ULTIMATE ANALYSIS OF COAL AND COKE –A TOOL AT RINL FOR PROCESS AND

PRODUCT CHARACTERIZATION ------S.K.SEET

BACKGROUND

Coal science and technology of coke making has made great advancement for last several decades; however views of
experts in this field still carry divergent opinions regarding structure –property correlation of coal and coal to coke
transformation mechanism. As a result the standards such as ASTM etc, is still using the conventional parameters such
as , volatile matter, calorific value & carbon content to classify coal for its relative merits & de-merits against a particular
use. Classification of coal for relative merits & demerits towards coking or carbonization is experiencing till now
maximum difficulty. Coke makers now are forced to use a long list of coal test parameters to characterize the coal to
assess its coking propensity for various commercial purposes.

Proximate and Ultimate analysis are mandatory coal testing parameters for coal characterization. They provide the data on
basic chemical constituents of the coal concerned. The data in isolation awaits for a further physico-chemical
characterization and research on the subject of coal structure-property correlation.

This study tried to extend the use of coal ultimate analysis in association with other coal parameters to the following areas.

1. Classify coal in line with various international systems and develop in house assessment module for coals
2. Explore correlation between ultimate analysis and vital coking parameters such as MMR, H/C ratios and final
coke strength parameters.
3. Incorporate ultimate analysis in coal blending process to formulate blends for RINL coke oven batteries
4. Estimate calorific value of coking coal, steam coal and other carbonaceous matters using the ultimate analysis
data to a fairly accurate manner so as to use them in several energy calculation schemes.

USE OF CHNS ANALYSIS IN COAL CHARACTERIZATION/CLASSIFICATION

There are several criticism regarding ASTM coal classification system by Rank and effort are still on by various expert to
devise more effective system of classification to categorize coal with respect to its coking propensity.European and
American researchers in the nineteenth and early twentieth centuries proposed several coal classification systems. The
earliest, published in Paris in 1837 by Henri-Victor Regnault (1810–1878), classifies types of coal according to their
proximate analysis (determination of component substances, by percentage), that is, by their percentages of moisture,
combustible matter, fixed carbon, and ash. It is still favored, in modified form, by many American coal scientists. Another
widely adopted system, introduced in 1919 by the British scientist Marie Stopes (1880–1958), classifies types of coal
according to their macroscopic constituents: clarain (ordinary bright coal), vitrain (glossy black coal), durain (dull rough
coal), and fusains, also called mineral charcoal (soft powdery coal). Still another system is based on ultimate analysis
(determination of component chemical elements, by percentage), classifying types of coal according to their percentages
of fixed carbon, hydrogen, oxygen, and nitrogen, exclusive of dry ash and sulfur. (Renaults had also introduced ultimate
analysis in his 1837 paper.) The British coal scientist Clarence A. Saylor developed this system in 1899–1900 and greatly
expanded it to include large numbers of British and European coals. Finally, in 1929, with no universal classification
system, a group of sixty American and Canadian coal scientists working under guidelines established by the American
Standards Association (ASA) and the American Society for Testing Materials (ASTM) developed a classification that
became the standard in 1936. It has remained unrevised since 1938.

1
 100
Carbon Oxygen
90
Hydrogen
80

70

60

50

40
coalification

C o m position(%,da f)
30

20

10

0
Tree Lignite Anthracite
Peat Bituminous

Figure-1: Process of coalification in relation to ultimate analysis of coal

Classification of coal by rank as per ASTM D 388-99(2004)

2
In 1949, the Coal Committee of Economic Commission for Europe set up the Classification of coals and “to make
recommendations…… on measures designed to achieve an international coal classification”. When taking this decision,
the Coal Committee had been conscious of the need for an internationally recognized and commercially operable system
of coal classification, absence of which had created confusion and inconvenience in several field of activity.

The need for coal classification stems from assessing coal quality towards its various utilization such as; coal
carbonization and combustion. These classifications mostly use rank, type and grade of coal. Rank is related to maturity of
coalification, grade is related to purity of coal with respect to minerals and ash and finally the type is related to
coking/caking potential. International classification of coal such as depicted below.

INTERNATIONAL COAL CLASSIFICATION REPRODURED FROM UN ECONOMIC COMMISSION FOR

EUROPE : International classification of Hard coal by type
Group Code number Sub-group
Determined by Determined by coking property
Caking
property
Group CSN The first figure of code number indicate class or rank of coal Sub Dilatometer LTGK
no The second figure indicate group of coal determined by caking property Grou
The third figure indicates subgroup indicated by coking property of coal p
no
3 >4 435 535 635 5 >140 >G 8
334 434 534 634 4 >50-140 G5-G8
333 433 533 633 733 3 .0-50 G1-G4
332 33 432 532 632 732 832 2 <=0 E-G
a 2
b
2 2.5 323 423 523 623 723 823 3 >0-50 G1-G4
to 322 422 522 622 722 822 2 <=0 E-G
4.0 321 421 521 621 721 821 1 Contraction B-D
only
1 1-2 212 312 421 512 612 712 812 2 <=0 E-G
211 311 411 511 611 711 811 1 Contraction B-D
only
0 0- 100 100 200 300 400 500 600 700 800 900 0 Non- A
0.5 A B softening
Class number 0 1 2 3 4 5 6 7 8 9
Class
parameter
V.M 0- >3-10 >10- >14-20 >20- >28- >33 >33 >33 >33
3 6.5-10 14 28 33

Calorific value - - - - - - >7750 >7200- >6100 >5700-

7750 - 6100
7200
The above classification system incorporates coal rank (VM & Calorific Value) and coking behavior (LTGK, CSN&
dilation) which are easily tested by all coal users and thus enabling them to classify coals. This classification approach
was tried for RINL coals and to characterize using a spreadsheet program depicted below.

3
IN-HOUSE DEVELOPMENT OF COAL CLASSIFICATION MODULE IN LINE WITH ECU International
classification of Hard coal by type(ECU)

ESTIMATION OF CODE OF THE CANDIDATE COAL AS PER ECE CLASSIFICATION International Comment on
Feed coal property Candidate Coal
code No (ECU) coking potential
IM 1.5
VM db 24.5 PEAKDOWN 434 Very Strong Coking
Ash db 9.0 GYMS 434 Very Strong Coking
C 80.8 A-MN 434 Very Strong Coking
A-MN Anglo
H 4.7 CAMBRIA 532 Coking Coal
S 0.52 West gilbert 532 Coking Coal
LTGK G7 Gregory 634 Coking Coal
CSN 8.50
BWS 532 Coking Coal
Max Dilation 35.97
Cumberland Wentz(NS) 632 Coking Coal
GY.NORTH 434 Very Strong Coking
Output Coal Data for coal Classification as per Economic Commission of Europe (ECE)
Burton Hard 432 Coking Coal
Vmdaf 27.6 Candidate coal Code No.
Norwitch Park 333 Coking Coal
GCV 8130 Eagle 532 Coking Coal
CSN 7
Cheviot 532 Coking Coal
LTGK G7 LTGK
Black Bear 432 Coking Coal
Max Dilation 36

1st figure 2nd figure 3rd figure

434 NZSSC
MCC
632
932
Coking Coal
Coking Coal
Class Group no Sub Gr. No.
Tuhup 532 Coking Coal
4 3 4 NZH 532 Coking Coal

Figure B: Coal characterization Classification

GERMAN SYSTEM OF COAL CLASSIFICATION

4
Figure-2: Classification of coal using ultimate analysis in German coal classification system

IN-HOUSE DEVELOPMENT OF COAL CHARACTERIZATION MODULE IN LINE WITH GERMAN

SYSTEM

The above classification correlates carbon content, volatile matter content, calorific value and the reflectance of coal. Thus
to identify coking potential of a candidate coal the above classification can be used using ultimate analysis. A program in
spreadsheet is developed which is depicted below.

Put Carbon (C) here

85
OUTPUT BELOW
COAL VM Carbon Hydrogen Oygen MMR
Bituminus High Volatile A 35 85 5.6 7.3 1.0-1.2

Oygen; 7.3
Hydrogen; 5.6

Carbon; 85

Candidate coal Class=Bituminus High Volatile A MMR=1.0-1.2

Figure B: Coal characterization Classification

COMPARISION BETWEEN SEVERAL INTERNATIONAL CLASSIFICATION SYSTEMS

International V.M C.V USA UK International
Class
0 no 0-3 Meta Athracite Code no
1A 3-6.5 Antharcite Non-caking 100
1B 6.5-10 Weakly caking 202
2 10-14 Semi-Anthracite
Low-Volatile Medium to 203
3 14-20
bituminous Strongly 204
4 20-29 Medium volatile coking 300
bituminous Very Strongly coking 400
High Volatile
5 29-33 Strongly coking 500
Bituminous A
Medium coking 600
Weakly coking 700
8450- High Volatile
6 >33(33-40) Very weakly coking 800
7750 Bituminous B
Non coking 900
7 >33 7750-
8 (34-44)
>33 7200
7200 High Volatile
9 (34-46)
>33 -6100
<6100 Bituminous C
Sub Bituminous
(34-46)

5
This comparison is quite helpful to correlate crude classification outcomes such as ; High-Medium Volatile Bituminous
etc,. to international class (4), code (434) and strength category (Very Strongly Coking).
EXPLORE CORRELATION BETWEEN ULTIMATE ANALYSIS AND VITAL COKING
PARAMETERS SUCH AS MMR, H/C

The literature references on reflectance and the ultimate analysis

The analysis of plant sample reflecting the relation

MMR
COAL BRAND H/Cdb Predicted LP MMR
Saraji (Australia) 0.65 1.428 1.5
PD(Australia) 0.67 1.365 1.4
GYMS(Australia) 0.7 1.193 1.18
A-MN(Australia) 0.69 1.193 1.17
BULI(Australia) 0.68 1.305 1.28
Gregory(Australia) 0.77 0.892 0.92
Norwich
Park(Australia) 0.62 1.69 1.65
TUHUP(Indonesia) 0.704 1.177 1.18
NZHard (Newzealand) 0.719 1.114 1.15
NZSS (Newzealand) 0.81 0.783 0.8
MCC 0.76 0.892 0.9

Prediction of MMM from the H/C ratio by the regression analysis

MMR=28.271*(H/C)daf2-48.835*(H/C)daf+22.005 -------------------------------------------------------------(1)

FEED INPUT In-house developed calculation module to predict MMR of coal from H/C ratio is
Cdry basis 81.5 arrived. It can be seen that the correlation is quite accurate with Coefficient of
H dry basis 4.5 Determination (R2) >0.9. To further normalize the prediction of MMR from the H/C,
Ash dry basis Taylor series interpolation on the above empirical equation is exercised. As can be seen
8.5
with the input data of %C, %H and % Ash of the coal the MMR is being predicted
OUT PUT
with a fair degree of accuracy. The applicability of the empirical relation is working
H/C 0.66
fairly well for coal of Australian origin. Deviation is observed for coal from Canadian
Predicted MMR 1.41
origins.
6
INCORPORATION OF ULTIMATE ANALYSIS IN COAL BLENDING PROCESS TO FORMULATE
BLENDS FOR RINL COKE OVEN BATTERIES

B-1 Prediction of Blend Property based on Ultim

COAL BRAND blend%
SARAJI CARBON 80.4
Norwitch Park 4.82
HYDROGEN
GC 0
SULFUR 0.64
PEAKDOWN 16
MMR 1.128 ASH 10.31
GYMS
VMdb 26.1 1.79
M.NORTH 55 NITROGEN
Coal ASHdb 10.3
TUHUP 0 OXYGEN(By Balance) 2.08
Coke Ash 14.0
TECH EAGLE H/C 0.719
Sulfur Coal 0.64
RIVERSIDE
CSR predicted-1 65 C-Aromatic 29.3
CAMBRIA CREEK
CSR predicted-2 65 Predicted CSR-Caro 64.3
WEST GILBERT 0
GREGORY 13
Sulfur coke 0.55 GCV 8751
BWS
Fuidity(ddpm) 687 Ammonia 0.32
NCA COAL
Blend Cost 7829 H2S 0.16
NEWZLAND 9
MCC 7
Sum 100

Table below:

Normal ranges of Ultimate analysisValues for coal

used for metallurgicalCoke production
Ultimate Analysis Figure:
% Value Range (daf) Predicted ultimate analysis values for the blend
of
coal
Total Carbon 80-87
Hydrogen 4-5 PREDICTION OF COKE STRENGTH
Nitrogen 1-2 FROM ULTIMATE ANALYSIS

Oxygen 3-4 Data analysis is done to develop a regression model

to predict CSR of coke from the coal carbon data.
Total Sulfur <0.8
After extensive data analysis it was found that
aromatic carbon (derived from total carbon C daf) correlates strongly with
CSR values obtained for RINL pilot oven straight coal testing. A regression analysis was done which reads as per
equation -2. The model is able to predict CSR with a better degree of accuracy (R-Square: 0.92). The relation between
actual and predicted CSR is also shown. The statistical significance of the below relation is also tested and is found to be
significant.

Regression of CSR of coke from carbon content of coal

CSR of coke= -0.059×(Carbon in coal)2 + 5.27× (Carbon in coal) - 44.48 -------------------------------------------(2)

7
 CSR- 80
Cdaf CSR-Actual
Coal Brand predicted
70
PD 92.6 74.0 73.12 60 R² = 0.07
GYMS 90.8 67.0 67.16
A-MN 89.7 67.0 65.58 50

Actual CSR
Gregory 88.1 51.0 51.20 40
Cheviot 88.6 49.0 49.00
30
Eagle 91.8 70.0 71.52
Norwich 20
Park 92.7 72.0 73.24
10
MCC 83.6 30.0 36.58
NZSSC 85.6 21.0 36.60 0
NZH 89.7 64.0 60.02 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0
TUHUP 92.7 65.7 73.10 Predicted CSR

PREDICTION OF AMMONIA AND H2S IN COKE OVEN GAS FROM NITROGEN AND SULFUR
OF COAL

Charge Coal Sulfur (Scb)=0.54

Coke sulfur (Sck)=0.40

1. H2S in coke oven gas = Ks*Scb*(34/32)

Value of Ks=0.17-0.19
2. Charge coal Nitrogen (Ncb)=1.8
Ammonia in coke oven gas =b*Ncb*(17/14)
Value of b=0.14

Based on the above formula the theotrically calculated ammonia and H 2S content in gm/nm3 is compared with that of the
actual. Figure Below is illustrating it. Thus Nitrogen and Sulphur content in blend coal or any candidate coal can be used
as a predictive tool to estimate the ammonia and H2S gas generation during coking.

4
3.6
3.2
2.8
2.4
2
1.6
1.2
0.8
0.4
0
1 4 7 10 13 16 19 22 25 28 31 34
Figure- Blend Sulfur-Actual H2S-Theritical H2S trend Figure- Blend N-Actual NH3-Theritical NH3 trend

8
CALCULATION OF HEATING VALUE (GCV) USING CHNS ANALYSIS DATA

The calorific value is usually expressed as the gross calorific value (GCV) or the higher heating value (HHV) and the net
calorific value (NCV) or lower calorific value (LHV). The difference between the gross calorific value and thenet
calorific value is the latent heat of condensation of the water vapor producedduring the combustion process.

The estimation of calorific value by bomb calorimeter is quite tedious and time taking process. Much care has to be
administered to arrive at acceptable analytical data for technological calculation. If a coal does not have a measured
calorific value, it is possible to make a close estimation of the calorific value (CV) by means of various formulas,
themost popular of which are (Selvig, 1945).

Heating value, Btu/lb = (14,544×C + 62,028×[H - (O ⁄8)] + 4,050×S ----EIGT

Heating value, Btu/lb= 144.4 × C +610.2× H-65.2× O-0.39×O2 ----------DULONG

where %C, %H, %N, %O, and %S are the respective carbon, hydrogen, nitrogen, oxygen, and organic sulfur contents of
the coal (all of which are calculated to a dry, ash-free basis).

In both cases, the values calculated are in close agreementwith the experimental calorific values.

In case of CHNS it is quite convenient and use of widely accepted empirical formula such as Dulog Formula is a
convenient means to arrive at heating value of various fuels in our plant.

9
CONCLUSION

Following can be concluded from the above study

1. CHNS analyzer can be used to estimate Gross Calorific Value (GCV) of coal, coke fractions, PCI, PCM,
Tar for thermal and energy calculation required for the plant. The data generated can be used for specific
energy consumption studies at departmental and plant level study.
2. With extensive analysis of available analytical data on various sources of coals, coal blends and coke
produced two significant empirical relations is developed. One is to predict coal MMR from its H/C
ration and other is CSR of coke from the Total Carbon value. Both relations will of great use in using
CHNS data for characterizing coal and coal blends.
3. Based on the CHNS data classification of coal in line with various international schemes can be made
which will provide additional / supportive information regarding coal from new origins. The evaluation
methodology as depicted in figure no- A & B will go well when conventional methods of assessing coal
is not able to provide significant insight.
4. The sulfur and nitrogen content of coal can be valuable tool to predict the amount of H 2S and NH3 gas
generation during coal carbonization in our battery.

________________________________ _______ END & THANK YOU___________________________________

10