Properties of Fly Ash-Extended Asphalt Concrete Mixes

Transcription

1 TRANSPORTATION RESEARCH RECORD Properties of Fly Ash-Extended Asphalt Conrete Mixes ABDULRAHMAN s. AL-SUHAIBANI AND EGON T. TONS Fillers added to asphalt onrete mixes have long been used as stiffening, extending, or otherwise void-filling materials, depending on their type and partile size. Fly ash, a by-produt of the oal-burning proess for power generation, is a filler that an serve one or more of the above-mentioned funtions. The role of fly ash in asphalt onrete mixes has been investigated. Three sizes of fly ash (oarse, medium, and fine) were used. The effet of these fly ash sizes on the resilient modulus, rutting potential, and water resistane of these mixes was investigated. It was found that the medium-size fly ash was the best size for an asphalt extender. Fly ash is a finely divided residue that results from furnaeburned pulverized bituminous oal. Eletrial generation is the prime onsumer of oal produed in the United States and, onsequently, the prime produer of fly ash. In the United States, the annual prodution of fly ash inreased from about million tons in 197 to about 6 million tons in 1979; the utilized amount during that time inreased from about million tons, or 1.6 perent of the prodution, to about 1. million tons, or 17. perent (1). From these figures, it is lear that there is a huge amount of unused fly ash that must be disposed of eah year. This not only osts power ompanies money but also reates a disposal problem. Asphalt pavement usually onsists of three omponents: asphalt, aggregates, and air. Beause nearly all asphalts used in road onstrution are of rude oil origin, the inrease in rude oil pries in reent years has resulted in an inrease in asphalt pries. The dwindling world resoures of oil have inreased onerns about the asphalt supply for highway pavements. In addition, a onsiderable portion of asphalt-type highways in the United States are deteriorating as a result of heavy loads and unfavorable weather onditions. All of this has led many researhers and agenies to look for ways of reduing the amount of the required asphalt and improving pavement resistane against loading and moisture damage. One method that has reently reeived muh attention is the partial substitution, or replaement, of asphalt with other materials. Some mineral fillers have been known to work as integral parts (extenders) of asphalt if thoroughly mixed with it, resulting in an inrease in binder ontent and, therefore, dereasing the amount of asphalt required for optimum onditions. Fly ash has been used suessfully as a filler for asphalt mixes for a long time. It has the advantage of inreasing the A. S. Al-Suhaibani, Civil Engineering Department, College of Enginee~ing, King Saud University, P.. Box 8, Riyadh 1121, Saudi Arabia. E. T. Tons, Civil Engineering Department, University of Mihigan, Ann Arbor, Mih. 81. resistane of asphalt mixes to moisture damage. In addition to filling voids, fly ash reportedly works as an asphalt extender (2). It is the main purpose of this investigation to study the possibility of using wasted fly ash as an asphalt extender in bituminous paving mixes. LITERATURE REVIEW As finely pulverized bituminous oal is burned, partiles of fly ash are suspended in the gas stream that reahes the boiler. As hot gases pass into the atmosphere, the partiles of fly ash ollet on the plates of eletrostati preipitators within the heating system. Fly ash is then proessed or filtered and finally aumulated and stored (,). Fly ash is haraterized by its low speifi gravity, whih is a funtion of its hemial omposition and varies between 2. and 2.6, averaging 2. (). The partile size distribution generally depends on the olletor used. It has been found that the ash olleted by the eletrostati preipitator ontains a greater perentage of very small partiles (<1. µ) (). In general, a typial fly ash partile size ranges between. µ and 1 µ (). Fly ash partiles are generally spherial (,6); however, a minor fration onsists of irregularly shaped partiles ( 6). Most types of fly ash have been used suessfully as mineral fillers in hot asphalt mixes ( -6 ). It has been reported that fly ash does not differ materially from Trinidad Asphalt's mineral filler (7). In another study, fly ash was used as a replaement for limestone dust in asphalt onrete mixes with good results (8). The suitability of fly ash as a filler in sheet asphalt mixes was investigated by the Detroit Edison Company and reported by Zimmer (9). It was found that stabilities of mixes ontaining fly ash are omparable to those ontaining limestone dust when proportions are based on weight, and that fly ash has virtually the same void-reduing properties as limestone dust. The resistane of asphalt mixes ontaining fly ash to water ation has been found to be equal or superior to those ontaining other types of filler (6,1). The strength retention for eight fly ashes ranges between 8 and 1 perent-all above the 7 perent figure onsidered to be the ritial minimum. A report by Rosner et al. indiates that the addition of up to 6 perent fly ash by weight of aggregate to asphalt onrete produes an aeptable mix (11). At the same time, asphalt requirements and voids in mineral aggregates (VMA) values are lower than those for mixes ontaining portland ement or hydrated lime. Tons et al. investigated the use of fly ash as a replaement for asphalt ement in asphalt onrete mixes (2). The three

2 12 types of fly ash used in the study were found to have positive effets on the physial properties of the evaluated mixes. Furthermore, the replaement of up to perent of asphalt ement by fly ash improved most of the mix's physial properties when a pratial asphalt ontent was used. Conerning other types of fillers, Kallas and Puzinauskas found that 9 of the 11 fillers investigated were effetive in replaing a portion of the asphalt required to produe minimum VMA (12). Other reports also point out that some mineral fillers an serve as asphalt extenders, and that they may atually derease the stiffness of the asphalt (1,1). Baghouse fines, also, were reported to at as asphalt extenders (1-18). Finally, it has been found that mixing fine fly ash partiles (passing Sieve #2S) with asphalt ement auses the highest inrease in visosity ompared with oarser fly ash sizes (19). STUDY OBJECTIVES This investigation had two objetives: 1. To study the effet of fly ash partile size, aggregate gradation, and binder ontent on the resilient modulus and rut-depth harateristis of asphalt onrete mixes. 2. To evaluate the use of fly ash as an asphalt extender in asphalt onrete mixes. EXPERIMENTAL WORK Materials Asphalt Cement AC-2 asphalt ement was used for this study. It was obtained from Marathon Petroleum Company in Detroit, Mihigan. The results of tests onduted on the asphalt are shown in Table 1. Fly Ash TRANSPORTATION RESEARCH RECORD 12 Fly ash used in this study was obtained from Consumer's Power Company, Mihigan. The fly ash was dry-sieved on Sieves #27 and #2S for ls min. Beause fly ash has pratially no partiles finer than 1 µ, silia fume (mirosilia), whih is 1 perent finer than O.S µ, was mixed with the fly ash fration passing Sieve #2S in a SO/SO ratio by weight to form the fine fration. This was done to study the stiffening effet of very fine materials. Table 2 summarizes the speifi gravities and partile size distribution of the three frations. Aggregate The oarse aggregate was a rushed gravel obtained from Thompson-MCully Asphalt Paving Company in Whitmore Lake, Mihigan. The fine aggregate was a onrete sand obtained loally. The oarse aggregate was sieved and divided into different sizes as follows: / in.- V2 in., V2 in.- /s in., :Ys in.-#, #- #8. Eah size was then thoroughly washed, dried, and stored until usage. The fine aggregate was divided into the-following sizes: #8-#16, #16-#, #-#SO, #S-#1, #1- #2. The oarse and fine aggregate frations were ombined with the desired proportions aording to the gradation urves A, B, and C. The three gradations were hosen in suh a way that the sand ontent was different for eah one; however, they were all within the speifiation limits of the Mihigan Department of Transportation. The fine gradation (C) represents the finest, and the oarse gradation (A) represents the oarsest. The medium gradation (B) falls between the two. An important property of these aggregate gradations is that they possess different surfae areas. This is important beause the larger the surfae area of aggregates, the larger the amount of asphalt needed to oat suh aggregates. The three gradation urves are shown in Figure 1, and their speifi gravities and absorptions are shown in Table. TABLE 1 SUMMARY OF ASPHALT CEMENT PROPERTIES Test ASTM Test Designation Test Results Speifi Gravity (77/77 F) Visosity x 1 6, poises (77 F) Penetration (77 F).1 lllill, 1 g' se Penetration (9 F).1 mm, 1 g, se Softening Point, of (R & B) P.I. D 7-76 D 7-77 D -7 D -7 D (1)* 122 (129)* (-.7)* * For Reovered Asphalt

3 Al-Suhaibani and Tons 12 TABLE 2 PROPERTIES OF FLY ASH FRACTIONS Grain Size Analysis Fly Ash Fration * Sieve size: # (9 µ.) # 1 (1 µ.) # 2 ( 7 µ.) # 27 ( µ.) 1 1 Hydrometer Analysis # 2 ( µ.) Speifi Gravity perent by weight silia fume en :.. "' Q.... : " u.. Q. " I I Aggregole Grodollon B A.I JP I/, /..,,,. I.' " Sieve Size, mm ' ' II " Ill 1'... 1// / l/ ' v FIGURE 1 Partile size distribution of aggregates. Test Variables and Speimens Preparation A set of variables was hosen for this study. In addition to the fly ash sizes and aggregate gradations already mentioned, two other variables-asphalt ontent and perentage of replaement-were used. Beause asphalt ontent is not onstant in the replaement proess, the term "asphalt equivalent," whih represents the asphalt ontent for ontrol mixes, was used instead. Three asphalt equivalents were used in this study-namely,,, and 6 perent. The asphalt ement at eah asphalt equivalent was replaed by an equal volume of fly ash in five replaement perentages:, 1, 2,, and perent. The asphalt ement was partially replaed by fly ash so that the total volume of binder (asphalt + fly ash) was kept onstant and equal to the original asphalt volume before the addition of fly ash. All speimens used were Marshall size and were ast aording to ASTM standards with blows on eah side of the speimen.. v 1 v 1 Either one or two speimens were prepared for eah mix, as shown in Table. For ontrol mixes (no fly ash), two speimens were prepared. This experimental design was used to redue the quantity of materials needed and the time neessary for preparing and testing speimens. CHARACTERIZATION OF MIXES Resilient Modulus Resilient modulus test before and after immersion was onduted on the same speimen; the nondestrutive nature of the test made this possible. Speimens were immersed in water at 1 F for 2 hr. Testing for resilient modulus was then done for the purpose of evaluating the effet of test variables (mainly fly ash size and quantity) on resistane of mixes to water ation. By testing the same speimen before and after immersion, errors owing to material variability were eliminated. At the same time, a redution in materials' quantities, time, and effort was ahieved. It should be mentioned that although only samples of the results are reported here, the onlusions were based on the results of the entire study (19). Dry Resilient Modulus Dry resilient modulus results for 6 perent asphalt equivalent are shown in Figure 2. These results show that, in general and for a given aggregate gradation and binder ontent, fine fly ash auses the most stiffening of the three fly ash frations. This is onsistent with the results of visosity testing of asphalt-fly ash mixes (19), whih show that the fine fly ash auses the highest stiffening effet among the three fly ash sizes. fly ash ranks seond; oarse fly ash gave the

4 126 TRANSPORTATION RESEARCH RECORD 12 TABLE SPECIFIC GRAVITIES AND ABSORPTIONS OF AGGREGATE GRADATIONS Parameter A B C Bulk Sp. Gr Bulk Sp. Gr. (SSD) Apparent Sp. Gr Perent Absorption TABLE MIX VARIABLES AND NUMBER OF SPECIMENS A 8 F M F M F M A!i!llllillt - l::!ll! inli!dt Perent of Asphalt Replaed perent perent perent perent perent Asllhalt - l::m!illilli!nt ::! Perent of Asphalt Replaed perent perent perent perent perent AllllllS!lt - fimiival!!lt 6 Perent of Asphalt Replaed perent perent perent perent 1 2 l perent l l 2 l :. l 2 l lowest modulus values. This kind of behavior is expeted beause the fine fration (ontains an appreiable amount of submiron partiles), when ombined with asphalt, aused its visosity to inrease to values higher than those aused by the other two frations. This translates into a high resilient modulus in asphalt-fly ash onrete mixes. As for oarse fly ash, given its large partile size, it reates more voids in the mix, hene produing a less stiff mix than the other two. For a given aggregate gradation and asphalt equivalent, dry resilient modulus inreases as the perent of replaement inreases for fine and medium fly ash, but it remains the same or dereases for oarse fly ash. The inrease in the resilient modulus in the ase of fine fly ash appears to be logial, given its high stiffening effet. The ineffetiveness or the negative effet of oarse fly ash is mainly attributable to the inrease in voids in mixes made with this fration as its quantity inreases. For a given aggregate gradation, inreasing asphalt equivalent from to 6 perent has little or no effet on dry resilient modulus. The details of this effet an be found elsewhere (19). In general, for a ertain asphalt equivalent, hanging the aggregate gradation from A to C shifted the resilient modulus values downward. This was more pronouned for mixes ontaining oarse fly ash and mixes with low replaement perentages. Grain interlok, whih is more pronouned for Gradation A, ould be the reason for this behavior. Estimates of the main effets of the four independent variables on dry resilient modulus are shown in Table. These effets seem to strengthen the previous onlusions. These

5 Al-Suhaibani and Tons 127., o..y. "'. x '" ::::J ::::J - I:..., : Q) '" Q) a: > ' Cl 2 : A Fl y Ash Size o+---.-~~~--..~~~--r~ : --m o+---.-~~~--..~~~--r~ ~~~~~~~~~~~~--. : C -< ,... f.. 1 = estimated main effet of the lth level of fly ash size, and E;iklm = an error term. These parameters indiate the effet of eah level of a given independent variable on the predited values. A plus sign indiates a positive effet; a minus sign indiates a negative effet. A very low value (approximately zero) indiates little or no effet. In other words, the estimated main effets show the ontribution of eah level of the independent variable toward the predited values. Table also shows the standard deviation of these main effets. Immersion Effet To study the effet of water on asphalt-fly ash onrete mixes, the index of retained stiffness (IRS) was used. IRS is the ratio of resilient modulus after and before water immersion. It indiates the resistane of a given mix to water ation. Figure shows a sample of the results of soaked resilient modulus. Corresponding IRS values are shown in Figure. Tables 6 and 7 show the statistially estimated main effets for soaked resilient modulus and IRS, respetively. These results show that, in general, IRS value gets lower as the aggregate gets finer and as the asphalt equivalent gets lower. However, muh lower IRS values were obtained for mixes made with Gradation A and fine fly ash at perent replaement. In fat, all mixes made with fine fly ash at perent replaement show lower IRS values than those mixes made with medium or oarse fly ashes. This was expeted, beause these mixes were not ompletely oated with asphalt, allowing water to easily strip asphalt films from aggregate surfae, thus reduing IRS values below those of other mixes. OT--~~~~--.~ ~~ Perent of Asphalt Replaed FIGURE 2 Dry resilient modulus versus perentage of asphalt replaed for 6 perent A.E. effets are represented by ex, (, "/, and A. in the predition model shown below. Yijklm = µ + CX; + fj + "lk + Ai + Eijklm (1) where Y,iklm = the predited value of the dependent variable for a given set of levels of the independent variables, µ = the grand mean (the onstant in Table ), a, = estimated main effet of the ith level of the asphalt equivalent, (i = estimated main effet of the jth level of the aggregate gradation, "lk = estimated main effet of the kth level of the perentage of asphalt replaed, Rut Depth Predition One of the major distresses that our in bituminous pavement is permanent deformation, or rutting. Permanent deformation has reeived onsiderable attention from researhers and pavement tehnologists. Several methods have been introdued in pavement design to address the problem of rutting. These methods are divided into two main ategories: indiret and preditive (2). Indiret methods are based on limiting layer thikness and omponent material strengths, stability, or density to a minimum, or limiting the vertial ompressive strain on the subgrade surfae to some maximum level. Preditive methods are based on prediting rut depth for a given set of onditions and redesigning the pavement if neessary until a satisfatory rut depth value is obtained. A preditive method that relates pavement permanent deformation to reep test results was developed by Shell Oil Company Laboratories. Beause of its simpliity as a tool for omparing different mixes and the availability of the needed equipment to run the reep test, this method has been adopted here to evaluate the rutting potential of asphalt-fly ash onrete mixes (21). To predit rut depth, the stati reep te~t-as suggested by Shell-was onduted on Marshall speimens representing

6 128 TRANSPORTATION RESEARCH RECORD 12 TABLE ESTIMATED MAIN EFFECTS FOR DRY RESILIENT MODULUS x 1 6 (kpa) Parameter Estimated Main Effets Standard Deviation Constant Asphalt Equivalent 6 Perent Perent Perent A B l Perent of Asphalt Replaed 1 Perent 2 Perent Perent Perent o. -'.l different mixes. Eah mix was represented by either one or two speimens (see Table ), whih were made and tested at random. The test was performed for 1 hr in a temperatureontrolled hamber maintained at about 1 F ( C). The time-dependent deformation was measured using a pair of linear variable differential transformers (L VDTs) that were onneted to a omputer to take readings, alulate the mix stiffness, and store the data. Time-dependent mix stiffness (Sm;.) was later plotted in a log-log sale against the stiffness of reovered asphalt (Sb;,). The sale, obtained from a Van der Poe! nomograph (22), was for the same onditions as those of Sm;x For rut depth alulation, pavement must be divided into layers and sublayers; then, for a given air temperature, the asphalt effetive visosity is alulated for eah sublayer. Effetive visosity, the slope of the Sm;x-Sb;, relationship, and traffi data were used thereafter to alulate the visous part of the asphalt stiffness (Sb;" vise). This value was then used to obtain the visous part of the mix stiffness from the Sm;x Sb;, relationship. The rut depth was then alulated using the following equation: where ilh; = rut depth in the ith sublayer, ilh = total rut depth, (2) () m = orretion fator for dynami effet, h; = sublayer thikness, smix,i = mix stiffness for the ith sublayer' obtained from the Sm;x-Sb;, relationship, and Z; proportionality fator between average stress and ontat stress. The fator Z in the equation is a funtion of mix stiffness in the high-stiffness region (short loading time), the sublayer thikness, the stiffness and the thikness of other sublayers, the modulus and thikness of the base ourse, and the subgrade modulus. The values of Z were alulated using the elasti layer theory and tabulated for different ombinations of the previously mentioned variables. The expeted rut depth in eah mix was alulated for the onditions shown in Table 8. The results of the alulated rut depth are plotted in Figures and 6. It is worth mentioning that most mixes made with Gradation C failed under the stati load, some after only a few minutes of loading, so no data were obtained for thse mixes. As the results show, the general trend for all mixes ontaining fine and medium fly ash frations is that the rut depth dereases as the replaement inreases. Mixes made with Gradation A and oarse fly ash show similar results. However, mixes made with Gradation B and oarse fly ash show an inrease in rut depth as the perent of replaement inreases. The reason that the oarse fly ash inreases the rut depth as the replaement inreases for Gradation B and not A is probably that the aggregates' interloking-and hene the mix's resistane to loading-is redued by inreasing the perentage of replaement in the ase of Gradation B, whih ontains a larger amount of sand. On the other hand, Gradation A ontains a small amount of sand; hene, even with

7 Al-Suhaibani and Tons 129 Aggregate Gradalion: A 1.2 : A '1 ---<> f ine '1 fine.. "' "' ' '1 x "' ;:, ;:, "' I:., ~ 2 "' Q) a: "' Q) "" " (J) '1 1 2 Aggregate Gradalion : B --e- fine _... ~:"'-~-- - -M - '1 ~ '1 Aggregate Gradalion: fine "' Cl>... ~ Cf) 'C.B Q) 2., a:... x Cl> "'.6.' O.B 1 2 : B ~-:--~.:.:~~-~~::.">-::_-: x;. fine ~- 1 2 : C '1 '1 2...,,: o+-~~--~-.--~--.~ ~ '1 Perent of Asphalt Replaed FIGURE Soaked resilient modulus versus perentage of asphalt replaed for 6 perent A.E..6.'1.2 ~ '1 Perent of Asphalt Replaed FIGURE Index of retained tiffness versus perentage of asphalt replaed for 6 perent A.E. the addition of oar e fly ash, the interloking among large aggregate partiles-and, in turn, the rut depth- is not muh affeted by the replaement proess. For a given aggregate gradation and asphalt equivalent fine, medium, and oarse fly a h frations gave the lowest, intermediate, and highest rut depth, respetively when ompared at the same perentage of replaemenl However there are a few exeptions. Ranking mixes ontaining the tb1 ee fly ash frations in the order mentioned is expeted given the stiffening effet of eah one as was shown in the disussion of resilient modulus results. The rut depth results for the three fly ash fration are generally the same in the ase of Gradation A but not B. This ould be auributed to differene in residual asphalt lmit weight and air void between mixes made with Gradation A and those made with B. When mixes are ompared at the same aggregate gradation, the result indiate, although not very obviou ly, that inrea - ing asphalt equjvalent from to 6 perent inreases rut depth for ontrol mixes (no fly a h added) and steepens lopes for urves representing the Tut depth ver u perentage of replaement. The fir t observation ould be explained as follows: for ontrol mixes the residual a phalt is more at high than that at low asphalt equivalent. Con equently the trengtb of bigh asphalt equivalent depend more on viso ity (resi - tane to flow) of the asphalt film than on partile-to-partile ontat. This results in weaker mixes and, onsequently, higher rut value in ase of mixes made with hjgh asphalt equivalent. The as oiation of steeper slopes with higher a phalt equivalent ould be attributed to the fat that, at a ertain perentage of replaement the higher the asphalt equivalent, the higher the amount of fly ash tliat should be added to the mix to replae asphalt at a given perentage of replaement. This, in turn produes mixes With lower asphalt/aggregate ratios, resulting in stiffer mixes as the perentage of replaement inreases in a e of high asphalt equivalent.

8 TABLE 6 ESTIMATED MAIN EFFECTS FOR SOAKED RESILIENT MODULUS x 1 6 (kpa) Parameter Estimated Main Effets standard Deviation Constant Asphalt Equivalent 6 Perent Perent Perent A B o... O.OJ Perent of Asphalt Replaed 1 Perent 2 Perent JO Perent Perent oarse TABLE 7 ESTIMATED MAIN EFFECTS FOR INDEX OF RETAINED STIFFNESS Parameter Estimated Main Effets standard Deviation Constant Asphalt Equivalent 6 Perent Perent Perent J A B Perent of Asphalt Replaed 1 Perent 2 Perent Perent Perent C I

9 Al-Suhaibani and Tons 11 TABLE 8 CONDITIONS FOR CAJ:,-CULATING EXPECTED RUT DEPTH (2) Condition Asphalt-fly ash onrete layer thikness Sublayer-1 thikness Sublayer-2 thikness Sublayer- thikness Unbound base ourse thikness Subgrade modulus Axle loads/lane/day Design period Traffi growth/year Ann Arbor weather data for Asphall Equiva lenl: 6 Measurement 18mm mm mm 1 mm mm 1 8 N/m 2 2, 1 years 2 perent E E :.. Q) Asphalt Equiv alenl: « '~: \ \ \ \ 'v Asphalt Equivalenl: JC / "" ' ' ' ' ' ', 'It'' --e ,...,,-.. M +-~~..--~---.~~-.-~~..,.-~ ~~~~~~~~~~~~----, - 6 Asphall Equivalent: Asphall Equivalenl:? E,,, ' ' E,,,, ' '...,, ' _... :.. ', Q) C ~ ::> : ~~..-~~..-~~..-~~~~~ 1 2 Perent of Asphalt Replaed FIGURE 6 Estimated rut depth versus perentage of asphalt replaed for Aggregate B ~.--~---.~~-.-~~..,.-~ ~~~~~~~~~~~~~~, 6 Asphall Equivalenl: e ~~~~~~~~~~~~~ Perenl of Asphall Replaed FIGURE Estimated rut depth versus perentage of asphalt replaed for Aggregate A. For and 6 perent asphalt equivalents, the use of Gradation B resulted in a higher rut depth for ontrol mixes than those produed by Gradation A. However, for perent asphalt equivalent, the rut depth is about the same for both gradations. It was expeted that ontrol mixes made with Gradation B should give higher rut depth than those made with A (as is the ase with and 6 perent asphalt equivalent), beause A produed mixes having higher unit weight and lower air voids than those made with B. So, the behavior of mixes made with perent asphalt equivalent ould only be attributed to experimental error. For a given asphalt equivalent, hanging aggregate gradation from A to B resulted in an inrease in rut depth for mixes ontaining oarse fly ash. This ould be explained by both unit weight and air voids, beause mixes ontaining Gradation A possess both lower air voids and higher unit weight than those ontaining Gradation B (19).

10 12 TRANSPORTATION RESEARCH RECORD 12 TABLE 9 ESTIMATED MAIN EFFECTS FOR ESTIMATED RUT DEPTH (mm) Parameter Estimated Main Effets standard Deviation onstant Asphalt Equivalent 6 Perent Perent Perent, A B Perent of Asphalt Replaed 1 Perent 2 Perent Perent Perent oarse Table 9 shows the estimated main effets of various variables on the estimated rut depth. Also shown are the standard deviations of these effets. These effets are onsistent with those mentioned earlier. A desriptive model of these effets was presented earlier. CONCLUSIONS It was the purpose of this study to investigate the feasibility of using fly ash as an asphalt extender and to study the effet of fly ash partile size and other variables on the replaement proess. The following i a ummary of the main onlu ions: 1. Fly ash having partile sizes between 1 and µ is the best as an asphalt extender beause oarser partiles tend to reate more voids among aggregate partiles and finer ones tend to stiffen mixes. In the latter ase, a mix with low workability is produed that also possesses large amount of voids after ompation. 2. It was strongly indiated that inreasing sand ontent inreases the sensitivity of rutting potential of different mixes to hanges in sizes of fly a h partiles.. For a given asphalt equivalent and aggregate gradation, the three fly ash frations are very similar in their effetiveness against moisttue damage exept for mixes having perent replaement.. At perent replaement, the fine fration aused the least resistane to moisture damage.. As far as this study shows, it is reasonable to replae up to perent of asphalt volume with medium fly ash (1 through µ in ize) for dry limates; however, for moi t limates, the replaement should not exeed perent. PRACTICAL APPLICATIONS The results of various tests performed on different mixes learly show that the replaement of up to perent of the asphalt volume by medium-size fly ash (pa sing Sieve #2) proved to be not detrimental to the mixes' performane. Beau e 8 perent of mo t fly ash produed by power plants pa e Sieve #2, fly a h should be quite ati fatory when u ed to replae up to perent of the asphalt volume. lt i also expeted that the e mixe should perform well if they are in tailed in the field. Furthermore, substituting fly ash for part of the asphalt an mean savings for the road builders, fewer disposal problems for the fly ash produers, and les damage to the environment. REFERENCES 1. Summary on Use of Fly Ash in Constrution. GAI Consultant, Nov E. Tons t al. Fly Ash as Asphalt Reduer i11 Bi111mi1111s Base Courses. Universiiy of Mihigan Ann Arbor, June W. L. Chilote. Fly A h: It Possibilities as a Soure of s in Portland Cement Conrete and Bituminous Conrete l>avement Mixtures. U. S. Navy Civil Engineers Corps Bulletin, Vol. 6, Ot. 192, pp S. Torrey. Coal Ash Utilization: Fly Ash, Bottom Ash, and Slag. Poll111io11 Tehnology Review, No. 8, J. H. Faber and A. M. DiGioia, Jr. Use of Ash in Embankment Constrution. In Tra11sportation Researh Reord 9, TRB, National Researh Counil Washington, D.C., 1976, pp J. P. Capp and J. D. Spener. Fly Ash U1ilizatio11: A w1111rary of Appliatio11s a/ill Tehnology. Information Cirular 88. Bureau of Mines, U.S. Department of the Interior C. M. Weinheimer. Evaluating fmportan of the Phy ial and Chemial Properties of Fly Ash in Creating Commerial Outlets

11 Al-Suhaibani and Tons for the Material. Transation of the Amerian Soiety of Mehanial Engineers, Vol. 66, No. 6, Aug. 19, pp L. J. Minnik. New Fly Ash and Boiler Slag Uses. Tehnial Assoiation of the Pulp and Paper Industry, Vol. 2, No. 1, Jan. 199, pp F. V. Zimmer. Fly Ash as a Bituminous Filler. Information Cirular 888. Bureau of Mines, U.S. Department of the Interior, 197, pp C. A. Carpenter. A Cooperative Study of Fillers in Asphalti Conrete. Publi Roads, Vol. 27, No., De. 192, pp J. C. Rosner et al. Fly Ash as a Mineral Filler and Anti-Strip Agent for Asphalt Conrete. Pro., 6th International Ash Utilization Symposium, Vol. 1, July 1982, pp B. F. Kallas and V. P. Puzinauskas. A Study of Mineral Fillers in Asphalt Paving Mixtures. Pro., Assoiation of Asphalt Paving Tehnologists, Vol., 1961, pp E. L. Dukatz, Jr. The Effet of Mineral Fillers on Asphalt and Asphalti Conrete. M.S. thesis. Department of Civil Engineering, Pennsylvania State University, Aug E. L. Dukatz and D. A. Anderson. The Effet of Various Fillers on the Mehanial Behavior of Asphalt and Asphalti Conrete. Pro., Assoiation of Asphalt Paving Tehnologists, Vol. 9, 198, pp R. G. Ward and J. M. MGougal. Bituminous Conrete Plant Dust Colletion System: Effet of Using Reovered Dust in Paving Mix. Final Report on Researh Projet 6. West Virginia Department of Highways, Charleston, De G. W. Maupin, Jr. Effet of Baghouse s on Compation of Bituminous Conrete. Report VHTRC 81-R9. Virginia Transportation Researh Counil, Charlottesville, May D. A. Anderson and J. P. Tarris. NCHRP Report 22: Adding Dust Colletor s to Asphalt Paving Mixtures. TRB, National Researh Counil, Washington, D.C., De D. A. Anderson et al. Dust Colletor s and Their Influene on Mixtures Design. Pro., Assoiation of Asphalt Paving Tehnologists, Vol. 1, 1982, pp A. S. Suhaibani. The Use of Fly Ash as an Asphalt Extender in Asphalt Conrete Mixes. Ph.D. dissertation. University of Mihigan, Ann Arbor, R. Hass and W. Hudson. Pavement Management Systems, reprinted. Robert E. Krieger Publishing Company, Malabar, Fla., Shell Pavement Design Manual-Asphalt Pavements and Overlays for Road Traffi. Shell International Petroleum Company Ltd., London, England, C. Van der Poet. A General System Desribing the Viso-Elasti Properties of Bitumens and Its Relation to Routine Test Data. Journal of Applied Chemistry, Vol., May 19, pp Climatologial Data, Mihigan. Vol. 89, No. 1-12, 198. Publiation of this paper sponsored by Committee on Nonbituminous Components of Bituminous Paving Mixtures. 1