

yields a very intense but localized heat . Anthracite is i ch eﬂy used in metallurgical Operations , and is not employed to any great extent for steam-raising purposes but w hen the SO employed no combustion chamber is required , and boiler surface may be placed Close to the mass Of burning

fuel .

I B - PETROLE UM AND TS Y PROD UCTS . Although liquid fuels have little application in the United K i for ingdom , ow ng to the limited supply available indus l R U tria purposes , in ussia and the nited States they are - Cr e largely used for heating and steam raising purposes . ud petroleum contains a large number Of liquid hydro-carbons C H + is r Oil or Of the general formula n 211 2 , and it the c ude a ﬁn a for the residues Obt ined in re ing it , th t is used steam R l raising purposes . In ussia especial y , the residue fr”om ﬁr ll Oil astatki is the st disti ation Of the crude , called ,

utilized also most entirely in this way . The following tests 1 show the difference in composition and heating value Of the Russian and American Oils

Caloriﬁc

P er cent. P er cen t. P er cent. Value .

B . Th. Un t . d ro en . Ox en . i s Carbon . Hy g yg

- Russian crude light 1 0 - he av y I 1 1 - astatki 1 2 - P e nnsylvaniacrude I 4 - W est Virginiacrude 5 I

When burnt in suitable furnaces with proper provision o f r the supply Of the Oil in the form Of a spray , petroleum

to l has . is superior ordinary coa as a fuel , and 33 per cent

1 The Calori c P ow er o F uels H . P e . ﬁ f , ool I I AND GAS FUEL , WATER ANALYSIS

al higher evaporative v ue than anthracite . Actual tests Of 0 have given an evaporation 2 8 lbs . Of water per lb . Of petroleum . Liquid fuel is e asily transported and stored ; yields no ashes on combustion ; and were the supply Of suﬂicient is crude Oil , there little doubt that it would be very largely used for steam-raising purposes in this and

ri . U other count es nder present conditions , however , the price Of petroleum is too high to enable it to compete with

coal as a fuel .

ARTIFICIAL FUELS . Gas-Coke and Coke-oven Coke — These two forms Of fuel are Obtained by submitting bituminous fuels to dry-distil lation— that is by heating in closed retorts or ovens in the

absence Of air . The moisture and volatile constituents Off and im Of the coal are drivean , only the carbon , ash , and purities remain as solid residue in the retorts or ovens . Gas-coke is the residue Obtained in the manufacture Of coal and Of Of r im gas , here the quality the coke is seconda y

portance to that Of the gas . The coke produced from coke

Ovens on the other hand , is the Chief product Of the coking use process , being designed for in the iron smelting industry , and in many cases the gases from coke ovens are allowed

to escape into the atmosphere . for - r The use Of coke steam raising purposes is ve y limited , since its price is comparatively high ; but it is sometimes to purchased and mixed with bituminous fuel , lessen the

trouble from smoke in very heavily worked boiler plants . Coke can be burned completely without any combustion

chamber . The following is an approximate analysis Of a typ ical coke :

latile matter. oke . Caloriﬁc alue . Ash . Vo - - C v 1 0 er e . 8 I O er e . 1 8 8 5 p er ce nt. 9 p c nt 9 p c nt

Ow to ing the loss of the volatile constituents Of the fuel , the percentage Of ash in coke is always higher than that of

the original fuel .

Coke contains as a rule from 2 0 per cent . up to 6 0 per Of cent . moisture , owing to its absorbent properties and to the fact that on drawing from the retorts or ovens it is

Often quenched with water . Coke retains also a large portion Of the sulphur impurity 1 2 FUEL of not the original fuel , and it is therefore a very suitable fuel for use in towns or cities for the reasons already given B itumin u Fuel under o s . P ea- — aﬁ l t Coke . This rti cial fue is Obtained by the dry f of a distillation o peat in retorts special design , after a l rge portion of the original water contents Of the peat have been Th removed by mechanical methods . s following is an analysis Of a peat-coke prepared by an electrical method

Moisture Ash Volatile matte r the e a on dri d s mple . Coke Fixe d Carbon

This test must not be taken as representative Of peat l ﬁ coke genera ly however, as it is possible by scienti c methods to obtain a much lower percentage of moisture ﬁ an d of volatile matter in the nished product . The ash test in this case has also been increased by the unwise f e addition o Chemicals to promote lectrical conductivity . A sample Of peat-coke from Argyleshire contained only

6 and 0 . ash 5 0 per cent . moisture 2 7 per cent while Nor - 8 w egian peat coke contained 4 2 per cent . moisture and

per cent . ash . Although peat-coke has n ot so far been manufactured is n ow upon a large scale , it being produced at several w C aa G small works in Nor ay , an d , Ireland , ermany , and - -fi other peat producing countries . When the coal eld s Of E A a a- is urope and meric are exhausted , pe t coke likely ’ to play an important rOle in the maintenance of the world s low a . O industries wing to its percent ge of volatile matter, - is and freedom from sulphur , peat coke a far more perfect fuel for household and industrial use than ordinary coal . B -F — B i - l riquette uels . r quette fue s are made chiefly from - semi anthracite and anthracite coal , by mixing the dust fine Of n and and coal obtained in the course mini g screening , i r and with tar or other bind ng mate ials , by moulding the

i . mixture nto bricks , in specially designed presses The resulting briquettes burn Slowly with evolution of little ﬁnd use G a E smoke , and a large in erm ny and other uropean l countries for domestic purposes . The fo lowing shows the composition of such a briquette-fuel made from German peat 1 3 AND GAS FUEL , WATER ANALYSIS

Moisture Ash oke C d r a e on y s mpl . Volatile matter Fixe d Carbon

As the cost of briquette-fuel delivered in manufacturing d Of is for istricts is higher than that coal , it little used indus - trial and steam raising purposes . - — Gaseous Fuel. There are many engineers who believe that the power and heat required by our industries in the - . future will be , Obtained from gaseous fuel Some Of the of of advantages this form fuel have already been mentioned . O r ﬁ as- ro the s are , that with slight modi cations in the g p d ucin g apparatus , the gas Obtained can be used either for - — heating or for power purposes . Large gas engines d riven by producer gas (Obtained from ordinary bituminous slack

. d . n w c osting only 63 per ton d . ) have o been running s G A atisfactorily for some years in ermany , in merica , and

K . in the United ingdom ; and as one h p . can be obtained f . O from one lb coal by this method , the advantage over w - the present methods of po er generation are undoubted . E ven with the existing power plants— consisting of steam boilers and reciprocating or turbine-engines— there is much to be said for heating by gas and in many works the instal lation of a large gas-producer plant would lead to fuel

e and n l l . conomy, i cidenta ly also so ve the smoke problem P roducer-gas varies in composition with the type of p ro d ucer of used , the method working , and with the composi tion of the fuel consumed . When steam is used in the - working Of producers , a mixture of ordinary producer gas - and water gas is obtained . Carbon monoxide and hydrogen are - as the essential heating constituents of producer g , and upon the percentage of these constituents present in the

as its e . g , th rmal value depends The earlier forms of gas-producers could only be worked - s atisfactorily with semi anthracite and anthracite fuel . Modern producers and methods of working permit the c heapest forms of bituminous slack to be employed with satisfactory results . The ammonia recovered in the form - o f sulphate as a by product is also of considerable value . The following is the composition of gas from a typical M - as i ond g plant , us ng cheap fuel 1 4

AND GAS FUEL , WATER ANALYSIS The sampling of peat and briquette fuels is best carried out by taking borings out of a large number of selected

blocks , and reducing these to powder by rubbing between A serrated iron plates . proper sampling plate and suitable — — hammer (see Fig . 1 ) are required for the preparation

I — S L L H ’ FIG . . AMP ING P ATE AMMER .

of fuel samples when this work is often carried out ; and in large works where fuel sampling is a regular portion of - ﬁ of the scienti c management the plant , a small mortar mill , for or large crushing and grinding mill , are essential the

speedy performance of the work . The ordinary type of coffee mill is u seful for crushing samples of fuel in the ﬁnal stage of their preparation for laboratory examination .

1 6 CHAPTER II .

The A rox mate A na of Fue pp i lysis l .

PREPARING THE SAMPLE .

H E sample of fuel after the operation of sampling as - de scribed in Chapter I should amount to about one i - . s pound , w th no particles that exceed i inch in diameter Thi should n ow be passed through a brass wire sieve (A) 2 of 1 -1 2 not (Fig . ) inch mesh , and the portion which will ﬁ pass through should be nely crushed by grinding in a NO . 3

Kenrick mill. Fig . 2 (B ) shows a small mill adapted for

F G 2 — LL N S S A D . I . . GRINDING MI , MORTAR IEVE

the preparation of laboratory samples of fuel . If the fuel or s be very wet , if the sample contain any very hard lump s of shale , this method of preparing the sample require modiﬁcation . The excessive moisture can be removed by spreading of on out the whole the fuel sample a piece of paper , and — tw o by leaving it exposed to the air for or three hours , in a warm place in the laboratory— before sieving and grind ing . The difﬁculty caused by very hard lumps of shale can be dealt with by picking these out of the sample and by crushing them separately in the steel mortar (C) shown in 1 7 C AND GAS FUEL, WATER ANALYSIS

C of l Fig . 2 . are must be taken that none the sha e is lost in a for l this oper tion , shale possesses little heating va ue , an d the proportion of it in a sample of fuel has great in for ash fi fl uence upon the tests and calori c value . Shale and of the f dirt are , in fact , the chief causes di ferences in l and l value of bituminous fue s , when the thermal resu ts are w or orked out upon pure coal combustible matter only, the variations are small and are chiefly those due to differences

in the Chemical constitution of the coal . For this reason it is wise to cover the mortar and the grinding mill with a c a rd or towel when crushing shale , since bits have a decided t to out to endency j ump and get lost during this operation . A of fter crushing in the mortar , the whole the shale must be

passed through the grinding mill, and the ground material r r of l h l etu ned to the general mass the samp e , w ich wi l now r occupy double the fo mer volume , and resemble in

appearance aheap of coarse gunpowder . The sample is then reduced by repeated mixing and q uartering— as described in the Appendix— until only about

1 20 grams (i lb . ) remain . This is divided into two equal of t p ortions , one hese being used for the following tests , and the other being reserved in a stoppered bottle for Use in c ase of any mishap in carrying out the tests with the ﬁrst

p ortion . The remaining portions of the sample are likewise not

t 1 . e of hrown away , but are retainead in the lb tin in cas accident or dispute . E ach s mple tin Should bear a label . of giving full particulars the date of sampling , the origin Of l the fue , and any other information regarding it that may be necessary for proper identiﬁcation of the sample and f test . In cases where a large number o samples are being d one m a ealt with at and the same ti e , consider ble care is n of ecessary to prevent confusion the samples , and the use F Of written labels is all the more urgent . or testing pur

l A B C. . p oses the samp es may be simply marked , , , etc ii the original sample tins bear corresponding letters— in il f addition to the full deta s o the fuel and date of sampling .

ESTI N E UE T G TH F L . The approximate analysis of fuel covers the estimation Of constituentS z— I 2 the following ( ) moisture , ( ) ash , a ﬁ 6 ( 3 ) volatile m tter , (4) coke , (5) xed carbon and ( ) sulphur . These w ill now be dealt with in the order named 1 8 FUEL

M — 1 . oist a ure . Ten grams of the fuel sample prepared as lready directed are weighed out upon a chemical balance a 1 l are tw o re ding to mil igram , and heated for hours at ° ° - 2 0 . 1 1 r . 3 F ( 0 C. ) in a copp er o aluminium air bath The loss of weight multiplied by ten gives the percentage of

moisture in the original sample . A desiccator must be for employed cooling the sample before weighing , as perfectly

dry fuel absorbs moisture from the air and gains in weight ,

b on . even while eing weighed the balance It saves time , l to however, if the crucib e be allowed go nearly cold in the air with the lid on , before being placed in the desiccator .

— L AND H S . . C FIG . 3 BA AN E WEIG T

On no account must the crucible be placed on the balance ll hot ll pan and weighed while sti , as this wi give increased

weighings . The crucible must be covered during heating in the air bath and the lid only removed when weighing

The balance shown in Fig . 3 is most suitable for fuel testing work it carries a load up to 50 grams and reads cor rectly to one milligram . The weights should always be 1 9 AND GAS FUEL , WATER ANALYSIS

in . placed the left hand pan , the fuel in the right The sample not of fuel should be weighed out directly upon the pan , n but in a tared porcelai crucible . (A NO . 1 B erlin crucible

holds I O grams of most fuels . ) A lead counterpoise can o be made f sheet lead with little trouble , and the operation of weighing out ten grams of the fuel sample with this dl crucible is then very rapi y performed , since the lead

counterpoise undergoes little change in weight , and only '

rarely requires fresh adj ustment . The lead counterpoise L e should be bent into the form Of the letter , to facilitat ff a removing on and o the pan of the b lance . The second Of tw o r weighing the fuel sample , after heating hou s at ° l l 2 30 E . can be most quick y performed by p acing the lead counterpoise and the ten gram weight on the left hand pan of d the 1 the balance , and by ad ing fractional parts of gram l to the right hand pan upon which the crucible is placed, unti ﬁ equilibrium is obtained . The nal adj ustment should always be made by means of the rider— (the little platinum weight which rides across the beam and records decimal fractions of one centigram)— with the front of the balance case closed ; t in order o avoid any disturbance by draughts . The total of the weights palaced on the right hand pan to restore equilibrium after he ting , then represents directly of the loss moisture , and multiplied by ten gives the per f l centage O moisture in the original samp e . The air-bath most suitable for heating fuel is Show n in f . o Fig 4 . It is made copper , and is provided with a ther ° mometer 0 0 F . reading up to 3 , and with an automatic gas ° f regulator which keeps the temperature within 5 F . o that desired . That shown in Fig . 4 is the Reichardt Muencke

of . is B as r type regulator The burner a unsen g bu ner , with to l to a rose p , and this must be p aced quite close the bottom U - of . the oven nless a gas regulator be used , the air bath - s requires constant attention , for the gas pressure varie f Of greatly at di ferent hours the day . The only precaution necessary with a gas-regulator is to see that there is no danger of lighting back when the main gas current is entirely b - off . cut , and only the y pass supply is in use A - -b water bath is less troublesome than an air ath , but ° the temperature attained in these cannot rise above 2 1 2 F and therefore they are of no use for determining the moisture of in fuels , the last traces which can only be got rid of by ° heating above 2 1 2 F . 2 0 FUEL

Should it be necessary to determine the total moisture or as is in very wet fuels , in the fuel delivered , it necessary to determine the loss which occurs during the grinding and s ampling operation . of For this purpose , the whole the sample contained in

— - H H H AND S- L . C FIG . 4 OPPER AIR BAT WIT T ERMOMETER GA REGU ATOR . the sample tin is placed on a porous plate and is weighed as l as a accurate y possible on a p ir of ordinary scales . The its or plate with contents is then placed in the sun , in a warm for tw o place on the boilers hours , and the fuel is twice turned over during this period . A second weighing of the 2 1 AND GAS FUEL , WATER ANALYSIS

f plate is made when cold . The loss o weight during this preliminary drying is then worked out as a percentage on the inal of l for h orig weight fue , and must be allowed in t e ﬁnal al c culation of the moisture , as ascertained by the second an d more accurate drying operation .

The remaining tests are all carried out with the 1 0 gram ° a of - 2 0 s mple fuel , after drying in the air bath at 3 F . °

1 1 0 C . a ill re ) This s mple must , however , be st further d uced in ﬁ neness before commencing these further tests . The sieve (D) and steel mortar and pestle (C) shown in

for . Fig . 2 are employed this work The former is of very fine as 1 -60 al to br s wire , inch mesh , equ holes to the square inch ; the latter is of cast steel and is 4 inches in Of 1 0 diameter . The whole the grams sample must be passed through this sieve . It will be found that this is most l r out r of quick y ca ried by repeated sievings , and retu n the r fi to larger pa ticles free from the ner portion , the mortar . - and fi r With semi anthracite shaly fuels , this nal g inding r Operation is troublesome , but it is essential that it be ca ried of out carefully , and that the whole the ten grams be passed 1 -60 through the inch mesh sieve , if correct results are to fi r of be attain ed . The nal pa ticles very hard coal or scale u to on no which resist cr shing the last , must account be AS C rejected . already stated in hapter I , correct sampling of too is the basis correct testing , and it is Often Slovenly and t l . o careless y performed If left untrained persons , it is to out of certain be badly carried , and the whole the test

$ results are then worthless . The fin ely ground sample is now carefully mixed on a f t a o o . OO sheet glazed paper , and is then transferred NO is (27? inch diameter) porcelain basin , which covered with ’ - one a clock glass , and is placed in the air bath for hour s - ° t e heating at 230 F . This is necessary to remove the moisture taken up during n the final grinding of the sample . The sample is ow ready for and the remaining tests , must be transferred to a stop s n pered gla s weighing tube , or kept covered in the porcelai l basin in the desiccator . Shou d the tests be delayed some re- - hours , a heating of the sample in the air bath for an hour ll fine wi be necessary , as coal in this state of subdivision is very hygroscopic . 2 2

AND GAS FUEL , WATER ANALYSIS

manot of Th y bake itself into a very hard mass coke . e fuel Should be turned over repeatedly with a platinum wire or l of S , patu a durinag this stage the heating process , to facilitate the esc pe of the hydrocarbon gases , and to pre f vent Sintering together o the mass . After the hydro c arbon gases have escaped , the heat of the ﬂame can be of gradually increased , until the bottom the basin where the

— . . F S F A S FIG 7 ARRANGEMENT O APPARATU OR S H TE T .

a l fi fuel rests is t a du l red heat . The fuel as the xed carbon burns away gradually becomes lighter in colour , but it is somewhat difficult to get rid Of the last few per cent . of f l . to a carbon In order e fect this , the b sin shou d be removed to on of from the flame , allowed cool a piece asbestos board , and the ash carefully crushed with a small agate or glass ash t pestle . The adhering o this is then brushed back into ’ the fine basin with a small camel s hair brush , the carbon d ust which collects round the edges of the basin is brushed down into the centre , and the whole is again raised to a red heat temperature , commencing with a small flame to avoid cracking the porcelain basin . 24 FUEL

The ﬁrst heating of the fuel to remove the hydrocarbon gases and to burn Off the greater part of the ﬁxed carbon of the coke requires generally from 3 0 to 60 minutes and afurther 1 0 to 2 0 minutes heating are required to remove

the last traces of carbon . From 1 } to 2 hours are therefore required for carrying out to the ash test , and any attempt reduce this time with b ituminous coal will produce incorrect results . The ash s hould be light in colour and quite free from black specks , l which denote unburned carbon . The basin after remova f l and rom the ﬂame is a lowed to nearly cool in the air , is then placed in the desiccator and its contents weighed when on to c old . The ash is carefully brushed out of the basin an of or on to for the p the balance the copper foil weighing , and the spatula or wire used for mixing the fuel must also be carefully brushed free from any particles of ash adhering

to it . The weight of ash multiplied by 50 gives the of percentage ash in the dry fuel . Ash which is red in colour denotes the presence of

iron impurities in the fuel , probably as coal brasses

( pyrites) . The behaviour of bituminous fuels on heating is quite d ifferent from that of semi-anthracite and anthracite fuels and after a little experience in testing coals from various s to ources , much information can be Obtained as the char a and a of cter qu lity the fuel , by careful Observation during the ash test . When a large number Of ash tests have to be performed d f to aily , it saves time to use a mu fle heated redness in -muffle e a a a gas furnac , the fuel s mples being cont ined in s hallow trays made from platinum foil . These must be numbered from 1 to 6 or 1 2 by notches cut with scissors in con their edges , and great care must be given to avoid Of the aro er fusion samples , and to the question of p p current f r O . too o air over the samples If the heating be rapid , if h too of fi l the draug t be great , some the ner partic es Of fuel or

ash . For may be carried away with the exit gases . this reason the porcelain basin method is the safest for ash deter minations non- r mufﬂ e , especially for those conve sant with f urnace management . The ash Obtained from each fuel test should be placed in a small paper bag and kept in the for a tin with the original fuel sample , reference in c se of t or rouble with the fuel Of dispute . 2 5 AND GA FUEL , WATER S ANALYSIS

Vol ti M n . . ale atter Coke and Fixed Carbo 3 , 4 , and 5 , These three constituents Of fuel are all determined by the One test which consists in heating one gram of the ﬁnely crushed an d dry sample of fuel in a covered platinum cru c - The ible until all the hydro carbon gases are expelled . l s crucible and its contents are then again weighed . The os in weight multiplied by 1 0 0 gives the percentage of volatile 1 00 matter , and the residue multiplied by gives the per of f sh as centage coke . On deducting the percentage o a

— . 8 . L C C L AND S FOR C K S FIG P ATINUM RU IB E UPPORT O E TE T .

2 of one s ascertained in test from the percentage coke , Obtain the percentage of ﬁxed carbon . Very careful manipulation is required in carrying out t e his test in order to Obtain correct r sults , and the following a is conditions must be strictly adhered to . The pparatus

8 . e shown in Fig . The platinum crucible Should measur x inch diameter and Should be provided with a ﬁ out of very closely tting lid . If the crucible has been bent to ﬁt its shape , causing the lid badly , it can be restored to l on r hot original circular form by ro ling a hard su face while , with pressure applied by a rounded stick or glass rod in its 2 6 FUEL

interior . The crucible and lid should be thoroughly cleaned r Off after each test , by bu ning the adhering graphite and l - sea . soot , and by po ishing inside and out , with wet sand The grey deposit formed by the burner ﬂame on the bottom of the crucible — (due to formation of a carbide of platinum) — s sea- as hould also be removed by sand after each test , a platinum crucible rapidly corrodes and deteriorates when

dirty and dented . A platinum wire triangle is made for holding this crucible t of by twisting ogether three p ieces thick platinum wire , and by mounting these in the centre of an ordinary pipe as 8 . f . 0 o su stem triangle , Shown in Fig N other method p in for porting the crucible the ﬂame is admissible coke tests , since it is absolutely essential that the crucible and its lid enti el su ounde m should be r y rr d by ﬂae during the test . The B r of . 1 unsen bu ner must be NO size , and should give a

steady colourless ﬂame at least 7 inches high . The tripod 8 so stand should be inches in height , and must be arranged of that the bottom the crucible , when supported by the plati l not 1 to num wire triang e , is more than 5 inches from the p he of the burner . If t place where the test is carried out to is at all subj ect side draughts , cardboard screens must be arran ged round the tripod and burne r to prevent access to a of air the crucible while the he ting is in progress . A lead counterpoise should be made for the crucible

and its lid to facilitate the weighing operation . This counterpoise will require adj ustment at short intervals of u l ll time , since the platinum cr cible wi l lose gradua y in

weight by the frequent cleanings . The one gram of fuel u must be very carefully weighed into the cr cible , with a f n n possible error o ot more than o e milligram . is The crucible then covered , inserted in the triangle , and placed upon the tripod stand in the centre of the gas has flame , which been previously ignited and surrounded - with any draught protecting device that may be necessary . The hydrocarbon gases commence to escape from under as u l the lid as soon the cr cib e attains a red heat , and these burn with a yellow and smoky flame for a period which of depends upon the percentage volatile matter in the fuel . As a r ule the gas evolution lasts from 1 to 1 5minutes with one bituminous fuels , and its duration and volume enables , of with a fair degree accuracy , to j udge the character and s n ature of the fuel . As soon athe last luminous candle 2 7 AND GAS FUEL , WATER ANALYSIS has disappeared from the B unsen flam e above the crucible lid l is , the crucib e removed from the flame very carefully , by lifting away the tripod which supports it without dis turbin to g the lid , and is allowed go nearly cold in the air , is before placing in the desiccator . When quite cold it t on l hen weighed with the lid still , and from the oss in of a fi weight , the percentage vol tile matter , coke and xed fi al carbon are calculated as already described . The n on Of weighing must be within e milligram . The deposit of soot found upon the inner surface the lid , and the deposit of graphite upon the inner walls of the crucible are ignored fi n of in the nal weighi g , since although they appear to be u considerable vol me and importance , they weigh only from i 3 to 5 mill grams . The coke left in the cr ucible after the expulsion of the on hydrocarbon gases is very hygroscopic , and this account n ot be it should exposed to the air before weighing . The appearance and character of the coke differs greatly with d f al as i ferent co s , and it may be Obtained either a powdery non- a adhering residue or as solid cake . With the South - e on ea Wales steam coals , which cake togeth r h ting , a slight of explosion Often occurs during the coke test , and some e the fu l is generally thrown out of the crucible . This effect appears to be due to the formation of an impermeable u of all l cr st coke around the fuel , before the volati e matter has escaped . The best way to obtain correct coke tests with this and l is to ' 0 other fue s which produce detonations , use only 7 0 for e f 1 or 5 gram the t st in place o gram . — 6 . S ulphur. The determination of sulphur in fuel in of a volves the performance an ordinary gravimetric nalysis , and some practice in quantitative analysis will be necessary a is before accur te results can be Obtained . It therefore useless to attempt this test without some training in the u l methods and apparat s used by chemists in analytica work . The directions for carrying out this determination are con sequently only given in condensed form . If a bomb calorimeter be used forfinding the calorific value Of e of fi the fuel , the d termination the sulphur is simpli ed , for the liquid Obtained by rinsing out the bomb after the l he exp osion contains all t sulphur in the form of sulphate . The whole of the rinsings are transferred to a porcelain ad n of . . basin , are evaporated to dryness with addition 5 c c 2 8 FUEL

- f s . o . pure hydrochloric acid p gr ) upon a water bath , in order to remove the nitric acid formed by oxidation Of

the nitrogenous constituents of the coal . The residue is i f . . r 1 0 . f 0 . o 2 o dissolved in c c water , c c pure hydrochlo c ﬁ reci i acid are added , and the solution is ltered before p p B tating the sulphur with barium chloride . oiling the solution for ﬁve minutes after the addition of the barium

chloride renders the precipitate more dense , and less likely

to pass through the ﬁlter paper . The remainder of the determination is carried out in the o ﬁ usual way . The weight f barium sulphate nally Obtained multiplied by ° 1 37 5 gives the corresponding weight of sul 1 0 0 an d phur , and this multiplied by divided by the weight of fuel used in the calorimeter gives the percentage of sulphur

present in the coal . r Should a bomb calorimeter not be available , the sulphu in the fuel must be oxidized by means of lime or magnesia and sodium carbonate , according to the method described L 1 of by Garrett omax . One gram the fuel is placed in a small platinum crucible and is intimately mixed with 4 grams of a mixture Of four parts of pure lime to one of l anhydrous sodium carbonate . The crucible is comp etely fi - lled with this lime soda mixture , and a larger platinum one crucible is then placed over the smaller , mouth down tw n w ae . o o wards The crucibles are inverted , and the sp c w fi - between the t o is lled with the same lime soda mixture . The mouth of the larger crucible is then covered with a all of a sm square thick asbestos board , and the whole is pl ced - D in a muffle furnace heated to bright redness . istillation Of the volatile portion of the coal occurs in about tw o as s minutes , and soon as a flame appears round the asbesto

board , this may be removed . To ensure complete oxidation Of a tw o the c rbon , the heating must be continued for hours . of e are to The contents the crucibl s then allowed cool , and l s brought into water , the sulphides oxidized to su phate of by means bromine . The solution is then made acid with ﬁ r hydrochloric acid , ltered , and the sulphur dete mined in the ﬁltrate by precipitation with barium chloride in the s of of u ual way . The calculation the percentage sulphur in the fuel is made as before .

ou al oc e h m cal I ndus tr . De e e 1 1 0 rn S i ty C e i y c mb r 5, 9 5, p .

2 9 CHAPTER III .

The Caloriﬁc Vauat on of Fue l i ls .

H E laboratory caloriﬁc valuation of fuels is carried out by burning a weighed quantity of the dry fuel in a c onﬁned Space surrounded by a known volume or weight of

water , and by noting the increase in temperature Of the latter

due to the combustion of the fuel . A very large number Of c alorimeters have been design ed and employed in the all of laboratory for carrying out this test , but these may

be brought into three classes , according to the character f an d method O supplying the necessary oxygen . E ither the oxygen is brought to the fuel in the form of a Chemical compound or in the free state and calorimeters using gas eous Oxygen for combustion of the fuel are differen iated l t by the pressure at which this oxygen gas is supp ied , the Fischer Mahler bomb calorimeters being the lead f ing representatives of these two di ferent types . In the following pages one calorimeter only of each class will be l described , and the method of use wil be explained in con

sid erable detail .

Of class 1 the Lewis Thompson . calorimeter has been l A se ected . lthough the results Obtained by this calorimeter modiﬁca have been often discredited , when used with the l tions suggested by the author , it yie ds concordant results 1 las with certain C ses Of fuel . O al as f c orimeters using oxygen g at normal pressure , the Darling Calorimeter has been described while Of class M D al has al 3 , the ahler onkin bomb c orimeter been de t

with . In each case the descriptions and details are based upon ’ al ow n lengthy tri s in the author s laboratory .

ENERA E U ATIONS FOR A ORI ETRIC OR K G L R G L C L M W . Wherever possible a room heated by hot water pipes or hot by air , and kept at the same temperature day and night ,

’ 1 S che urer Ke stner s tests with this form of calorimeter conﬁrm this Opinion . 3 0

AND GAS FUEL , WATER ANALYSIS

occurs D is the holder for the same and E is a mercury ° 0 to thermometer graduated in tenths , and reading from ° 45 C . The method Of use is as follows — Two grams Of the ﬁnely l Of powdered and dried samp e coal , prepared as described C out in hapter II , are weighed and intimately mixed with - 1 twenty two grams of the oxygen mixture . This mixing is best carried out upon a sheet of glazed white paper and

— L S H S C L L FIG . 9 . EWI T OMP ON A ORIMETER ( ORIGINA FORM) .

n o an d the must be continued until black specks are visible , o a mixture is f uniform grey colour . The mixture is now to C Of transferred the copper cylinder , with the aid a dry l its Off funne having leg cut , and is pressed down by means 11 of a glass rod having a ﬂattened glass button on its end . Of to C When the whole the mixture has been transferred , a D Of this is pl ced in the holder , a few strands fuse are em in Of bedded the top the mixture , the projecting free ends a r are ignited , and fter quickly cove ing with the copper B A cylinder , the whole is placed in the vessel , containing of has grams Of water , the temperature which been most carefully noted . If the ignition and combustion are

1 This mixture contains thre e parts of potassium nitrate to one of d b a ae an e e d r . pot ssium chlor t , must quit y 1 This is easily made by he ating the end of aglass rod to re dne ss n th n ﬂ ad by pressing e hot e d on aat surface of cold iron . 3 2 FUEL

l for successful , the test is now nearly comp eted , after the whole Of the combustion gases have bubbled up through A is to ta the water in , it only requisite Open the p at the B a to C top end Of , allow the w ter access the cylinder , and A to thoroughly mix the water in , by plunging the cylinder to ﬁ al of up and down in it , Obtain the n reading the ther

— L S H S C L . I O . FIG . EWI T OMP ON A ORIMETER ( MODIFIED FORM )

ometer i s to com m , which g ves the increa e in heat due the bustion Of the two grams of fuel . The calculation which yields the caloriﬁc value of the fuel is then Simply Obtained by use Of the equation

water equivalent of the apparatus in Centigrade units ; an d F: x for i t the gain in temperature . The value Of th s for of is apparatus , and in fact all forms calorimeter best 33 D AND GAS FUEL , WATER ANALYSIS

obtained by testing a standard sample of fuel or pure cellu l of al ﬁ x as ose known c ori c value , and by using the unknown e of in the above quotation in place of , the value which can

then be inserted . The improvements in the Lewis Thompson calorimeter and method of use devised by the author are as follows 1 is t . The weight of water used for each test increased o

of see . grams , by use a larger vessel ( Fig A thin $ena glass beaker replaces the thick vessel shown in

Fig . 9 . This change lessens the rise in temperature and

thus diminishes radiation losses . h 2 . Radiation losses are further minimized by use of t e h B . 1 0 u t e tin cylindrical shield in Fig , which surro nds i - u beaker , and prov des a non cond cting air j acket between the water and the outside air . ﬁ 3 . A copper gauze ring is provided to t on to the copper i cylinder and nearly touch the beaker at all points , . Th s breaks up the bubbles that pass through the water on

to . making a test , and causes the waste gases be better cooled r 4 . The usual the mometer is replaced by one which allows - h the temperature to be read to 1 1 oot of a degree Centigrade . This improvement in reading Off the increase in temperature must of course be accompanied by increased care in the - and . A one e weighing , mixing measuring operations litr

ﬂask is used for measuring the water used for the test .

5. The oxygen mixture is dried before use (one hour at ° f u to i 1 1 0 C . or a s e ch test) , and the amo nt be employed calculated from the combustible actually present in the 1 0 0 — x a , fuel . The formul used for this calculation is W in which W the we ight of oxygen mixture required for 2 of x grams the fuel , and the percentage of ash in the l dry samp e .

6 . The mixing of the oxygen compound with the 2 grams of fuel an d charging of the copper cylinder with the same a u a out l d are most c ref lly c rried , and no loss of fuel is al owe to occur in this Operation . The cylinders are numbered according to their cap city as measured by water contents , and as far as possible it is planned that the fuel and oxygen

to . mixture shall ﬁll the cylinder to within 3. inch of the p The precise degree of pressure it is necessary to employ in ﬁlling the cylinder with this mixture can only be learned by experience ; if too loosely packed the combustion is too 3 4 FUEL rapid if too tightly packed the charge will probably become u A exting ished before all is burnt . combustion Should las t 1 minute .

. is of 7 The ignition fuse made of four strands lamp wick , one s inch in length , dipped in pota sium nitrate solution and

. a one dried The four str nds are twisted together , and quarter inch of the ends are then Spread out in the form of of the feet , and the little igniter is placed upon the top charged cylinder with the ends lightly embedded in the i l mixture , wh ch must be left oose and porous at this place .

8 . The water equivalent of the apparatus is determined r by actual t ial with a standard fuel as already described . B D i . in C an d n . are a 9 The copper vessels , , Fig 9 pl ced a m C a is the c lori eter after is charged , and the temper ture B read when equilibrium is established . The tap of must , of is course , be closed during this period . The charge then a an d B C and D for ignited by withdr wing disconnecting , i of a very brief per od time . O as 6 perating described in and 7 , the author is seldom troubled with failures to ignite or with incomplete com b ustions . Using this modiﬁed form and method of working with L e e a has a d the ewis Thompson calorim t r , the uthor obt ine are u results which within 2 per cent . of the calc lated of a values the fuel , when testing bituminous fuels cont in ing over 25 per cent . of volatile matter . The conditions u is a nder which the test m de must , however , be kept as far as possible the same as those under which the t an a wa er equivalent is determined , and y wide vari tion in o a these conditions will lead t discrep ncy in the results . - With semi bituminous and anthracite coals the Lewis . e unsatisfac Thompson form of calorimeter yi lds , however ,

u of u . tory res lts , owing to the amount nconsumed carbon C a a oke and peat lso cannot be burned in the c lorimeter , so that its ﬁeld of usefulness is somewhat narrow and restricted . ﬁ s of all of r For correct calori c te ts these classes fuel , eithe r of the calorimete s about to be described must be used .

E DAR I N A ORI ETER TH L G C L M .

The Darling Calorimeter is based upon the principle of combustion in a current of oxygen gas at normal pressure , 35 AND GAS FUEL, WATER ANALYSIS and the apparatus is a modiﬁed and improved form of the - n C . 1 1 w well k own William Thomson alorimeter Fig . Sho s D the essential parts of the arling calorimeter . A small n u 1 1 ickel or platinum cr cible , 1 inches high by inch in dia

— . L C L L F I . G I I DAR ING A ORIMETER ( ORIGINA FORM) . m of eter , is held by brass arms in the centre the perforated - A l- brass base plate of the apparatus . ﬂanged glass bel j ar a r as provided with rubber cork , car ying a br s oxygen supply and the p ipe , two copper conducting wires is clamped to base-plate by means of rubber rings and by a brass ring p late held in position by screws and nuts . It is advisable to use a rubber ring above and below the glass ﬂange on the 3 6 FUEL bell-j ar to avoid breaking this when tightening the milled r on head screws . A brass gauze ing is shown in position - 1 1 an a i l the bell jar in Fig . , but this is ddition to the orig na a us . 1 2 r apparatus . Fig shows another fo m of the pparat

— F IG 1 2 1 . C L . . DAR ING A ORIMETER ( MODIFIED FORM) . in which the bras s oxygen supply tube is replaced by a tube of hard potash glass widened out to form a cup at its or lower end , and connected above to the oxygen cylinder - f gas holder by a glass T piece . In the first form o the Darling apparatus ignition of the charge occurs by means o f a fine platinum wire fixed between the lower ends of the u are con two copper cond cting wires , which themselves n ected to four bichromate cells or to some other source of 1 electrical energy . Fig . 3 Shows a convenient B ichromate B a attery m de up of four pint cells . In the second form of of a of apparatus , ignition occurs by means a sm ll piece glowing charcoal ( 0 1 to 0 2 gram) this being dropped 3 7 AND GAS FUEL , WATER ANALYSIS d ow n the oxygen supply tube by removing the small cork w hich Closes the vertical end of the glass T piece . This method of ignition is preferable to that recommended by D of has arling , as the use sulphur for ign iting the charge many disadvantages . In both forms , a Sheet tin cylinder surrounding the glass vessel diminishes the radiation losses and is a useful addition to the original apparatus .

— . CH Y F LLS I . FIG . 3 BI ROMATE BATTER O FOUR CE

The method of carrying out a determination with the Darling calorimeter is as follows grams of water are me asured carefully into the glass 1 2 one- e vessel shown in Fig , using a litr ﬂask and graduated

cylinder for this purpose . From half to one gram of the is dray sample of fuel weighed out into the crucible , this is pl ced in its holder , and the apparatus is clamped together

as shown in Figs . 1 1 and 1 2 . If electric ignition is to be employed the ends of the two copper conducting wires 38

AND GAS FUEL , WATER ANALYSIS mature loss of heat it is advisable to support the crucible s of as u upon an abestos tray . The current oxygen g thro gh s the apparatu is then slackened , and the temperature of the water is noted every half minute until a fall occurs . A di u ufﬁ ccor ng to the inventor , minutes sho ld s ce to a of has burn one gr m fuel , but the author never succeeded r l in car ying out atest in this time . The caculation is x t ) , made as before by use of the equation C w

C ﬁ u of t the in which equals the calori c val e the fuel , equals w w is gain in temperature of the ater , the weight of fuel

x i . used , and is the water equ valent of the apparatus Although a value for x is given with each apparatus and a use sold , it is far better s fer to determine it by of a as ul r standard combustible such cell ose or pu e carbon , as described in the previous section of this chapter . The Darling calorimeter only yields entirely satisfactory results with fuels low in ash and in volatile matter . When fuels containing a high percentage of volatile matter are burned in the Darling apparatus there is a tendency for some of the volatile hydro-carbons to escape unconsumed and ash for when fuels high in are tested , there is a tendency portion of the ﬁxed carbon to remain unburned in the of a crucible . In the case coke this unconsumed c rbon may

u . amount to 1 0 per cent . of the total comb stible matter

‘ The ﬁrst of these difﬁculties may be overcome by mixi ng one-third of its weight of china clay with the fuel before

al i . burning in the c or meter This remedy , however , is only for a partial , it incre ses the amount of ash remaining in the of fi crucible , and therefore the risk of some the xed carbon remaining unconsumed . The second difli culty can only be overcome by making of small pellets or briquettes the fuel sample , and using these for combustion in the calorimeter in place of the finely

. 1 u powdered sample Fig . 4 shows the pellet mo ld used by ﬁ the w riter for this p urpose . Two grams of the nely pow dered sample are mixed in asmall agate or glass mortar f of - . . o 1 with one c c a per cent . solution pure gum arabic in u be water . Very bit minous fuels can moulded without -a a is e adding gum rabic . This p ste then pr ssed into three i ut cylinders in the mould Shown . E ach of these s o while a six still moist into two h lves with a sharp knife , and the little pellets of fuel obtained in this way are then dried at 40 FUEL

° - 2 0 C. on e . or 3 for hour in the air bath Two , three four of ellets accordin re uired are for these p , g to the weight q , then used

al r h a out as . the c orimet ic test , w ich is c rried already described e When fuels containing less than 1 0 per cent . of volatil matter are tested in the Darling calorimeter some difﬁculty will be met w ith in the absence of an electric current in for causing them to ignite , neither charcoal nor sulphur will fﬁ always su ce to start the combustion . The best plan in such cases is to use a small weighted portion of some pre

— — I . L . I . F FIG . 4 BRIQUETTE MOU D . FIG S WASH BOTTLE OR REGULATING OXYGEN S LY UPP . viousl u u y tested bit mino s coal , either making a mixed pellet of the two coals or using both in the powdered form and sprinkling a little of the standard bituminous fuel on the f surface of the other in the platinum crucible . The di ferent caloriﬁc value of the added fuel must of course be allowed for in calculating the result of the test . B y using only half a gram of fuel an d a thermometer read to 1 -1 ooth C ing of a degree entigrade , the time required for burning the fuel and the gain in temperature can be reduced - r d ue to i o one half , and the erro s rad ati n losses corresp ond in gly diminished . When working with this small amount and of fuel , very careful accuratae weighing and measure ments are required to obtain reli ble results . The combustion of the fuel can be effected much more quickly when the fuel is in briquette form , since the cur rent of oxygen gas can be increas ed greatly without danger of blowing the fuel out of the crucible . 4 1 AND GAS FUEL , WATER ANALYSIS

The oxygen used in working the Darling apparatus should as - either be stored in a g holder over water, or should be pas sed through a glas s tower ﬁlled with pumice soaked in water , in order that it may take up before use all the water it can carry at the temperature of the laboratory .

6 — - F IG . 1 . HL K C L MA ER DON IN BOMB A ORIMETER .

It is also useful to have a wash-bottle through which the r as - i oxygen bubbles , inse ted between the g holder or cyl nder as of t and the calorimeter , in this way the speed the curren 1 can be more easily noted and controlled . Fig . 5 shows as - a suitable form of w h bottle .

When used with the precautions described above , the Darling calorimeter can be made to yield accurate results 42 FUEL w f o . B ut l ith all classes fuel as usua ly employed , the com bustion is in many cases imperfect , and unless allowance be m for r - a and ﬁ ade the unbu ned hydro carbon g ses xed carbon , the ﬁ al o w calori c v ues btained are too lo .

E MA E R-DON K I N B O B A ORI ETER TH HL M C L M .

1 6 Figs . to 1 9 show the component parts of the Mahler D onkin bomb calorimeter . The combustion of the fuel in this calorimeter is carried out in oxygen gas at from 20 0 to 00 1 6 4 pounds pressure . Fig . shows the most important

I — SS L AND X N L S FIG . 7 . WATER VE E MI I G PADD E .

— p ortion of the apparatus namely the bomb . This is a ay e A heavy gun met l c lind r ( ) , provided with three screwed s u for D t d pins bolting down the cover ( ) , and lined inside with gold to protect the metal from corrosion by the acid s formed during combustion of the fuel . The cover is pro vid ed with amilled head screw valve (B ) for regulating the i i a C nlet of oxygen gas , w th an insul ted condaucting wire ( ) , which runs through the cover and termin tes below , and with three heavy nuts for bolting down the cover on to the c hi ylinder . T n lead wire is used to make the j oint between is u to the cylinder and cover , and the bomb tested p ’

. 1 pounds pressure before leaving the makers works Fig . 7 4 3 AND GAS FUEL , WATER ANALYSIS

of r shows the rest the apparatus , namely a water containe and paddles for quickly obtaining a uniform temperature . A water-j acket and also an air-space surround the inner vessel containing the measured volume of water and the

dl ﬁt . e bomb , and the mixing pad es round the latter Thes be an d or paddles can worked up down , in a circular manner , by the wooden nob which proj ects six inches above the

— FIG . 1 M H A D . 8 . B OMB COVER WITH PLATIN U DIS N IGNITI ON WIRE

a is fi . water container, when all the appar tus tted together n of fine The ig ition the fuel occurs by electric heating , a platinum wire being used as bridge between the lower ends 1 8 of the two electrodes fixed on to the bomb cover . Fig . a r shows the cover lone supported on a t ipod stand , with the small platinum dish and the platinum wire ign ition l r oop . It is well to make a trial test of the elect ical con nections in s u s thi position , Since if the platin m wire doe not heat to redness it is comparatively easy to find out the on is on to the cause , whereas later , when the cover bolted c i ylinder and the bomb is mmersed in the water , a failure 4 4 FUEL

n P t o ig ite causes much trouble . latinum wire must be ﬁ used for ignition purposes . The author uses the nest m 2 . ade , of which twelve inches weighs grains Iron wire c annot be used, for it produces fused iron oxide on ignition and this speedily corrodes the gold plating of the bomb at A 1 those places where it falls . small platinum dish inch in a d iameter and } inch deep is preferable to a crucible for

— - H XY S LY C C S AND SS . I 1 . F G . 9 BOMB WIT O GEN UPP ONNE TION PRE URE GAUGE

r holding the fuel , since it is easier in this case to a range that the platinum loop shall dip in the fuel . For this pur p ose the wire is wound seven oreight times round the straight and li e lectrode , the loop is lengthened as required by pul ng one or more of these coils down with a pair of small forceps . The weight of fuel used for carrying out a test with the bomb al the c alorimeter varies from 4 0 to 1 0 gram . The sm ler u q uantity of fuel used , the smaller is the rise in temperat re o f and athe water , the less is the time required for completing test , both of which reduce the error due to radiation 45 AND GAS FUEL , WATER ANALYSIS

O n s losses . n the other hand this method of working dema d great care in all weighing and measuring Operations ; and some degree of skill in manipulating the apparatus is necessary, before correct results can be obtained when work ing with 5 gram of fuel . A fter weighing out the fuel roughly into the platinum dish ,

— 2 0 . FIG . OXYGEN CYLINDER WITH REDUCING VALVE AN D PRESSURE GAUGE . it is advisable to heat the same for one half-hour in the air ° - 2 0 C. re bath at 3 , and to weigh carefully before making the test with it . The dish is then transferred to the holder u nder the cover , the platinum wire ignition loop is lowered e carefully until its lowest s ction j ust dips into the fuel . and

is c . c . the cover placed on the bomb and bolted down . Ten 46

AND GAS FUEL , WATER ANALYSIS

- ﬁn gas tight . It is well to have a e screw regulating valve e 1 0 A and a pr ssure gauge reading up to 2 t . attached to the

n . oxygen cylinder as show in Fig 20 . Otherwise it is diffi as an d cult to control properly the flow of g into the bomb , too rapid a r ush of oxygen may cause some of the fin ely powdered fuel to be blown out of the platinum dish . The gas must be admitted to the bomb slowly . The amount required , as indicated by the pressure , varies with the amount of fuel use d . For half a gram of fuel a pressure f 0 ffi r 0 0 o 22 pounds su ces , for one g am of fuel 4 pounds is ressue necessary . With this higher p there will be some ffi di culty in keeping all the j oints tight , but slight leaks M -D may be ignored . In the ahler onkin bomb as originally designed , a considerable loss of pressure and of gas occurred also on closing the milled head screw valve on the top of w as the bomb , since to close this it necessary to loosen the screw nut connecting the oxygen supply pipe above it . The author has therefore had a small back pressure valve inserted in his bomb in the portion of the cover marked

B (Fig . and this comes into action and closes the exit i pipe as soon as the exterior pressure is relieved . This s a most use ful addition to the earlier design .

The milled head screw valve is now closed , the bomb disconnected from the supply , and the opening in the top of the bomb is closed by the screw nut shown in the fore

1 6 . n w ground in Fig . The bomb is o carefully transferred to the w ater vessel which has been previously filled with ae c . c . of w t r at the laboratory temperature , the dl mixing pad es are placed in position , and the conducting m a e wires fro a bichrom te battery of four c lls , or from some other source of electrical energy (tw o amperes at 1 2 volts e r are required) are connect d to the te minals on the bomb . 1 e a Fig . 2 shows the general arrangem nt of the appar tus when ready for firing the charge of fuel. The screw nut pro jecting above the water in the calorimeter vessel is used as the second terminal on the bomb . The mixing paddles are now worked gently until the thermometer dipping in the water of the inner vessel ceases to rise or fall , and the fuel is then fired by depressing for three seconds the last of the zinc plates of the bichromate battery . Within half a minute , if the ignition has occurred properly , the ther meter mo will begin to rise , and the mixing paddles are now i worked steadily unt l the mercury again becomes stationary . 48 FUEL

ﬁ th The nal temperature is en recorded , and the test is

ﬁnished . tw o For work with this calorimeter , the author uses Specially 2 made mercury thermometers 4 inches in length , graduated ° °

0 0 . in 5 ths and ranging from C to 25 C . With the aid I -I Ooth as of a magnifying glass , of a degree can e ily be read on these thermometers . The calculation is carried OIit as before by aid of the x ) t equation C in which C represents the w

ifi x l calor c value , the water equiva ent of the apparatus , t w the gain in temperature , and the weight of fuel as used in grams . The value of x is best determined before n athalene by making a test with pure cellulose , p or carbon , and by inserting x as the unknown value in the above P th n h fi equation . ure nap ale e as a calori c value of calories , pure cellulose of calories , and pure amorphous carbon of calories . The Mahler-Donkin bomb yields accurate results with all fi w classes of fuels , and when once tted up ith all the necessary r accesso ies , tests can be made with it in from one half to one hour . The whole of the apparatus should be dis mantled and cleaned after each test , especial attention being as given to the bomb , cylinder and cover , the gold plating if is of these requires very careful cleaning and drying , it P to last any length of time . ortion of the thin platinum wire used for ignition is generally fused during the com bustion , but the heat absorbed by this is included in the value of x used for the wate r equivalent of the apparatus .

When testing anthracite or coke , the platinum wire igni n tion sometimes fails to ig ite the charge , and to overcome this difﬁculty ° I o gram of a standard bitumin ous fuel may e be used on top of the weighed portion of cok or anthracite , as n ig ition charge . The thermal value of this addition must be allowed for , in calculating the results of the test .

49 I CHAPTER V .

he P rat aA at on of the Te t e T c ic l pplic i s s R sults .

NGINEE RS who have never made use of the results

obtained by laboratory tests of fuel , either in relation to their supplies of fuel or in the working of their boilers , are disposed to be somewhat scornful as to the value of fuel test ing to the practical engineer . In their view the results obtained - r are from steam raising t ials alone worth depending upon . That this V iew is superficial and incorrect is proved by the fact that engineers who have adopted fuel testing as an aid n and in the work of boiler ma agement rarely relinquish it , w as their kno ledge of its usefulness extends , they place more and more reliance upon the laboratory test results , - a com and less upon the steam r ising trials , in forming their f parative j udgments of di ferent fuels . When one remem bers that the essential conditions obtaining during steam raising tests are never exactly the same ; that all fuels are treated alike although different fuels demand different methods of firing and different treatment as regards draft to obtain their maximum evaporative effect ; and that the fireman or engineer who is in charge of the boilers may very probably be receiving a bonus or present from one of the fi fi . fuel supply rms , this loss of con dence is not surprising The value of comparative steam-raising tests as a guide in the placing of fuel contracts , or in the control of the fuel - supplies is , in fact , greatly over rated , and the larger number are of these tests , as they at present carried out , are not worth the time and expense entailed in obtaining them . In the following pages the author will deal with three practical and scientific applications of the test results as an d obtained in the laboratory examination of fuels , will Show how these can be utilized in

P . 1 . lacing fuel contracts C 2 . ontrolling the fuel supplies . fﬁ 3 . Checking the evaporative e ciency of the boilers . 50 FUE L

ACIN UE ONTRACTS . I . PL G F L C

Contracts for the delivery of large quantities of fuel n an involvi g the expenditure of m y hundreds of pounds , ought always to be based upon the laboratory test results the for reasons already given . The sampling and testing of each sample load of fuel should be conducted as describe d C I ﬁ s in hapters II and II , both approximate and calori c te ts being carried out . l If possible , a bomb ca orimeter should be employed for i as u the latter tests , since in th s c e the res lts obtained are - can reliable for all classes of fuel . Steam raising tests of course be made at the same time but the author has given reasons for doubting the value of these unless much more

fi . scienti cally carried .out than is usual at present The laboratory test results of the various sample loads of fuel are i then grouped as show n in Table I (see Append x) . If the to as contract is be b ed solely on cheapness , the decision is easily made by comparison of the figures given in col umn 0 fi u 1 . These figures are obtained by dividing the calori c val e multiplied by by the price per ton in pence of the fuel As u delivered in the bunkers . a r le , however , the percent ages of ash and of volatile matter will have some influence H upon the choice of fuel . aving eliminated from the list w to ash e or those which are unsuitable o ing high cont nts , e to too high a percentage of volatil matter , the selection the from those remaining in the list , of fuel which yields the greatest number of calories per penny of cost is an easy P i matter . rovided that correct conditions of stok ng and draft are employed in burning the fuel under the boilers , the fuel which gives the greatest number of calories per penny of cost will prove the most economical in actual use . That wide variations exist in the value of different fuels when j udged by this method is proved by the follow ing fi e a tests an d s gur s , b sed on actual price

I a e er e . No . c lori s p p nny of cost S outh Wale s N 2 O ' ’ ste am coals aa No . 3 . L e N . . 1 nc shir -( o 4 N . . bituminous . o 5 a ( co ls No . 6 . The calorie used in this table is the pound Centigrade ﬁ 1 8 unit , and the above gures must be multiplied by to 5 1 AND GAS FUEL , WATER ANALYSIS

B . U obtain the corresponding value in ritish Th nits . If burnt under correct conditions to obtain the maximum fﬁ e N 6 o . e ci ncy , it is therefore evident that the coal marked in the above list would prove the most economical in actual

. 1 2 n use , for it yields double the heat units of Nos or per pe ny f r o actual cost . If burnt in fu naces designed for South Wales s a team co l , with a small combustion area , it is , however , 6 p ossible that this No . coal would not Show this superiority , and the author must once again emphasize the necessity for aa n d pting the furnace desig , dimensions and methods of ﬁ ring , to the type of coal which is to be burned , if good

results are to be obtained . The discrepancies which occur between the laboratory test results and those obtained in actual practice are chiefly due to the neglect to observe this ' essential condition for obtaining good results under the r boilers . The subject of fu nace design and control of the a the combustion process , however , is somewh t outside and scope of this work , readers must refer to the book i 1 named below for an adequate discussion of t . H aving selected a fuel from the small or large number of is which sample waggons have been submitted for trial , it advisable that the contract should be based upon the results of the laboratory tests , and that some kind of a sliding scale fi relation between calori c value and price should be arranged . The exact details of this will vary in the individual cases but the specification reproduced below is an example from actual practice of a fuel contract based on laboratory w a tests , ith a sliding scale arr ngement between calo

riﬁc value and contract price of the fuel . The fuel in this a s . d as . c e was a South Wales ste m coal , delivered at 9 7 per ton .

SUPPLY OF LARGE WELSH SMOK ELESS STEAM COAL . P F AT ND D T S ECI IC ION A CON I IONS .

T be Lae We S e e S e a a e e o rg alsh mok l ss t m Co l , fr e a1 er e . from dust , to cont in not mor th n 5 p c nt of a and ae a m ar ae sm ll , to h v minimu c lo iﬁc v lu of (thirteen thousand three hundre d) Bri tish T e aU er aas e a e b h rm l nits p pound of co l , m sur d y acalori meter approved by the Borough E lectrical

E ngineer .

1 e P even tion and F uel E conom e ha S mok r . y , by Booth K rs w

ae 1 0 . Const bl , 9 5 FUEL

D the i ae the a e uring cont nu nc of contr ct , t sts will be made once eve ry w eek from samples d e e e r haw e e and the ea liv r d du ing t t k, if m n of four tests during any month show that the caloriﬁc ae e the aa the ae er a v lu is b low st nd rd , r t p ton p id to the contractor for the whole of the coal delive re d a abe es athe a during th t month , sh ll l s th n contr ct price as set forth in the P rice S chedule at ch tae d he reto .

2 he Notw ithstanding Condition No . t Council rese rve to the mse lves the right to reject any the a e e e the e w portion of co l d liv r d , t sts of hich S how that the caloriﬁc value is less than

British Thermal Units p er pound of coal .

S hould any dispute arise as to the accuracy of e w e a a a t sts , ights , qu ntity of dust or sm ll , q u lity , e ae w a e e the e the or oth r m tt rs h tso v r , d cision of Borough E le ctrical E ngineer shall be ﬁnal and

binding .

D PRICE SCHE ULE .

S cale of re duction in price for low e r caloriﬁc value than T e aU British h rm l nits .

P l vered . er Poun d . D e i

For aminimum of B TU the contract price Below but not less than a1 e r r duction of 5d . p e ton Below but not le ss than a e r r duction of 1 od . p e ton Be low but not le ss than a

e 1 5 . d e r duction of 3 . p r ton

Clause 4 of this contract is obviously unfair to the Colliery C ompany , since an independent chemist or fuel expert B ut ought to be appoin ted as referee in case of disputes . otherwise the speciﬁcation has much to recommend it . Another method of working out the price variation for f ﬁ al has di ferences in calori c v ue , been recently adopted by the city of Chicago in the purchase of tons of coal . The stan dard price in this cas e w as per ton of

. i B . lbs for a coal test ng Th . Units in the undried A state . llowing for the moisture and cost of removing the 53 AND GAS FUEL , WATER ANALYSIS

B . . U ash , this worked out to Th nits for one cent .

. b If the B . Th . Unit per cent came out above or elow this limit , the price of the coal varied accordingly the principle underlying the contract being that in all cases one cent .

. U of should purchase B . Th nits heating value in u the f el . Thus at times the seller , and at times the buyer , was favoured by the price variation , and the price asked f u and paid or the fuel was an absol tely fair one . The author believes that in time this last described method of buying fuel will apply to all large fuel contracts . The present method is grossly unscientific and unfair alike to l b the buyer and sel er , and the sooner it is replaced y another the better it will be for all concerned . In drawing up fuel contracts it would also be well to stipulate that the moisture and s ulphur contents shall be kept well within the percentage show n by the test of the f sample waggon . Owing to various causes the percentage o moisture and sulphur in coal and slack may at times rise to fi u a very high g re , and apart from the foolishness of f s . 1 0 3 . or an d paying from 7 to per ton this water sulphur , there is the question of the heat w hich will be carried off to the chimney when the water is converted into steam in the As n u . f rnace of the boiler regards ash conte ts , it would also be well to reserve the right to rej ect deliveries of fuel showing a percentage much above that embodied in the B ut al o contract . the increase in ash will ways be acc m anied p by a corresponding diminution in thermal value , and the variation in ash contents is there fore covered by the sliding scale arrangement based on the result of the calorific tests . Though the freedom to reject coal below the stipulated thermal value may be reserved to the buyer as in Clause 3 of the contract printed above , in most cases it is least to e to a e troublesome accept d livery , and cl im the d duction in price under the sliding scale arrangement . Otherw ise much trouble and expense may be entailed in connexion - w with re loading and rail ay charges .

ONTRO I N T E UE UPP IES . II . C LL G H F L S L The laboratory testing of fuel can be employed most usefully to check the quality of the supplies of fuel from day to day and from week to week . When the contract has been ﬁ based upon calori c value , this testing is of course essential . 54

AND GAS FUEL , WATER ANALYSIS

fi matter to arrange that every fth , tenth or twentieth bucket shall be tipped into a sample storage bin . If fuel from more than one colliery is being delivered and used at one e tim in the works , it will of course be necessary to have an equal number of sample storage bins . At the end of each day the fuel in these bins is crushed and sampled in accordance with the instructions given in A i the ppend x , and the one pound samples are sent to the laboratory or to the fuel expert for examination an d as . r test It will not , a rule , be necessary to car y through a complete approximate analysis and calorific test as d e in C scribed hapters II and III , with each of these daily samples ; the determination of the moisture and ash in each being sufficient to enable one to check roughly the quality of the coal being supplied under the contract . The daily sam le tin samples are , however , retained each in their own P , an and once a week or oftener , average sample is made up 0 1 0 0 from these by weighing out 5 or grams of each , and by thoroughly mixing the same on the sampling plate , or A on a Sheet of glazed paper . complete approximate analysis and calorific valuation is then carried out with this U sample . pon the results of these weekly tests complaints to the Colliery Company as to the quality of the supplies and the variations in price of the fuel delivered under the con tract are base d . A book should be kept for entering up both the daily and u a weekly f el tests . The results obt ined during the de livery of coal under any one contract will prove of great a in fi v lue xing the terms for the next contract . Special samples of single waggons or trucks of fuel that appear to be of bad quality should be taken occasionally , for at times the Colliery Company may become negligent in their supervision of the sorting and cleaning of the coal a out brought up from the pit , and , as alre dy pointed , much can as dirt and shale be sometimes sold fuel . The author a 6 . ash has tested fuels cont ining only per cent of , from e to 2 the same pit that produc d slack testing up 5 per cent . ash r w as e of , and some yea s ago he ask d to report upon a - sample of so called coal w hich contained no less than 37 70

. ash per cent of . These ﬁgures prove the fallacy of the opinion held by many engineers that coal from one pit or from one seam is of ﬁxed and unvarying composition when delivered into 56 FUEL

u and is r u the tr cks , that it alone necessa y in drawing p fuel contracts to Specify the pit or seam from w hich the supplies are to be drawn . raw Fuel , in fact , appears at present to be the only material u sed in large quantities , the sale and delivery of which is still , in the maj ority of instances , conducted by rule of thumb methods . It is satisfactory to note , however , that in are re this matter also , the methods of the past being placed gradually by the more scientiﬁc and exact methods

C . of procedure , outlined in this hapter

K V A V OF III . CHEC I NG THE E APOR TI E EFFICIE NCY TH E OI E RS B L . With the aid of the figures obtained by the calorific test w of the average eekly sample of fuel , as described in Section C II of this hapter, it is possible to make out a balance sheet for the working of the boiler plant , and to calculate the evaporative efficiency of the same with a high degree of exactitude . B y evaporative efficiency is signified the percentage ratio and between the total heat value of the fuel , that utilized o a in the conversion f water into ste m . The other tests required for ascertaining this efficiency by n calculation , are those showing the compositio and tem t - i po ature of the exit gas es . The methods of obta ning these ll C C wi be dealt wiath in hapters IX to XII , and in this hapter the method of pplying the results will alone be dealt with . There are four directions in w hich the heat produced by f combustion of fuel is dissipated in the furnace o a boiler . fi fin fi In the rst place , a certain de ite and xed loss occurs by radiation from the exposed surface of the boiler and of can the surrounding brickwork . This loss be roughly a fi ascertained by dr wing the boiler res , Closing the steam entirel off valve and feed water valve , y cutting the draft , and f 0 0 lbs by noting the time required for afall o 1 or 2 . in the As u pressure of the steam . a r le , radiation losses account 0 a for from 7 5 per cent . to 1 per cent . of the he t present in fi the fuel . The gure when once ascertained for any one boiler may be used as a constant in all the future heat balance calculations . The second direction in which heat is dissipated or lost in boiler working is with the waste gases passing away from flues the boiler to the chimney . The method of ascertain 57 AND GAS FUEL , WATER ANALYSIS

l ing and calculating this oss will be described in Chapter XII . 0 0 This loss may amount to from 1 per cent . to 3 per cent . of the heat value of the fuel burned .

The third direction of heat loss is by the ashes and Cinders . w The loss here is t ofold , being partly due to the heat carried by the ashes and partly that due to the unburned carbon in i O the C nders . nce a week the ashes may be sampled and the unburned carbon determined by the method used for ee estimating the amount of ash in the fuel . (S Chapter

II) . The total heat loss in the ashes and Cinders should not m . a exceed 5 per cent , but it y rise much higher with chain a e w e a fire gr te stok rs h n b dly worked , or with untrained - men and hand fired boilers . The fourth direction in w hich heat is dissipated is through the boiler plates . It is this portion of the heat alone which is the active agent in water evaporation and gives the e vaporative efficiency of the boiler . It is ascertained by adding the three items of loss named above together and r by subtracting the total from the the mal value of the fuel , as determined by the laboratory examination of the same . The follow ing is an example of this method of calculating the working efficiency of boiler installations

FUEL .

6d . . Lancashire slack at 7 s . per ton delivered in bunkers H eat units produced per pound by calorimeter test ,

B . U . . Th nits

L tc . 1 0 e r e . 1 . e b aa e oss s y r di tion , ( p c nt ) the C e a ae 1 0 e r 2 . Losse s w ith exit gase s up himn y (c lcul t d) p ° d e at 6 F . n a cent CO ; a g s s 4 5 L b a i n C e an d ae an d ea 3 . osses y unburnt c rbon ind rs sh s , h t

carrie d aw ay in same ( 5 p er ce nt . ea e b ffe e e 4 . Conve rte d into st m in boil rs ( y di r nc )

fﬁ ° 1 1 1 oo The e ciency of the plant , therefore , equals 9 7 / 3 , 5 1 o r 67 per cent . Items and 3 are based on estimates or upon previous test results , but these two items under ordinary working conditions should not vary more than

5 per cent . Should the analysis of the exit gases show that carbonic o as r xide g is present , a fu ther entry will be necessary in 58 FUEL the above heat balance to show the thermal units lost by u L this fail re to burn all the carbon to carbon dioxide . et x us take , as an example of this additional calculation , e it gases containing 2 per cent . by volume of CO and 8 per

e . C0 . c c nt 2 Since arbon , when burnt to carbon dioxide , yields the same volume of gas as when burnt only to carbon - monoxide , the above percentages Show that two tenths or one-ﬁfth of the carbon present in the fuel has escaped com l ete . n p combustion Now , as fou d by the approximate the ﬁ in l analysis , percentage of xed carbon present the fue

60 . r was per cent , and the the mal units liberated when carbon is burned to CO are less than when carbon is burned to C0 2 and respectively) . The loss of heat due to incomplete combustion in the case under consideration is therefore

O 6 B . o x x . U fﬁ 5 Th nits , and the e ciency of the boiler is reduced to 58 per cent . in consequence of this incomplete combustion of a portion of the fuel .

This additional loss is a considerable amount (9 per cent . ) of the total heat energy of the fuel , and the calculation shows the necessity for keeping a close check upon the pre sence of CO in the exit gases . Since the automatic recording instruments now in general use for checking the composi tion of the exit-gases from steam-boiler plants do not record as it e a the presence of this g , is evid nt th t independent tests Orsat with an apparatus ought at times to be made , in order to check the presence or absence of CO . This gas is most likely to be present when attempts are being made to work and fi aa with a minimum of air thick res , in order to att in high percentage of C0 2 in the exit gases . Thus efforts to a e to e reduce the he t loss s due excess of air , unless und r a ’ Chemist s control , may result in very considerable heat losses in another direction . The need for expert scientific conatrol of the testing work is therefore apparent , and autom tic a is ffi testing appar tus , though useful , not an e cient substitute for the trained chemist . The subj ect of gas-testing w ill be dealt with at length in C e a hapters IX to XII , but it may be point d out here th t as fuel and gas analysis enable one with ease and a fair degree of accuracy to determine the thermal efficiency of a is a the whole boiler inst llation , it deserving of more tten tion than has yet been accorded to it by enginaeers in charge of boiler plants . In fact , these test results en ble one , at a 59 AND GAS FUEL , WATER ANALYSIS

moderate expense , to achieve daily or weekly what at present is only carried out at long intervals of time and with much disturbance to the usual routine , under conditions of as r working which cannot be regarded no mal . The money now expended upon special steam-raising and efﬁciency r a t i ls , if expended in the way suggested upon regular testing ’ e work , would lead , in the author s opinion , to far mor fﬁ valuable results , and to a far better average of e ciency than is attained under the present system .

60 P ART I I W ATE R .

U AND GAS F EL , WATER ANALYSIS

- e particular water for steam raising purpos s , without some preliminary treatment for removal of its impurities . The source from which the water is drawn has much inﬂuence upon the character and amount of its impurity . B elow, the general characteristics and composition of rain , - so river and well waters are given , far as these have any bearing upon the use of such waters for steam-raising pur poses .

- RAI N WATER .

This is the purest natural water , for it contains only the impurities taken up by the water during its passage from r the clouds to the ea th , namely oxygen , carbon dioxide ,

u . r r sulphur dioxide , soot and d st In count y dist icts rain water is often entirely free from the three last-named r impu ities , and contains only comparatively small amounts o f oxygen and carbonic acid gas . To obtain an adequate in supply of such water for boiler feed purposes , however , volves the creation of large reservoirs fed by large catchment a s is not are s , and in most ca es this possible for the ordinary

or . works factory , owing to the capital expenditure involved Rain-water collected in town s or upon the outskirts of in d ustrial r is so dist icts , on the other hand , contaminated with s ulphurous acid , soot , dust and other impurities , that ﬁltration and Chemical treatment are necessary before such s - e water can be safely u ed in steam boilers . For this r ason rain t water is not of en employed for feed water supply , except in those special cases where these two obj ections lose much of their force and weight . It is quite possible , however , that rain-water might be used much more extensively than is a for - at present the c se , steam raising purpose ; and that s ome thought and ingenuity expended by engineers in arranging catchment areas and storage reservoirs would be well repaid in the greater freedom of the boilers from scale An and corrosion due to the impurity of the feed water used . average rain water will contain less than two parts of dis s olved or suspended matter per of water .

URFACE AN IV ER ATERS S D R W .

Rain-water after it has travelled any distance over the s urface of the ground either in country or town districts becomes further contaminated w ith suspended and dissolved 64 WATER

matter , and therefore less suitable for use in boilers without H some puriﬁcation treatment . ere again the appearance of the water may be wholly deceptive and chemical examina tion will alone reveal the nature and amount of the impuri ties present . In mountainous districts the amount of l impurity taken up from the soil is sma l , and the surface a L in w ter collected in the catchment area of ake Vyrnwy ,

North Wales , or in the catchment area of Thirlmere , in the L D for of L ake istrict , supply of the cities iverpool and

anchester respectively, is remarkably pure and free from Missolve n d As o n salts . e comes dow to lower levels and am approaches the neighbourhood of towns , the ount of dissolved and suspended matter in the surface water in of L creases considerably . The following tests the iverpool water supply are interesting in this respect

Parts p er of w ater. - - Vyrnwy— Total solids in solution 4 46 H ardness 1 7 0 - Rivington 1 0 36 3 9o

in r Vyrnwy is a mountainous district No th Wales , while B Rivington is the old catchment area near olton . The inﬂuence of the neighbourhood of town s upon the purity of the supply is shown by the increase of dissolved matter from to 1 0 3 6 pts . per When the catchment area is formed by peaty land a large ﬁ its amount of organic matter nds way into the water , due

to the gradual decay of the plant life which forms the peat . In such cases the surface water will be coloured brown and

may contain large amounts of organic acids . Surface waters - coming from peat or moor lands , even when colourless , should therefore always be examined by a competent a- r for chemist before use in ste m boile s , these organic acids

are not detected by the ordinary reagent , litmus . Surface waters collected in chalky districts may contain a or large amounts of lime in solution , either as carbon te

sulphate . In some cases the dissolved impurities present in such surface waters may exceed in amount those found

r . in well wate s The impurities present in surface waters , f i in fact , vary greatly in di ferent d stricts , and the geological formation of the surrounding country has much inﬂuence s upon the nature and amount of the impurities . A a s general rule , surface waters collected in country district a r are purer th n river or well waters , and such wate s may

often be used in boilers without any preliminary puriﬁcation . 65 AND GAS FUEL , WATER ANALYSIS

River w ater on , the other hand , is usually contaminated with the drainage and refuse of the villages and towns its situated along banks and the water , except when taken ’ from near the river s source , is more impure than the surface

drainage water which ﬂows into it . The dissolved salts e may in some exceptional cas s be less , but the suspended impurities and organic matter are nearly always greater in

amount . As as the river grows in width and volume the impurities ,

a rule , increase in even greater ratio , and attain a maximum

at the point where the river meets the tidal waters of the sea . H ere sodium and magnesium chloride are added to the list

of dissolved salts present in the water . It is therefore dangerous to use water from tidal rivers within ﬁve miles of their mouth for steam-raising purposes without daily sampling and testing for detection of these impurities ; chlorides b eing most dangerous salts to have present inside - a steam boiler .

PRI N AN E ATER S G D W LL W S .

The water that has passed through any considerable ’ r is as r thickness of the ea th s crust , a general ule , highly and charged with dissolved gases salts . The greater the

depth from which the water ﬂows or is pumped , the larger

will be the amount of these impurities . Carbonate of lime with the corresponding salt of magnesia is insoluble in of pure water , but the presence carbonic acid in the water

renders these salts soluble . This acid is formed in the water by the carbon dioxide dissolved in its percolation through - ’ the upper soil and sub soil of the earth s crust . Sulphate of l ime on the other hand , is slightly soluble in pure water , and grams of water can dissolve 2 7 2 grams of calcium C of sulphate . arbonates and sulphates lime and magnesia ’ exist in large masses in the earth s crust . The nature and amount of the dissolved impurities present in well and s e f spring water therefore vari s in di ferent districts , and water drawn from 30 0 feet below the surface in some dis tricts will be much purer than water from a Similar depth in other districts . Since the percolation through the soil re w moves all suspended impurities , and charges the ater with l l dissolved gases , wel and Spring waters are genera ly clear and sparkling . Owing to the amount of gaseous and solid WATER

m r i pu ities they contain , these waters are , however , most d unsuitable for boiler feed purposes , and they Shoul never l be W be employed if an alternative supp y available . hen the local conditions are such that well or spring water must be al used , a treatment by chemic s for removal of the greater portion of the impurities is essential to the economical working of the boiler plant . All s fi water from whatever source , containing over fteen sub parts of solid matter per should , in fact , be mitted al aal fi to chemic and mech nic puri cation treatment , and as far as possible the boiler should be fed with the purest water which can be obtained in this way . The rm of so m fo ation scale inside boilers leads to any , and ‘evils so ffi of l diminishes the e ciency and life the boi er , that in most cases the capital outlay upon and cost of working a softening n pla t is a wise expenditure of money , and is soon repaid by h t e savings in fuel and repairs .

The salts found in solution in natural waters may , in fact , be divided into two classes— scale-forming and non -scale ﬁ t- forming. To the rs named class belong calcium car

bonate , magnesium carbonate , calcium sulphate and mag nesium hydrate . To the second class belong the following : C salts alcium nitrate ; magnesium sulphate , nitrate and

chloride ; sodium sulphate , nitrate and chloride . The carbonates of lime and magnesia produce what is known as temporary hardness since it can be removed of by boiling , while the sulphates and hydrates lime and manent hadness magnesia produce what is known as per r . It is important to note that the accumulation of the non scale-forming salts in a boiler produces evils nearly as great C a as those caused by scale . hlorides decompose at temper

tures which may be attained in the boiler , and produce free chlorine , a most corrosive gas . The sulphates and nitrates , on the other hand , when present as concentrated solutions attack the brass and gun-metal ﬁttings of the boiler and cause much trouble from leakage . When using feed water containing considerable amounts of these non-scale-forming off ire salts , it is therefore necessary to blow the boiler at

e . quent int rvals This , of course , means loss of heat and s diminution of its working eﬂi ciency . For these rea ons in nearly every large works it will be found that economy is p romoted by the use of a purer water or by the erection of a w - ater softening plant . 67 AND GAS FUEL , WATER ANALYSIS

A P I NG B OI ER EE D ATER S M L L F W .

The sarnpling of water does not present such difﬁculties am of as the s pling fuel , but some thought and care must be expended upon the matter if fair and representative samples

are to be obtained . The water in large tanks or reservoirs can be adequately n of sampled by stoppi g the ﬂow water from the inlet pipe ,

using a wooden plunger to mix the water, and after waiting w one ten minutes , ithdrawing or more samples of the water r as in Winchester qua t gl s bottles , provided with well

fitting glass stoppers . These bottles are rinsed out twice and fi with the water to be sampled , are then held , until lled , in a vertical position with the mouth one ianch below the surface of the water in the reservoir or t nk . In large reservoirs it is advisable to take samples from more than one position filling an aliquot portion of the bottle from each n the selected poi t , and a boat is requisite in these cases for

work of sampling . If the supply of natural water is known to be quite con

stant in character and composition , the sample may be taken

either at the inlet or exit pipe of the tank or reservoir , but in most cases it is advisable at the same time to sample the whole body of water as describe d above and to compare the test results . The sampling of water as it ﬂows from the reservoir or tanks to the boilers can also be carried out by automatic

apparatus , and below a description is given of an apparatus ’ which is in use at Messrs . $oseph Crosﬁeld Son s Soap

Works , Warrington , for this purpose . l Fig . 2 2 is a sectionaelevation of portion of this sampling A in apparatus . is an iron tank 24 inches diameter by 36 i -off nches in height , provided with a run tap at the bottom . Above this tank a -inch wrought iron pipe branches from

the main supply pipe to the boilers , and delivers the water A « — n -flow into the tank , by the l inch diameter dow pipe i k B C . mar ed , and cock Th s cock is opened and closed D of rod ffi by the lever , formed solid iron , and su ciently to n n heavy fall into the position Show by the dotted li e , E when the support at the point is withdrawn . F is a wooden float 8 inches in diameter and 1 5 inches in thick rl H ness , which ordina i y rests on the support , and by means of an elbow lever and pivoted j oint at G provides a support 68 WATER

E for the cock lever at . When the tank is quite ﬁlled with F off its D water the ﬂoat rises support , the lever is released E of at and drops into the position the dotted lines , thus automatically closing the cock C and cutting off the supply A of water to the tank . B y arranging that the pressure of water in the i -inch is of pipe not too great , and by providing a series these

— . 2 2 . C S L S FIG AUTOMATI WATER AMP ING APPARATU .

tanks each with its float and cock , it is possible to collect of any number of samples water , each representing the flow from the softening apparatus or to the boilers for an equal period of time . as w 2 2 The supply pipe must be bent , sho n in Fig . , between e a in fill ach successive t nk order to make these in rotation , 60 AND GAS FUEL , WATER ANALYSIS anad by aid of a cock on the -inch pipe where it j oins the m in to the boilers , the flow of water can be regulated so that each of the tanks requires any pre—determ ined period of time for filling . A series of twelve tanks may thus be ’ made to contain samples of twelve or twenty-four hours al work , and from these twelve tanks either a gener sample may be prepared by putting twelve equal portions into a or Winchester quart bottle and vigorously shaking , any particular one may be sampled and tested as desired . This apparatus is of especial use in checking the work of softening apparatus of the automatic and continuous type , since for the reasons given in the next chapter one can never be certain that the amount of softening chemicals is n am correct , u less regular and systematic s pling and testing are carried out . The sampling of water from springs and artesian wells f r demands a di ferent method , for here it is impo tant that the water should lose none of its dissolved gases , before testing . In this case , therefore , the Winchester quart glass bottle is filled directly from the spri ng or well as C with little delay as possible . are must be taken also r fi to select a bottle with a pa ticularly close tting stopper , which should be tied down afte r the sample is taken . Samples of w ell and spring water should not be exposed to - an d sun light should not be violently shaken before testing , since either of these may influence considerably the amount of gaseous matter retained in solution in the water .

7 0

AND GAS FUEL , WATER ANALYSIS

rs lu Such wate colour neutral litmus paper a slightly b e tint , o and give a pr nounced canary yellow tint to methyl orange , n al the i dicator usu ly employed for this work . In order to s al e timate the total kalinity of the water 2 0 0 c . c . of the be C ﬁl sample taken as descri d in hapter V are tered , if sus n s pended matter be prese t , and are measured in the ﬂa k

o 0 0 . sh wn in Fig . 23 . This 2 c c . are then emptied into a t as 22 . whi e porcelain b in of about oz capacity . The ﬂask r n l s is i sed out with disti led water , the wa hings are added to the liquid in the basin , and two drops of a con

FIG 2 — S I L SK S . 3 . MEA UR NG F A . centrated sol ution of methyl orange are also added to the A of water . standardized solution of hydrochloric acid -ﬁfth r 1 one no mal strength is now added from a burette , r i drop by drop , with constant stir ing unt l the colour of the liquid in the basin chan ges from canary yellow to orange ﬁ a and nally to faint pink , this last change being t ken as the end point of the reaction .

The number of c . c . of the hydrochloric acid solution used from the burette are then read off and the result is calcu lated as parts of calcium carbonate per of water . In the case of water from peaty lands it will often be found that these give a aslightly acid reaction , due to the decomposition of the org nic matter of the peat , but that these organic acids are not revealed either by litmus paper

. s a to or by methyl orange Such water , in fact , ppear be - r alkaline with the last named indicator . It is necessa y in s al the case of these water , therefore , to use an coholic solu

1 A normal solution contains the molecular weight of the chemical - ra . . w ae . A one ﬁfth ma e in g ms, in c c of t r nor l solution yi lds aa end ea in e e e an d e e ae aw eae one sh rp r ction th s t sts , is pr f r bl to k r , but the volume of solution used must be read off by aid of the

Me niscus ( see Fig . 7 2 WATER

of - al as i tion phenol phth ein ndicator , and to heat the e 200 . . ﬁl s orc c c of water after tration and mea uring , in the p lain basin before titrating with the standard alkali solution carbonate of soda . P - henol phthalein is colourless with acids , but yields a of purple red tint with the slightest excess alkali , and the stan dard sol ution of alkali is therefore dropped in from the burette with constant stirring of the liquid in the basin ,

— - . 2 . S S FIG 4 WATER TE TING APPARATU .

until a permanent purple red colour is obtained . The f number of c . c . o the standard alkali solution used is then off a as read and from this the result is c lculated , not acid , but as the alkali equivalent of the acid present in the water .

Fig . 24 shows the apparatus used for making these tests .

. . . and 0 . in The burette holds 5 c c , each c c divided tenths , - u is provided with a glass stop cock . The standard sol tions of acid and alkali employed are contained in the glass bottles shown on either Side of the burette and titration basin . 7 3 AND GAS FUEL , WATER ANALYSIS

These standard solutions of acid and alkali are made as follows grams of pure anhydrous carbonate of soda are weighed out and dissolved with the aid of heat in 250 c . c . l of disti led water . This solution is then made up to nearly

c . c . in the large measuring ﬂask shown in Fig . 25 and ° is 1 e the cooled to 5 C. b fore ﬁnally diluting to the mark on neck , and transferring to the stoppered glass bottle shown on i the right in F g . 24 . 1 is S tandard Alkali N This solution labelled 3 the exact strength being placed on the label afterwards when the 1 h s . a . solution been titrated with the 3 N acid solution

0 . The latter is made by taking 4 c . c of pure concentrated

— ’ 2 L S L S K . FIG . 5 ONE ITRE MEA URING F A

hydrochloric acid and diluting this to c . c . after cooling ° 1 to 5 C. This solution is now transferred to the other storage n t n d A id N bottle show in Fig . 2 4 and is labelled S adar c 1, . It is checked by weighing out very carefully 4 0 gram of the pure carbonate of soda (after drying half an hour in the covered platinum crucible at a dull red heat) dissolving this n a of hot in the porcelai b sin in a small quantity water , diluting with alarge volume of cold water and titrating n with the standard acid solutio , using methyl orange as indicator . The change from canary yellow to faint pink with orange as an interm ediate stage— indicates the end f o . 0 0 . the reaction It will be found that about 2 c c . of 7 4 WATER

the standard acid solution are required to neutralize the 0 of for 4 gram sodium carbonate , and the test is repeated the sake of accuracy . Finally the stan dard alkali solution ff e o 2 . . is itself checked , by measuring 5 c c in the pipett

FIG 2 FIG 2 . . 6 . . 7 - Y M N1 U S F . c c . . E S C TWENT FIVE . PIPETTE READING O BURETTE

6 l Shown in Fig . 2 and by titrating this with the norma

acid as before . The following are examples of this method of checking the standard solutions 1 ‘ 1 1 . H so i n . f r . Cl lut o . 0 of Na o 5 4 grm 2C0 3 used each test

- 1 . . No . I re quire d 9 8 c c - 2 1 . . No . 9 9 c c - Me an 1 9 8 5 c . c .

1 0 0 f 1 c . c . o It was calculated from this that the {3 acid

2 o 1 rms . Na r 0 0 rms . solution equalled 5 g 2C0 3 o 1 9 g CaCOa ﬁ w , and these gures were placed upon the label belo

the description . 1 1 N o u i n C s t o . . . a 2 5 2 0 3 l 25 c c . used for each test .

1 The e a a H Cl for a ae ch mic l symbol for hy drochloric cid is , c rbon t N 0 n ae a aC ad a e CaCO . of sod , 2 3 , for c rbon t of lim 3 7 5 AND GAS A FUEL , WATER NALYSIS

- N . 1 e e 2 . . H o r quir d 4 0 c c Cl solution . - 2 . No . 2 4 1 c c . - Mea 2 0 . n 4 5 c c .

I 0 0 —— From this it w as calculated that 1 c . c . of the alkali g n s — 1 olution equalled c . c . of the acid solution , or 9 37

rms rr s . N C or C . . a a z Oa 1 8 2 n COa g , 7 g These values were therefore placed upon the label of the an s t dard alkali sol ution . In testing waters the bas in and burette must be well di ll e washed and rinsed with sti ed water b fore each test , and the burette must also be rinsed with the standard test s olution before it is ﬁlled for the actual titration . The level of the solution is best read off by holding a black strip of eh the r paper b ind meniscus , and by reading from the lowe

c . 2 . urve of this , as shown in Fig 7 The following is an example of the calculation for water showing akalinity Wate used 0 0 r 2 . . , c c

l z 0 . nuetrai e . Standard acid solution required to it , 3 7 c c 3 7 x x 0 1 9 1 parts calcium carbonate 20 0

of per water . In acid water from peaty land gave the follow ing re s u{ts

Wate used 0 r 2 0 . . , c c

al z . . Standard alkali solution required to neutr i e it , c c x x 0 1 937 1 0 6 53 parts sodium carbonate 20 0 per of water .

. USPENDE D O ID MATTE R II S S L .

i is Water from rivers and surface dra nage generally,

often charged with suspended matter , and it is at times n to f A -n ecessary determine the amount o this . 4 i ch

1 The factor 9 50 can be used in these calculations in place of - 1 0 0 0 0 0 x 0 1 9

7 6 WATER

circle of specially prepared ﬁlter paper is folded in radiating ﬁltrations is the rib fashion to promote quick , and dried in ° ’

0 . stoppered drying tube for one hour at 2 3 F . Fig 4 - 8 . 2 shows the air bath used , and Fig shows the tube ﬁl of employed for drying ters , the stopper being course

removed during the application of heat . The tube and ﬁlter are placed in the desiccator and are weighed together n whe cold, the stopper being inserted in the tube to prevent ﬁ absorption of moisture by the lter while being weighed .

— - FIG 2 8 F R . . . 1 LTE DRYING TUBE

The ﬁlter paper is now removed with small forceps from of the the tube , is placed in a funnel suitable size with

f 0 0 . a 2 . o least possible h ndling , or rubbing the folds , and c c Th of the water are measured and poured through it . e large sample of water must be well shaken before with fi s 0 0 c . c. drawing 2 for this test . The lter with its content fi is nally washed with cold distilled water , dried in the an d funnel as far as possible by suction , is then removed carefully from the funnel and placed in the drying tube for ° fi in the nal heating at 230 F . The dry g must be continued n until a consta t weight is obtained . The gain in weight multiplied by 50 0 gives the parts of suspended matter p er of water . The following is an example of the figu res obtained in this test 7 7 AND GAS N FUEL , WATER A ALYSIS

2 -2 2 Drying tube and ﬁlter and residue 1 1 grms . - alone 1 2 0 67

- Gain in weight 0 55 grm .

' ° 0 55 x 50 0 parts of solid suspended matter per f o water . The colour of the dried residue upon the ﬁlter paper will of give some indication its composition , iron oxides being n generally red , and clay or alumina compou ds being slatey

grey in colour .

’

III . TOTAL SOLI DS .

If the suspended matter has been separately determined ﬁ sam le by the method j ust described , the ltered P of water

is used for this test . Should the suspended matter n ot un ltered w ater have been separately determined , the ﬁ , well

shaken , is employed for the total solids determination .

T 0 . his is carried out by evaporating 25 c c . of the water to

d ryn ess in a tared platinum basin . Fig . 29 shows the s apparatus used for this test . The bain is 2 4 inches in

0 . . diameter and holds 5 c c . of water The water bath is of

c 6 c . c . opper , 5 inches in diameter , holds , is provided of r of with rings va ious sizes , and is supported upon a piece

iron guaze resting on a large tripod stand . A bamboo or glass triangle with a long leg ﬁtted into a c ork which slides upon the upright of a retort stand is em

ployed as show n In Fig . 2 9 to carry a piece of glazed paper i n order to protect the water in the platinum basin from A n e dust or dirt . large cork o the upp r side of this dirt s hield serves to hold it in its place . The platinum basin must n ot be allowed to touch the copper ring of the water b of ﬁl ath , but must be supported by three squares ter

p aper , folded to form thick wedges and inserted between the basin and the e dge of the ring in such a way that an annular space is left for the escape of the steam from the

water in the bath . The platinum basin must be thoroughly c leaned and dried by heating for ﬁve minutes in the air °

0 . A 0 bath at 23 F . before weighing bout 4 c . c . of the

2 0 . sample of water measured in the 5 c c . ﬂask shown in 2 are Fig . 3 then poured carefully into the basin , already placed in position on the ring of the water bath with the aid of the ﬁ to lter paper strips already referred , and the e vaporation is commenced . 7 8

AND GAS FUEL , WATER ANALYSIS

r stage of cooling , before weighing , as pe fectly dry salts are always hygroscopic . The increase in weight of the basin , multiplied by 40 0 gives the parts of solid matter per of water . The following is an example of the ﬁgures obtained in the test :

n n d P latinum basi a residue from 2 50 c . c . of e e w ae 2 6 -1 6 rms unﬁlt r d t r 5 g . - P latinum basin alone 2 6 0 2 3

We e e ° 1 2 ight of r sidu 4 grm .

1 2 x I OO OOO 4 ’ 568 parts of solid matter for of

water . Should it be desired to find what proportion of this s residue is present as organic matter , the ba in and its con tents are heated to redness for five minutes upon a platinum B wire support over a large unsen burner , the change in colour being noticed during the further heating . Organic and compounds darken char under this treatment . The n basin is the allowed to cool , the residue is moistened with a solution of ammonium carbonate in water to reconvert and of the oxides into carbonates , the excess ammonium carbonate and water is then driven off by application of a C gentle heat . are must be taken during this operation that s - the ba in and its contents do not approach a red heat , otherw ise the calcium and magnesium carbonates will be again reconverted into oxides . Finally the basin is cooled in the desiccator and again weighed . The loss in weight 0 0 by this treatment , multiplied by 4 , gives the parts of organic matter per of water . The following is an example of the figures obtained in this test

° - Basin and re sidue after dry ing at 300 F . 2 6 1 4 2 Basin and re sidue after ignition and treatment - w ith ammon . carb . 2 6 1 20

- Loss by ignition 0 2 2

0 2 2 x 1 0 0 0 0 0 8 8 0 parts of organic matter per

8 0 WATER

A E PORAR A ER A E T H R DNE ss . N . N N N A IV . D V T M Y D P M

The temporary hardness of water is caused by the pre sence of carbonates of lime and of magnesia in solution ; per manent hardness is due to the corresponding sulphates and C f r oxides of these two elements . The lark test o hardness is carried out by shaking the water before and after boiling

with a standard soap solution , and by measuring the

volume of this required to produce a permanent lather . The Clark or E nglish degree of H ardness represents the grains on e of 1 of calcium carbonate in gallon water , while the French degree of hardness represents the parts of calcium a G c rbonat e in parts of water , and the erman degree of hardness represents the parts of calcium oxide (CaO) in f n parts of water . The Clark test or hardness is ot

very satisfactory , although still generally employed in this r country . The results obtained va y within wide limits , and are much too dependent upon the personality of the For C a tester . this reason the lark test with so p solution is now giving place to the more scientiﬁc ones described of n below , in which the weight calcium and mag esium car bonates and of calcium and magnesium sulphates as de termined by the approximate analysis are used for calculat of ing the degrees of hardness the water . E The alkalinity , test already described , yields the nglish f of degrees o hardness , if the parts calcium carbonate per

. of of water as ascertained by the c c . standard acid

solution used , be multiplied by 7 . The following is an example of the ﬁgures obtained in this test

f ﬁ . . 0 0 . o 2 c . c ltered water took 3 7 c c of standard acid

s . olution , using methyl orange as indicator x 3 7 E nglish degrees of hardness .

0 1 9 x The expression can be reduced to the 20 0 factor — to simplify the calculation for temporary n . f c c . o hardness . The standard acid solutio used , multiplied b 6 6 E y 5, then gives the temporary hardness in nglish

degrees .

1 This ﬁgure w as chose n by Clark because in his d ay the gallon e e are ai in n o e all . w as the standard of volume . Th r gr ns g on 8 1 G AND GAS FUEL , WATER ANALYSIS

The P ermanent H ardness of water can be approximately ascertained by aid of the results obtained in the tests for al ﬁ has tot solids , if the ltered water been used for carrying out n e this test , and no orga ic matter be pr sent to complicate

the calculation . The results of this test expressed in parts per m al E ultiplied by 7 , will then give the tot hardness in nglish r and f deg ees , the di ference between the total hardness and

the temporary hardness yields the permanent hardness . Taking the ﬁgures given in s ections 1 and 3 of this chapter as an example we have

6 8 . Total solids , 5 per of water 68 x CaCOa 5 parts per of water , or total

hardness in English degrees . The temporary hardness of this water as found by test 1 w as , however , degrees and less gives

as the permanent hardness expressed in English degrees .

It is safer , however , to make a special test for the sul

phates and oxides which produce perman ent hardness . This is carried out as follows

. ﬁ n d 1 0 0 c c . of the ltered sample of water are taken a evaporated to dryness on the water bath in the platinum f basin with an addition of 25 c . c . o the standard sodium

carbonate solution . This treatment decomposes the sul n phates ad oxides and converts them into carbonates . The residue is dissolved in distilled water and is gently f heated . The carbonates o lime and magnesia present in

the residue are left insoluble , and are now removed by ﬁ ﬁltration . The lter and its contents are well washed with

warm water to remove the excess of alkali . The ﬁltrate and washings are collected in the porcelain basin Shown in

. 2 Fig 4 , and after cooling are titrated with the standard n acid solution using methyl orange as indicator . The amou t of acid solution required to neutralize the c arbonate of soda n ﬁ to how remaini g in the ltrate , enables one calculate much has been used for decomposing the sulphates and oxides . The following is an example of the ﬁgures obtained in this test

1 00 c . c . of the ﬁltered water were evaporated with 25 c . c . of n al sta dard kali solution .

ﬁ . The ltrate and was hings required 2 2 5 c . c of standard acid solution to neutralize the excess of alkali . s 0 l l to Now a1 0 c . c . of the akai solution is equivalent 8 2 WATER

. . . . al . c c 2 . of the acid solution , 5 c c equ s c c ,

and c . c . This is the volume of acid corresponding to the carbonate of soda which has disappeared in this test . The calculation for the permanent hardness is therefore as follows I 55 x 0 1 parts CaCOaper of 1 n

water . 0 1 9 x The expression can be reduced to the 1 0 0

1 . . . factor 3 3 to simplify calculation The c c used , multi plied by then yield the permanent hardness in iE nglish

degrees . It is possible to make the standard solutions of acid and al of r s can alk i such a strength , that the deg ees of hardne s f . o be read off directly from the c c . solutions used in making

the test .

It is , however , somewhat troublesome to make a solution f of hl 1 00 . . o hydroc oric acid , c c which shall equal exactly

° 1 a CaCO of 1 c . c . a , 43 gr m of , and another carbonate of soda

of which shall exactly neutralize 1 c . c . of the stan dard acid in . M solution oreover , such weak solutions yield very fi de nite end reactions , and hydrochloric acid solution is liable to change in strength when above or below a definite o limit f concentration . The writer has therefore thought it best to Show how solutions of approximately fifth normal

strength may be applied to water testing without alteration . As al out for ready pointed , by the use of single factors the ﬁ al al xed v ues of the calculation , the labour in c culating the temaporary and permane nt hardness of the water is reduced to minimum . In works where many water analyses are made in the r to cou se of a day it will , of course , be found advisable make these solutions in large quan tities at once and of such a

Strength that calculation is altogether avoided . In such

cases when using 1 00 c . c . of water for the hardness tests the solutions should have the following values

° a . . n 1 C CO 1 00 c . c sta dard acid solution 43 gram a

1 0 0 . . an 1 00 c . c . standard alkali solution c c st dard acid o s lution . 8 3 T AND GAS FUEL , WA ER ANALYSIS

No simple apparatus can be put into the han ds of the ﬁreman or boiler engineer which will enable him to deter

mine for himself the degree of hardness of the feed water . The use of slips of neutral litmus paper (this is sold in books by chemical apparatus dealers) will afford some indication l of the character and purity of the water supp y, but exact al r chemic tests carried out as described , are necessa y in order to obtain any information of real value to the boiler e ngineer . Special testing sets provided with tabloid chemicals and p rinted directions which can be read and carried out by any ﬁ are of reman or stoker , mere inventions the unscrupulous for the delusion of those engineers w ho know little of the c omplex chemistry of the subj ect .

. ESTS FOR OIL VI T .

n Oil is a most da gerous impurity in boiler feed water . When using water from condensers for boiler purposes it is

F IG — S N L 0 . . . 3 EPARATING FUN E

e therefore always advisabl to test the feed occasionally , in

order to find if any oily or fatty matter be present . The sample of water should be cooled before testing and then shaken vigorously with one-fifth of its volume of petrol - a . 0 orof ether in the type of sep rating funnel shown in Fig 3 . The liquid is then allowed to stand fifteen minutes in the u an f nnel supported in an upright position in a retort st d, 8 4 WATER

a is off the w ter run by the separating cock , and the petrol or ether solution which floated on the water and contained all oil n n the and fat , is run into the porcelain basin show I lS Fig . 24 . This is placed upon the water bath which . filled with hot water but has no lighted lamp beneath It , ad since petrol or ether vapour is a most inflammable gas -n I Is all lights must . be extinguished in the room where t present . The petrol or ether evaporates quickly and leaves the oily or fatty matter as a residue in the basin . If desired this test can be made quantitative in character transferrm 0 . f 2 . o for by using 5 c c the water the test , and by g the solution of the fat to the tared platinum basin towards the end of the evaporation . The large porcelain basin and the separating funnel must in this case be rinsed out with s to I the saame spirit , and the rinsing added the solution n the pl tinum basin . u for The oily resid e in this , after heating some time at °

1 . ﬁ a e 2 2 F . , is n lly w ighed To distinguish between animal or vegetable and mineral oils the residue in the basin should be heated on the water A bath with a strong solution of caustic potash . nimal and e t vegetable oils saponify (i. e . yi ld a soap) under such trea

n . ment , while mineral oils remain u altered The presence of oily and fatty matter in boiler scale can be detected by heating the sample of scale to redness in the n i platinum basin . If charri g occ urs or if a d sagreeable c ro odour be noticed , some organi fatty compounds are p bably present . Fats and oils of animal or vegetable origin are more dan gerous lubricants for engines than hydrocarbon s fﬁ oil of the para n class , since the former contain oxygen . and yield organi c acids when decomposed at the high tem r r p e atu e occurring in the boiler .

8 5 VII CHAPTER .

The P rat aA t o n T u c ic l pplicai s of the est Re s lts .

I . FREQUE NCY OF SAM PLI NG AND TE STI NG .

H E - u feed water from whatever source , sed for the o boilers ught to be sampled regularly , at shorter or in to of longer tervals according the nature the supply , and these samples ought to be t ested in accordance with C the instructions given in hapter VI . In the case of surface and river waters the amount of impurities present in the water will be high during the dry a ﬁ the season , and also fter the rst heavy rains at end of A such a season . t such times therefore the samming and l testing of the water should occur at least once week y , or oftener if this be possible with the laboratory staff available D d afor the work . uring the winter and spring after the ry ccumulations of the summer have been washed away , surface and river waters do not show such large amounts of n an d dissolved and suspe ded matter , at these periods of the year a monthly examination of the feed water will fﬁ su ce .

Well waters , and waters drawn from large catchment and ﬁ i areas in mountain districts stored in arti c al lakes , do u not , on the other hand , Show s ch marked variation in the amount of impurity at the various seas ons of the year ; or and in the case of such waters , a monthly quarterly fﬁ examination of the water is su cient .

IL. II . O If any portion of the boiler feed be drawn from the h to and condenser water t is ought be specially sampled daily ,

tested for oil in the manne r already described . In skilled hands the test will not require more than twenty minutes out r to carry , and the results obtained may save ve y heavy

expenditure upon boiler repairs at some later date . If oil be discovered in appreciable quantities the water should be 8 6

AND GAS FUEL , WATER ANALYSIS

1 0 8 47 x 1 0 x 1 0 0 will require 6 9 lbs of 98 per cent . 1 6 x 98

alkali .

M . IV . S USPE NDE D SOLI D ATTER

Q Water Which contains more than ﬁve parts of suspended ﬁ matter per of water , should be ltered or allowed

to . to settle in large tanks , before admitting the boilers Sand ﬁlters with an upward ﬂow are the most suitable for

removing suspended impurities , since these can be quickly cleansed by washing out in the reverse direction with clean

. use al ﬁl water When softening plant is in , speci tration is for unnecessary , the suspended matter in the untreated water is carried down with the carbonate of lime ; and the only dan ger in such cases is that this lime is not allowed time

to settle . The water from the softening plant may there at of fore times contain suspended carbonate lime , the

remedy for which is a larger softening plant . AS a rule the precipitation of lime and other insoluble matter in a water hot softening plant occurs more quickly with feed water ,

and it is advisable , wherever possible , to use heated water in to the softening plant , and take the necessary steps to

avoid loss of heat during the settling process . Water in a vessel exposed to the outside air will necessarily lose much an d heat by radiation , the convection currents set up by this loss will interfere with the settling of the insoluble

matter .

OTA O I DS . V . T L S L

The total solids in feed water should n ot exceed 20 parts per and where this limit is exceeded the water to of f must be submitted the action so tening chemicals . The principal salts found in feed water are the carbonates

and sulphates of lime and magnesia , while the chlorides and sulphates of these elements and of sodium occur seldom

and in smaller amount . The danger of large amounts of sodium chloride being present at times of ﬂood tide in

water drawn from tidal rivers , has already been referred to ; and for boiler plants located in such positions it is necessary to have some means for quickly noting the change i n the character of the water supply .

The chemist in charge in such cases , either Should test the 8 8 WATER

at ull tide for the water daily f chlorides , or Should provide ﬁreman in charge of the boilers with asalinometer of a of a i reliable type . A test based on the use c psules of S lver nitrate containing a weighed quantity of this salt with e for potassium chromate as indicator , may also be d vised showing when the feed contains more than 1 0 parts of

sodium chloride per of water . As regards the treatment by chemic als for removal of

the total solids , this subj ect will be dealt with at length a e in the next chapter , and it is only necess ry to state her that carbonate of soda is the chemical usually employed

in conj unction with lime .

VI . TEMPORARY AN D PERMANE NT HARDNESS .

Waters testing above 1 5 English degrees of temporary and permanent hardness ought not to be used without - chemical treatment in steam boilers . The heat losses and d isadvantages resulting from the of so formation scale are great , that in nearly all cases in large works a positive saving results from the installation 1 i. e . a . P of softening app ratus ermanent hardness , that due to sulphates of lime and magn esia is the more dan erous r g , since these salts fo m a more compact and adherent al a sc e than the corresponding carbon tes . The removal of hardness by chemicals will be dealt with in Chapter V may aof III , but it be remarked here that the remov l r is n f b and pe manent hardness o ly e fected y substitution , that an equivalent amount of another salt remains in solu i tion n the water . The lime and magnesia sulphates are l converted into the insoluble carbonates , whi e the sulphuric of i acid radicle remains in the water , in the form sod um

sulphate , a very soluble salt . Where waters with a high permanent hardness test are for - al l s employed boiler feed purposes , the tot dissolved sa t will therefore always remain high even when a softening

use of . plant is in , owing to the presence this sodium sulphate In such cases it is advisable to blow off the boilers oftener w than would other ise be required , since a concentrated solution of sodium sulphate is corrosive in its action upon ﬁ the gun metal and brass ttings of the boiler .

1 A scale inch in thickne ss is the limit that ought to be allow ed

to form on the boiler plates . 8 9 AND GAS FUEL , WATER ANALYSIS

In all cases it is in fact advisable to test the water re in and off maining the boiler at weekly intervals , to blow the boiler when the total dissolved matter exceeds 250 parts per of water . A The ppendix contains some typical tests of feed waters .

9 0

AND GAS FUEL , WATER ANALYSIS

1 a to ﬁ e or 5 degrees of tot l hardness , ought be puri ed befor

use . The treatment of the water with softening chemicals h To s ould occur in special apparatus outside the boiler . r car y out the softening treatment inside the boiler , is likely

to aggravate some of the evils it is desired to avoid .

The proper control of the feed water of the boiler , can only be exercised by aid of systematic and constant sampling and testing of the water as it enters and leaves the softening

. O ad aapparatus ccasional tests of the feed water before n fter softening treatment are useless . In the ﬁrst place the character of the supply may change n from week to week and even from day to day . In ma y localities the rainfall is the chief agent in causing this variation in the amount of suspended an d dissolved im u in al the p rities present the natur water supply , in others the has dictum of city water engineer , who , as a rule , more d n than one available source of supply . Secon ly, the softe ing apparatus even when of the automatic type requires h c ecking , for owing to this variation in the composition of

’ naturalsu l s the pp y , there is grave danger of exces ive chemi a and c ls being used at times in the softening operation , the these , as already stated , may be almost as harmful as

original impurities of the water . With pure water a boiler may run from to 0 fﬁ h urs without cleaning , and the evaporative e ciency u al re d ring this period , owing to the absence of Sc e , will

main high . With dirty and impure feed water on the other 0 0 hand , the limit of safe work is 5 or less hours , and the evaporative efﬁciency will fall rapidly towards the end of i this comparatively brie f working period . It s evident of therefore that in the maj ority works , a water softening of i plant will prove a wise expenditure capital , for it w ll increase the capacity of the existing boiler plant and reduce

the fuel consumption at one and the same time , with only

a slight addition to the daily working expenses .

. E E ISTR OF T E OFTENING ROCESS II TH CH M Y H S P . The chief salts that can be removed from water by the use of softening chemaicals are the carbonates and sulphates of lime and magnesi . Taking these seriatim we have the following 02 WATER

1 i m abonat — C m a CaCO . Calc u C r e he ic l symbol s , always al a CaH C0 present in natur waters as bic rbonate , ( 3 ) 2 This salt can be removed n t e w te reci i (a) B y boili g h ar. This method yields a p p tate of calcium carbonate which may either form a hard scale , or a soft mud , according to the slowness or rapidity o f separation . The chemical equation for this Change is

H CO 3 0 H 0 60 . C3 . ( 3 ) 2 C C 3 2 2 (insoluble )

k ime o lime to the w ate (6) B y adding slaed l r r. This method yields double the amount of calcium carbonate aas ﬂocculent which occurs in method a aprecipitate , buat it avoids the evolution of carbonic acid g s . The chemic l e quation is H aC0 H H C0 C3 O ZC 2 0 . Ca. ( 3 ) 2 ( ) 2 3 2 (insoluble )

( c) B y adding caustic sodato the w ater. This chemical C0 a combines with the second 2 molecule of the bicarbon te , and precipitates the calcium carbonate as before , while sodium carbonate remains in solution . The chemical equation is as follows Na H a NaOH 3 0 CO 2 0 . C (H C0 3 ) 2 Z C C 3 2 3 2

‘ M i m C bon te — C a M C . anes u ar a O 2 g hemic l symbol ag always present in natural waters as bic rbonate . H Mg ( C0 3 ) 2 . This salt can be removed B (a) y boiling . 1 (b) B y adding slaked lime or lime water .

( c) B y adding caustic soda . The chemical equations are similar to those given fo M Ca calcium carbonate , with the substitution of g for in the a B a a u molecular formul e . oth c lcium and m gnesi m car bonate are loose ﬂocculent precipitates when ﬁrst caused to use separate from water by the of lime water , but boiling

a . renders these precipitates dense and cryst llaine i a C C S O . al um ul h te . . c S 3 C p hemical symbol , . This salt can be removed from water bonate (a) B y heating w ith sodium car . The chemical eq uation for this change is

1 Magnesium carbonate is converte d into magnesium hy drate by hi e ae es the re heating with lime water . T s furth r ch ng complicat ae i al r e a e ae b e e ae e t. ctions for soft ning w t r y lim , wh n m gn s um s ts pr s n 9 3 AND GAS FUEL, WATER ANALYSIS

NaC Na aS 0 S O a . C O. , 3 g . C COs ( insoluble )

n w i h baium c b (b) B y heati g t r ar onate . This chemical one an d the produces two insoluble compounds instead of , sulphate is removed from the water as barium sulphate . The chemical equation is aO B aCO CaCO B aS O C S 4 s a .. (insoluble ) ( ins oluble )

ul hate . C M Manesium S S o . 4 . g p hemical symbol g , This salt can be removed from water by heating it with sodium carbonate or with barium carbonate . The chemical M Cain equations are as in 3 , with the substitution of g for As a r . the molecul r fo mulae before , the magnesium car is be bonate converted into the hydrate , if excess of lime present . h a It will be observed that w en saodium c rbon ate is the chemical used for removing sulph tes , calcium and mag nesium are carbonates again the precipitated salts , but that an equivalent amount of sodium sulphate remains in solu tion . ﬂocculent The precipitated carbonates , as before , are when ﬁ n rst separated , but become de se and crystalline on boiling . B a th rium carbonate on the other hand , removes both e base and the acid radicle of the calcium or magnesium sulphate , and for this reason would be generally employed , if cheaper and more easily applied to this purpose . As barium salts are extremely poisonous their use is attended a one with some d nger in aalarge works , where only water supply is available for ll purposes . M nesium Chloride and Manesium Nitrate 5. ag g . These two salts are oc casionally present in natural waters in ap n preciable amou ts . They can be decomposed in the same manner as the sulphate , namely , by heating the water with sodium carbonate . The chemical equation for the changes are as follows

N3 M N I MgC12 2C0 3 gC0 3 Z aC . M N0 N3 C0 M CO NaN ( 3 ) 2 + 2 3 g 3 Z oa. g ‘ H ere again an equivalent amount of sodium chloride or n is itrate remains in solution , and the change from a dan n s gero s salt to a less dangerous one , for magnesium ul l ri d e o phate , ch o de and nitrate are all salts which may c m 9 4

AND GAS FUEL , WATER ANALYSIS

H 0 = C8 C (33 8 0 2 NaH 0 50 H . ( 3 ) 2 4 + O 2 C3 C 3 Na 2 4 Z 2O

Carbonates and sulphates do not , however , occur often in C ais o water in molecular proportions . austic sod als d earer than caustic lime . The calculations of the necessary amount of chemicals are madeafrom the chemical equations given in section 2 of this ch pter . These show that one molecule of calcium c arbonate or magn esium carbonate requires one molecule of calcium hydrate or two molecules of sodium hydrate to n precipitate it , and that one molecule of calcium or mag e nesium sulphates requires likewise one molecule of sodium or barium carbonate for its decomposition . The molecular weights are as follows S a CaC lts O3 1 00 pre se nt in Mg CO . 84 w ae as C S t r a 4 1 36 impuritie s MgS O . 1 2 0 CaO 6 or CaOH 5 ( )2 S e NaOH 40 or 2 Na( oH ) oft ning ‘ e a N 1 0 6 Ch mic ls aeCO . B ac O I 3 9 7

Since the English degrees of temporary and permanent hardness in water are expressed in grains of calcium car er of i is bonate p gallon , and the molecular weight th s salt

1 0 0 ﬁ . just , the calculations are considerably simpli ed It is only necessary to multiply the degrees of temporary or permanent hardness by 56 or by to ﬁnd the grains of lime or sodium carbonate requiraed for each gallon of water . The further calculation is equ lly simple , Since the result multiplied by 58—23(that is divided by gives the pounds avoirdupois required to soften gallons of water .

Fig . 3 1 is a diagram showing the pounds of softening chemi c of als required per gallons water , for the various ° degrees of hardness from 1 to For the other chemicals the method of calculation is simi 8 0 lar , the factor for caustic soda being and for barium c arbonate in place of those given above . It must be remembered in making these calculations that commercial c c hemicals are seldom pure , and that some allowan e for impurities and losses in making up the solutions must be

1 On e pound avoirdupois contains grains . 9 6 WATER

ﬁ made , in the nal calculations for water softening . For a l 0 is caustic lime n al owance of 2 per cent . advisable ; for sodium carbonate 5 per cent . , and for caustic soda an allow ance of 1 0 per cent . is It , however , unnecessary to purchase high grade chemi for h cals softening purposes , and if a chemist be in c arge of the n can ff in softeni g plant , a considerable saving be e ected the as annual expenditure upon chemicals , by purch e of low grade

— F IG . I . T S L Y S C LC L S 3 DIAGRAM O IMP IF WATER OFTENING A U ATION . or t 0 damaged produc s from the manufacturers , at 5 per

c . ent below the usual market prices . Thus in the manu facture of caustic soda an impure product known in the trade as caustic bottoms is produced which sells at less of a than half the price white caustic sod , and yet contains o 0 . Na 0 . ver 5 per cent of zO as compared with 7 per cent in the standard product . Should it be desired to use caustic s oda in the softening plant instead of caustic lime , this 9 7 H AND GAS FUEL , WATER ANALYSIS

impure product is quite good enough for the purpose . da Carb”onate of soda is usually bought in the form of a so ash can , and here again a low grade product be bought with advantage . Sulphides in solution , howeaver , may be dangerous , and should be removed by chemic l treatment ,

before using the soda solution in the softening apparatus .

TH E IV . SOFTENING APPARATUS . All the well-known and patented forms of water-softening apparatus are based on the chemical reactions given in f section 3 . They di fer one from the other simply in the al of mechanic arrangement the plant , and in the devices used for rendering the apparatus automatic in its action . The best illustrated description of water softening apparatus M is to be found in a paper read by essrs . Stromeyer B aron before the Institution of Mechanical Engineers in

D 1 0 . w as ecember , 9 3 This paper printed in full in many the o r of technical j u nals of that date , and readers desiring further information upon the subj ect of water-softening apparatus are referred to it .

The use of secret compositions for putting into boilers , to in order prevent scale and grease troubles , is unwise and A h s should never be practised . part from t e excessive price paid for such compositions , the impurities of a boiler feed water ought to be removed before the water enters the boiler , ul and sho d not be separated inside it . The addition of caustic soda to the water as it enters the boiler is another practice which must be con demned except when carried out under expert control , for under certain conditions this use of s caustic soda does more harm than good , and also increase of the tendency the boiler to produce wet steam . The water therefore must be treated with the softening

l . chemicals in some apparatus , outside the boi er plant If a chemist be available to take charge of and control the operations , the writer considers the intermittent type of apparatus in which the chemicals are allowed a consider in n able time to act large ta ks , in which the precipitate can

. An r settle by gravitation , to be the best y chemical enginee n - i the could desig such a water soften ng installation , and payment of heavy royalty fees or of an excessive ﬁrst cost for the plant would thus be avoided . aIf no chemist be available to control the work of the pl nt , one of the automatic and continuous types of soften 98

AND GAS FUEL , WATER ANALYSIS

of calcium or sodium hydrate in the treated water can also use a ll r be detected by the of turmeric p per, which wi tu n

reddish brown when these chemicals are present . This effect will of cours e signify either that an excess of lime and soda is being used in the softening plant or that the water is being passed too rapidly through the plant and that the chemical reactions described in section 2 have not had

time to be completed . The time element is in fact much more important than many makers of automatic softening a is apparatus have re lized , and it this time factor which gives the -non-continuous tank system of softening a decided

advantage . When taking samples of water from the tanks shown in 6 mix Fig. 22 (p . 9) it will be necessary to thoroughly

— . 2 . GLAS S $AB FO R X S L S FIG 3 E AMINING WATER AMP E .

n the water with a wooden plu ger , for a sediment of calcium ma carbonate y be found at the bottom of each vessel , due

to unsatisfactory working of the softening plant . When

such a sediment is found , either the rate of flow through the softening apparatus is too rapid or the filtering chamber is fi out of order , and an investigation must be made to nd H aand remedy the fault withouat delay . aving selected cer t in samples for further ex mination , the temporary and permanent hardness of these must be determined by the C methods described in hapter VI . The difference between the tests of the untreated and treated water will Show the ffi of n working e ciency the softe ing plant , and will also 1 0 0 WATER provide the necessary data for any alterations in the amount l . O or of chemicals used nce weekly , daily if this be possib e , an average sample of the water flowing from the softening s plant is tested for total solids and for oil . A already pointed out the use of sodium carbonate or caustic soda has the disadvantage that sodium sulphate remains in solution in the treated water ; and thus the . use of these chemicals will not cause any reduction in the total dissolved solids but i f n actually Increase them . The soften ng o a water containi g 1 0 s rm i 1 0 er degree of pe anent hardness , w ll add over p cent . to the amount of dissolved salts in the softened water . For this reason , boilers fed with hard waters that have been softened by use of sodium carbonate , require frequent off a con blowing , otherwise the sodium sulph te becomes centrated to a dangerous extent and attacks the boiler

fi n s . AS tti g regards the presence of oil , this will not often a w be found in the w ter flo ing from the softening apparatus , since the precipitation of calcium carbonate in this will r il . S o ca ry down the particles of oil hould be found , n to al oil however , it will be ecessary erect a speci separator for treating the water as it flows into the softening t apparatus , a separator working on the electroly ic principle being most suitable for works where electric power is generated .

I OI

CHAPTER IX .

The Chemical and P hysical Characte ristics of the Wate G ae — Sam n the Gae s s s pli g s s .

E I CA ARACTERISTICS I . CH M L CH .

H E burn ing of ordinary bitumin ous fuel in the furnaces of steam boilers is in reality a chemical oxidation of and of the da the hydrogen carbon fuel , this oxi tion being fli n l n su cie t y i tense to produce light and heat . When this oxidation orcombustion of the fuel has been properly carried a as of e n out , one obt ins products the ch mical cha ges only d u e n carbon dioxi e and aqueous vapo r , these two b i g diluted , the i an d um x however , with n trogen uncons ed o ygen of u the air . S lphur dioxide gas is also generally present as AS an impurity derived from the sulphur in the coal . al h n s s ready pointed out , it is t is i gredient of the waste ga e and s not the soot which destroys vegetation , and damage stone work in all great industrial centres where coal is burned l on a large sca e . The chemical products of the perfect combustion of bitu — C s minous fuel in air are therefore arbon dioxide , aqueou u and as vapour , and sulph r dioxide , with oxygen nitrogen diluent gases derived from the air . Should the combustion process have been improperly r t u car ied out , par icles of nconsumed carbon , and carbon monoxide gas will be present in the exit gases farom the furnace of the boilers ; while with very bad man gement , entirely unconsumed hydrogen and hydrocarbon gases n may be fou d in addition to those already named . ' Semi-anthracite and anthracite fuels contain less hydro u b as gen , and when heated prod ce less hydrocar on g i than bitum nous coal . The waste gases from boiler fur fi of u naces red with these classes fuel , as a r le , therefore , contain little soot or undecomposed hydrocarbons but on a ma the other h nd , the proportion of carbon monoxide y n i fli ien nsu c c . be high , owi g to y of the air supply Coke also is a fuel which may produce large amounts of 1 0 5 AND GAS FUEL , WATER ANALYSIS c arbon monoxide when the management of the furnace and l of the air supply is at fault . Sulphur dioxide is genera ly present in large amounts in the waste gases from furnaces

ﬁred by coke manufactured from cheap classes of fuel . d It may be pointe out here , that with coke and anthracite . the appearance of the waste gases gives no indication of the an d an d n s c omposition of the gas , a clear appare tly sati factory waste gas may contain large amounts of carbon monoxide and be associated with very bad conditions of c ombustion in the furn aces of the boiler . Chemical tests of the exit gas are therefore required to d iscriminate between good and bad work with coke and anthracite . - s When semi anthracite or bituminous fuel are employed , i fﬁ ll s as nsu ciency of air supply wi cau e , before . carbon mon

oxide to be present in the exit gases , but in this case , owing to the larger amount of hydrocarbons liberated on heating l as n the coal , smoke wil be produced , and will serve a sig al i that the combustion of the fuel s incomplete . With these cheaper and more generally employed classes o f al fuel , a chemic examination of the waste gases therefore is not required to prove the existence of bad management

o f the furnaces and incomplete combustion . The freedom from smoke of the gases issuing from the chimney top is s ufﬁcient to indicate whether good or bad conditions 1 i e xist within the furnaces . It s somewhat depressing to

have to state that , although bituminous coal of all grades can be burned entirely without smoke production in the - furnaces of steam boilers , in nine cases out of ten the correct c onditions for perfect combustion are absent , with the result that soot and carbon monoxide with unconsumed hydrocarbon gases are allowed to escape in large quantities

into the atmosphere .

A ARA ER C . SIC CT ISTI S II PHY L CH .

Temperature — The temperature of the waste gases de p ends very largely upon the amount of excess air allowed

to pass into the boiler furnaces , and upon the cleanliness of of the boiler . A large excess air causes a low initial temperature in the furnace and a low heat transmission to

1 E e air ma ea a ea e xc ss of y l d to cl n chimn y top , but does not ﬁ p roduce ef ciency in steam raising . 1 0 6

AND GAS FUEL , WATER ANALYSIS

values are too high unless a very excessive amount of

aqueous vapour be present in the exit gases . The import ance of speciﬁc heat is recogn ized w hen ' calculating the heat

lost with the exit gases . Chapter XII will deal with this an d al n subject , in the c culations there given a consta t value of 2 40 for the speciﬁc heat of the dry exit gases will

be employed , this being the condition in which they are

generally an alysed . S pecific Gravity — The specific gravity varies like the fi speci c heat , with the chemical composition of the exit an a l d c n al . gases , be c cu ated when this is known The following are the figures for the specific gravity of the gases likely to be present - Carbonic acid (CO 1 52 9 - Carbonic oxide (CO ) 9 67 - N e N 1 E e e ta itrog n ( )2 9 7 xp rim n l O e O I -1 0 w xyg n ( )2 5 figures ith H e H -0 6 i - ar 1 0 0 0 . y drog n ( )2 9 A e a H O qu ous v pour ( 2 ) - S ulphur dioxide ( 8 0 2 ) 2 2 1 7

A large percentage of carbonic acid or of sulphur dioxide in the exit gases is therefore accompanied by a high speciﬁc

gravity . The variations in speciﬁc gravity are used in one C0 r d form of 2 testing appa atus , to in icate the percentage r of C0 2 . Fu ther details of this instrument will be given

in Chapter XI . nd — Refractive I ex . The refraction index of furnace gases has not until lately been regarded as of any value as a guide An to their chemical composition . apparatus based on accurate observations of variations in the refractive index i n H a has has now however been des g ed by F . ber , and quite recently been placed upon the market by Carl Zeiss h l of $ena . T is instrument alows percentage variations w ° 1 fra 0 . C re c do n to per cent of O. to be easily detected , the tive index of the furnace gas being compared with that of or aa A f air some other st nd rd gaseous mixture . di ference f of 1 per cent . of carbon dioxide makes a di ference of 0 0 0 0 0 1 the co-efﬁcien t s 5 in refraction , and thi corresponds d of to ivisions upon the scale the instrument . Further details of this instrument will be given in

Chapter XI . Colou — of r. The colour the gases passing away from the furnace or combustion chamber of a boiler is some guide 1 0 8 WASTE GASES

e G to the completeness of the combustion proc ss . ases in

w hich all the carbon has been burnt to C0 2 and the hydro

gen to H 20 are colourless and transparent while incan descent particles of carbon produce yellow tongues of flame an d unconsumed hydrocarbon gases appear as dark streaks e l upon the bright background of the furnac wa ls . It is advisable to use a dark blue glass when examining the a fi furnace g ses in this way , and to study rst the various B u phases of combustion , with a nsen burner flame held a B ag inst a cherry , red background of card or paper . y varying the amount of air admission at the base of the burner all degrees of combustion c an be obtained with the hydrocarbon gases supplie d to the burner and in this way an artificial reproduction of the conditions prevailing in a r a boiler furn ace may be obtained . It is unfo tun te that the combustion process in the furnaces of the ordinary internally fired Lancashire boilers cannot be studied in this a e way , and th t the method is only applicable to the furnac - gases from water tube boilers . In these latter and in ex ternall fi L y red ancashire boilers , slight holes closed with sheet mica can easily be made at suitable points in the r b ickwork setting , and a valuable check upon the com bustion process can be exercised by observations made

through blue glass in the manner already described . The o bservations should be made as far from the boiler furnace as n possible , but before the gases come into co tact with the w - ater cooled plates or tubes of the boiler .

III . SAMPLI NG THE WASTE GASES . For regular and systematic control of the work of the

boiler plant , it is necessary that frequent tests should be e e h for mad of the exit gas s from eac boiler , both tempera

ture and for chemical composition . The best place for taking the sample of the gases is at a point about 1 8 inches or two feet on the boiler side of the damper which is used to cut off the boiler from the main Laa flue to the economizers or chimney . If nc shire boilers are installed and a separate damper be provided for each flue side , two sampling holes will be required . The holes should be made large enough to take a thirty inch length of 1 - % inch wrought iron pipe , provided with a flanged top which can be closed with a bolt or plug when the pipe 1 0 0 AND GAS FUEL , WATER ANALYSIS

ﬁ - is not in use . This pipe is set with re clay in the brick of work setting the boiler , and is always ready for use when ﬁ required . When xing new boilers in position it will save time and trouble if a number of these sampling holes be

provided in the brickwork setting , at points suitable for e testing purposes . Similar holes ought to be provid d at one or more places in the maain ﬂue , and in the breast of the chimney , since the chemic l composition of the waste gases at these points serves to indicate the amount of air leakage through faulty brickwork that is occurring between the

boiler and the chimney . The sampling hole on the chimney breast must be made at a height of at least 1 5 feet from u bafﬁe a the the gro nd , as w lls are generally provided at chimney base to direct the gases from Opposite ﬂ ues in an U a pward direction , and a s mple of the mixed gases cannot

therefore be obtained at a lower level . ﬁn s Foar withdrawing the sample of gas from the es , tube of h rd (potash) glass must be employed in 3 feet lengths , 1 - n A e and of 5 inch exter al diameter . rrangem nts for attempting to sample the gases across the whole width of the ﬁne by use of a perforated or split wrought iron tube

are of little use , and it is best to simply insert the hard glass 8 so 1 . tube , that its far end proj ects about inches into the flue A large bung 1 7} inches in diameter is fixed upon the glass fi the sampling tube , so that it is rmly held in position in e flanged end of the wrought iron pipe , and no air leakag A into the flue occurs at this point . second or third length of hard glass tube is now j oined to the first by thick red rubber tubing of good quality (black rubber is not suited

for work with hot gases) , and the sampling apparatus is t then connected o the free end of the sampling tube or tubes . No ﬁltering arrangement for retaining soot is used at this stage of thae asampling , since it is better to have the freest possible p ss ge between the ﬂue and the sampling vessel . Soot and moisture will collect upon the inner walls of the l glass tubes , and these must be wel washed after each e sampling operation , and allowed to drain and dry befor use again . Three methods for collecting the sample will now be described na-sam les h i h (a) . S p p of t e w aste gases taken by ad of t e ’ ’ d method and H oni man s s- — H ni n s ry g gaburette . o gma original gas-burette w as designed for testing the amount of carbonic 1 1 0

AND GAS FUEL , WATER ANALYSIS a e w for pparatus must be increas d correspondingly , to allo the increased volume of air requiring displacement . - D n s The glass stop cocks on are then closed , the spri g clip are of A placed on the rubber at each end , and the sampling a c an pparatus then be disconnected , and the sample taken away for analysis . Since no liquid is used in this method of collecting gas s amples , no absorption of carbonic acid gas can occur before s i l testing . A a general rule the soot w l be removed by l of d eposition upon the inner wal s the sampling tubes ,

before the gas arrives in D or A . With very smoke laden gases , a little glass wool must be inserted in the rubber tube D c onnecting to the tubes that pass to the ﬂue , in order to

ﬁlter out the particles of unconsumed carbon . s extendin ove one hou — W Averae sam le r r. (b) . g p g hen an average sample of the waste gases over aperiod less than is o ne hour is required , the apparatus shown in Fig . 34

— . S L S FO R K S L S FIG . 34 AMP ING APPARATU TA ING AVERAGE AMP E ONE H OVER OUR .

A - employed . is a water j et air pump which works upon the B inj ector principle , and draws air through the arm at a rate

dependent upon the pressure of the water supply . The arm B is connected to the limb marked C of aStead gas sampling E of D bottle , , and the other limb this is connected to the d h glas s tube or tubes lea ing to t e ﬂue . The sampling vessel E ﬁ a is lled up to the point where it j oins the horizont l limb ,

with mercury , or with water saturated with carbonic acid . The pump is now set to work and gas is drawn from the im C D ﬂue at a constant rate through the horizontal l bs , 1 1 2 WASTE GASES

n of the gas sampler . One of the lower cocks of this is the to al A opened low a slow trickle of mercury to pass out . mercury bottle provided with a funnel is placed below to

receive this and store it for the next sampling operation . The rate of flow of the mercury must be arranged so that the sampling bottle will still contain some of the metal at the end of the period over which the sampling is to extend . Du w fin ring this time , gas dra n from the es will have been C D ro passing continuously through , , and a small p portion of this gas will have been coutinually passing E into to replace the mercury which has flowed away . At of E the end the period , therefore , will contain a fair of all average sample the flue gases , and on closing the cooks the apparatus can be disconnected , and the vessel E containing the gasasample can be taken to the laboratory for chemical examin tion . n If the waste gases are very smoky , a small bottle co tain fi ing glass wool must be used to lter out the soot , this being r D inse ted between and the first length of sampling tube . to Tahe wool must be renewed for each test , as owing the l rge volume of gas passing through the apparatus , a large al quantity of soot will soon be collected . The mercury so will u out hlter req ire frequent cleaning . This is best carried by in g through a large dry filter paper having a pin-hole in the - apex , and by then washing the mercury with tap water and fil drying with ter paper or blotting paper . When quite and clean no reaction occurs between the dry flue gases . the r mercu y , and the sample can be kept for some hours in the E h vessel without c ange . Average sample extending over tw elve hours — When an average sample of the flue gases extending over a period is w longer than one hour required , the apparatus sho n in

Fi . g 34 cannot be employed , for the weight and cost of the mercury required would be prohibitive . It is then necessary to use an apparatus depending upon difli the flow of water from one vessel to another , and a culty arises due to the variation in the head of water at f di ferent stages of the sampling operation . The head being l greatest when the sampling vessel is quite ful of liquid , the outflow of water and inflow of gas will start at a maximum and gradually become reduced as the water is displaced by T gas . he sample of gas obtained in this way cannot there fore be regarded as atrue average of the waste gases in 1 1 3 1 AND GAS FUEL , WATER ANALYSIS

ﬂue the , during the twelve hours over which it has been

collected . - i M . Fig . 35 shows a gas collect ng apparatus in use at essrs ’ Crosﬁeld s S oap Works at Warrington which is free from

this defect , the rate of ﬂow from the upper to the lower

vessel being constant for all variations in the water level . The water from the upper gas collecting vessel A passes - B through the intermediate water seal bottle , and then into d eliverv the lower reservoir D by the delivery tube C. This al l D of its tube f ls as the water evel in rises , on account

To m s

— S L S FO R K N V S L S . M FIG . 35 AMP ING APPARATU TA I G A ERAGE A P E L H S OVER TWE VE OUR . attachment to the ﬂoat E by a cord passing over the pulley

The head of water is thus seen to depend upon the vertical di A stance between the level of water in the vessel , and the

C . l position of the delivery tube , and a study of Fig 35 wil Show that this is constant for all positions of the water level A A in . The gases will thus be drawn into at a constant rate during the whole of the twelve hours , and the sample therefore may be regarded as a true average of the flue gases in dur g this period . be n Two precautions must observed in usi g this apparatus . The water in A must be covered with a film of oil to pre vent C0 2 absorption and the tube G leading to the flue u m st be connected not direct to the flue , but to a T 1 1 4

CHAPTER X . T f he App roximate A nalysis o the Waste G as es .

AVING obtained a satisfactory sample of the waste gases by one of the methods described in the previous chapter , the sample is brought to the labora tory for further examination . AS a general rule it is advisable to test the gas sample as soon as possible after its withdrawal from the ﬁnes or from the col lectin g vessel , since there is always some danger of air leakage at the glass cocks or clips of the retaining vessel . It is also best to make the chemical tests in the laboratory , away from the dust and heat of the boiler e house , for no gas testing apparatus will give accurat results when operated in a dusty atmosphere and subj ected E of to rapid changes of temperature . ach the sampling methods described in Chapter IX permits the ﬁnal sample to be transferred to the laboratory for the chemical examina ’ a n G s . ion . The Stead s Sampli g vessel shown in Fig 34 is used to w ithd raw asample from the larger vessel shown in H n Fig . 35 by placing it upon the bracket and by connecti g K G is it to the tube , while the clip at closed and the bottle

B is raised to relieve the suction in A . It is important to note that neither the soot nor the moisture in the waste gases are estimated by the ordinary a a a is methods of an lysis , and that speci l apparatus required when the amounts of these are to be separ tely determined . The soot and moisture collect to a large extent upon the inner walls of the glass tubes and vessels through which n ﬁn the gases are draw from the es , and the water used in the measuring burette during the test of the sample of e course condens s the remaining moisture . To prevent any soot from getting into the gas -testing u one apparatus , it is usual to place a pl g of glass wool in of the connecting tubes and to renew this as often as required . Two methods of testing the samples of waste gases will now be described in detail 1 1 6 WASTE GASES

. ESTS W IT T E H ONIGMAN GAS URETTE I T H H B .

S ee . . ( Fig 33 , p The burette A is first immersed for two minutes in the F -fifths fil trough , which is four led with water , in order to bring the gas contents of the burette to a standard tempera on n ture . The spring clip the zero end of the burette is the i removed , and the burette is held upright in a vertical pos tion with its lower end immersed in the water contained in the E glass j ar . The zero mark of the burette is placed level E is l w with the brim of , and the upper clip opened to a lo of the excess of gas to escape into the air , water being poured E A into from the trough , until the gas contents of measure

1 c . exactly 0 0 c . The burette is then removed from the j ar E and is placed in a slanting position in the trough , with its upper end resting on A 1 one of the ends of the same . small gram piece of solid caustic potash is taken in the forceps provided for the purpose , and is inserted in the rubber tube at the lower end of the burette . It is then pushed up into the same by the rod short length of glass , this being left in the rubber tube as stopper .

The burette is now removed from the trough , and is gently r of shaken , the pu pose being to dissolve the small lump caustic potash in the water contained in the burette , and to bring the resultant solution into use as an absorbing liquid . ’ Two minutes agitation is usually suﬂicient for this pur and is pose , the burette then again immersed in the water contained in the zinc trough , in order to bring back the con ﬁn tents to the original temperature . The al operation is to withdraw the glass rod from the rubber tube on the lower avertical t end of the burette , to hold the latter in posi ion in off of l the glass j ar , and to read the volume gas after level ing the liquid within and outside the burette . The loss in volume equals the percentage by volume of C0 2 contained in the gas sample .

This apparatus yields accurate results , and tests with it ﬁve but ra can be completed within minutes , some little p c tice is necessary to obtain this accuracy and speed . The precautions necessary in its use are as follows

1 . The rubber tubing at the upper end of the burette 1 1 7 AND GAS FUEL , WATER ANALYSIS

must not col be immersed in the water , since if any liquid lects inside it the expulsion of the excess gas above 1 0 0 c . c . fﬁ is rendered exceedingly di cult . 0 2 . N air must be allowed to enter the burette at the lower end when removing the spring clip . To expel the air in the rubber tube it is advisable to squeeze the ﬁngers along the rubber towards the Open end before the clip is r m Opened , keeping the end of the ubber tube eanwhile under water . B C0 3 . oth before and after the absorption of 2 , the gases must be reduced to the standard temperature— that of the

water in the trough . 0 l 4 . N air must be al owed to enter the tube and burette l when it is being placed in the j ar for leve ling purposes , or

when inserting the caustic potash in the solid form . This last operation must be carried out under water ; and speed

is necessary , as this chemical dissolves very rapidly in water .

. A 5 fter each test , the burette , j ar , and trough must be thoroughly washed out with clean water in order to remove

all the caustic potash solution . The burette itself should fi of be lled and emptied three times , the simplest method fil n ling being to immerse it , in a slanting position in the zi c fi trough , after the latter has been emptied and re lled with i clean water . The burette must be well drained before us ng fl for another test of the ue gases . By using the tabloids of pyrogallic acid which are now r prepared and sold for photographic pu poses , the use of the f r H onigman burette can be extended to tests o oxygen . These tests are made by pushing the tabloid of pyrogallic on e e of acid , and additional piec caustic potash , into the r a burette from below , after abso ption of the c rbonic acid

as . gas . The remainder of the test is then carried out before

ESTS W IT T E PROV E D OR OF RS AT PPARATU S O . II . T H H IM F M A

When the percentage of oxygen and of carbonic oxide are t h of Orsat required in addition to at carbonic acid gas , the 6 apparatus shown in Figs . 3 and 37 is the most convenient r for general use . The w iter does not recommend that this apparatus should be carried to the boilers and the sample of gas drawn directly into it from the ﬂues by use of the

G. ﬁnger pump For snap samples and tests of this kind , the

H onigman burette is the most useful apparatus . The Orsat I 1 8

AND GAS FUEL , WATER ANALYSIS

A L ﬁ drawn into . The cock at is now closed , and the nal adj ustment is made by raising the bottle F cautiously and as K allowing the excess of g to escape by . It is advisable to allow the gas sample to remain under pressure in the gas burette A for two minutes before commencing the test . This gives time for the last traces of moisture and sulphuric as acid g to be removed from the sample , and prevents error

— ’ - . . F O RS AT S S S S FIG 37 IMPROVED FORM O GA TE TING APPARATU . a e C . An from these c us s in the 0 2 test y leakage at the taps or j oints can also be discovered by this means before the test is made . Orsat aa The pp ratus comprises a measuring burette A ,

1 0 0 . . a holding c c , and gradu ted in tenths of a c . c . for the - ﬁ 2 . . T B rst 5 c c of its contents a glass piece , with three a - a C D ﬁl gl ss stop cocks two bsorption pipettes , , , led with caustic potash and pyrogallic acid solution respectively ; 1 2 0 WASTE GASES

’ and fi E a modi ed . form of Winkler s combustion pipette and e for determining the carbonic oxide , methane , oth r c a a - F and ombustible gases . The w ter spirator bottle , fin er- G the rubbaer g pump , are the remaining accessories of the app ratus . The measuring burette A is enclosed in a water-j acket to preserve the gas from temperature changes during the C D test , and the absorption pipettes and should have caoutchouc balloons fixed on their free limbs to preserve the contents from contact with the oustid e air . The pipettes C D fi a and are lled with n rrow glass tubes , in order to pro V ide a large surface moistened with the absorbing reagent as and a when the g is passed into them , the caustic pot sh or a a pyrog llic acid solution isadiaspl ced . The method of using this pp ratus is simple in principle , o but details of manipulation , which can only be understo d a u by actual pr ctice with it , req ire to be attended to in order

o a 1 . t a A 0 0 c c . obt in ccurate results . fter of the gas have A been carefully measured ina, and time has been given for this to assume the temper ture of the surrounding water , note being taken of any alteration in volume during this A C interval , connexion is made between and , and the gas is made to travel backwards and forwards three times by F alternately raising and lowering the reservoir bottle . The C ﬁ liquid in is nally drawn up to the mark , the cock is closed ,

1 0 . and the reduction in volume of the 0 c . c represents the per cent . of carbon dioxide gas in the gas sample . With CO fresh caustic potash solution the absorption in 2 is rapid ; of with a solution containing 30 or more per cent . carbonate

C is . the absorption of O. slow It is necessary , therefore , a k a to repe t the absorption procedure , and to ma e two me sure as b e all ments of the residual g , in order to certain thata the C0 2 is removed . The caustic potash solution must lso be CO renewed , when its absorption power for . has become too Slow for quick work . a For the determination of the oxygen , the s me process a D a an is repe ted in the absorption pipette , which cont ins H a alkaline solution of pyrogallic acid . ere again at le st two absorptions and measurements of the residual gas must a r b be m de , in order to be ce tain that all the oxygen has een removed by the pyrogallic acid solution . The volume of the residual gas should now be about 8 1

c . c . If more than this , there is reason to suspect the pres 1 2 1 AND GAS FUEL , WATER ANALYSIS

as or ence of carbonic oxide g , of unconsumed hydrogen and hydrocarbon gases of the methane or ethylene type . The amount of these gases is determined by burning in the E combustion pipette with an excess of oxygen . For this f 0 . . o l purpose 5 c c only the residua gas are used , and are

0 . . . K mixed with 5 c c of the outside air drawn in by , care being taken that these gas measurements are correct . This air provides the oxygen required for the further combustion of the carbonic oxide , hydrogen and hydrocarbons . The is and as total volume of gas carefully measured , the g is w E dra n over into the combustion pipette , the platinum coil meanwhile being kept at a bright red heat , by means of an electric current of two amperes at 1 0 volts from any con ni As ve ent source of electric energy . the combustible gases pass over this coil they are burnt to CO. and H 20 resp ec i l w t ve y . The gas is now dra n back into the measuring u re burette , and the diminution in vol me is noticed and

. ﬁ corded If no diminution occurs , it signi es that no car bo i n c . oxide , hydrogen or hydrocarbons have been present a to Should any diminution have occurred , it is necess ry distinguish between that due to carbonic oxide , and that d u e to hydrogen or hydrocarbons . For this purpose the residual gas is ﬁrst passed over into the absorption pipette C , and the carbon dioxide gas produced by the combustion of the carbon monoxide is determined as before by the loss of As of volume . one unit volume carbon monoxideapro duces the same volume of carbon dioxide , the decre se in

u CO a C b - x vol me due to . bsorption in , multiplied y ( being 5% the volume of the original gases left after aborp tion of the CO. and 0 ) represents directly the percentage of carbon monoxide in the original gas sample . The percentage of hydrogen is now found by calculation , according to the following method

Contraction on combustion 2 3 c . c . Ca CO 1 6 . rbon monoxide found by . absorption c c . L 6 — ° 8 . 1 6 o c c . or oxygen corresponding to c . c . carbon 2 monoxide . 8 1 5ac . c . or contraction corresponding to the unknown percent ge of hydrogen and hydrocarbons . B ut this contraction was due to the union of two volumes 1 2 2

AND GAS A FUEL , WATER NALYSIS

as is p ared above , and when cold poured upon the weighed am l ount of pyrogallic acid contained in a sma l ﬂask . This must be kept corked while solution occurs , and the solution must then be quickly transferred to the second absorption p ipette . The solution of pyrogallic acid in caustic potash is of a d as be ark brown tint , and deepens rapidly in colour , it c omes further oxidized . It absorbs oxygen rapidly when Orsat warm , but only slowly in the cold , and the apparatus s hould therefore be kept in a room rather above the normal temperature . The Orsat apparatus will only yield correct results when k - ept clean and in good order , with all the stop cocks and At j oints well attended to . monthly intervals the glass s top-cocks ought to be removed from their sockets and thoroughly cleaned with soft paper . A thin coating of re ﬁ or u ned lard , other suitable l bricant , may be employed n to make them turn easily after cleaning . A y mishap with is s . C the p gr . caustic potash solution in , whereby it

' w -c dra n up past the stop ock , must be followed by the a a of remov l and w shing this . Inattention to this precaution may result in much delay and expense , owing to the stop c in a ock sticking fast its se t , as a result of the wetting with c a austic pot sh solution .

M . ESTS FOR OISTURE AN OOT III T D S .

Owing to the high specific heat of aqueous vapour ( see p . the percentage of moisture present in the waste gases from boilers is of more importance than is generally r ecognized , and it is unfortunate that so little accurate information is available relating to this constituent . This absence of accurate data is partly accounted for by the d ifficulty of obtaining correct results in the tests made for moisture . Condensation of the moisture occurs upon the inner walls of the sampling tubes before the weighed absorp tion apparatus is reached , and the soot and dust present in the chimney gases also add to the difficulty of the deter mination . The following method is as satisfactory as any that can be devised , the soot and dust being determined at the same

8 . t a ﬁl . ime by actu l weighing in the ter tube , shown in Fig 3 1 2 4 WASTE GASES

as w is s The gl s tube Sho n of hard potash glas , 9 inches in i length , and about 11 nch in internal diameter . The lower end is rounded off and is provided with an 1 Opening 1 ; inch in diameter . The upper end is pro vid ed w ith a tubulus and ring , by the aid of which it

ﬁ 7 is rmly attached to a screwed 2 inch wrought iron pipe , ﬁr by means of a cap , which holds the glass ring mly against

the end of the iron pipe . Asbestos packing may be used is fi to render this j oint gas tight . The glass tube lled through fi the lower Opening with long and silky asbestos bre , ex all fine cluding dust , and after heating carefully over a

8 FIG . L . . 3 FI TER

B its are unsen burner flame , the tube and contents allowed as r to cool , and are weighed accu ately as possible on a - balance or scales reading to 1 1 0 0 th of a gram . The filter tube is now attached by the screwed cap to the fi wrought iron carrier , and this is xed in the sampling hole of the flue in such a position that the lower opening of the filter tube is about the centre of the flue and of the passing a th . e g ses From the increase in the weight of this tube ,

can . C of amount of soot and dust be calculated are must , l of course , be taken to c ean the exterior the tube thoroughly before the second weighing . The projecting end of the iron carrier tube is attached 1 25 AND GAS FUEL , WATER ANALYSIS

by thick r ubber tubing to the moisture absorption apparatus - w . . sho n in Fig 39 The wash bottle contains 1 50 c . c . of c old water and serves to condense the main portion of the

moisture , while the two calcium chloride tubes retain the

— FIG . 39 . MOISTURE ABSORPTION APPARATUS . remaining portions of moisture present in the waste gases The complete apparatus is weighed before and after making as as ll the test , with much accuracy the scales a ow , and from the gain in weight the moisture present per cubic foot f o waste gas is calculated . If any condensation of moisture o ccurs between the soot ﬁlter and the water absorption

t o h ot . ﬁ lter and ﬂ ue

8 0011011 30:

— - S FO R T S L S 0 KI . FIG . 4 . WATER A PIRATOR A NG AMP E

i a . n pparatus , both tests will be spo led For this reaso it is necessary to have the wrought iron pipe carrying the ﬁlter h entirely sunk in the ﬂue , and to use only a very short lengt

o f rubber tube for connecting it to the second apparatus . The aspiration of gas through the smoke ﬁlter and water 1 2 6

CHAPTER XI .

Th e Use of C onti nu ous and R e c ordi ng G as

Te t n A aratu s i g pp s .

LARGE number of automatic and recording gas testing instruments of various types have been and - devised are now being sold to steam users , for checking the work of the ﬁremen engaged upon the boiler plant .

These instruments are costly to instal , and require skilled to a attention maint in them in satisfactory working order . When placed in the charge of men who will give the necessary time and trouble to the work , they yield in most cases , however , satisfactory results . The tests obtained with these automatic and recording - gas testing instruments can only be regarded as supple mentary to the gas—testing methods described in the pre vious chapter , for neither hydrogen nor carbonic oxide are a gas recorded by the pparatus . The heat losses due to the presence of these two constituents in the waste gases so from the boilers are , however , great that no method of control which ignores their possible presence can be con id ered a s s tisfactory . This matter will be further dealt with in Chapter XII . Even the carbon dioxide test is liable to be incorrect , and independent check tests once a week or Orsat oftener with an apparatus are necessary , to ascertain the margin of error with the automatic apparatus . The automatic gas-testing instruments may be divided into two classes— those depending for their action upon carbon dioxide absorption in caustic potash , and those based upon the observation of some physical characteristic of the a w a fi fi w ste gases , hich v ries in de nite ratio and within xed m f the li its , with di ferent percentages of the same gas . To first of these classes belong the Ados and Uehling ; to the K ell A n e r r dt H abe . O second the , and r instruments f thes various types the Ados and the K rell are the most widely U K used in the nited ingdom . 1 28 WASTE GASES

I. THE CONNECTIONS B ETWEEN THE TE STING I NSTRUMENTS AN T E UES D H FL .

As generally arranged , one automatic instrument with a recording attachment is installed in a closed cabin or case in - the boiler house , and connection from this instrument to the side fines of each boiler of the installation is made by one of t long main i inch wrought iron pipe , with branch connec E tions of the same diameter . ach branch pipe to the boiler A fil fi or fl ue is provided with a cock . ter lled with glass wood wool is employed on the main suction pipe to retain dirt and soot . B y means of a fan or pump , gas is drawn continuously from the flues through these connecting pipes to the testing appa ratus , and by Opening only the cock leading to the flue of the as boiler or boilers under test , the g sample can be varied

as required . M aany obj ections , all bearing upon the accuracy of the results obt ined , may be urged against this method of obtaining

the sample . The distance of the testing apparatus from the boiler ﬂues and the presence of a very large number of cocks ﬁl and soot ters on the one line of pipes , may cause air leak

CO . age , and consequently incorrect 2 tests The use of iron to as absor pipe leads corrosion by the chimney g es , and to p

tion of some portion of the constituents . In some cases so serious is this corrosion and formation of salt deposits within fﬁ the pipes , that great di culty is experienced in keeping these

open for the passage of the gas sample . The large volume of stale gas contained within the long length of sampling pipe , ﬁl and in the soot ters , is another cause of incorrect results , since the speed of the fan or pumps is not sufﬁciently high

to clear this gas out in reasonable time . Six times the volume of gas must be drawn through these pipes and ﬁlters before

all the stale gas can be considered to have been removed . The following recommendations are therefore made for connecting the automatic recording apparatus of any type to the boiler ﬂues and only when these have been strictly

followed can confidence be placed in the test results . 1 1 A water jet pump of the type shown on p . 2 should be flu employed for drawing the gases from the boiler es . This pump should be fed from a water tank placed 20 or more l feet above the ground , fed from the genera water supply ll fil r by a ba tap , and provided with a tering a rangement at 1 29 K U AND GAS F EL, WATER ANALYSIS

- - the run off pipe which carries the water to the water jet

pump . This plan ensures a constant head of clean water r all and constant work of the pump at a unifo m rate , with freedom from stoppage d ue to the collection of grit or dirt

in the water jet . The connecting pipe to the ﬂues should either be of iron ix or s . enamelled inside , of feet lengths of glass tube If - glass be used the pipes must be placed over head , and not

laid along the boilers , and a T piece provided with a glass - fl stop cock must be used for each branch to the boiler ues . The vertical branch into the flue must be of hard potash 4 All al . glass , } inch extern diameter connections of glass tube must be made w ith the best red rubber tubing and fastened with wire ; and once a week the tightness of the system must be tested by closing all the cocks and working - 20 the water j et pump , until a inch vacuum is registered 1 by the pressure gauge attached to the system near the pump . The dependent glass tubes into the flues must be fixed in C flanged guard pipes by cork bungs , as described in hapter w IX , and once a week they must be withdra n and cleaned , ll both interna y and externally , from the adherent soot . The main lengths of enamelled iron or glass tubing must

be set with a fall away from the testing instrument , and these

should also be cleaned monthly , or oftener if required by

blowing steam through them . The vessel or vessels which act as soot-ﬁlters must be - placed close to the water j et pump , and must be designed

of ll n . without any water seal , and sma dime sions They

Should be cleaned weekly or oftener if required . is one n Finally , it advisable to have automatic i strument l instal ed for each six boilers , and to place the cabin or cup in a of board a position central to this b ttery boilers , in order that the connecting pipes to the furthest boiler of A the system may not exceed 30 feet in length . connection

should also be made to the nearest point in the main flue , A r leading to the chimney . separate appa atus should be all at inst ed for testing the exit gases the chimney , and from the flue midway between the economizers and the B chimney . y this plan of arranging the testing system , a battery of 24 boilers would be provided with five testing instruments , and continuous tests of the gases from any

1 - A 2 9 inch vacuum can be obtaine d with these wate r je t air m e afﬁ the e e e ae aaae . pu ps, wh n su ci nt pr ssur of w t r is v il bl in supply 1 3 0

AND GAS FUEL, WATER ANALYSIS

1 tw o cylindrical vessels shown on the right han d of Fig . 4 contain samples of the flue-gases draw n by suction from the fi i fl ues . While one of these is ll ng , the other is delivering a sample of the gas into the measuring and absorption 1 00 pipettes of the apparatus . c . c . of the gas are used for CO ab each test , and the percentage of . is determined by

— W F IG 2 AD os C S NE . . 4 . RE ORDING APPARATU ( FORM )

sorption with caustic potash solution , exactly as in the two forms of testing apparatus already described and illus trad - te . The absorption takes place in the disc like glass 1 vessel shown on the left in Fig . 4 , in order to obtain a large surface of KOH solution in contact with the exit gases .

. The volume of the gas remaining after the absorption of the £30 2 is recorded by means of a ﬂoat and lever upon paper 1 3 2 WASTE GASES a carried on a revolving drum , vertical line being marked

by the pen attached to the lever for each separate test .

h i . as 1 0 . . e . W en the volume of remaining g is 0 c c ( . when no C0 l 2 is present) , this line extends the ful width of the lined 1 0 C r . O paper ; when per cent of , is present , the ve tical 1 line extends only up to line 0 . The clear space above C0 the line thus measuresathe percentage of 2 present in the ﬂue gases at any p rticular time during the day , and as ll the frequency of testing can be varied at wi , the average ’ of the day s work can be ascertained by means of this ll apparatus . It is customary to arrange the contro ing mechanism so that samples of gas are draw n from the ﬂues

and tested every ten minutes . The Ados apparatus KOH requires careful adj ustment and control , and the solution must be renewed at intervals depending upon the

— . . Y C L C S F TH E D S TUS FIG 4 3 T PI A RE ORD O A O APPARA .

frequency of the tests . Fig . 43 shows a record made with

this instrument over a period of 1 2 hours . 6 The in as i — ( ) Uehl g G Compos meter. This instrument A an d is an merican invention , is not much known in u this country , altho gh a pyrometer based on the same principle is widely used in iron works in America and U K om simeter w in the nited ingdom . The gas c p o is sho n

in Fig . 44 . Its working is based upon the law which ﬂ w A governs the o of gases through small apertures . con stant suction is applied by means of a steam jet at one end of a gas tube provided with an inlet smaller than its out and let , connected to the exit ﬂue . In this tube the chim ney gases are submitted to the action of a caustic potash s s solution . The variations in pressure of the ga e within this tube give the measure of the absorption which is occur T i CO ese . he r ng , and therefore of the amount of . pr nt 1 3 3 AND GA FUEL , WATER S ANALYSIS exact measurements of pressure variations by means of a mercury column is consequently all that is required to

i CO . instu e y eld the percentage of , This m nt can be pro vided i and with a record ng pen drum , in order to obtain c n n s CO o ti uou records of the , present in the waste gas es . Since the Uehling apparatus requires a caustic potash

— . . HL C S FIG 44 UE ING RE ORDING APPARATU .

n solution , and takes no accou t of the carbonic oxide present ,

it is open to the objections already mentioned . ’ ( c) Arndt s E conometer or Gas B alance — The two instru ments described above depend for their action upon the absorption of carbon dioxide gas by solutions of caustic

potash . The three remaining recording instruments depend for their action upon the variation in some physical property 1 3 4

AN GAS FUEL , WATER D ANALYSIS

in A mittent , as the dos apparatus . The possible pres in ence of carbonic oxide the exit gases is , however , ignored , all r as in other fo ms of automatic recording apparatus . Since this gas is lighter than air ( 1 litre weighing only grams) its presence in any quantity will render determina tions of the CO. from the readings of the Arndt apparatus incorrect . e Re d — d The r ll cor er. a ( ) K This appar tus is shown in Figs . 6 l . a A 4 and 47 In principle it is lied to the rndt gas balance , since its action depends upon the relative weights of the

as . f flue g es and of air In this case , however , the di ference f is measured by a delicate di ferential gauge , attached to two one tubes of equal height , containing air and the other the fl sample of gas from the ues . The pressure gauge readings are taken by means of a slightly inclined glass tube filled with coloured alcohol , which proj ects from the pressure gauge of box . The position the end of this thread of coloured and alcohol varies with the pressure in the gauge box , by means of a scale marked in percentages of CO. correspond r ing to ce tain pressures , the percentage of this gas present in the flue gases at any given moment can be read off directly on the scale . This instrument is usually provided with accessories by which the position of the thread of coloured alcohol is continuously photographed on a record ing cylinder and paper . A continuous record of the per

0 . centage of C 2 present in the waste gases is thus obtained ﬁ The defects of this apparatus are rst , that no account is kept of the carbonic oxide which may be present in the s exit gases this gas , if present , will render all the reading CO for . incorrect , for the reasons already given when dis i A . s cussing the rndt instrument The second defect , that

the necessity for developing the records photographically , adds considerably to the trouble and expense of keeping

the apparatus in working order . The Krell apparatus is now installed at several electric light stations in this country . - — in e The H abe L ew 0 Testin I ns ument. ( ) r o e C 2 g tr This strument differs from all others in that the variations in the refractive index of the waste gases are made use of to determine the percentage of C0 , by direct observation . is one The method therefore an Optical , and resembles in some respects the optical methods for determining tempera

ture . 1 36 WASTE GASES The instrument consists of the following essential parts a prism , a telescope and a microscope . The gas to be tested ﬂows through prism which

6 — G S . AND K LL C S . FI 4 4 7 . RE RE ORDING APPARATU

hollow , is made of varnished metal , and is provided with T windows for the incident and refracted rays of light . he gas used for comparison ﬂows through a casing . When observations are made with the telescope and microscope I S 7 U AND GAS F EL , WATER ANALYSIS

the ﬁ in position , observer sees in the eld of vision a scale which is partly light and partly dark . In pure air the ° boundary of the dark zone would coincide with 0 on the

° scale in air containing 9 per cent . C0 2 this boundary ° would coincide with 1 ° o on the scale and in air containing 0 of 9 per cent . C0 2 the boundary of the dark zone would be 1 0 0 at on the scale . The instrument can be graduated so as to yield the per C0 in centage of 2 the waste gases by direct observation , air being generally used as the basis of the comparison .

If desired , the instrument can be made to record the read

as K . ings by a photographic attachment , in the rell recorder

FIG 8 — H -L G AS - S S . 4 . ABER OEWE TE TING IN TRUMENT .

The refractive indices upon which the instrument is based are as follows

H y droge n Aqueous vapour Oxy ge n Dry air Nitroge n Carbon monoxide Methane Carbon dioxide Acety lene S ulphur dioxide E thyle ne

The method is only applicable with success to gas mixtures c ontaining two gases whose indices of refraction differ by ﬁ considerable amounts , and the gures given above Show that the presence of hydrogen , aqueous vapour , and sul 1 38

CHAPTER XII .

The P rac tical Applicati ons of the G as-Te st e u t R s l s .

I . CALC ULATION OF THE TOTAL HEAT LOSSES DUE To THE VO U E AN PECI FIC EAT OF T E XIT ASES L M D S H H E G .

N order to be able to calculate the heat lost daily with the is waste gases from any boiler plant , it necessary to obtain ﬁ CO gures Showing the mean temperature , and the mean , CO 1 1 2 and contents of these gases for the or 24 hours . These returns may be based either upon the average of a - stated number of snap tests , or upon the results obtained of with one the automatic recording instruments . As C already pointed out in hapter XI , the automatic gas CO testing apparatus only records percentages of 2 , and a of is is neces where any app ratus this kind installed , it still sary to make occasional tests of the gases with the modiﬁed Orsat e apparatus , in order to be certain that no carbonic oxid is h present in the waste gases . This precaution is all t e more necessary since the tendency where automatic appara has is too e tus been installed , to work the furnaces with littl rather than with too great an excess of air , in order to obtain carbonic acid gas averages above 1 0 per cent . A simple nt calculation will show that it is better to have 1 0 per ce . C0 and no C0 in the exi as n 1 c n C0 and t es tha er e t. 2 g , 4 p 2

2 e n C0 . p r ce t. This is a fact which is not sufficiently recog nized by boiler engineers who have installed automatic apparatus , and trust to the records of this alone in control of fi ling the working their boiler plant . The modi ed form of Orsat ri C ou ht therefore apparatus desc bed in hapter X g , to be installed in any works using over fifty tons of fuel per week . Tests of the waste gases Should be made with as this apparatus daily , both a check upon the automatic recording apparatus , and as a guide to the economic work in g of the whole plant .

1 O the c e i a f ar di xide and O a C , is h m c l symbol or c bon o , C th t for a i c rbon monox de . 1 4 0 WASTE GASES

i A r r 2 1 . contains , in round numbe s , per cent by volume f o oxygen and 79 per cent . of nitrogen . When this oxygen m is employed for burning solid carbon , it yields the sa e

ab . volume of c r on dioxide gas If, therefore , the whole of the oxygen of the air could be used up in burning fuel on - the grates of steam boiler furnaces , and if this fuel con f as h ained nothing beyond carbon and , a maximum per

c entage of 2 1 per cent . C0 . by volume in the waste gases

would be attainable . Two facts render this maximum impossible of attain ﬁ ment . In the rst place , all solid and liquid fuels contain not hydrogen , and the oxygen likewise contained in fuel is suﬂi cient to combine with this hydrogen , and to yield with

it aqueous vapour . Therefore some portion of the oxygen of the air is required to burn the excess hydrogen contained

in the fuel .

Further , under actual conditions of work , it is impossible to burn solid fuel properly with only the theoretical amount

An 0 . of oxygen . excess varying from 5 per cent to 1 0 0 per

cent . above this theoretical amount is required to obtain r pe fect combustion of the fuel . The hydrogen present varies considerably in different as classes of solid fuel in bituminous fuels it ranges , a rule , 6 B i from 3 to per cent . y calc ulat ng the weight of oxygen e and required to burn the hydrogen present in the fu l , by

deducting from this the oxygen contained in the fuel , one D obtains the oxygen required from the air . educting the of 2 0 8 ercentae equivalent volume this from , we have the p g b volume o en ai b e o c mbustion o the cabon y f oxyg valal f r o f r . This ﬁgure for most bituminous fuels is about 1 9 2 per

cent . , and this therefore represents the maximum per cent ae of CO g . which could be obtained by perfect combustion ,

with the theoretical amount of air . Now supposing we have an average test for the day of the waste gases of 1 0 per °

. C f 6 0 o . cent 2 , and an average temperature 4 5 F at the ﬁ to base of the chimney , how are these gures be used to B i on 8 calculate the ritish Th . Un ts given p . 5 as of P the thermal loss , under these conditions work The calculation is as follows - Ten per cent . CO, equals an air consumption of I O As 1 or times that theoretically necessary . 2 lbs . o f al 1 air are theoretic ly required for the combustion of lb . 1 lb of 2 x 2 s . of fuel , this excess air rep resents 3 1 4 1 U AND GAS I F EL , WATER ANALYS S

m lb u 2 lbs . e er . of air consu ed p of f el , yielding 4 of wast s ga es . ﬁ be 0 2 0 The speci c heat of these gases may taken as 4 .

The loss of heat with the waste gases per lb . of fuel burnt is therefore 1 B 24 x x 400 ritish Th . Units . C0 the The calculation for any other percentage of 2 in ex m it gases is similarly made , the maximu percentage of C 1 2 for ﬁ O. ( 9 ) being used as a means nding the excess of CO air represented by any given percentage of ,

ALCU ATI ON OF T E EAT OSSES DUE TO I PERFECT II . C L H H L M O B T O C M US I N .

The gases d ue to the imperfect combustion of fuel are s carbon monoxide , free hydrogen , and certain hydrocarbon of the methane and ethylene series . Since the heat pro d uced by the combustion of fuel is dependent upon the com plete oxidation of the carbon and hydrogen to carbon dioxide gas and aqueous vapour respectively , it is evident that incomplete oxidation must be accompanied by loss of heat , and that the amount of this loss will depend upon the of r car volume the unbu ned hydrogen , hydrocarbons ; and b ni The o c oxide gases . following methods can be applied for calculating these losses from the results of the app roxi mate analysis of the w aste gases by the improved Orsat apparatus described in Chapter X ; the percentage of in of carbon the fuel and percentage by volume hydrogen , carbonic oxide , and carbon dioxide in the waste gases ﬁ u f r al being the test g res required o the c culation . (a) Losses Due to Carbonic Oxide — Fuel contained 80

. 2 0 per cent carbon , waste gases contained per cent . car

bonic 8 0 . . e oxide , with per cent of carbon dioxide Sinc carbonic oxide yields on combustion its own volume of al of carbon dioxide , the tot volume carbon dioxide , if the oxidation of the carbon had been complete , would have been 1 - - 2 0 8 0 or 0 0 per cent . Therefore 2 1 0 ths or 1 5th of i uiv the carbon had escaped complete oxidation . This s eq a

8 0 1 1 6 . lent to x /5 or lb of carbon . has Now it been proved by experiment that when I lb .

° 1 hah ir e athe It is assume d t t t e at mperature is 65 F. and th t ° f ea the exit a is 6 = n et loss o h t in g ses 4 5 65 400 F. 1 4 2

AND GAS I FUEL , WATER ANALYS S

1 1 0

’ 55

These tests prove conclusively that free hydrogen may e xist in the exit gases from boiler furnaces in the presence of b and considerable percentages of car on dioxide , that the older V iew that the hydrogen must be ﬁrst completely burnt to aqueous vapour before the combustion of the carbon i to occurred is incorrect . It is therefore always adv sable test for the presence of hydrogen in the exit gases from boilers worked under Close supervision as regards air excess , since with the escape of 1 per cent . of unconsumed hydrogen considerable heat losses may arise . The method of calculation is less simple than in the case of carbonic oxide , since in the absence of any test for the to aqueous vapour , one is unable state the loss as a pro portion of the total hydrogen present in the fuel . It is therefore necessary ﬁrst to calculate from the volume per of u centages the gas analysis , the weight of the vario s gases

s fi 1 0 8 . by aid of the p . gravity gures given on p . Taking the first sample given above by S od eau we have the following figures

en ta Perc ge by Eq uivalent by

ol . v ume w eight.

Carbon dioxide Carbon monoxide H y droge n Oxyge n Nitrogen

- - 1 0 0 0 0 1 0 2 58 5

The ﬁgure 0 45 in 1 0 2 5 85 of waste gases is equivalent to

0 . 439 per cent hydrogen by weight . B y application of the method described in Section I of this Chapter we ﬁnd that a percentage by volume in these 0 C . 0 C . O gases of 9 per cent 2 and per cent , corre s ond s of of to p to an air supply lbs . air , and a produc b l s . e tion of of wast gases per pound of fuel burnt . The weight of hydrogen which is escaping combustion per pound

1 44 WASTE GASES

8 1 or 00 . ow . as 943 lb N lb of hydrogen g , when burned H 0 B ri T c h . U completely to 2 , produ es tish nits of heat 1 and the loss of heat in the particular case named ° 0 0 8 x 86 B above is therefore 945 or 5 ritish Th .

U . of nits , equivalent to 4 per cent the total heat value of Al h the fuel . though t e percentage of free hydrogen in s is the wa te gases by volume may be small , it thus seen to be of considerable importance in its effect upon the heat its l losses , and presence ought a ways to be made the subject of special tests . 0 H eat Losses Due to Unburned H drocarbon Gases ( ) y . n w il Si ce the hydrocarbon gases , methane , ethylene , etc . , l have been burned to carbon dioxide and aqueous vapour resp ec tivel C y , in the combustion pipette described in hapter X , and have been returned In the test as carbon monoxide and to cal free hydrogen , it is unnecessary make any separate culations to the of for the heat losses due presence these gases . ﬁ The error in the nal result is negligible , for both elements exist in combination in the gas eous state in the hydro carbon gases , and the thermal energy required to render them free is probably small . The thermal values given for the combustion of free hydrogen and of the gaseous atom of carbon in carbon monoxide are therefore applicable and the calculations given under aand 6 include the losses due to the possible percentage of these hydrocarbon gases .

A C U ATION OF E EAT O E DUE To AIR III . C L L TH H L SS S EAKA E L G . — O (a) At the Dai ers . Tests of the C . present in snap samples of the exit gases drawn simultaneously from the sampling hole j ust within the damper (on the furnace side of it) and from the nearest sampling hole on the chimney

of . side it , at a distance of at least two yards , are made If the tests differ by less than half of one per cent . the air leak age may be regarded as slight . ’ ll f In some cases , however , it wi be found that a di ference 2 0 the of of or more per cent . exists between tests the two samples . The excess of air and heat loss represented by this difference is then calculated by the method given under

1 This assumes that the latent he at of the steam produce d is re If the h aaa as a covered . water passes away from t e pp r tus v pour ° i i s . h s e is e . Th. U at 2 1 2 F. t ﬁgur r duced to 5 B n t 1 45 AND GAS S I FUEL , WATER ANALY S

is S ection I of this Chapter . It advisable when making use of s e air at the ers thi m thod to check the leakage damp , es fi e are rn an d to make the t t when the boiler r s bu ed red , to see that no interference with the fires occurs during the B lin samplin g Operation . oth samp g tubes should be in serted an equal distance into the fines an d it is of cours e ne cessary to take the second sample of gas from the flue at a point where the exit gases from other boilers have no a to t had a ch nce to mix with it , and change its com

position . Air leakage at dampers and through the ﬁnes of incom pletely damp ered off boilers is a proliﬁc cause of bad draught

in most boiler plants , and attention to this defect is urgently n ﬂues needed . Since the cold air drawn i to the in this way does not pass through the furnace of the boiler it may be

argued that the heat loss is negligible . In most modern an boiler pl ts , however , economizers or feed water heaters are employed for extracting some portion of the heat from

the waste gases and the dilution with cold air , drawn in

through the dampers of the boilers , diminishes greatly the fﬁ e ciency of this accessory plant . The use of gas testing

apparatus for discovering and checking this leakage is ,

of n . therefore , considerable importa ce 6 Throu h the B rickw ork o the B oiler S ettin E cono ( ) g / g, mizer and Flues — The method of making the test is imu neous described under (a) . S lta snap samples are drawn f of from di ferent sections the ﬂue , and are tested for the per

centage of CO.. The gradual reduction in the percentage C0 a the of 3 , as the s mpling approaches the chimney , is s measure of the air leakage . The calculation of the air exces of ribe and the heat loss , is carried out as already desc d in

detail under S ection I of this Chapter . In nearly every case to u where special attention has been devoted this s bject , considerable air leakage through the brickwork of boilers and ﬂues has been fo und and the casing in of the rough in has setting of these , glazed bricks or in sheet iron , been

found to result in considerable economy . It must be pointed out here that leakage through the brickwork setting of a i or s of bo ler , through the economizer wall is a direct source n ﬂ ues s heat loss , si ce the cold air which leaks into the in thi way is raised to the boiler or economizer temperature by n n heat abstraction from the boiler plates ad eco omizer tubes . r r is The ordina y direction of heat t ansfer therefore reversed , 1 46

AND GAS FUEL , WATER ANALYSIS quickest method of ascertaining and controlling this factor in good boiler working .

D E V . CALCULATION OF THE HEAT LOSSES U To AQUEOU S V APOUR .

As on . 1 2 pointed out p 4 , the amount of aqueous vapour in the waste gases from steam boiler plants is rarely to d etermined , and the heat losses due this constituent fi of the nal exit gases are therefore largely unknown . The r following calculations however prove their impo tance , and the necessity for keeping the moisture in the coal as as as n low as possible , well the advantage of usi g fuels con taining a low percentage of hydrogen . The superior steam raising efficiency of South Wales steam coal is no doubt due to its low percentage of hydrogen . s C Taking the coal , the analysi of which was given in hap find w as ter I , we that there per cent . moisture and

. as per cent hydrogen present in the fuel delivered . Since hydrogen yields nine times its own weight of

1 00 . aqueous vapour on combustion , every lbs of this coal x or f would produce 9 ) lbs . o aqueous

1 . cor vapour in the furnace gases , and lb of the coal would

n . resp ondi gly produce 449 lb of aqueous vapour . This to aqueous vapour had , however , all be raised from the tem p erature of the external air to the temperature of the exit to gases . In addition to the loss of heat due the high ﬁ speci c heat of aqueous vapour , there was also that due to the fact that the water passing off as aqueous vapour from the furnace of the boiler had ﬁrst to be raised to ° 6 . 6 B . U 2 1 2 F . and then vaporized Now 9 ritish Th nits

of 1 . heat disappear in the vaporization of lb of water , n Th and this value is k own as the latent heat of steam . e calculation for the heat lost with the moisture present in these gases is therefore made as follows : the temperature 66 ° of F . the exit gases being taken as 5 , and the external ° air temperature at 65 F . °

H 6 . 1 . eat required to raise the moisture from 5 F to ° 6 66 B x 2 1 2 . U . 2 1 2 F . ( 5) ritish Th nits H to 2 . eat required convert this moisture into steam at ° 6 B 6 . 2 1 2 F x 9 433 ritish Th Units . 1 48 WASTE GASES

° H 2 1 2 . 3 . eat require to raise this steam rom F d f . to ° l U i . 665 F- X (665 2 1 2 ) x 5 40 1 0 9 B ritish Th . n ts The total heat loss due to the presence of this aqueous vapour in the exit gases is therefore

66 : U . x 433 x 1 0 9 60 8 B ritish Th . nits With the fuel under consideration this loss was equivalent 8 8 of . or to 4 per cent . the total heat value of the fuel 1 260 0 a an d The chief loss is due to the latent heat of the ste m , this heat cannot be recovered from the waste gases unless they ° 1 . be cooled below 2 2 F . before passing up the chimney S ince the eﬂiciency of chimney draught is dependent upon the ° temperature difference of at least 400 F . between the vapour

within the shaft and the air outside it , it will be seen that l until fan draught is more general , the loss of this heat is large y be unavoidable . It can , however , minimized by the use e of dry fuel , and the prevention of all unnec ssary steam i escapes into the furnaces of the boilers . The fuel tests g ven in the Appendix Show that great differences exist between I the moisture tests for various classes of coal , and that n

washed Slacks of bituminous character , the moisture may ﬁ rise to a very high gure . It is therefore questionable whether the freedom from ash obtained by washing is not purchased at too high a cost and it would be well to stipulate in buying such washed slacks or peas that the moisture shall not exceed 5 per

cent . in the fuel as delivered .

A C U ATION OF T E EAT B A ANCE . VI . C L L H H L The method of ascertaining the evaporative efﬁciency of the boilers by means of a heat balance was described In i of Chapter IV . It is unnecessary to repeat here the deta s l. -raIsm this method of calculating the efﬁciency of the steam g C plant . The second entry in the heat balance given in hap

ter IV may , however , be expanded to include all the heat

S C . losses discussed in ections II , IV and V of this hapter The heat balance is then set out as follows

1 The e ea ea i e i he e e a e and at sp ciﬁc h t of st m r s s w th t t mp r tur , ° - is 60 0 . 665 F . about F AND GAS UEL , WATER ANALYSIS

H EAT VALUE OF FUEL IN CALORIMETER .

B . Th. Units.

Losses aiai e tc . by r d t on , w aste ases Losses by g . (a) d ue to exce ss air ( b) d ue to unburne d carbonic oxide ( c) d ue to unburne d hydroge n and hydro carbons ( d ) d ue to moisture Loss es by unburne d carbon In Cinders and as h Losses by heat carrie d In Cinders and as h B alance heat converted in to steam

f al 1 - the The di ference between the tot of the items 4 , and heat value of the fuel as ascertained by the calorimeter the n s al m give the tot units of heat converted into steam , and fro this the evaporative efficiency of the boiler or boilers is . i as e the obtained , by express ng the result a p rcentage of r total heat available . It is impo tant that the heat valu e undried uel an d the of the f be used in these calculations , not d uel as ll r s . heat value of the y f , usua y returned by the chemi t al The former can be c culated when the latter is known . in r the a of u Fig . 49 shows g aphic form percent ge f el wasted by heat losses in the exit gas es at various tempera s A tures . It is ba ed upon a table given in the ppendix o f 1 the book named below . As k an illustration of the use of the diagram , we may ta e as 8 er . CO an d the case of waste g es containing p cent 2 , ° having a temperature of 60 0 F . At the point where the w lines representing these values intersect in the diagram , e - find fi u 2 of the g re 4 9 , which represents the percentage loss fuel under these conditions . If the percentage loss is not u marked for the partic lar case under consideration , the value 15 obtained by interpolation . Taking a cas e in which - as as 1 0 . C0 s the w te g es contain 5 per cent 2, and pa s away °

60 0 F . 1 0 c again at , the percentage fuel loss at per ent . ° nd 1 r - 0 a 2 e . CO C 600 . is 2 2 is 1 O 1 . , at F , at p cent . it 7 1 0 - e The loss at this temperature with 5 per c nt . CO, is I 7 ° I evrdentl 20 -2 0 1 1 - therefore y 9 4 per cent . 4 )

1 S moke P revention an d F uel E conom o an d ershaw y, by B oth K

ae . Const bl , 1 90 4 1 50

a n o m —an B 0 m n m 0 0 u o 0 m v 0 m w m z m au w u A w N N 5 9 2 ma E

N m m m v 0 e O e a 0 o 0 3 m 9 R vi i ~ m m 0 a s “ A n w § ﬁ § 6 m

0 0 m A 0 0 ” o « m m m M m 3 m0 0 m o e v m S 3 u 8 E m m m m m n 9 u o E m o 0 . . . . . 3 “ R “ “ £ “ d d x § ﬁ N d ﬁ 0

m O 0 o m c 0 0 m 0 m Q o m mm . . 9 h 9 . 9 m k 0 d w h e m m A u u m o : m m

0 o m m m 0 O m m m w v m w 9 m o m . 9 . . e . w w m». m m b 0 b n m b a u m u

o C N O h 3 u w E B . m m G b A “ u ” w m B u e o e 0 E — g . a fl 0 m B S w z w S S 3 o o o 5 w 0 o S m m n 5 w w 5 i c s 0 5 0 s o o E m 0 0 n o h w - 0 m m 0 2 2 ” a0 0 mm 8 B a v w 0 c m 0 u m u 0 m E m h 0 a S o 5 N a0 v 8 0 a5 5 c m2 h o a0 h 0 $ aa 0 w ao n 8 a 0 am 5 0 s 3 m m —u 2 5 -a0 w o 8 fi 3 k 0 0 n H m 0 a0 s 3 0 3 »? fl m z 2

1 56 GOUTAL’ S FORMULA FOR CALCULATION OF TH E CALOR IFIC ALUE OF TH E FUEL FROM TH E RESULTS OF V , TH E APPRO MATE ANALY S S XI I . Goutal in 1 9 0 2 publishe d 1 the formulagive n below for calculating the caloriﬁc value of the fuel from the results of the approximate aa the ae ae ae ae e e n ly sis, formul b ing b s d upon v ry l rg numb r of t sts F Th h e of various rench coals . e author as used this formulafor som ea an d has ae the e ae the a a y rs , comp r d r sults obt in d with it with ctu l

F IG 0 . . 5

th m caloriﬁc values obtained w ith the bomb calorimete r . In e a orit a the e e e are aan d the amabe j y of inst nces diff r nc s sm ll , formul y e aaae e e a the a ae w e us d with dv nt g for d t rmin tion of c loriﬁc v lu , h n ae e e b e . a ae w e e no dir ct t st is possi l It is mist k , ho v r, to pl c too eae ae the a ae ae an d all ae gr t r li nc upon c lcul t d v lu , in c s s of import ae the e e one the a e ai e e nc , dir ct t st with of pprov d forms of c lor m t r should be made . The formulais as follows P 8 2 aV C ,

1 — m Rendus . 1 Co ptes S ep t. 9 0 2 . 1 57 APPENDIX in which P is the caloriﬁc value of the fuel in Centigrade units C is the percentage of ﬁxed carbon V is the percentage of volatile mat ter ; and (a) is avariable de pe nding upon the percentage of hydro n carbons ad of hyd rogen in the actual combustible of the coal . Using the formula V x 1 00 C + V the following values are obtained for a 1 V . 1 0 . 1 . 20 . 2 . 0 . 5 5 5. 30 . 35. 38 4

a 1 . 1 . 1 1 . 0 1 0 . 1 0 . . . 80 . 4 5 3 7 9 3 9 8 . 9 4 8 5 Figure 50 is adiagram giving the curve based upon the above e and hi mabe e ad aae al ain the ﬁgur s, t s y us d with v nt g in c cul t g r e ae of af an ae f V 1 e w ee the i i s co r sponding v lu ( ) or y v lu o ( ) , b t n l m t

5 and 40 . The following are afew results selected to show the fairly close agreement between the calculated and observed the rmal values for d ifferent coals tested by the author (results in Ce ntigrade units )

TESTS OF WATER FROM AR OUS SOURCES V I .

Res ults expressed in parts per

Total Temp orary Permanent Total Descrip tIon . S olid s . H ard ness . H ardn ess. H ardness .

LIV ER P OOL COR P OR ATIO N

VVATE R . H Lake Vyrnwy P Rivington e D ud low Lane e Green Lane mWindsor o Waterloo District

D S VARIOUS I TRICTS . > S pring water 0 9 S urface water o S urface water HmF rom aB oiler in need of w ashing out

T NO ES . Lake Vyrnwy is in North Wale s Rivington is acatchment area near Bolton . S a e and e e the i a the mpl s 3 . 4 5 w r from old pump ng st tions of

L e a and are a e e aer . S a e 6 iv rpool Corpor tion , s mpl s of w ll w t s mpl the a V Ri and e ae e is combin tion of yrnwy , vington w ll w t r suppli d n i n e e t to Liverpool ad district for domestic uses . It s a xc p ionally - S . 8 s pure water . ample No 9 contained 8 part of organic matte r - h p er an d gave an acid reaction with phenol phthalein . T e a e e 1 0 -8 ar NaCO er ae cidity r quir d 47 p ts . . p of w t r, to

ne utralize it . n S ample NO . 1 0 contained 4 3 2 parts of orgaic matter p er ae and e the ae i e of w t r, show d d ng r of running bo l rs too long without ff e eai hi th e w . blowing o . In t s case e ﬁttings w r al ays l k ng

1 60 TY P CAL TESTS OF E T ASES I XI G .

P er cent. CO , P er cent. CO

Carbon d ioxid e . ( arbonic oxid e ( ) C ) .

NOTES .

All the above tests were made upon samples drawn from the fl ues le e a aw the a e ei ae of boi rs und r ctu l orking conditions , s mpl s b ng t k n fl from the side ues on the furnace side of the dampers . Tests 1 to 4 show low CO . percentage s and excess oxy gen whil e tests 8 an d 9 indicate that eve n with acomparatively large excess of oxy gen it is possible to produce carbonic oxide under bad con e e e the fire w as d itions of firing . In th s two cas s it is possible that ae an d hr h in e and athe air ex e s thick in pl c s burnt t oug oth rs, th t c s w as rush mg in at thes e thin and bare places in the grate .

Te s 6 and w ai d O e e ae but N . h st 5, 7 sho f rly goo C . p rc nt g s, o 5, wit - er e a o e i a the ae ae i i 2 0 0 . es ts p c nt of c rbonic xid , ndic t d ng r of tt mpt ng

aa CO e e ae e e e e . to tt in high . p rc nt g s , xc pt und r sci ntiﬁc control The thermal losses d ue to carbonic ox ide formation are discussed in C I hapte r XI (p .

1 61

INDEX

’ a s e a e 8 1 F uel : e ae ae s Cl rk t st for h rdn ss , loss s with w st g s , e aaer 1 2 1 1 Cok , ch r ct istics of , 5 e 2 6 e ae 1 t st for , t sting , v lu of , 5 , 54 , 57 ash i ae 2 F ae e e a e a a Colour of , sign ﬁc nc of , 5 urn c t mp r tur , c lcul ae ae i ae 1 w st g s s , sign ﬁc nc of , tion of , 47 1 0 8

e e s Combustion , compl t , product 1 0 of , 4 , 5 Gae L a 2 rr tt om x , 9 e e incompl t , products of , 4 , 9 Gas-coke aae 1 2 Ch r ct ristics of , 1 0 5 s 1 0 product of combustion , 5 e e proc ss , ch mistry of , 3 , 4 , Gaseou s uel A aae f dv nt g s of , 3 , 1 0 5 I 4 a e a ae Comp rison of th rm l lu s , v aae i 1 I ch r ct rist cs , 4 , 5 an d e e 1 pric s of fu l , 5 e aae 1 th rm l v lu of , 5 Con tin uous C0 2 testing appara Gae e ae 6 s ous impuriti s of w t r , 3 , tus e 1 2 8 , d scription of , 9 1 e s fl ue 1 2 conn xion to , 9 Ge a e e e a e d e rm n d gr of h rdn ss , a e ai 1 Contr cts for fu l , pl c ng , 5 n fi ition , 8 1 t e e Con rol of fu l suppli s , 54 ’ Goutal s aI formul , 57 Crosfield S 68 ons , Crosfie ld Dae 1 1 l , 5 H

H ae -L e e CO aaa b r o w 2 pp r tus ,

’ 1 6 Da a e e 3 rling s c lorim t r , 35 a n ater a d e H rd ness i w D e a a aw ae Cl rk issolv d s lts in n tur l t rs , ee 8 1 6 66 6 1 6 gr , 3 , , 7 , 9 , 9 ’ a e f 8 1 Cl rk s t st or , definitions of Fre nch and Ge a e ee 8 1 rm n d gr s of , E e e f a thyl n , orm tion of, from H e thrin ton a g , 7 co l , 9 H eaaae e w t b l nc for boil r orking , E e air ea e d ue xc ss , h t loss s to , 8 1 5 . 49 H eat losses in w aste gases D ue air e aae 1 to l k g , 4 5 d ue a e to c rbonic oxid , 59 , Fee - ae a i ae ana 1 0 1 2 d w t rs , pprox m t 4 , 4 1 d u e e e air 1 0 6 lysis of , 7 to xc ss , F e a e 2 6 d ue e 1 ix d c rbon , t sts for , to hydrog n , 4 3 F e de e e a e d efi d ue a ae r nch gr of h rdn ss , to hydroc rbon g s s , n ition 8 1 I , 4 S F uel A a e 6 8 I O d ue e 1 8 n lys s of , , 7 , , , to moistur , 4 1 1 H ea e d ue ae i e , t loss s to sc l on bo l r a ae aa 1 ae 1 0 pproxim t n lysis of , 7 pl t s , 7 ai a 1 2 e e e e rt fici l , 4 , producing l m nts in fu l , 3 a a 1 ae e contr cts , pl cing , 5 v lu of fu ls , 3 ’ a ae a e H oni man s as- a a contr cts , typic l x mpl s g g s mpling p 2 aa 1 1 0 1 1 of , 5 , 53 p r tus , , 7 1 64 INDEX

H dr carb n ases F ai Non- ae-f i a ae y o o g orm t on sc l orm ng s lts in w t r , a 6 of from co l , 9 7 ea e d u e 1 h t loss s to , 4 5 H yd rogen in coal S igniﬁcance 0 1 , 4 1 of Oil Dae e e - ate ng rs of , in f d w r , ea e d ue 1 h t loss s to , 4 3 86

‘ e s ee - ae and t st for , in f d w t r e ae 8 8 86 1 0 1 I boil r sc l , 4 , 5, , - O -filters 1 0 1 i ie ee ae ai il , Impur t s of f d w t rs , d n O a ae in ae e e 6 , , g rs of , 3 rg nic m tt r w t r t st for 8 0 I ncomplete combustion Orsat as-e i aaa 1 1 8 1 0 g t st ng pp r tus , ducts of , 4 , 9 , 5 hea e d ue 1 2 t loss s to , 4 Intermittent ty pe s of water P e i aaa 8 soft n ng pp r tus , 9 P e a haae 6 t , c r ct ristics of , 4 , P eat-coke aae 1 Ch r ct ristics of , 3 use as a i e a1 K subst tut for co l , 3 P ermanent hard ness of w ater e C0 e aaa 1 6 Kr ll 2 t sting pp r tus , 3 De ni i 6 fi t on of , 7 e 8 2 8 t sts for , , 9 L P e e i i 1 1 trol um , const tut on of , 5, Le -T a e e 1 P a i a wis hompson c lorim t r , 3 hy sic l const tution of co l , 4 L e aae 6 P a e a I ignit , ch r ct ristics of , 4 , l cing fu l contr cts , 5 L i e as a P e 1 1 1 0 iqu d fu ls , cl sific tion of , 4 ool , , 7 L ead ue e e P e ai e a e anal oss of h t , to incompl t r p r ng fu l s mpl s for y 1 2 sis 1 1 8 2 2 combustion , 59 , 4 , 7 , ,

M R

Mae i a ae e a Rai - ae aae ~ gn s um c rbon t , r mov l n w t r , ch r ct ristics of , ae of from w t r , 9 3 64 ae e a Re ae i A aa sulph t , r mov l of from cords m d w th dos pp r ae atus 1 w t r , 9 4 , 33 Mae -D ai R ea hl r onkin bomb c lor id l , 9 e e R e - ae aae m t r , 4 3 iv r w t r , ch r ct ristics of , Me hae a of a t n , form tion from co l , 66

n al H ea e Moisture i co : t loss s 8 d ue 1 8 to , 4 S a e 1 1 6 1 e 1 1 2 1 mpling fu ls , 5, , 55, 55 t sts for , 7 , 9 , ae ae 1 0 , 9 M - as 1 1 w st g s s ond g , 4 , 5 ae 68 w t r , S a C0 e i aaa rco 2 t st ng pp r tus , 1 3 1 Natural uels O i 6 S ae - i a ae 6 f rig n of , 4 , c l form ng s lts in w t r , 7 ai a S ae i e ae eat cl ss ﬁc tion of , 3 , 4 c l on bo l r pl t s , h i e d ue 1 0 const tution of , 4 loss s to , 7 1 65 INDE X

S e e s s in oi e Te e a e f oi e f r ae cr t commis ion b l r mp r tur o b l r u n c s , e a ai f 1 hous , 55 c lcul t ons o , 47 - S emi Anthracite Characteris Thermal losse s se e of 1 0 e tics , loss s s f ae see H eat ae product o combustion of , v lu v lu 1 0 T ta ee ae e s 5 o l solids in f d w t r , t st S e ai ae 8 8 8 1 0 1 mok form t on , c us s of , 9 , for , 7 , , 1 06 T ae e 6 8 1 0 ypic l t sts of fu l , , 7 , , , S a a e ae ase 1 6 n p s mpl s of w st g s , 5 a 1 0 1 1 0 ae 1 60 t king , 9 , w t r , S o tenin aaratus : Ad van ae ae 1 f g pp w st g s s , 44 ae 6 1 2 t g s of , 7 , 9 , 9 r cont ol of , 9 9 e 8 ty p s of , 9 Ue hli C0 aaa 1 ng 2 pp r tus , 33 S oftening reagents Calculations

for , 9 5. 9 7 e r 6 V ch mist y of , 9 3 , 9 S Ors at aaa olutions for pp r tus , V ae ae e e s f ol til m tt r in fu l , t st or , e ai 1 2 pr p r ng , 3 2 6 S e r and ae oot , th o y c us s of , 9 , 1 06 W S ae aae i ti pring w t r , ch r ct r s cs and ae 6 66 Waste ases A x ae an d ng rs of , 3 , g ppro im t S aa ae a 1 1 6 t nd rd solutions for w t r lysis of , e 8 ae ae a e 1 0 1 1 0 t sting , 7 4 , 3 v r g s mpl s of , 9 , , ’ S ea as - a aaa 1 1 1 2 6 t ds g s mpling pp r tus , 4 , 1 0 8 colour of , S ea a haae i i 1 0 t m co ls , c r ct r stics of , composit on of , 5 1 0 e e e 1 06 n c ssity for t sting , S ul hur in coal P e e e e a e e 1 0 8 p r s nc of , 9 , r fr ctiv indic s for , , 1 0 1 38 1 0 e i a 1 0 8 products of combustion , 5 sp c fic gr vity , e 2 8 2 e i ea 1 0 t sts for , , 9 sp c fic h t , 7 S ul huric acid A e e a e 1 06 p mount pro t mp r tur of , d uced e e 1 2 , 9 t sts for moistur , 4 e a ae ae e 1 2 1 26 r mov l of from w st g s s , t sts for soot , 4 , aaa 1 1 6 1 9 ty pic l n ly sis of , 44 , S ae ae haae i Water A i ae aa urf c w t rs , c r ct r stics pprox m t n lysis 6 1 of , 4 of , 7 S uspe nde d solid matter in fee d che mical and physical char ae e 6 8 8 acteristics 6 6 66 w t rs , t sts for , 7 , of , 4 , 5, i e a a6 impur ti s of n tur l , 3 a e a a6 h rdn ss of n tur l , 7 e e and ae e n c ssity v lu of t sting , Temporary hard ness of w ater D efinition 6 ae and e s of , 7 oil in , d ng rs of t st e 8 1 8 8 8 86 1 0 1 t sts for , , 9 for , 4 , 5, ,

1 66