Sampling of Coal Dr kalyan sen Director Central

Sampling of Coal Dr kalyan sen Director, Central Fuel Research Institute, Dhanbad, 2003

Quality Monitoring (QM) of Coal is an essential requirement for process control, plant performance or for any commercial transaction between Consumer and Producer QM requires proper implementation of standard sampling, preparation and test procedures 11/4/2020 2

Objective To collect a representative portion of fuel (coal) for determination of quality parameters. 11/4/2020 3

Purpose ¨ Commercial Transaction ¨ Location ¨ Parameters Normally at Loading point Ash, TM, 60% RH moisture, GCV ¨ Quality Parameters for Power Generation § Location § Parameters 11/4/2020 Normally at unloading point Ash, TM, GCV, VM S, N, Maceral composition 4

Steps for quality assessment © Sampling © Sample preparation © Analysis 11/4/2020 5

H Coal is heterogeneous material H Very difficult to achieve highest level of Sampling precision H In terms of variance D 80 % is from Sampling D 20 % is from preparation & analysis H Overall precision is influenced primarily due to Sampling H Utmost importance need to be given for 11/4/2020 6 Sampling

Sampling methods depend on < mechanical or manual sampling < sampling from moving belt < sampling from stationary lot (wagon, stockpile, etc. ) 11/4/2020 7

Sampling types Moving Stream v Auto-mechanical sampling system v Manual 11/4/2020 Stationary lot, Wagon, Stockpiles, etc. F Auto-mechanical Auger F Manual 8

Sampling variance is a function of product variability i. e. different results can be obtained from usame increments for different coal udifferent 11/4/2020 increments for same coal 9

The objective is to reduce the sampling variance as far as practicable 11/4/2020 10

Any Sampling scheme normally conforms with the national or international standards (BIS/ISO/ASTM, etc. ) Constraint - technical, cost and time Thus modifications in sampling procedures are necessary with mutual agreement between parties 11/4/2020 11

Precision u measures the closeness of data in given condition u indicates the reproducibility of the results u measures the chance error as expressed by variance SMALLER THE RANDOM ERROR, PRECISE IS THE METHOD A commonly accepted index of precision is twice the population standard deviation 11/4/2020 12

Precision depends on u Variability of coal u number of samples from a lot u number of increments comprising each sample u mass 11/4/2020 of sample related to the nominal top size 13

Precision PL = 2 * Sqrt [VI/m*n + VPT/m] n m VI VPT PL 11/4/2020 = = = no. of increments no. of sub lots Primary increment variance Preparation & testing variance Overall precision at 95% confidence level 14

Bias Systematic error which leads to the average value of a series of results being persistently higher or lower than those which are obtained using a reference sampling method which is intrinsically unbiased 11/4/2020 15

Reference method of sampling is ‘Stop Belt Method’ (free of Bias) 11/4/2020 16

Design of Sampling Scheme Basic Principles Both for Mechanical & Manual systems 11/4/2020 17

General scheme for sampling. . . u Decide purpose of sampling e. g. plant performance, process control, commercial transaction u Identify the quality parameters, viz. , general analysis, TM, size, washability, etc. u Define the lot u Define the precision required u Decide whether continuous or intermittent sampling is required 11/4/2020 18

General scheme for sampling . . contd. . u Determine the number of sub-lots, increments to achieve the required precision. u Determine the nominal top size of the coal u Determine the min. mass/ increment and the min. mass of the total sample u Decide on the method of combining the different increments for gross sample u Decide on drawing common or separate samples, for analysis 11/4/2020 19

General principle of Sampling u Primary increments should account for the Variability u Equal probability to all particles to be selected and included in the sample u Largest particle of the lot should pass freely through the sample device u Sufficient mass of the sample to enable particles to be present in the same ratio as in the lot 11/4/2020 20

SAMPLING FOR COMMERCIAL TRANSACTION 11/4/2020 Joint Sampling • Washed coking coal • Power coal 21

Joint sampling u at loading point - by customer and producer on mutually agreed methods u at both ends - mean value u bonus/penalty to producer for values beyond agreed tolerance limits u requires periodic testing . . . Unfortunately rarely practiced in India 11/4/2020 22

Reasons for discrepancies in results u level of precision not defined u Non-identical procedures for sampling at both ends u manual sampling results in large human error u deviation 11/4/2020 in procedures from agreed one 23

Primary requirements for development of a methodology u testing for estimation of the variances, Vi and Vpt u decision on level of precision of the ash value u calculation for no. of sub-lot and increment /sublot at desired precision from known values of variances u estimation of precision for the existing procedure u estimation of min. mass/ sub-lot form the std. Table u estimation of min. mass/ increment 11/4/2020 24

SAMPLING SCHEME is designed based on the above test The procedure can significantly reduce the discrepancies in the results at both ends 11/4/2020 25

Sampling of washed coking coal Samples are drawn from the u Automatic mechanical Sampler (AMS) u Conveyor Belt For day to day quality monitoring, samples are reduced by offline and/or manual means to analyze ASH & TOTAL MOISTURE 11/4/2020 26

Sampling of Power Coal Best option: AMS at loading/ unloading point AMS for coal x 200 mm or above is not a proper choice to ascertain quality parameter Suggestion: sampling on crushed coal below 50 mm or preferably at 20 mm 11/4/2020 27

Where AMS is non-existing/ non-functioning , sampling may be done for the time being, at loading point from the wagon by manual means Wagon top sampling is difficult, because • segregation occurs because of large size • impractical to collect sample from the full depth • introduces bias due to manual operation Suggestion: smaller size (< 50 mm) of the sample 11/4/2020 28

Periodic tests ¨ Estimate of overall precision ¨ Estimation of increment variance ¨ Estimation of preparation & testing variance If the overall precision is different from the desired one then, number of increment & sub lot to be modified accordingly. 11/4/2020 29

Alternative Long Term ¨ Auto mechanical sampling system at Sampling points ¨ Auto mechanical Augers from the wagons 11/4/2020 30

Alternative Short Term ¨ Estimation of variance of sampling, sample preparation & testing ¨ Design a practical procedure for routine implementation ¨ Perform periodic test for checking 11/4/2020 31

Conclusion: • Choice of Sampling methodology depending on the purpose • Efforts to reduce the sampling variance to a min. possible limit • Sampling on mechanically crushed coal below 50 mm • Preferable size is 20 mm ( feed to most power plants) 11/4/2020 32

Conclusion…………contd. • Replacement of manual sampling method by AMS • In absence of AMS, manual wagon top sampling of this size would give better results • In absence of AMS, manual sampling from wagon top can be done as an temporary option, following the prescribed methodology 11/4/2020 33

Ash Sample Preparation P Gross Sample P Air dry & Crush to 12. 5 mm P Air dry (if required oven dry at 35 deg but not more than 2 h at a time) P Reduce the sample to 2 kg P Divide into 4 parts P 2 parts (Reserve); 1 part (212 micron) & 1 part (212 micron) P 1 part preserve as check sample P 1 part divide using laboratory divider (2 samples) P 11/4/2020 Sample A 1 Sample A 2 34

Total Moisture Sample Preparation Gross Sample Record Initial wt. Air dry Record wt. Stage 1 loss% Y Size 12. 5 mm above ? Crush to 12. 5 mm N Oven dry 35 deg not exceed 2 h Record wt. Stage 2 loss% Reduce 2. 5 kg Crush to 2. 8 mm Divide into 4 parts Sample A 11/4/2020 Stage 3 Moist. 10 g, 108 deg, 2 h Sample B RESERVE 35

Common Sample Preparation Common Bottle Sample Record Initial wt. Air dry Record wt. Stage 1 loss% Oven dry 35 deg not exceed 2 h Record wt. Stage 2 loss% Divide into 2 parts Divide Crush to 2. 8 mm Sample A Crush to 2. 8 mm Divide into 2 parts Sample B RESERVE Stage 3 Moist. 10 g, 108 deg, 2 h 11/4/2020 Sample C Ash 36

THANKS 11/4/2020 37