Interlaboratory Proficiency Test 07/2018

Transkriptio

1 REPORTS OF THE FINNISH ENVIRONMENT INSTITUTE Interlaboratory Proficiency Test 07/2018 Gross an net calorific values in fuels Mirja Leivuori, Minna Rantanen, Eliisa Hatanpää, Riitta Koivikko, Keijo Tervonen, Sari Lanteri an Markku Ilmakunnas Finnish Environment Institute

2

3 REPORTS OF THE FINNISH ENVIRONMENT INSTITUTE Interlaboratory Proficiency Test 07/2018 Gross an net calorific values in fuels Mirja Leivuori, Minna Rantanen, Eliisa Hatanpää, Riitta Koivikko, Keijo Tervonen, Sari Lanteri an Markku Ilmakunnas SYKE Helsinki 2018 Finnish Environment Institute

4 REPORTS OF THE FINNISH ENVIRONMENT INSTITUTE Finnish Environment Institute SYKE Proftest SYKE Layout: Markku Ilmakunnas The publication is also available in the Internet: hela.helsinki.fi/syke ISBN (pbk.) ISBN (PDF) ISSN (print) ISSN (Online) Author(s): Mirja Leivuori, Minna Rantanen, Eliisa Hatanpää, Riitta Koivikko, Keijo Tervonen, Sari Lanteri an Markku Ilmakunnas Publisher an financier of publication: Finnish Environment Institute (SYKE) P.O. Bo 140, FI Helsinki, Finlan, Phone , syke.fi. Year of issue: 2018

5 ABSTRACT Interlaboratory Proficiency Test 07/2018 Proftest SYKE arrange the proficiency test (PT) for measurement the gross an the net calorific value, the content of ash, carbon, nitrogen, hyrogen, moisture, sulphur, an volatile matter in peat, woo pellet (not sulphur) an coal samples in September In total, there were 26 participants in the PT. The participants coul also calculate the emission factor for the peat an coal samples. The robust mean, meian or mean of the reporte results by the participants was use as the assigne value for measurements. The evaluation of performance was base on the z an E n scores. In total, 89 % of the reporte results were satisfactory base on s when the eviations of 1 30 % from the assigne values were accepte. In measurement of the gross calorific value from the peat sample 92 %, from the woo pellet sample 83 % an from the coal sample 88 % of the results were satisfactory. In measurement of the net calorific value from the peat sample 82 %, from the woo pellet 73 % an from the coal sample 79 % of the results were satisfactory. All results evaluate base on E n scores were satisfactory. The evaluation of performance was not one for the measurement of M a in all samples an N in the woo pellet sample. Warm thanks to all the participants of this proficiency test! Keywors: Proficiency test, interlaboratory comparison, coal, peat, woo pellet, calorific value, emission factor, ash, moisture, carbon, sulphur, nitrogen, hyrogen, volatile matter, environmental laboratories TIIVISTELMÄ Laboratorioien välinen pätevyyskoe 07/2018 Proftest SYKE järjesti syyskuussa 2018 pätevyyskokeen kalorimetrisen ja tehollisen lämpöarvon sekä tuhkan, veyn, hiilen, typen, rikin, haihtuvien yhisteien ja kosteuen määrittämiseksi turpeesta, puupelletistä (ei rikkiä) ja kivihiilestä. Lisäksi osallistujilla oli mahollisuus arvioia/laskea turve- ja kivihiilinäytteien päästökerroin. Pätevyyskokeessa oli yhteensä 26 osallistujaa. Vertailuarvona käytettiin osallistujatulosten robustia keskiarvoa, keskiarvoa tai meiaania. Pätevyyen arviointi tehtiin z- ja E n -arvojen avulla. Koko tulosaineistossa hyväksyttäviä tuloksia oli 89 % z-arvolla arvioituna, kun vertailuarvosta sallittiin 1 30 % poikkeama. Kalorimetrisen lämpöarvon tuloksista oli hyväksyttäviä 92 % (turve), 83 % (puupelletti) ja 88 % (kivihiili). Tehollisen lämpöarvon tuloksille vastaavat hyväksyttävien tulosten osuuet olivat 82 % (turve), 73 % (puupelletti) ja 79 % (kivihiili). E n - arvolla arvioiut tulokset olivat kaikki hyväksyttäviä. Tulosten arviointia ei tehty testinäytteien kosteuspitoisuuen määritykselle ja puupelletin typen määrityksille. Kiitos pätevyyskokeen osallistujille! Avainsanat: pätevyyskoe, vertailumittaus, kalorimetrinen lämpöarvo, tehollinen lämpöarvo, päästökerroin, tuhka, kosteus, hiili, rikki, typpi, haihtuvat yhisteet ja vety, turve, puupelletti, hiili, ympäristölaboratoriot SAMMANDRAG Provningsjämförelse 07/2018 Proftest SYKE genomföre i september 2018 en provningsjämförelse som omfattae bestämningen av kalorimetriskt och effektivt värmeväre, svavel, väte, kol, kväve, askhalt, flykthalt och fukthalt i torv, trä pellet (inte svavel) och stenkol. Det var en möjlighet att beräkna emissionfaktor i torv och stenkol prover. Totalt 26 eltagarna eltog i jämförelsen. Som referensväre för analyternas koncentration använes et robusta meelväret, meelväret eller meian av eltagarnas resultat. Resultaten väreraes me hjälp av z och E n vären. I jämförelsen var 89 % av alla resultaten acceptabel väreraes me z vären, när en total eviation på 1 30 % från referensväret tilläts. Av et kalorimetriska värmeväret var 92 % acceptabla (torv), 83 % (trä pellet) och 88 % (stenkol). För resultaten av et effektiva värmeväret var 82 % (torv), 73 % (trä pellet) och 79 % (stenkol) acceptabla. Alla resultaten var acceptabel väreraes me E n vären. Det var inte gjorts värering till fuktighalt i alla prover, beräkning av väte i torv provet och nitrogen i trä pellet. Ett varmt tack till alla eltagarna i testet! Nyckelor: provningsjämförelse, kalorimetriskt och effektivt värmeväre, emissionfaktor, svavel, väte, kol, nitrogen, askhalt, flykthalt fukthalt stenkol, torv, trä pellet, miljölaboratorier

6

7 CONTENTS Abstract Tiivistelmä Sammanrag Introuction Organizing the proficiency test Responsibilities s Samples an elivery Homogeneity Feeback from the proficiency test Processing the ata Pretesting the ata Assigne values Stanar eviation for proficiency assessment an result evaluation Results an conclusions Results Analytical methos Gross an net calorific value Measurement of ash, carbon, hyrogen, moisture, nitrogen, sulphur, an volatile matter Uncertainties of the results Estimation of emission factor Evaluation of the results Summary Summary in Finnish References APPENDIX 1 : s in the proficiency test APPENDIX 2 : Preparation of the samples APPENDIX 3 : Homogeneity of the samples APPENDIX 4 : Feeback from the proficiency test APPENDIX 5 : Evaluation of the assigne values an their uncertainties APPENDIX 6 : Terms in the results tables APPENDIX 7 : Results of each participant APPENDIX 8 : Results of participants an their uncertainties APPENDIX 9 : Summary of the z an E n scores APPENDIX 10 : s in ascening orer APPENDIX 11 : Analytical measurements an backgroun information for calculations APPENDIX 12 : Results groupe accoring to the methos APPENDIX 13 : Eamples of measurement uncertainties reporte by the participants Proftest SYKE CAL 07/18 5

8 6 Proftest SYKE CAL 07/18

9 1 Introuction Proftest SYKE carrie out the proficiency test (PT) for analysis of gross an net calorific value in fuels in September 2018 (CAL 07/2018). In the PT, gross an net calorific value, C, S, H, N, moisture content of the analysis sample (M a, ), ash content as well as volatile matter (V b ) were teste in peat, woo pellet (not S) an coal samples. Finnish Environment Institute (SYKE) is appointe National Reference Laboratory in the environmental sector in Finlan. The uties of the reference laboratory inclue proviing interlaboratory proficiency tests an other comparisons for analytical laboratories an other proucers of environmental information. This proficiency test has been carrie out uner the scope of the SYKE reference laboratory an it provies an eternal quality evaluation between laboratory results, an mutual comparability of analytical reliability. The proficiency test was carrie out in accorance with the international guielines ISO/IEC17043 [1], ISO [2] an IUPAC Technical report [3]. The Proftest SYKE is accreite by the Finnish Accreitation Service (FINAS) as a proficiency testing provier (PT01, ISO/IEC 17043, This proficiency test has been carrie out uner the accreitation scope of the Proftest SYKE. 2 Organizing the proficiency test 2.1 Responsibilities Organizer: Proftest SYKE, Finnish Environment Institute (SYKE), Laboratory Centre Ultramariinikuja 4 (formerly Hakuninmaantie 6), FI Helsinki, Finlan Phone: , proftest@environment.fi The responsibilities in organizing the proficiency test were as follows: Mirja Leivuori coorinator Riitta Koivikko substitute of coorinator Keijo Tervonen technical assistance Markku Ilmakunnas technical assistance Sari Lanteri technical assistance Co-operation partner an analytical epert: Eliisa Hatanpää, Eurofins Environment Testing Finlan Lt, Vantaa (formerly Ramboll Finlan Lt, Ramboll Analytics), eliisahatanpää@eurofins.fi. Also Minna Rantanen was the analytical epert in this PT. Proftest SYKE CAL 07/18 7

10 Subcontracting: The peat, woo pellet an coal samples were homogenate an ivie into sub-samples at the laboratory of KVVY Tutkimus Oy (Tampere, Finlan, T064 accreite by FINAS, Samples were teste by Eurofins Environment Testing Finlan Lt, Vantaa (T039 accreite by FINAS, s In total 26 participants took part in this proficiency test, of which 12 were from Finlan an 14 from abroa (Appeni 1). Altogether 69 % of the participants use accreite analytical methos at least for a part of the measurements. The samples were teste at the laboratory of Eurofins Environment Testing Finlan Oy, Vantaa an their participant coe is 6 in the result tables. 2.3 Samples an elivery Three ifferent fuel samples were elivere to the participants: peat (B1), woo pellet (B2) an coal (K1) samples. Gross (q V,gr, ) an net (q p,net, ) calorific value, C, S, H, N, moisture content of the analysis sample (M a, ), ash content as well as volatile matter (V b ) were teste in peat, woo pellet (not S) an coal samples. The material for the peat sample (B1) was collecte from the Finnish marshlan. The material was air rie an groun by the mill with 500 µm sieve before homogenization an sample iviing. The peat sample was prepare by Eurofins Labtium Lt in Jyväskylä (Finlan). The woo pellet sample (B2) was provie by Vapo an it was pre-treate (grining) by Eurofins Labtium Lt. The raw material for woo pellets was spruce sawust. The material was first crushe with a cutting mill an then groune by the mill with 1000 µm sieve before homogenization an sample iviing. The woo pellet sample was prepare by Eurofins Labtium Lt in Jyväskylä (Finlan). The coal sample (K1) was prepare from Russian steam coal by the Helen Lt (Finlan). All samples were homogenize an ivie into sub-samples at the laboratory of KVVY Tutkimus Oy. The sample preparation is escribe in etails in the Appeni 2. In the cover letter elivere with the samples, the participants were instructe first to store the samples close for one ay after their arrival an then to measure the moisture content of the analysis sample (M a ) as the first measurement. The samples were instructe to be homogenize before measurements an to be store in a ry place at room temperature. Further, the moisture content of the analysis sample was instructe to be measure on every ay of measurements. This was important as it eliminates the influence of humiity on the measurements. 8 Proftest SYKE CAL 07/18

11 s coul also estimate/calculate the emission factor (as receive), EF, for peat an coal samples. For this estimation/calculation the total moisture contents of the samples as receive (M ar ) were given: peat B %, coal K % The samples were elivere to the participants on 3 September The samples arrive to the participants mainly latest on 6 September One participant informe the arrival of the samples on 12 September 2018, but the tracking system of the elivery showe the sample arrival at the service point on 7 September The samples were requeste to be measure an the results to be reporte latest on 25 September All the results were reporte accoringly. The preliminary results were elivere to the participants via ProftestWEB an on 1 October Homogeneity Homogeneity of the samples B1, B2 an K1 was teste by measuring the gross an net calorific value an ash content as uplicate eterminations from five subsamples (Appeni 3). Moreover, the other measurans were teste from two subsamples as uplicate measurements. Accoring to the homogeneity test results, all samples were consiere homogenous. Particle size istribution was also teste from one sub sample of peat (B1) an coal (K1). The requirement of particle sizes given in the international stanars was fulfille (Appeni 2). 2.5 Feeback from the proficiency test The feeback from the proficiency test is shown in Appeni 4. The comments from the participants mainly ealt with sample elivery an participants reporting errors. The comments from the provier are mainly focuse to the lacking conversancy to the given information with the samples. All the feeback is valuable an is eploite when improving the activities. 2.6 Processing the ata Pretesting the ata The normality of the ata was teste by the Kolmogorov-Smirnov test. The outliers were rejecte accoring to the Grubbs or Hampel test before calculating the mean. Also before the statistical results hanling some outliers were rejecte in cases, where the result iffere from the ata more than s rob 5 or 50 % from the robust mean. The rejection of results was partly base on the rather strict requirements for the reproucibility given in the stanars for analysis escribe in the cover letter of the samples. The uplicate results were teste using the Cochran test. If the result was reporte lower than the limit of etermination, it has not been inclue in calculations. Proftest SYKE CAL 07/18 9

12 More information about the statistical hanling of the ata is available in the Guie for participant [4] Assigne values Mainly the robust mean value of the participant results was use as the assigne value for measurans of the test samples, when there were at least 12 results (n(stat)12). In calculation of the robust mean the outliers are normally not rejecte, but they are iterate before the final calculation of the robust mean. Also the mean an the meian values of the ata were calculate an they mainly iffere only slightly from the robust means use as the assigne values (Table 1). In cases, where the number of results was lower than 12, the meian of participants results was use as the assigne value: the peat sample B1 all measurans with the eception of Ash (robust mean) an H (mean value). The meian was use as the assigne value for the woo pellet sample B2 measurans: C, H, N, q p,net,, V p an for the coal sample K1 measurans: N, q p,net,. The assigne value of emission factor EF in the peat sample B1 was base on the meian value of the results. For nitrogen (N ) in the pellet sample (B2) the informative assigne value is given, but ue to the high eviation of results the performance evaluation was not one. In cases, where the number of results was less than 6 (n(stat)<6), the performance evaluation was one using E n score, if the assigne value an its uncertainty was set i.e. for emission factor (EF) for the peat sample (B1). When the robust mean was use as the assigne value, the uncertainty was calculate using the robust stanar eviation. When the meian or the mean value was use as the assigne value, the epane uncertainty was estimate base on the stanar eviation [2, 4]. When using the robust mean, the mean or the meian of the participant results as the assigne value, the epane uncertainties of the assigne values for calorific values were between 0.2 % an 0.3 %. For the other evaluate measurans the uncertainty varie from 0.4 % to 7.8 % (Appeni 5). After reporting the preliminary results no changes have been one for the assigne values Stanar eviation for proficiency assessment an result evaluation The requirements for the reproucibility of the use stanar methos were informe in the cover letter of the samples an they were use for estimation of stanar eviation for proficiency assessment in this PT. The reproucibility require for the stanar methos was mainly fulfille for gross calorific values. The stanar eviation for the proficiency assessment (2 s pt at the 95 % confience level) was set to 1 30 % epening on the measurements. Stanar eviation for proficiency assessment was not set for analysis moisture content M a, (all samples), for N in the woo pellet sample (B2) an for EF in the peat sample (B1), an thus the results have not evaluate. After reporting the preliminary results the stanar eviation of the proficiency assessment (s pt ) has been cross-checke for S in the coal (K1) sample an change from 17 to 15 %. 10 Proftest SYKE CAL 07/18

13 This cause minor numerical changes for values, but no changes to the participants performance evaluation. For other measurans an samples no changes have been one for the stanar eviations of the proficiency assessment. Aitionally, when the number of reporte results was low an the uncertainty was set for the assigne value, an the participant reporte measurement uncertainty, the performance was estimate by means of E n scores ( Error, normalize, Appeni 9). These are use to evaluate the ifference between the assigne value an participant s result within their claime epane uncertainty. E n scores are calculate: ( ) =, where i = participant s result, pt = assigne value, U i = the epane uncertainty of a participant s result an U pt = the epane uncertainty of the assigne value. E n scores of -1.0 < E n < 1.0 shoul be taken as an inicator of successful performance when the uncertainties are vali. Whereas scores E n 1.0 or E n -1.0 coul inicate a nee to review the uncertainty estimates, or to correct a measurement issue. The reliability of the assigne values was teste accoring to the criterion u pt / s pt 0.3, where u pt is the stanar uncertainty of the assigne value an s pt is the stanar eviation for proficiency assessment [3]. When testing these reliabilities the criterion was mainly fulfille an the assigne values were consiere reliable. The reliability of the stanar eviation for proficiency assessment an the corresponing was estimate by comparing the eviation for proficiency assessment (s pt ) with the robust stanar eviation (s rob ) or stanar eviation (s) of the reporte results [3]. The criterion s rob (or s)/ s pt < 1.2 was mainly fulfille. In the following cases, the criterion for the reliability of the assigne was not met an, therefore, the evaluation of the performance is reuce in this proficiency test: Sample B1 K1 Measuran N V p 3 Results an conclusions 3.1 Results The summary of the results of this proficiency test is presente in Table 1. Eplanations to terms use in the result tables are presente in Appeni 6.The results an the performance of each participant are presente in Appeni 7. The reporte results with their epane uncertainties (k=2) are presente in Appeni 8. The summaries of the z an E n scores are shown in Appeni 9 an the s in the ascening orer in Appeni 10. If the participant Proftest SYKE CAL 07/18 11

14 i not report the requeste parallel results for measurans, the evaluation scores are not available. When neee the participant can calculate their own s [4]. The robust stanar eviations an stanar eviations of the results varie from 0.3 to 13.3 % (Table 1). The robust stanar eviation or stanar eviation was lower than 2 % for 42 % of the results an lower than 6 % for 92 % of the results (Table 1). For Ash an S the robust stanar eviation of the results was higher than 6 % (B2 an B1, respectively, Table 1). The robust stanar eviations an stanar eviations were approimately within the same range as in the previous similar proficiency test Proftest SYKE CAL 7/2017, where the eviations varie from 0.3 % to 30.7 % [5]. Table 1. The summary of the results in the proficiency test CAL 07/2018. Measuran Sample Unit Assigne value Mean Rob. mean Meian srob srob % 2 spt % n (all) Acc z % Ash B1 w% B2 w% K1 w% C B1 w% B2 w% K1 w% EF B1 t CO2/TJ K1 t CO2/TJ H B1 w% B2 w% K1 w% Ma, B1 w% B2 w% K1 w% N B1 w% B2 w% K1 w% qp,net, B1 J/g B2 J/g K1 J/g qv,gr, B1 J/g B2 J/g K1 J/g S B1 w% K1 w% Vb B1 w% B2 w% K1 w% Rob. mean: the robust mean, s rob : the robust stanar eviation, s rob %: the robust stanar eviation as percent, 2 s pt %: the stanar eviation for proficiency assessment at the 95 % confience level, Acc z %: the results (%), where z 2, n(all): the total number of the participants. 12 Proftest SYKE CAL 07/18

15 In this proficiency test the participants were requeste to report replicate results for all measurements. The results of the replicate eterminations base on the ANOVA statistics are presente in Table 2. The targets for the repeatability are the ones recommene in the international stanars or technical specifications relate to the measurements of fuels.in particular, in measurements of the calorific values, the requirement for the repeatability is ± 120 J/g. In this proficiency test the requirements for the repeatability of the measurements of the gross calorific value were 0.55 % for the sample B1, 0.59 % for the sample B2 an 0.42 % for the sample K1 an in measurement of the net calorific value 0.58 %, 0.64 % an 0.43 %, respectively. In each case, the obtaine repeatability of the measurement of the gross calorific value an the net calorific value was lower than the repeatability requirement (Table 2, the column s w %). The estimation of the robustness of the methos coul be one by the ratio s b /s w. The ratio s b /s w shoul not ecee the value 3 for robust methos. Here, however, the robustness eceee the value 3 in many cases (Table 2). For the gross calorific value, the ratio s b /s w, was 1.9 (the sample B1), 1.8 (the sample B2) an 5.5 (the sample K1), for the net calorific values 3.4, 3.8 an 5.5, respectively. For the calorific values the ratio s b /s w was mainly within the same range than in the previous similar proficiency test CAL 07/2017, with the eception of somewhat lower values for the peat (B1) an woo pellet (B2) an higher ratio in the coal sample (K1) [5]. Table 2. The summary of repeatability on the basis of replicate eterminations (ANOVA statistics). Measuran Sample Unit Assigne value Mean sw sb st sw% sb% st% sb/sw Ash B1 w% B2 w% K1 w% C B1 w% B2 w% K1 w% EF B1 t CO2/TJ K1 t CO2/TJ H B1 w% B2 w% K1 w% Ma, B1 w% B2 w% K1 w% N B1 w% B2 w% K1 w% qp,net, B1 J/g B2 J/g K1 J/g qv,gr, B1 J/g B2 J/g K1 J/g S B1 w% K1 w% Vb B1 w% B2 w% K1 w% Ass.val.: assigne value; s w : repeatability stanar error; s b : between participants stanar error; s t : reproucibility stanar error. Proftest SYKE CAL 07/18 13

16 3.2 Analytical methos The participants were allowe to use ifferent analytical methos for the measurements in the PT. A questionnaire relate to the use analytical methos was carrie out along the proficiency test. The summary of the answers is shown in Appeni 11. The use analytical methos an the results of the participants groupe by methos are shown in more etail in Appeni 12. The statistical comparison of the analytical methos was possible for the ata where the number of the results was 5 (several cases in this PT). In those cases the comparison is base on the graphical result evaluation Gross an net calorific value The analytical methos base on ifferent stanar methos were use for the measurements in this PT. The use analytical methos of the participants are shown in more etail in Appenices Mostly, stanar methos were use for measurement of calorific values (q V,gr,, q p,net, ) (EN [6], EN ISO [7], ISO 1928 [8], Appeni 12). One participant use stanar ASTM D 5865 [9]. One participant (13) use other stanar metho (EN 15400) an one reporte to use isoperibolic calorimeter (participant 22). In the calculations of gross calorific value (q V,gr, ), various correction factors were use. Fuse wire, ignition, aci, moisture, nitrogen an sulphur corrections were most commonly use in several ifferent combinations epening of the test material (Appeni 11). For the calculation of net calorific value (q p,net, ), ifferent combinations of correction factors were use as well epening of the test material (Appeni 11). Mainly nitrogen plus oygen (N+O) an hyrogen (H) content was use for corrections. Base on the statistical comparison an the graphical evaluation no ifferences between the use methos in gross an net calorific value measurements coul be conclue (Appeni 12) Measurement of ash, carbon, hyrogen, moisture, nitrogen, sulphur, an volatile matter In the PT mainly the following stanar methos or technical specifications were use for measurements of ifferent parameters: Measuran Metho Ash EN [10], ISO 1171 [11], EN ISO [12], ASTM D 7582 [13] C, H an N ISO [14], ASTM D 5373 [15], EN ISO [16] Ma (analytical moisture content) EN [17], ISO 589 [18], DIN [19], ASTM D 7582 [13], EN ISO [20], ISO [21] S EN ISO [22], ASTM D 4239 [23] Vb, (volatile matter) EN [24], ISO 562 [25], EN ISO [26] 14 Proftest SYKE CAL 07/18

17 However, in some cases also other international an national stanars or technical specifications (e.g. EN 15403, ASTM D 7582, EN 15414, EN 15402, ISO 19579) or internal methos (e.g. participants 1, 22, 15) were use. The ash content was etermine mainly gravimetrically by heating at the temperature 550 C (Samples B1 an B2), at the temperature 750 C or 815 C () or at the temperature 950 C (Samples B2 an K1). Ash content was measure also using TGA for samples at the temperatures between 550 C an 815 C (Appeni 11). In the international stanars EN an EN ISO the ashing temperature is mentione to be 550 C for soli biofuels [10, 12]. While in ISO 1711 for soli mineral fuels it is mentione to be 815 C [11]. Base on the graphical result evaluation, clear ifferences between the use methos in measurements coul not be conclue (Appeni 12). Moisture content was etermine gravimetrically by heating in air or N 2 atmosphere at the temperatures of C. Moisture content was measure also using TGA at the temperatures of C (Appeni 11). Most of the participants conucte CHN analyses from air rie samples, one participant use rie B1 sample, an two participants use rie B2 an K1 samples (Appeni 11). Base on the graphical result evaluation, clear ifferences between the use methos in CHN measurements coul not be conclue (Appeni 12). Also for S an V b no clear ifferences between the use methos were notice (Appeni 12). In the PT also information of etection limits for nitrogen an sulphur was collecte (Appeni 11). The etection limits varie for N: w% an for S: w%. 3.3 Uncertainties of the results At maimum 88 % of the participants reporte the epane uncertainties (k=2) with their results for at least some of their results (Table 3, Appeni 13). The range of the reporte uncertainties varie between the measurements an the sample types. Several approaches were use for estimating of measurement uncertainty (Appeni 13). The most use approaches were base on IQC ata an metho valiation ata. Three participants reporte the usage of the MUkit measurement uncertainty software for the estimation of their uncertainties [27]. The free software is available on the webpage: Generally, the use approach for estimating measurement uncertainty i not make efinite impact on the uncertainty estimates. The estimate uncertainties varie highly for all the teste measurans (Table 3). Especially, very low or high uncertainties can be consiere questionable. It was evient, that some uncertainties ha been reporte erroneously for the measurans (incluing calorific values, Appeni 13), not as relative values as the provier of this proficiency test ha requeste. It is evient that harmonization is still neee for the estimation of the epane measurement uncertainties. Proftest SYKE CAL 07/18 15

18 Table 3. The range of the epane measurement uncertainties (k=2, U i %) reporte by the participants. Measurement Uncertainty B1,% Uncertainty B2, % Uncertainty K1, % Ash C EF H N qp,net, qv,gr, S Vb Estimation of emission factor Aitionally, the participants were aske to estimate the emission factors for the peat an coal samples istribute in the PT by taking into account their own net calorific values an the total moisture values as receive, which was informe in the cover letter of the samples. The calculation of the emission factor of the woo pellet sample (B2) was not one as it is a CO 2 neutral fuel. In this PT, very few participants reporte their results for the emission factor (5-8). Due to the low number of the reporte results, the peat sample (B1) was evaluate base on E n score (Appeni 9). 4 Evaluation of the results The evaluation of participants was base on the s an E n scores, which were interprete as follows: Criteria Performance z 2 Satisfactory 2 < z < 3 Questionable z 3 Unsatisfactory -1.0 < En < 1.0 Satisfactory En or En 1.0 Unsatisfactory In total, 89 % of the results evaluate base on s were satisfactory when accepting the eviation of 1 30 % from the assigne value (Appeni 9). All results evaluate base on E n scores were satisfactory (Appeni 9). About 69 % of the participants use the accreite methos an 92 % of their results were satisfactory. In the previous similar proficiency test CAL 07/2017 the performance was satisfactory for 89 % of the results when eviation 1 30 % from the assigne value was accepte [5]. 16 Proftest SYKE CAL 07/18

19 Table 4. Summary of the performance evaluation in the proficiency test 07/2018. Sample Satisfactory Accepte eviation from Remarks results (%) the assigne value (%) Peat, B Very goo performance. Only approimate assessment for N. In the CAL 07/17 the performance was satisfactory for 100 % of the results, when accepting % eviation from the assigne value [5]. Woo pellet, B Difficulties in measurements for C an qp,net, < 80% satisfactory results. In the CAL 07/17 the performance was satisfactory for 85 % of the results, when accepting % eviation from the assigne value [5]. Coal, K Difficulties in measurements for qp,net, < 80 % satisfactory results. Only approimate assessment for Vb. In the CAL 07/17 the performance was satisfactory for 88 % of the results [5]. The summary of the performance evaluation is shown in Table 4. The percentage of the satisfactory results varie between 83 % an 95 % for the teste sample types. The criteria for performance evaluation is mainly set accoring to the target value for reproucibility recommene in international stanars or technical specifications for measurement of the calorific values an other eterminants. The reproucibility require in the stanars was fulfille for the gross calorific values. For the net calorific value increase reproucibility from the value for the gross caloric value was use. There was no criterion for reproucibility for the net calorific value in stanars methos. Peat In the previous similar PT (CAL 07/2017) 100 % of the results were satisfactory for the peat sample (B1) when accepting % eviation from the assigne value [5]. In this PT the number of satisfactory results is slightly lower (95 %, Table 4). The number of satisfactory results of the gross an net calorific values for peat sample was lower for the gross calorific value an the net calorific value when compare to the previous similar PT (100 % for both) [5]. The results of analysis moisture (M a ) have not been evaluate, but the assigne values are presente (Table 1). The results of EF were evaluate base on the E n scores, which were all satisfactory (Appeni 9). Woo pellet In the previous similar PT CAL 07/2017 the satisfactory results of the woo pellet sample (B2) were in total 85 %, when accepting eviation % from the assigne value [5], thus the performance in this PT was slightly lower (83 %, Table 4). The satisfactory results varie between 73 % (C, q p,net, ) an 95 % (Ash ) for the woo pellet sample (Table 1). In the measurement of gross an net calorific values 83 % an 73 % of the results, respectively, were satisfactory when accepting eviations of 1.4 % an 1.7 % from the assigne values (Table 1). The number of satisfactory results of the gross an net calorific values for woo pellet was at Proftest SYKE CAL 07/18 17

20 the same level for the gross calorific value an lower for the net calorific value than in the previous similar PT CAL 07/2017 (83 % for the both) [5]. The estimation of EF was not one as it is a CO 2 neutral fuel. Also the results of analysis moisture (M a ) an nitrogen (N ) have not been evaluate, but the assigne value is given (Table 1). Coal In the previous similar PT CAL 07/2017 the satisfactory results of the coal sample (K1) were in total 88 % [5], thus the performance was at the same level in this PT (89 %, Table 4). In the measurement of gross an net calorific values, 88 % an 79 % of results, respectively, were satisfactory, when accepting the eviations of 1 an 1.2 % from the assigne values (Table 1). These were lower than in the previous similar PT CAL 07/2017 (94 % an 92 %, respectively) [5]. From the calculate emission factor results 88 % were satisfactory an it was lower than in the previous similar PT CAL 07/2017 (100 %) [5]. The results of analysis moisture (M a ) have not been evaluate, but the assigne value is given (Table 1). 5 Summary Proftest SYKE carrie out the proficiency test (PT) for the analysis of the gross an the net calorific value as well as for content of ash, carbon, hyrogen, nitrogen, sulphur, analytical moisture content an volatile matter in fuels in September Three types of samples were elivere to the participants: peat, woo pellet (not sulphur) an coal. In total 26 participants took part in the PT. The participants also ha the possibility to estimate or calculate the emission factor for peat an coal samples. The robust means, meians or means of the results reporte by the participants were use as the assigne values for measurans. The uncertainty for the assigne value was estimate at the 95 % confience level an it was less than 0.4 % for calorific values an at maimum 8 % for the other measurans. The evaluation of the performance was base on the s, which were calculate using the stanar eviation for proficiency assessment at 95 % confience level. In some cases the number of the reporte results was low an the performance was evaluate by using E n scores (EF in the peat sample). The evaluation of performance was not one for the measurement of M a in all samples an N in the woo pellet sample. In this proficiency test 89 % of the ata was regare to be satisfactory when, epening on the measuran an sample, the result was accepte to eviate from the assigne value from 1 to 30 %. About 69 % of the participants use the accreite methos an 92 % of their results were satisfactory. In measurements of the gross calorific value from the peat, woo pellet an coal samples, 92 %, 83 % an 88 % of the results were satisfactory, respectively. In measurements of the net calorific value from the peat, woo pellet an coal samples, 82 %, 73 % an 79 % of the results were satisfactory, respectively. In general, the results were in the same range as in the previous similar Proftest SYKE proficiency test, CAL 07/2017 [5], but the performance in the gross an net calorific 18 Proftest SYKE CAL 07/18

21 value was somewhat lower for peat an coal samples an also for the net calorific value for woo pellet sample in the present PT. The evaluation of ata base on E n scores for the peat sample show satisfactory performance for all results. 6 Summary in Finnish Proftest SYKE järjesti syyskuussa 2018 pätevyyskokeen kalorimetrisen ja tehollisen lämpöarvon sekä tuhkan, hiilen, veyn, typen, rikin, kosteuen ja haihtuvien yhisteien määrittämiseksi turpeesta, puupelletistä (ei rikkiä) ja kivihiilestä. Lisäksi osallistujilla oli mahollisuus laskea päästökerroin turve- ja kivihiilinäytteistä. Pätevyyskokeeseen osallistui yhteensä 26 laboratoriota. Osallistujien pätevyyen arviointi tehtiin z-arvon avulla ja sen laskemisessa käytetyn kokonaishajonnan tavoitearvot olivat määrityksestä ja näytteestä riippuen välillä 1 30 %. Turvenäytteen päästökerrointulokset arvioitiin käyttäen E n -arvoa tulosten vähyyen vuoksi. Testisuureen vertailuarvona käytettiin osallistujien ilmoittamien tulosten robustia keskiarvoa tai niien meiaania ja keskiarvoa, jos tuloksia oli vähän (n<12). Vertailuarvon epävarmuus oli lämpöarvomäärityksissä alhaisempi kuin 0,4 % ja muien määritysten osalta korkeintaan 8 %. Tulosten arviointia ei tehty testinäytteien kosteuspitoisuuen määritykselle eikä typen määritykselle puupelletistä. Koko tulosaineistossa hyväksyttäviä tuloksia oli 89 %, kun vertailuarvosta sallittiin 1 30 % poikkeama. Noin 69 % osallistujista käytti akkreitoituja määritysmenetelmiä ja näistä tuloksista oli hyväksyttäviä 92 %. Kalorimetrisen lämpöarvon tuloksista oli hyväksyttäviä 92 % (turve), 73 % (puupelletti) ja 79 % (kivihiili). Tehollisen lämpöarvon tuloksille vastaavat hyväksyttävien tulosten osuuet olivat 92 % (turve), 83 % (puupelletti) ja 88 % (kivihiili). Hyväksyttäviä tuloksia oli lähes saman verran kuin eellisessä vastaavassa pätevyyskokeessa CAL 07/2017 [5]. Turve- ja hiilinäytteen osalta kalorimetrisen ja tehollisen lämpöarvon menestyminen sekä puupellettinäytteen tehollisen lämpöarvon menestyminen olivat jonkin verran alhaisempia kuin eellisellä kierroksella. E n -arvolla arvioiut turvenäytteen päästökertoimen tulokset olivat kaikki hyväksyttäviä. Proftest SYKE CAL 07/18 19

22 REFERENCES 1. SFS-EN ISO 17043, Conformity assessment General requirements for Proficiency Testing. 2. ISO 13528, Statistical methos for use in proficiency testing by interlaboratory comparisons. 3. Thompson, M., Ellison, S. L. R., Woo, R., The International Harmonize Protocol for the Proficiency Testing of Analytical Chemistry laboratories (IUPAC Technical report). 4. Proftest SYKE Guie for laboratories: Running proficiency test ( 5. Leivuori, M., Rantanen, M., Koivikko, R., Tervonen, K., Lanteri, S., Ilmakunnas, M., Proficiency test 07/2017. Gross an net calorific value in fuels. Reports of Finnish Environment Institute 35/ pp. ( 6. EN 14918, Soli Biofuels. Metho for the etermination of calorific value. 7. EN ISO Soli biofuels- Determination of calorific value. 8. ISO 1928, Soli mineral fuels - Determination of gross calorific value by a bomb calorimetric metho, an calculation of net calorific value. 9. ASTM D Test metho for gross calorific value of coal an coke. 10. EN 14775, Soli biofuels. Determination of ash content (withrawn). 11. ISO 1171, 2010 Soli mineral fuels - Determination of ash. 12. EN ISO 18122, Soli biofuels - Determination of ash content. 13. ASTM D 7582, Stanar Test Methos for Proimate Analysis of Coal an Coke by Macro Thermogravimetric Analysis. 14. ISO 29541, Soli mineral fuels - Determination of total carbon, hyrogen an nitrogen content - Instrumental methos. 15. ASTM D 5373, Stanar Test Methos for Instrumental Determination of Carbon, Hyrogen, an Nitrogen in Laboratory Samples of Coal an Coke. 16. EN ISO 16948, Soli biofuels - Determination of total content of carbon, hyrogen an nitrogen. 17. EN , Soli biofuels. Methos for the etermination of moisture content. Oven ry metho. Part 3: Moisture in general analysis sample (withrawn). 20 Proftest SYKE CAL 07/18

23 18. ISO 589, Har coal - Determination of total moisture. 19. DIN 51718, Determining the moisture content of soli fuels. 20. EN ISO , Soli biofuels - Determination of moisture content - Oven ry metho - Part 3: Moisture in general analysis sample. 21. ISO 11722, Soli mineral fuels - Har coal - Determination of moisture in the general analysis test sample by rying in nitrogen. 22. EN ISO 16994, Soli biofuels - Determination of total content of sulfur an chlorine. 23. ASTM D 4239, Stanar Test Methos for Sulfur in the Analysis Sample of Coal an Coke Using High - Temperature Combustion an Infrare Absorption. 24. EN 15148, Biofuels, Soli fuels, Biomass, Fuels, Chemical analysis an testing, Volatile matter etermination, Gravimetric analysis (withrawn). 25. ISO 562, Har coal an coke - Determination of volatile matter. 26. EN ISO 18123, Soli biofuels - Determination of the content of volatile matter. 27. Näykki, T., Virtanen, A. an Leito, I., Software support for the Nortest metho of measurement uncertainty evaluation. Accre. Qual. Assur. 17: Mukit website: Magnusson, B. Näykki. T., Hovin, H. an Krysell, M., Hanbook for Calculation of Measurement Uncertainty in Environmental Laboratories. NT Technical Report 537. Nortest. 29. Ellison, S., L., R. an Williams, A. (Es). (2012) Eurachem/CITAC guie: Quantifying Uncertainty in Analytical Measurement, Thir eition, ISBN ISO/IEC Guie 98-3:2008. Uncertainty of measurement - Part 3: Guie to the epression of uncertainty in measurement (GUM: 1995). Proftest SYKE CAL 07/18 21

24 APPENDIX 1 (1/1) APPENDIX 1: s in the proficiency test Country Bulgary Czech Republic Estonia Finlan France Lithuania Republic of Irelan Republic of Korea Romania Spain Sween AES-3C Maritza East 1 EOOD; Testing Laboratory "Energy Materials" ALS Czech Republic s.r.o. Enefit Energiatootmine AS Chemical Laboratory Eurofins Ahma Oy, Oulu Eurofins Environment Testing Finlan Oy, Vantaa, Inustry an Power Plant Chemistry Eurofins Labtium Oy, Jyväskylä Finnsementti Oy Fortum Waste Solutions Oy, Riihimäki Helen Lt Kuopion Energia Oy / Tuotanto-osasto KVVY-Botnialab, Vaasa Kymen Ympäristölaboratorio Oy Kymenlaakson ammattikorkeakoulu Luonnonvarakeskus Kokkolan laboratorio SSAB Europe Raahe, Raahe ArcelorMittal Fos sur Mer SOCOR Dechy France Aioma servisas, Biofuel researh laboratory Cement testing laboratory Co Akmenes cementas Lithuania Eenerry Power Lt Intertek KIMSCO Ulsan Testing Center, South Korea Holcim(Romania) SA Ciment Campulung ICSI Ramnicu Valcea Romcontrol Laboratorio Central e Calia - LCC RISE Research Institutes of Sween AB 22 Proftest SYKE CAL 07/18

25 APPENDIX 2 (1/2) APPENDIX 2: Preparation of the samples Sample B1, peat Sample B1 was prepare from peat taken from Finnish marshlan. The peat was air-rie (35 ºC) an groun in a mill with a 500 µm sieve at the laboratory of Eurofins Labtium Lt (Jyväskylä, Finlan). The rie an sieve sample was mie by a mechanize sample mier an istribute to sub-samples of ca. 30 g using a rotary sample ivier equippe with a vibratory sample feeer at the laboratory of KVVY Tutkimus Oy (Tampere). The particle size istribution of peat was measure by the laboratory of Eurofins Labtium Lt using laser iffraction (Malvern)., woo pellet was prepare from spruce sawust. The woo pellets were first crushe with a cutting mill an then groun by the mill with 1000 µm sieve at the laboratory of Eurofins Labtium Lt. The sieve sample was mie by a mechanize sample mier an istribute to subsamples of ca. 30 g using a rotary sample ivier equippe with a vibratory sample feeer at the laboratory of KVVY Tutkimus Oy (Tampere)., steam coal fuel was a Russian steam coal. The coal was rie at room temperature an groun to particle size < 212 µm at the the Helen Lt (Finlan). The rie an sieve sample was mie by a mechanize sample mier an istribute into subsamples of ca. 30 g using a rotary sample ivier equippe with a vibratory sample feeer at the laboratory the laboratory of KVVY Tutkimus Oy (Tampere). The particle size istribution of coal was measure by the Helen Lt, Power Plant Chemistry using laser iffraction (Malvern). Proftest SYKE CAL 07/18 23

26 APPENDIX 2 (2/2) Particle size To test the particle size of peat (B1) an coal (K1) samples, they were teste using laser iffraction (Malvern). Figure 1 is showing the istribution of particle size for the samples B1 an K1. For peat sample B1 the mean size of particles was 92 µm an ca % of the particles were smaller than 550 µm. For coal sample K1 the mean size of particles was 60.9 µm an 99 % of the particles were smaller than 212 µm. The requirements of particle sizes given in the international stanars were mainly fulfille for the teste material [6, 8]. a) The particle size istribution of peat B1. b) The particle size istribution of coal K1. Figure 1. The particle size istribution of the fuel samples a) the peat (B1) an b) the coal (K1) sample. 24 Proftest SYKE CAL 07/18

27 APPENDIX 3 (1/1) APPENDIX 3: Homogeneity of the samples Homogeneity was teste from uplicate measurements of calorific value (Table 1) an ash content in five samples, which were homogenise before sampling. Aitionally, the other measurans from two samples was teste. Criteria for homogeneity: s anal /s h <0.5 an s sam 2 <c, where s h % = stanar eviation for testing of homogeneity s anal = analytical eviation, stanar eviation of the results within sub samples s pt % = stanar eviation for proficiency assessment s sam = between-sample eviation, stanar eviation of the results between sub samples c = F1 s all 2 + F2 s anal 2, where s all 2 = (0.3 s h ) 2, F1 an F2 are constants of F istribution erive from the stanar statistical tables for the teste number of samples [2, 3]. Table 1. Results from the homogeneity testing of the peat (B1), pellet (B2) an coal (K1) samples. Measurements Mean sh% spt% sh sanal sanal/sh Gross calorific value, J/g Net calorific value, J/g Gross calorific value, J/g Net calorific value, J/g Gross calorific value, J/g Net calorific value, J/g Peat (B1) Is sanal/sh<0.5? ssam ssam 2 c Is ssam 2 <c? yes yes yes yes Pellet (B2) yes yes yes yes Coal (K1) yes yes yes yes Conclusion: In each case, the criteria were fulfille. Thus, all the samples coul be regare as homogenous. Also the results of the other teste measurans confirm the homogeneity of the samples. Proftest SYKE CAL 07/18 25

28 APPENDIX 4 (1/1) APPENDIX 4: Feeback from the proficiency test FEEDBACK FROM THE PARTICIPANTS Comments on technical eecution Action / Proftest SYKE 1 The participant informe receiving the samples on 10 th September. 22 The participant informe receiving the samples on 12 th September. The use istributor (Posti) i not eliver the samples accoring to the agree scheule. Accoring to the istributor's (Posti) tracking system the samples arrive to the participant on 7 th September. The provier recommens to check the internal package elivery proceures. Comments to the results Action / Proftest SYKE 5 The participant reporte only one result (Ash, Ma ), though replicate results were requeste. The provier recommens the participant to follow the given guielines. 14, 24 The participant i not eliver the results to Proftest The provier accepte the results. SYKE by selecting Sen results on ProftestWEB. 23 The participant reporte erroneously their results of The provier oes not correct the results gross calorific value for net calorific value. after elivering the preliminary results. Their correct values were: The erroneous results were hanle as B1: J/g, J/g outliers in the statistical treatment. They B2: J/g, J/g K1: J/g, J/g. i not affect to the assigne value evaluation. If the gross calorific values value ha been reporte correctly they woul have been satisfactory, with eception of coal sample. The participant can re-calculate the s accoring to the Guie for participants [4]. 14 The participant contacte the provier ue to their performance in the PT. The participant requeste further information for the carbon an calorific value results. The provier iscusse with the participants about their performance, but clear eplanation for the reason was not conclue. FEEDBACK TO THE PARTICIPANTS Comments 5 The participant reporte only one result instea of replicate results for some measurans. The results have been eclue from the calculation of the assigne values, an results are not evaluate. The participants shoul follow more carefully the instructions given by the provier. 1, 6, 13, 14, 17, 22, 25 For these participants the eviation of replicate measurements for some measurans an samples was high an their results were Cochran outliers. The provier recommens the participants to valiate their accepte eviation of replicate measurements. 4, 12, 14, 22 It was evient, that some uncertainties ha been reporte erroneously for the measurans (incluing calorific values), not as relative values as the provier of this proficiency test ha requeste. The provier recommens the participants to valiate the calculation of measurement uncertainties an follow more carefully the instructions given by the provier. 26 Proftest SYKE CAL 07/18