Variability of Ignition Furnace Correction Factors (2017)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

65 AASHTO T 308 procedures require the determination of asphalt content and aggregate correction factors for each asphalt mixture and for each ignition furnace used. In some instances, when numerous asphalt mix designs and several ignition fur- naces are available, correction factors have been shared, in some laboratories, between ignition units, even when that practice is not permitted by AASHTO T 308. The experimental plan in this project was designed, con- ducted, and analyzed to evaluate different ignition furnaces for a number of factors. These collected data were used to determine how these factors affect the results and what improvements could be made to minimize variability when different units are used. Three experiments were conducted as part of this study: 1. A sensitivity study at the NCAT lab, 2. Round-robin testing, and 3. Troubleshooting of outliers from the round-robin study. The following sections summarize the findings for each experiment. 6.1 Sensitivity Study at NCAT Lab The following sections discuss conclusions from the sen- sitivity study. 6.1.1 Factors Affecting Asphalt Correction Factors 1. Ignition furnace type: Type of ignition furnace affects the correction factors determined for Mixes 1, 3, and 4. For Mix 1, the maximum difference was equal to 0.15. For Mixes 3 and 4 with aggregates that experience addi- tional mass loss during ignition, causing higher correction factors, the type of furnace had a bigger impact on the test. This finding suggests that sharing correction factors between different furnaces may be possible when aggre- gates with low correction factors are used. 2. Test temperature: Test temperature was found to affect the asphalt correction factors. A reduction in the test tem- perature from 1,000 to 800°F (538 to 427°C) reduced the average AC correction factors for all aggregates that did not contain lime. The reduction in temperature seemed to decrease the mass loss during the test, and testing at these lower temperatures was shown to effectively remove the asphalt. For Mix 2, when lime was added to the same aggregate used for Mix 1, higher correction factors (more negative) were obtained when the temperature was reduced to 800°F (427°C). 3. Air flow: Air flow did not significantly affect the correc- tion factor for the aggregates with low mass loss (Mixes 1 and 2). Air flow was found to affect the correction factors for aggregates with higher correction factors (Mixes 3 and 4). It was also found that a restriction in air flow had a bigger impact for the Troxler unit. 4. Asphalt content: Based on the results of this study, changing the asphalt content from optimum minus 1% to optimum plus 1% may affect the asphalt correction factors when high-loss aggregates are used with a Troxler unit. 5. Sample mass: Changing the sample mass from the mini- mum mass recommended in the standard to the maximum mass allowed (500 additional grams) did not affect the cor- rection factors for any of the mixtures under evaluation. 6. Temperature rate (Thermolyne unit only): The rate of increase in temperature did not affect the measured cor- rection factor. It did take an average of 7 extra minutes to conduct the tests when temperature rate was reduced (as a result of a heating element being disconnected). 7. Burning profile (Troxler unit only): The burning pro- file affected the measured asphalt correction factors. For Mixes 1, 3, and 4, the Option 1 burn profile yielded the C h a p t e r 6 Summary of Findings

66 lowest correction factor, and Option 2 yielded the highest correction factor. The change in asphalt correction factor caused by changing the burning profiles was more pro- nounced for Mix 4. 6.1.2 Factors Affecting Aggregate Gradations After Ignition 1. Furnace type: The correction factor measured for each type of furnace used was found significantly different for 33% of the combinations tested. For the combinations that were found to be significant, the results indicated that for sieve sizes equal to or larger than #8, the difference of the average percentage passing (D% passing) was equal to or less than 2.6%; for sieves smaller than #8 and larger than #200, the D% passing did not exceed 2.4% for any of the combinations; and for the #200 sieve, the D% passing was less than or equal to 0.5%. For comparative purposes, the AASHTO T 308 standard permits a siev- ing difference of ±5% for sieves larger than or equal to #8, ±3% for sieves larger than #200 and smaller than #8, and ±0.5% for sieve size #200. Even though the change in gradation was statistically significant for many of the combinations tested, there was no practical difference in the sieve results. 2. Test temperature: Temperature appeared to affect aggre- gate gradation, but only for 16% of the different combi- nations. The differences found in the D% passing were all less than or equal to 0.8% for sieve sizes equal to or larger than #8, 0.4% for sieves smaller than #8 and larger than #200, and 0.5% for sieve size #200. These are not practical differences. 3. Air flow: Air flow was found to affect the aggregate grada- tion for five out of 36 different combinations. The differ- ences found in the D% passing were all smaller than the allowable sieving differences recommended in the AASHTO standard for any sieve size and are not considered practically different. 4. Asphalt content: Asphalt content affected the aggregate gradations for 19% of the different combinations. For all combinations, the differences in gradation found at the two asphalt content levels were relatively small, were within the allowable AASHTO standard sieving difference, and were not considered practically different. 5. Sample mass: Sample mass was found to be significant for 17 out of 36 different test combinations: nine of these combinations were for Mix 4, and eight were for Mix 2. For all 17 combinations that were found to be statistically significant, the differences in gradation found at the two sample mass levels were within the allowable AASHTO standard sieving difference and were not considered prac- tically significant. In general, although some variables were found to be sig- nificant for different sieve sizes, the differences in percentage passing were not practically different. The analysis presented in Table 38 compared the average gradation for all the com- binations used in the study to the blank sample gradations determined per the AASHTO T 308 standard. It was shown that the average sample gradations were very close to the blank samples for the different mixtures. The aggregate cor- rection factors for each sieve size for each mixture were deter- mined, and the highest gradation correction factor did not exceed a value of 2.4%. 6.2 Round-Robin Study The main purpose of the round-robin study was to identify laboratories with asphalt correction factor results considered to be outliers so that an investigation could be conducted to determine the possible causes of the differences. Other objectives were to validate the precision of the ignition test and recommend potential changes to the existing precision statement if needed. The study included a total of 28 furnaces; 18 laboratories had one furnace and five laboratories had two different units. The brands of the furnaces were Thermolyne (17), Troxler (8), and Gilson (3). A total of four mixtures with a 12.5 mm NMAS were used. The selection of the outliers was conducted in two steps; the first step was to identify extreme outliers or laboratory results that were not consistent with the general trend found for all the laboratories. Once these extreme outliers were eliminated, ASTM E691, “Standard Practice for Conduct- ing an Inter-Laboratory Study to Determine the Precision of a Test Method,” was used to select additional laboratories that were deemed outliers. This two-step process was con- sidered necessary to identify as many outliers as possible for additional investigation. After the outliers were identified, an additional laboratory was selected for further assessment because it yielded significantly lower correction factors. Since one of the goals of this study was to minimize differences in correction factor, it was considered important to investigate the reason for these low test results. Six laboratories were selected for further evaluation. Three of these laboratories had a Troxler unit, one had a Thermolyne unit, and two had Gilson units. Following the procedure recommended in ASTM E691, within-laboratory and between-laboratory precision val- ues were determined for each mixture. These results were presented in Table 45. The results showed that for Mixes 1

67 and 2, the within-laboratory and between-laboratory stan- dard deviations were similar: 0.089 and 0.074 for the within- laboratory standard deviation and 0.131 and 0.111 for the between-laboratory standard deviation. These numbers are close to the standard deviations recommended in AASHTO T 308, 0.069 and 0.117 for within-laboratory and between- laboratory, respectively. As the correction factors increased for Mixes 3 and 4, their corresponding standard deviations also increased. The within-laboratory standard deviations for Mixes 3 and 4 were found to be 0.112 and 0.178, respec- tively, and the between-laboratory standard deviations were 0.264 and 0.403. This suggests that different precision state- ments may be necessary for aggregates with high breakdown potential. An additional assessment was conducted using the results of the round-robin study to determine the effect of lime on asphalt correction factors. Mixes 1 and 2 used the same mix design, and the only difference was that Mix 2 included 1% lime. From the round-robin test results, it was found that the addition of lime caused a significant variation in correction factor, from 0.12% with no lime to -0.23% with 1% lime, resulting in a difference of 0.35. 6.2.1 Troubleshooting Outliers from Round-Robin Study NCAT staff visited the six laboratories identified as out- liers to further explore the possible causes of the differences for their test results. All visits were conducted during the months of December 2015 and January 2016. The causes of the differences are summarized in the following. 6.2.1.1 Lab 4-TX The results from the RRS for this laboratory showed sig- nificantly different average correction factors for Mixes 1 and 2. It was found that the results that were reported from the furnace tickets were significantly different from the results from weighing the samples outside the furnace. For some of these tests, the tickets showed that the test ran for exactly 15 min, which is much less time than normally required. When the results obtained from external weighing were used, the results were found to be similar to the results found for the other laboratories. 6.2.1.2 Lab 23-TX The average correction factor results for Mix 4, in this laboratory, were significantly higher than the average correc- tion factor from all of the laboratories. During the labora- tory visit, it was found that the factory-programmed default burn profile for this furnace was different from other Troxler furnaces. It was also found that the Option 1 burn profile was exactly the same as the default burn profile. The research team adjusted the burning profiles to match the factory pro- files. Tests were conducted after this adjustment, and the results were similar to the average test results from the other laboratories. 6.2.1.3 Lab 17-TH The results from the Thermolyne furnace in this labora- tory were higher than the mean AC content from the other laboratories for Mix 3. During the laboratory visit, several tests were conducted using samples from Mixes 3 and 4, but the results from the RRS samples tested by this laboratory could not be replicated. The NCAT test results at the time of the visit were lower than the RRS results from this laboratory and more in line with the mean values for all laboratories. It was speculated that the high test results obtained during the round-robin study were due to an error during testing because no issues with the equipment or the equipment setup were found. 6.2.1.4 Lab 16-GS The test results for this laboratory showed higher cor- rection factors for Mix 4. The research team found that the Gilson unit required several temperatures to be programmed, with the most important being the preheat temperature and the burnout temperature. It was suspected that during the RRS testing, the technician had only changed the preheat temperature to 900°F (482°C) and left the burnout tem- perature set to 1,000°F (538°C) when testing Mix 4. This was confirmed by testing a sample of Mix 4 using a preheat temperature of 900°F (482°C) and a burnout temperature of 1,000°F (538°C). A sample was then tested using 900°F (482°C) for both temperatures to confirm that this would yield lower results. 6.2.1.5 Lab 19-GS The test results for this furnace yielded lower than aver- age results for Mixes 3 and 4. During the visit, it was noticed that the input parameter for the afterburner temperature was set to 1,000°F (538°C) instead of the recommended 1,562°F (850°C). Tests were conducted with this incorrect setting, and the results obtained were similar to the results obtained dur- ing the round-robin testing. Additional tests were conducted with the correct afterburner temperature, and the average AC content results did increase to be similar to the average of all laboratories. These results seem to explain the reason for the

68 low correction factors that were obtained during the round- robin study. 6.2.1.6 Lab 21-TX The test results for this laboratory showed higher correc- tion factors for Mixes 3 and 4. After the results had been sub- mitted, this laboratory decided to replace the Troxler furnace that was used during the round-robin study. The main rea- son for this was that rain water had consistently leaked in through the exhaust vent and caused damage to the existing furnace, causing it to periodically malfunction. This prob- lem more than likely explains the difference in the RRS results from this laboratory. Tests were conducted with the new Troxler unit, and the results were closer to those of the other laboratories.