ASPHALT ROUND 1 PROFICIENCY TESTING PROGRAM. April 2009 REPORT NO. 605 ACKNOWLEDGEMENTS

Transcription

1 ASPHALT ROUND 1 PROFICIENCY TESTING PROGRAM April 2009 REPORT NO. 605 ACKNOWLEDGEMENTS PTA wishes to acknowledge gratefully the technical assistance provided for this program by Mr Hugo Van Loon at The Department for Transport, Energy and Infrastructure, Government of South Australia and Mr Rod McArthur at Topcoat Asphalt Contractors Pty Ltd. This assistance included preparation, distribution and homogeneity testing of the sample, along with input into the design of the program, technical advice and discussion in the final report. COPYRIGHT PROFICIENCY TESTING AUSTRALIA 2009 PO Box 7507, Silverwater NSW 2128

2 CONTENTS 1. FOREWORD 1 2. FEATURES OF THE PROGRAM 1 3. FORMAT OF THE APPENDICES 1 4. STATISTICAL DESIGN OF THE PROGRAM 2 5. OUTLIER RESULTS 3 Table A: Summary Statistics 3 Table B: Outlier Results 4 6. TECHNICAL COMMENTS 5 Table C: Particle Size Distribution Upper and Lower Limits 6 7. REFERENCE 6 APPENDICES APPENDIX A Summary of Results Bitumen Content Maximum Density Bulk Density Grading A1 A4 A7 A9 APPENDIX B Homogeneity Testing B1 APPENDIX C Instructions to Participants and Results Sheet C1

3 1 1. FOREWORD This report summarises the results of a proficiency testing program covering a series of tests on asphalt samples conducted in December This exercise was conducted by Proficiency Testing Australia. The aim of the program was to assess laboratories' ability to competently perform the tests examined. 2. FEATURES OF THE PROGRAM a) A total of 55 sets of samples were dispatched to 52 laboratories included in the program. Two laboratories were unable to return results in time for inclusion in the final report. b) The following tests were performed by participants on the samples: Bitumen Content; Maximum Density; Bulk Density; and Grading. c) The samples were analysed for homogeneity. Based on the results of this homogeneity testing it was concluded that the samples were sufficiently homogeneous. Therefore, any results later identified as outliers could not be attributed to any significant sample variability. d) Laboratories were requested to perform the tests according to the "Instructions to Participants" and to record their results on the accompanying "Results Sheet", both of which were distributed to participants with the sample. e) Each laboratory was randomly allocated a unique code number for the program to enable confidentiality of results. Reference to each laboratory in this report is made by its code number. Please note that some laboratories reported more than one set of results and, therefore, one code number (with letter) could appear several times in the same data set. 3. FORMAT OF APPENDICES APPENDIX A Appendix A lists all the results reported by the participating laboratories for each test. Calculated robust z-scores, z-score charts and a table of robust statistics - number of results, median, normalised interquartile range (IQR), robust coefficient of variation (CV), minimum, maximum and range, are presented for the following tests: Bitumen Content; Maximum Density; and Bulk Density.

4 2 Number of results, median and normalised IQR are presented for the Grading test. No z-scores were calculated for these tests. Upper and lower limits were determined by multiplying the normalised IQR by 3 then adding or subtracting this result from the median. Results reported outside of these limits are classed as outlier results. Scatter plots display the distribution of results reported. Calculated upper and lower limits (median ±3 * IQR) can be above 100% or below 0% which is why scatter plots are used for these tests. Scatter plots are used to set more realistic upper and lower limits i.e. within Zero to 100%. APPENDIX B Appendix B contains details of sample preparation and homogeneity testing data. APPENDIX C Appendix C contains a copy of the Instructions to Participants, the Method Codes and the Results Sheet supplied to participants. 4. STATISTICAL DESIGN OF THE PROGRAM Robust statistical procedures were used to generate the z-scores and summary statistics for the results, including: number of results; median; normalised IQR; minimum; maximum; and range. Robust z-scores were calculated for each laboratory based on the results submitted for each test. Robust z-scores were not calculated for the Grading test. Outliers were determined by setting upper and lower limits.

5 3 5. OUTLIER RESULTS In order to achieve the program's aim of assessing laboratories' testing performance, a robust statistical approach, which uses z-scores to assess participants performance, has been utilised. The z-score is a measure of how far the result(s) is from the consensus value - a normalised value which gives a "score" to each result relative to the other results in the group. Therefore a z-score close to zero means that the result agrees well with those from other laboratories. An outlier will be any result(s) which has an absolute z- score value greater than three. For the Grading test, any results reported that are outside the upper and lower limits are deemed outliers. TABLE A: Summary Statistics Test No. of Results Median Normalised IQR Bitumen Content (nearest 0.1%) Maximum Density (0.001 t/m³) Bulk Density (0.001 t/m³) %Grading 19.0 mm (1%) mm (1%) mm (1%) mm (1%) mm (1%) mm (1%) mm (0.1%) mm (0.1%) mm (0.1%) mm (0.1%) mm (0.1%)

6 4 Laboratories who reported outliers are listed in Table B. Any laboratories reporting outliers are encouraged to investigate the cause/s of the outlier. TABLE B: Outlier Results Test Lab Code Bitumen Content (nearest 0.1%) Maximum Density (0.001 t/m³) Bulk Density (0.001 t/m³) none %Grading 19.0 mm (1%) N/A 13.2 mm (1%) N/A 9.5 mm (1%) 1, 5, 9, 14 and 23A 6.7 mm (1%) 2, 9 and mm (1%) 2, 9 and mm (1%) 2, 9 and mm (0.1%) 2, 6 and mm (0.1%) 2, 6 and mm (0.1%) mm (0.1%) mm (0.1%) 4

7 5 TECHNICAL COMMENTS General Comments Overall, participating laboratories performed at an above average level. For the 14 major tests (including sieve sizes), 53 participating laboratories reported a total of 722 results. There were 25 outlier results reported, which means that only 3.46% of the total results reported for this program were outlier results. This low percentage figure is considered to be a good result. Bitumen Content Laboratories 14 and 30A received an absolute z-score greater than 3. Laboratory 30A did not receive an outlier as it used method RC210.06, which utilises the Ignition Oven Method. This method usually gives a higher result for Bitumen Content, which can be seen in the homogeneity binder content results where the Ignition Oven Method was used. In addition, absolute z-scores > 2 were obtained by laboratories 2 and 45A. Laboratory 45A also used the Ignition Oven Method. One outlier out of 53 (1.89%) is a low figure. Maximum Density One laboratory (code 35) reported an outlier for Maximum Density. In addition, absolute z-scores > 2 were obtained by laboratories 11 and 15A. Whilst these are not classified as outlier results, any laboratory that obtained an absolute z-score > 2 is also encouraged to review their results. Bulk Density No laboratories reported outliers for this test, with five laboratories reporting an absolute z-score > 2. The percentage of outliers for this test is 0%, which is a good result. % Grading Upper and lower limits were set for each sieve size, with any results falling outside these limits considered outliers. Scatter plots (see Appendix A) highlight the upper and lower limits and display the spread of results. Upper and lower limits for each sieve size are presented in Table C.

8 6 TABLE C: % Grading Upper and Lower Limits Sieve Size Lower Limit (%) Upper Limit (%) 19.0 mm mm mm mm mm mm mm mm mm mm mm No outliers were recorded for sieve sizes 19.0mm and 13.2mm. For the 9.5mm sieve size, outliers were reported by laboratories 1, 5, 9, 14 and 23A. Laboratories 2, 9 and 14 reported outliers for each of the sieve sizes 6.7mm, 4.75mm and 2.36mm. Laboratories 2, 6 and 14 reported outliers for each of the sieve sizes 1.18mm and 0.600mm. Laboratory 6 reported outliers for sieve sizes 0.300mm and 0.150mm. Laboratory 4 reported an outlier for the 0.075mm sieve size. Overall, It is considered that good results were obtained for the % Grading tests. 7. REFERENCE [1] Guide to Proficiency Testing Australia (2008). This document is located on the PTA website at under Documents.

9 APPENDIX A Summary of Results

10 Lab Code A1 Bitumen Content (nearest 0.1%) Result Robust Z-Score Method Code A MAT-TP470 AS B MAT-TP470 AS Q30A Code Q308A 8 12A Q308A Q308A Q308A (1) 15A Q308A 15B AS AS (1) WA A Q308A (Q308A) QDMR Q308A Q308A A * RC AS Q308A (8) 36A MAT TP B MAT TP C MATTP470 36D MAT-TP470

11 A2 Bitumen Content (nearest 0.1%) Lab Code Result Robust Z-Score Method Code 37A AS/NZS A B A AS (3) WA A AS A Q308A (8) AS (3) Q308A (8) Note: denotes an outlier (i.e. z-score > 3) * denotes an z-score > 3 which is not considered an outlier due to a different method being used (refer to page 5). Statistic Summary Statistics Bitumen Content No of Results 53 Median 5.00 Norm IQR 0.19 Robust CV 3.7% Minimum 4.2 Maximum 5.6 Range 1.4

12 A3 Bitumen Content (nearest 0.1%) C 39B 29 15A 3A A A A A D 37A 38 44A 3B 31 39A 2 45A 30A 19 15B B Laboratory Code Robust Z-Score

13 Lab Code A4 Maximum Density (0.001 t/m³) Result Robust Z-Score Method Code A TP435 AS B TP435 AS (P&S T35-Water Displ) (Q307A) Q307A (QDMR307A) Q307A 14 12A Q307A Q307A Q307A (15) 15A Q307A 15B AS (16) WA A Q307A (Q307A) QDMR Q307A Q307A DTEI MAT-TP Q307A 30A AS AS Q307A (15) 36A TP-MAT-470

14 A5 Lab Code Maximum Density (0.001 t/m³) Result Robust Z-Score Method Code 36B TP MAT C MATT TP D MAT-TP435 37A AS/NZ A A (Q307A) WA A AS A Q307A AS (15) Q307A (15) Note: denotes an outlier (i.e. z-score > 3) Summary Statistics Statistic Maximum Density No of Results 51 Median Norm IQR Robust CV 0.5% Minimum Maximum Range 0.057

15 A6 Maximum Density (0.001 t/m³) 15A A A 45A A A A B 31 36D 36C A B A A 11 30A 36B Laboratory Code Robust Z-Score

16 A7 Bulk Density (0.001 t/m³) Lab Code Result Robust Z-Score Method Code A AS B AS / Prep 23 Test Q306B and Q305 for Preparation QDMR 306B A Q306B Q306B Q306B (19) (23) 15A Q306B 15B / / AS (19122) WA A Q306C A AS / AS AS / / Q306B (23/19) 36A B AND 22 TPMAT 428 AS AND 22 TPMAT 428 AS

17 A8 Bulk Density (0.001 t/m³) Lab Code Result Robust Z-Score Method Code 36C AS TP MAT D AS TP MAT A AS/NZS A & / WA A AS A Q306B (23/19) AS (19) Q306A (23/19) Note: denotes an outlier (i.e. z-score > 3) Statistic Summary Statistics Bulk Density No of Results 44 Median Norm IQR Robust CV 0.7% Minimum Maximum Range 0.091

18 Lab Code A9 Grading (nearest 1%) 19.0 mm 13.2 mm 9.5 mm 6.7 mm 4.75 mm 2.36 mm 1 # A B A A B # # # A # A # # A B C D A

19 Lab Code A10 Grading (nearest 1%) 19.0 mm 13.2 mm 9.5 mm 6.7 mm 4.75 mm 2.36 mm A B A # A A # Note: # denotes that the laboratory left this space blank Summary Statistics Statistic 19.0 mm 13.2 mm 9.5 mm 6.7 mm 4.75 mm 2.36 mm No of Results Median Norm IQR * Norm IQR Upper Limit Lower Limit Note: 1. The lower limit is calculated by: median 3 * norm IQR. 2. The upper limit is calculated by: median + 3 * norm IQR. 3. denotes an outlier (i.e. result lies either below the lower limit or above the upper limit).

20 Lab Code A11 Grading (nearest 0.1%) 1.18 mm mm mm mm mm A B A A B A A A B C D A

21 Lab Code A12 Grading (nearest 0.1%) 1.18 mm mm mm mm mm A B A A A Summary Statistics Statistic 1.18 mm mm mm mm mm No of Results Median Norm IQR * Norm IQR Upper Limit Lower Limit Note: 1. The lower limit is calculated by: median 3 * norm IQR. 2. The upper limit is calculated by: median + 3 * norm IQR. 3. denotes an outlier (i.e. result lies either below the lower limit or above the upper limit).

22 A13 Grading Methods Lab Code Method Lab Code Method AS A AS B AS A RC AS Q308A 35 Q308A (8) A AS Q308A 8 36B # 12A Q308A 36C AS Q308A 36D # 14 Q308A 37A AS B 8 38 # A B A 8 19 # 41 AS AS (3) 20 AS (1) WA / WA A RTA T607 23A Q308A 45A # 24 8 (Q30A) 46 Q308A (8) 25 QDMR Q308A 8 (wire mesh 300mm Air Sieves Asphalt) 47 AS (3) 26 Q308A 48 Q308A (8) Note: # denotes that the laboratory left this space blank

23 A14 98 Grading 9.5 mm (nearest 1%) Result A 3B A 15B A A A 41 36D 39A 35 36B 37A 39B 40A 36C 44A 45A Laboratory Code Number

24 A15 Grading 6.7 mm (nearest 1%) B A Result A A 13 15A 15B A A B 36A 36C 38 36D 40A 37A 39B 39A A Laboratory Code Number

25 A16 Grading 4.75 mm (nearest 1%) A 13 12A A 3B B A A B 36A 36C 36D 37A 38 39A 39B 40A A 45A Laboratory Code Number Result

26 A17 Grading 2.36 mm (nearest 1%) B A D 44A Result A A A 15B A B 36A 36C 38 39A 40A 39B A A Laboratory Code Number

27 A18 46 Grading 1.18 mm (nearest 1%) Result A 3B A A 15B A A B 36D 37A 36A 35 36C 40A A 39B 42 44A A Laboratory Code Number

28 A19 Grading mm (nearest 1%) Result A 3B A 9 15A 13 23A A 15B B B 32 37A A 34 40A A A 36D A 42 36C Laboratory Code Number

29 A Grading mm (nearest 1%) Result B 3A A 14 15A 15B A A 36A B 36D 36C 37A 38 39B 40A 39A 42 44A 43 45A Laboratory Code Number

30 A21 Grading mm (nearest 1%) A 41 Result A 3B A A 15B A B 36D 36B 35 37A 39A 36A 36C 44A 40A A Laboratory Code Number

31 A Grading mm (nearest 1%) 8.5 Result B 1 3A A 15A 13 15B A A B D 35 40A 36B 42 36A 37A 39A 36C 44A 43 45A Laboratory Code Number

32 APPENDIX B Homogeneity Testing

33 B1 HOMOGENEITY TESTING 10 numbers were randomly generated and the samples labelled with these numbers were tested for homogeneity. Statistical analysis of the results indicated that no noticeable sample variability existed. Therefore, it was concluded that any outlier results subsequently identified could not be attributed to sample variability. Sample Number Binder Content Grading (%) 13.2 mm 9.5 mm 6.7 mm 4.75 mm 2.36 mm Sample Number Grading (%) 1.18 mm mm mm mm mm

34 APPENDIX C Instructions to Participants and Results Sheet

35 C1 PROFICIENCY TESTING AUSTRALIA ASPHALT PROFICIENCY TESTING PROGRAM ROUND 1 INSTRUCTIONS TO PARTICIPANTS Please note your confidential Lab Code is located on the top right corner of the result sheet. To ensure that results from this program can be analysed properly, participants are asked to adhere carefully to the following instructions. 1. One Asphalt sample is provided to each laboratory. Depending on the sample order, some laboratories may receive more than one sample. 2. The following tests are to be conducted: Bitumen Content and Grading Maximum Density Bulk Density 3. Participants can perform any or all of the above mentioned tests. If participants perform any of the tests they will be included in the final report and be issued an individual summary sheet. 4. Participants should use the routine test methods which would normally be used to test samples. The Australian Standard method AS 2891 is the preferred method. Please identify the method used on the Results Sheet by using the Method Codes listed, otherwise state the method used. 5. The results for all determinations are to be recorded on the results sheet to the accuracy and reporting basis indicated. 6. Fill in all the grading results on the results sheet. E.g. if all the sample passes through a grading then the percentage grading is 100%; do not leave the box blank or record Testing may commence as soon as the sample is received. All laboratories must return the results sheet no later than Friday 19th December 2008 to: Ray Sultan Proficiency Testing Australia PO Box 7507 SILVERWATER NSW 2128 Phone: Fax: rsultan@pta.asn.au

36 C2 METHOD CODES AND DESCRIPTIONS Bitumen Content and Grading Method Description Code AS Reflux 1 AS Centrifugal Filter 2 AS Pressure Filter 3 RC RC Ignition Oven Method 5 RTA T607 6 RTA T609 7 QDMR Q308A 8 QDMR Q308D 9 DTEI MAT TP DTEI MAT TP WA Austroads AG: PT/T Other (please specify on results sheet) 14 Maximum Density Method Description Code AS Water Displacement 15 AS Methylated Spirits 16 Other (please specify on results sheet) 17 Bulk Density Method Description Code AS Waxing 18 AS Presaturation 19 AS Mensuration 20 Other (please specify on results sheet) 21

37 C3 PROFICIENCY TESTING AUSTRALIA ASPHALT PROFICIENCY TESTING PROGRAM ROUND 1 TEST (report to) RESULTS SHEET Result Lab Code Method Bitumen Content (nearest 0.1%) Maximum Density (0.001 t/m³) Bulk Density (0.001 t/m³) Grading 19.0mm (1%) 13.2mm (1%) 9.50mm (1%) 6.70mm (1%) 4.75mm (1%) 2.36mm (1%) 1.18mm (0.1%) 0.600mm (0.1%) 0.300mm (0.1%) 0.150mm (0.1%) 0.075mm (0.1%) Date of tests: Signature: Please return to: Ray Sultan, Proficiency Testing Australia PO Box 7507, Silverwater, NSW, 2128 Fax: By Friday 19 th December 2008

38 ..End of Report..