FRANK H . REED, HAROLD W. JACKMAN,
AN DP . W . H E N L IN E
S ta te G eological S u r v e y , U rb a n a , III.M
ID W E S T E R N by-product coke ovens use annually m illions of tons of high-volatile bitum inous coals which are shipped from th e A ppalachian coal fields in W est Virginia, P ennsylvania, an d eastern K entucky. Illinois has larger re
serves of high-volatile bitum inous coal th an any state east of th e R ocky M ountains; only Colorado exceeds Illinois in re
serves. A lthough th e A ppalachian coals are principally of higher ran k th a n th e m idw estern coals, th e reserves of these high- ra n k A ppalachian coals are being depleted rapidly.
T his growing scarcity of th e best A ppalachian coals and the critical tran sp o rtatio n problem now confronting th e nation, prom pted th e Illinois S tate Geological Survey, aided by th e Office o f P roduction R esearch and D evelopm ent of W PB, to begin a research program to stu d y th e possibilities of substitu tin g mid- w estern coals for a portion of th e eastern high-volatile coals now being carbonized in th e Chicago and St. Louis districts. Freight r ta te to th e G a ry -E a st Chicago-C hicago district by rail favor th e southern Illinois coals over th e A ppalachian coals by more shan one dollar p er ton. W h en a com bination of rail haul from
th e A ppalachian fields to a L ake E rie p o rt an d th en lake barge to the Chicago district is used, th e so u th ern Illinois coal can still be delivered w ith a saving in freight of 50 cents p er to n . I t costs over 2 dollars per to n m ore to deliver A pp alach ian coals b y rail to th e St. Louis d istrict th a n to deliver so u th ern Illinois coals to the sam e point.
Illinois coals contain m ore m oisture th a n do th e A ppalachian coals, however, an d a correspondingly low er am o u n t of fixed carbon on th e as-received basis. T h u s a lower yield of coke p er to n of coal is o b tained from th e Illinois coals. Previous com m er
cial and sem icom m ercial experience h as show n th a t “ a fairly satisfactory grade of m etallurgical coke” (4) can be m ade from coal of certain so uthern Illinois m ines. U ndoubtedly th e m e th ods of coal p rep aratio n an d th e resulting q u a lity of coal shipped from these m in es are b e tte r to d ay th a n th ey w ere during th e la s t w ar, or during th e period from 1921 to 1934 w hen F ra n k lin C ounty, 111., coal was used in th e R o b erts ovens a t G ra n ite C ity, 111.
560
T h e comparison of costs to be obtained by coking A ppalachian coals alone or in com bination w ith Illinois coals in any given p lan t can be determ ined only by commercial operation over an extended period of tim e. T he su itab ility of th e coke for blast furnace operation, th e yield of coke from th e coal, and th e am ount and evaluation of th e by-products are among the princi
pal item s which m ust be considered, along w ith freight rates, cost, and uniform ity of coal supply, to determ ine th e over-all economic picture. However, th e cost of experim entation w ith various blends of coal in commercial coke ovens is no t only exceedingly high, b u t it also interferes w ith regular production. Conse
quently, only a m inim um of such experim entation is carried out.
T he first objective of th e present research program was, there
fore, th e design and construction of a small-scale slot-type coke oven in which blends of coals could be carbonized under condi
tions approxim ating those obtained in commercial ovens closely enough to produce cokes w ith physical and chemical properties directly com parable to those of cokes produced from th e same blends in commercial ovens. R epresentative d a ta selected from the first fifty runs are show n in T able I to indicate th a t th is ob
jective has been attain ed .
T his p aper gives briefly th e construction and operation of the experim ental slot-type oven, shows th e duplicability of operating result«, and compares results obtained w ith th is experim ental oven and w ith commercial slot-type ovens.
DESIGN O F OVEN
The prim ary objective in design of th e experim ental slot-typ*
oven was to construct a u n it which would duplicate essentially a small section of a comm ercial oven, and in which th e process of
coking coal would be controlled rigidly. This was based on th e assum ption th a t th e coking process is a complex chemical reac
tion. Thus, duplication of operating conditions from one run to another w ith identical blends of coal should produce batches of coke w ith identical physical and chemical properties. In prac
tice “identical blends” are not obtainable, b u t blends with sim ilar average physical and chemical properties can be obtained; there
fore th e cokes produced from such blends should have sim ilar physical and chemical properties.
Only in th e w idth of the oven was an a tte m p t m ade to dupli
cate any size dimension of a commercial oven. T he average w idth of m ost commercial slot-type ovens ranges from 13 to 21 inches. The average w idth of th e experim ental oven is slightly above th e lower lim it of this range. Thus, th e oven was de
signed so th a t it could be operated to give th e sam e h eat penetra
tion (average w idth of oven in inches divided by coking tim e in hours) and final tem perature as obtained in commercial practice.
Figure 1 shows th e slot-type experim ental oven being dis
charged and th e coke being quenched. T he uniform oven wall tem perature up to the top of th e charge and th e slightly cooler space above for gas collection are apparent. Also visible are th e coal-charging hole, th e door w ith opening for leveling bar, and other details of oven construction, including buckstays and angle- iron supports in th e side walls.
Figure 2 is a diagram showing cross-sectional views from th e front and side of the oven. As in all slot-type coke ovens, h eat is applied from flues placed on both sides of th e oven cham ber, 1.
T he inside of th e oven cham ber was designed to tap er in w idth from 13.25 inches a t th e back to 13.75 inches a t th e front. On account of irregularities in th e shapes received, th e oven as
con-FRONT VIEW SIDE VIEW
y / / A \ \ \\ V K s
- m
cx¿i i x i [ x a c x j i x i I X .
: - - - :
-s ' - 51/
2"-FIRE BRICK V /////A VERMICULITE
N \ \ \ \ S l HIGH-TEMPERATURE SILICON c a r b i d e
INSULATORS
CA ST SHAPES | COMMONS
Figure 2. Sketch o f Slot-T ype Experim ental Coke Oven
562 I N D U S T R I A L A N D E N G I N E E R I N G C H ^ M + S T R Y
W5l.
37, No. 6 S E C T IO N A - AA N G L E A N G L E
SECTION B - B [•— 13 !■
SE C T IO N C - C
H I G H - T E M P E R A T U R E I N S U L A T O R S V ////A V E R M I C U L I T E
H E A T I N G F L U E O P E N I N G S S I L I C O N C A R B I D E
F ig u re 3. D e ta ils o f S lo t-T y p e Oven C o n stru ctio n
stru cted averages 14 inches in w idth instead of 13.5 inches as planned. T he charging space in th e oven cham ber is 36 inches in length and 35 inches in depth, and holds approxim ately 10 cubic feet of coal per charge.
T he side walls, 4, and floor, 5, of th e oven are m ade of silicon carbide tile, 2 inches thick. E ach side wall consists of a single tile, an d th e floor is form ed from tw o tiles laid end to end with overlapping joint. T h e walls are anchored a t th e back of the oven an d left free to expand vertically and horizontally. T hey are held in place a t th e top an d bottom by th e surrounding brick
work, an d are fu rth er supported on each side by tw o rows of long firebrick, 6, which touch th e oven walls and are, in tu rn , strength
ened by steel angles, 7, running th e full length of th e outside walls of th e oven. These supporting firebrick are spaced from front to back of th e flues, leaving 4.5 inches betw een bricks, so th a t ap
proxim ately 50% of th e flue space is left open (Figure 3, section C -O . These flue openings are staggered in th e tw o rows of sup
p orting brick in each flue.
T his leaves th e th ree sec
tions of each flue closely interconnected and al
lows th e h e a t to equalize from to p to b o tto m of each oven wall. T h e oven cham ber is su r
rounded on th e sides and top by v en n icu lite in sulation, 8. T his insula
tion acts n o t only as a h e at baffle b u t, being soft, as a cushion ag ain st th erm al expansion or swelling pressures w hich otherw ise m ig h t crack th e silicon carbide walls.
T he to p of th e oven cham ber, 9, is cast of re
fracto ry concrete. C oal is charged th ro u g h a 6- inch pipe, 10, extending th rough th e casting, an d a 6-inch b lan k flange, 11, serves as a charge hole cover. G as escapes from th e oven th ro u g h a 3-inch pipe, 12, extend
ing th ro u g h th e to p an d connected to th e by
p ro d u ct recovery equip
m ent. T h e b a ck of th e oven cham ber consists of p e r m a n e n t brickw ork, w hereas th e fro n t is covered by a refracto ry concrete door, 13, w hich is raised or low ered by a chain h oist an d is m udded into place before th e oven is charged.
A fter charging, th e coal is leveled th ro u g h a rec
t a n g u l a r opening, 14, in th e door located 35 in c h e s a b o v e t h e c h a m b e r flo o r. T his le v e l b a r o p e n in g is t h e n b r i c k e d a n d m udded. B etw een th e d o o r a n d t h e c o a l charge a tem porary brick wall, 15 (9 .inches in d ep th ), extends from th e floor to th e coal level. T his wall, w hich consists of one layer of firebrick next to th e charge an d one lay er of insulating brick next to th e door, is rem oved before a coke charge is pulled, and is replaced im m ediately a fter th e oven is discharged. T he oven stru ctu re is held to gether by tie rods, 16, extending th ro u g h th e to p brickw ork an d foundation. T hese rods are anchored to heavy buckstays, 17, a t each com er of th e oven.
Figure 3 gives m ore details of th e oven brickw ork construction.
H orizontal sections A -A , B -B , an d C-C, w hich refer b a ck to Figure 2, show th e brick arran g em en t ju s t below floor level a t th e oven floor, and a t a p lane betw een th e low er an d m iddle flue sections. T he b ack view shows th e arrangem ent of th e openings for heating u n its an d therm ocouples in to th e h eating flues an d th e oven cham ber. T herm ocouples are n ever placed in all of th e holes show n during an y one run, b u t th e holes are b u ilt in to th e oven to be available w hen and if desired.
I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y 563
e Average of 1-month commercial operation.
/ Pounds of coal per cu. ft. as charged to oven.
o Screen analysis of coal as charged to oven.
Series 2________ _______Series 3 Series 4
C21 C3c S23 S24 C4<¡ S34 S35 C5* elements (2, Figure 2), 67 inches long and having a middle heating section 36 inches in length and 1.25 inches in diam eter, designed parallel across one secondary of the three-phase tap
transformer. In the center and bottom flue sec
tions the Globars are connected in the same m anner across the other two secondaries. W ith this hookup are formed, essentially, three independently con
trolled single-phase circuits of approxim ately 16.7 kv.-amp. each, which have proved adequate to provide uniform tem perature in th e oven chamber.
Each Globar element is supported a t the front and rear of the oven in special insulating shapes (3, Figure 2) as provided by th e vendor of the Globar heating elements.
Flue tem peratures are controlled by a Wheelco Capacitrol which is actu ated by a No. 8 gage, chromel-alumel therm ocouple placed in th e center flue section on one side, and adjacent to b u t not touching the oven wall. T his therm ocouple acti
vates the three-phase prim ary circuit. N o ap
preciable tem perature difference has been found to exist between th e two flues, and th u s it was not approxim ately to th e center of the oven. Another is placed in the exact center of the coal charge, and a fifth is located in the free gas space above the coal. Tem peratures recorded a t these points show the progress of tem perature change throughout the car
bonizing period.
Figure 5 is a photograph of a typical tim e-tem perature chart.
T his chart records a run in degrees centigrade, and the tem pera
tures should not be confused with those on the Fahrenheit scale used otherwise throughout this paper. T otal elapsed tim e for the carbonization period is 13 hours 10 m inutes. Curve (1) indicates
1 5 0 -A M P . 3 - POLE CONTACTOR
564 I N D U S T R I A L A N D E N G I N E E R I N G C HeI TiS T I Y vSl 37, No. 6
th e tem perature of th e gas in th e free space above th e coal charge.
T h e therm ocouple in use a t th e beginning of th e operation was evidently defective and was reading too low. I t was replaced after 1.75 hours by a new therm ocouple. C urves (2), (3), and (4) represent, respectively, th e tem peratures of therm ocouples placed a t th e top, middle, and bottom of th e charge next to th e side wall. These curves are representative of norm al charts. A t the early p a rt of th e run th e highest tem perature along th e oven wall is recorded a t th e to p of th e charge and th e lowest tem perature a t th e bottom . All tem peratures rise in practically parallel stra ig h t lines for ab o u t 11 hours, after which th e tem perature a t th e center crosses over and becomes th e highest. C urve (5) indi
cates th e tem perature a t th e approxim ate geometric center of th e charge. H ere th e tem perature rem ains constant a t about 125° C. (257° F.) for th e first 5 hours, rises 100° C. (180° F.) in th e n ex t 2 hours, th en 700° C. (1260° F.) in th e next 5 hours, and finally reaches th e tem perature of th e charge a t th e side wall.
C urves (4) and (5) are recorded by th e No. 4 position on th e tem perature recorder by connecting it altern ately w ith therm o
couples 4 and 5. T he circled points on th e ch art indicate th e con
tin u atio n of th e curve for therm ocouple 4, as shown by occasional recordings, while therm ocouple 5 is being recorded continuously.
OPERATION O F OVEN
In operating th e sm all scale slot-type oven, various charging tem peratures and final flue tem peratures have been tried. In all com parative tests an a tte m p t is m ade to duplicate th e average h e at p enetration through the charge and th e average final coke tem perature as attain ed by commercial batteries. As th e silicon carbide walls of th e experim ental oven have a higher therm al conductivity th an th e silica brick walls of large scale ovens, it is
possible to o b tain approxim ately th e sam e average p e n etratio n ra te a t a m uch lower flue tem p eratu re in th e experim ental oven th a n is required in com m ercial ovens. R esults th a t duplicate closely those of com m ercial practice have been o b tained b y charg
ing th e oven a t an in itial flue tem p eratu re of 1600° F . and^raisrng th is tem p eratu re 30° F . p er h o u r to a m axim um of 1850°. T h e coking tim e under these conditions is found to be 12.75 to 13.5 hours, or th e average p en etratio n is 1.04 to 1.10 inches p er hour, depending upon such factors as bu lk density, m oisture co n ten t, and plastic characteristics of th e coal. T h e final average coke tem perature is 1770-1800° F.
T he sam e by-product recovery equipm ent is used on th e slo t oven th a t was em ployed w ith th e experim ental sole-flue oven (3).
W hen a test ru n is sta rte d , th e ta p transform er is se t a t approxi
m ately 70% of ra te d capacity on all h eatin g sections, an d th e gas exhauster is sta rte d w ith th e by-pass open. A charge of 10.1 cubic feet (approxim ately 500 pounds) of coal is dum ped from a n overhead hopper thro u g h th e charge hole an d leveled in th e oven.
Gases are v ented to th e outside u n til th e level a n d charge holes are sealed. T he gas is th en pum ped th ro u g h th e purifying equip
m ent b y th e exhauster. T h e in itial quenching effect of th e coal on th e flue tem p eratu re is insignificant, an d it is believed th a t th e u n it has sufficient capacity to m ain tain a m uch higher in itial charging tem perature, although no in itial tem p e ratu re above 1700° F . has been tried. As in operating th e sole-flue oven (S), a constant pressure of 0.02 inch of w ater is m ain tain ed in th e oven cham ber an d approxim ately 0.5 inch of w ater a t th e m ete r outlet.
T he settin g of th e ta p transform er for each of th e th ree h eatin g sections is changed from tim e to tim e as required to m ain tain a uniform wall tem p eratu re from th e b o ttom to th e to p of th e oven
charge.
R a te of gas evolution is essentially co n stan t u n til th e p lastic zone reaches th e cen ter of th e charge. T his ra te increases g radually while th e center is heating, th e n g radually decreases. T h e B .t.u . valu e of th e gas decreases slow ly from approxim ately 850 u n til th e p lastic zones m eet, increases slightly over a sh o rt period, th e n drops sharply to a b o u t 300.
In all tests coking is continued u n til gas evolu
tio n has dropped to th e ra te of 50 cubic feet p er hour. T h e oven is th e n opened, th e b rick re
tain in g wall is rem oved, an d th e coke is p ulled an d quenched. T h e p y ro m eter tem p e ratu re con
troller is se t b ack to 1600° F ., w here i t is m ain tain ed during all idle periods, th e pow er is c u t to 25% of ra te d capacity, a n d th e fire b rick re ta in ing wall is replaced inside th e oven door. T a r yield, gas m ake, a n d coke yield are com puted as in th e sole-flue oven te s ts (S).
S h a tte r an d tu m b ler te s ts of th e coke pro
duced are m ade in accordance w ith sta n d a rd m ethods ad o p ted b y th e A m erican Society for T esting M aterials (1, 2).
C O K IN G R E S U L T S ON D U PL IC A T E S A M P L E S T he ab ility of th e experim ental oven to re produce coking results u n d er closely controlled operating conditions is show n in T ab le I I ; re
sults are given of duplicate ru n s on each of tw o coal blends. B lend A is a m ixture of Illinois and eastern coals, an d blend B is all e astern coal. O perating conditions, such as coking tim e ra te of h e a t p en etratio n , a n d final coke tem peratures, were k e p t c o n stan t th ro u g h o u t th ese four runs. T h e degree of p u lverization an d
June, 1945 ■ ■ I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y 565 6 the cubic feet of gas evolved during each hour of the coking period and th e B .t.u. value of the gas are p lotted from experim ental d a ta taken during both sets of duplicate runs. B .t.u . values are n o t shown for th e gas beyond th e tenth hour. G as evolved during th e balance of the coking period is very high in hydrogen, and the calorimeter is n o t adjusted to read accu
rately in th is low range. These curves are typical of th e results obtained under norm al operating conditions.
Because of the close control of operation pos
sible with th e experim ental oven, w hich cannot be realized in a gas-heated commercial size
oven, the results on the experim ental oven have been shown to be more dependable and m ore easily duplicated th a n those obtained from individual ovens of a commercial b attery .
EX P ER IM EN TA L AND COM M ERCIAL R ESU L TS The extent to which th e design of th e oven is successful in permitting the duplication of commercial results is shown in Table I, which gives th e results of four representative series of tests comparing experim ental and commercial operations.
In evaluating coke q u ality th e producers consider th e sh a tte r is due to th e inconsistency of coking results from individual com
In evaluating coke q u ality th e producers consider th e sh a tte r is due to th e inconsistency of coking results from individual com