A Sustainable and Cost-Effective Pavement Preservation Method: Micro-Milling and Thin Overlay

Transcription

1 Tsai A Sustainable and Cost-Effective Pavement Preservation Method: Micro-Milling and Thin Overlay Yichang (James) Tsai Professor School of Civil and Environmental Engineering Georgia Institute of Technology 0 Atlantic Drive, Atlanta, GA 0 Phone: (0) -0 Fax: (0) - james.tsai@ce.gatech.edu Submission Date: //0 Word Count:,0 Tables and Figures: Total:,00+ 0 =,0

2 Tsai ABSTRACT Because of funding shortages, the Georgia Department of Transportation (GDOT), like most of its counterparts, is actively searching for improved and innovative pavement preservation methods that can help stretch limited funding to preserve its deteriorating pavements. In 00, GDOT developed a new pavement preservation method (using micro-milling in conjunction with a thin overly) that can cost-effectively replace only the deteriorated thin open-graded top surface layer (/ to / ) without removing the large sound underlying layer. GDOT, working with Georgia Tech, has developed processes, including a Ridge-to-Valley Depth (RVD) indicator (for characterizing the micro-milled surface texture) and quality control/acceptance process, to successfully implement the new method on two large-scale projects (0 lane-miles on I- near Perry and lane-miles on I- near Savannah). The new RVD indicator and its measures on these two projects were presented in earlier papers; however, the total experience of implementing these two projects has not been comprehensively presented, especially the construction processes and their quality controls. This paper presents the development of the new method, including the challenges, the construction steps, the experiences learnt during the course of the projects, and the recommendations based on the projects' results and experiences. The new method has produced satisfactory pavement performance. According to the pavement condition evaluation performed by GDOT, the pavements on the I- and I- projects are still in good condition after and years, respectively. The average saving is approximately $,000 per lane-mile, and the total savings on both projects is more than $ million, which is significant when compared to traditional pavement resurfacing method. The long-term performance monitoring on these two projects is currently underway, and their long-term cost-effectiveness will be analyzed. INTRODUCTION Pavements are one of the largest roadway infrastructures in the United States (U.S.), and they are deteriorating. Because of funding shortages, transportation agencies in the U.S. have faced the challenge of stretching their limited budgets to preserve their pavements and are searching for cost-effective, sustainable pavement preservation methods that can achieve this goal. The Georgia Department of Transportation (GDOT), one of the state DOTs that has an effective pavement resurfacing program focusing on thin overlay (), has been actively searching for alternative pavement preservation methods that maximize the return on its investment. In 00, GDOT challenged its common practices for preserving its interstate highways and developed a new pavement preservation alternative. Georgia s interstate highways were commonly constructed with a ¾ in. of Open Graded Friction Course (OGFC) or -/ in. of Porous European Mix (PEM), a coarser OGFC, on top of dense graded asphalt pavements. When the open-graded layer wore out after to years and the underlying layer is still in good condition, the common practice in Georgia was to mill out this layer together with an approximately ¾ in. to ¼ of the dense graded asphalt beneath it using conventional milling (,, ). A thin, dense graded layer (e.g., Hot-Mix Asphalt (HMA) or Stone Matrix Asphalt (SMA)) was placed on top of the milled surface, and then a thin open-graded layer is placed on top of it. Replacing only the open-graded surface layer directly over the milled surface has rarely been done due to two concerns: ) the potential for poor bonding between the open-graded surface layer and the rough milled surface and ) the potential for water entering the open-graded layer and trapped in the valleys of the rough milled surface texture. These situations could, potentially, cause delamination of the open-graded surface layer. Therefore, to prevent potential

3 Tsai delamination GDOT s practice was to mill to ¾ to ¼ in. of the dense graded asphalt mixtures while it may be still in good condition. However, the replacement of the dense graded asphalt layer is expensive, and, with a limited budget, GDOT sought opportunity to save money for preserving more pavements. To address the concern of a rough milled surface, GDOT developed a new pavement preservation method. The method uses micro-milling in conjunction with a thin overlay to costeffectively replace only the top, worn-out open-graded surface layer when the underlying layer is in good condition. A micro-milling operation can produce fine and smooth texture on the milled surface because of the dense spacing and the large numbers of teeth on the milling drum. Figure (a) shows the pavement textures on conventional-milled and micro-milled surfaces. Several states, including Virginia, Massachusetts, Utah, and New York, use micro-milling to provide a comfort ride on the milled surface before it was overlaid (, ). The use of micro-milling allows the open-graded layer to be placed directly on the fine, micro-milled surface without encountering the potential delamination problems occurring with conventional milling. Figure (b) shows the new design for using micro-milling with a thin overlay compared to the conventional milling and overlay method. Thus, it is critical to control the quality of the micromilled surface to ensure successful implementation of the new design. Conventional milling Micro-milling (a) Pavement texture on milled surfaces 0 (b) Design for overlay with conventional milling and micro-milling FIGURE Conventional-milling vs. micro-milling.

4 Tsai The new method had been applied on two interstate highways, I- near Macon in 00 and I- near Savannah in 0. Recent survey based on GDOT s Pavement Condition Evaluation System (PACES) () shows the pavements are still in good condition. During the course of the projects, GDOT, working with Georgia Tech, conducted research and various tests, and closely observed the construction to develop the processes for successful implementation of micro-milling in conjunction with a thin overlay on its interstate highways. The successful experiences of developing and implementing the new method are presented in this paper. This paper is organized as follows. Section presents the challenges and the motivation for developing an alternative pavement preservation method along with key technical considerations. Section introduces the two micro-milling projects on I- and I-. Section presents the construction processes of micro-milling and thin overlay technique, and Section discusses the micro-milling specifications developed by GDOT for achieving a smooth, uniform micro-milled surface. Finally, conclusions and recommendations for future research are presented. GDOT S MICRO-MILLING AND THIN-OVERLAY PROJECTS The first micro-milling and OGFC overlay project in Georgia was on I- near Perry, Georgia, in 00. The project was approximately. miles long ( lanes in each direction). The average annual daily traffic (AADT) was approximately 0,000 with % truck traffic. The project was constructed in, and it was last resurfaced in with the resurfacing practices commonly used at the time, which were to mill and replace / of OGFC and. of dense-graded HMA. The pavement condition evaluation in 00 showed that the top OGFC layer had deteriorated. Although load cracking and raveling appeared on the OGFC (top) surface, the underlying HMA was in good condition. Core samples were taken at the project to verify that the HMA layer was, indeed, mostly intact. Therefore, the decision was made to apply micro-mill and overlay only to replace only the top worn-out layer (, ). The project was micro-milled and paved in 00. Various tests, including a bond strength test, a tack coat application rate, and a pavement texture measurement on the milled surface, were conducted on this project for developing the quality control processes. The second project was located on I- near Savannah, Georgia. The project was approximately miles long ( lanes in each direction). It was widened and resurfaced in. The top three layers were / in. of OGFC,. in. of SMA, and in. of an Asphaltic Concrete B ( mm) Mix. The pavement condition evaluation in 00 showed that the top OGFC layer had worn out after years of service. There was severe raveling on the top layer, especially on the outside lane. In addition, some longitudinal and transverse cracks were observed. Despite the distresses observed on the top layer, the underlying SMA was still sound. This was also verified by the core samples. Therefore, the decision was made to micro-mill and inlay only the top layer (/ in. of OGFC) without disturbing the SMA layer (, ). The project was micro-milled and paved in late 0, and, due to the cold weather, the construction was stopped and resumed in March, 0. The projects on I- and I- have estimated savings of $. and $. million, respectively by not milling and replacing the. in. of SMA layer beneath the OGFC layer. Table shows the cost breakdown for the conventional milling and micro-milling based on the unit prices reported by GDOT in 0 (). While the micro-milling is slightly more expensive

5 Tsai ($,0 per lane-mile) than the conventional milling, the saving of the. in. of SMA is significant, $, per lane-mile. This results in an estimated saving of $, per lane-mile. In addition to the cost saving, the new method provided additional construction work flexibility and less traffic congestion and interference because the micro-milled pavement surface can be opened to traffic immediately after micro-milling. TABLE Costs Breakdown for Micro-milling and Conventional-milling 0 Based on GDOT s surface distress survey, the pavements on these two projects are still in good condition since the micro-milling and thin overlay was applied. GDOT has performed asphalt pavement condition evaluations on state highways annually since based on its PACES (). PACES survey involves a visual inspection of the severity and extent of ten types of distresses (e.g., load cracking, block cracking, reflective cracking, raveling, etc.). A PACES rating scale of 0 to 0 is then computed based on the distress conditions and used for planning statewide resurfacing program. A rating of 0 or above represents the pavement is in good condition. A rating of 0 triggers the need for resurfacing. Figure shows the PACES rating for the I- project. A rating of 0 was reported in 0 with minor rutting and reflective cracking. While it is difficult to estimate the remaining service life based on the limited data, Figure shows the deterioration of the micro-milling and thin overlay in the first six years is comparable to that of the overlay with conventional milling.

6 Tsai 0 0 FIGURE PACES rating on the I- project. MICRO-MILLING AND THIN-OVERLAY CONSTRUCTION PRACTICES The design of micro-milling and thin overlay puts more stringent requirements on the construction processes than does the conventional milling and overlay process. Therefore, various processes were developed as a part of the micro-milling projects to ensure good performance of the rehabbed pavement. This section presents the pertinent processes GDOT developed in the I- and I- projects, including pavement surface preparation to minimize the premature development of reflective cracking on the new open-graded surface layer, micromilling operation, quality control for the micro-milled surface texture, and paving operation. Pre-treatments Applying appropriate pre-treatments to repair the underlying layers prior to construction is critical to the life and performance of micro-milling and thin overlay because untreated distresses on the underlying layer can propagate through the thin PEM layer within a short period of time, develop into early cracking and/or raveling, and, consequently, reduce the pavement service life. Especially, water can go freely through the open-graded layer and become trapped in the cracked and/or raveled areas, which damage the base layer. To maximize the pavement service life and performance, the surface preparation was included for a) identifying areas with severe distresses deeper than the micro-milling depth and b) applying proper treatment(s) on them before the micro-milling operation. On the I- project, GDOT s engineers conducted a field observation to identify the areas with distresses deeper than / in., including load cracking Severity Levels and and raveling Severity Level, based on the GDOT s pavement condition evaluation system (PACES). The areas that required treatment were labeled and deep patched, as shown in Figures (a) to (e), and deep patched before micro-milling. The areas were milled inches deep to remove the distresses using a conventional milling operation, and a / in. thick of mat was placed in the areas to prevent the cracks from propagating onto the surface. Finally, the areas were paved using dense-grade materials (e.g., SMA).

7 Tsai (a) Field survey to identify the severe distresses (b) Mark the location for preparation 0 0 (c) Example of the mat (d) Place the mat (e) Deep patched areas FIGURE Pavement surface preparation processes. Micro-milling Operation The quality of a micro-milled surface is the key to the new design. Because water can go through the PEM layer on the top and enter the surface of the milled layer, the ridge to valley depth of the texture on the milled surface cannot be too large or it will trap water, which could lead to delamination if the water were trapped in the milled layer. Therefore, additional processes, including the quality control process using RVD measurements, were designed to meet stringent requirements. The following describes the micro-milling operation, including performing a micro-milling on a test section, checking the quality of micro-milled surface, and cleaning the milled surface. Micro-milling on a test section A test section is designed to ensure the micro-milled surface conforms to the requirements specified in GDOT Special Provision Section () before putting the micro-milling into production. Prior to construction, the contractor performs micro-milling on a,000-ft test section with a uniformly textured surface and cross section, approved by the project engineer, to meet the requirements, including RVD and HCS IRI. When any of these requirements are exceeded in the test section, work is halted, and the contractor must submit a plan detailing what steps/actions will be taken. If approved by GDOT s engineer, the contractor will construct another 00-ft test section to achieve the requirement. The failed test section will also be re-milled to achieve compliance with the requirements. The contractor cannot proceed with micro-milling unless the test section meets the requirements. The contractors on both I- and I- projects did not initially meet the requirements on the test section. The test sections helped them adjust the operation, such as milling speed and milling drum speed (revolution per minute, RPM) for the successful implementation of the entire project.

8 Tsai 0 0 Micro-milling operation The micro-milling drum, with more than 000 teeth, can mill to / in. accuracy and produce a finer milled surface than can a convention milling drum. Special attention must be paid to the micro-milled surface and milling controls, including the milling machine speed, the speed of cutting drum (RPM), and the condition of the teeth, to meet stringent requirements. During the micro-milling operation (as shown in Figure (a), the micro-milling depth is measured frequently using a gauge, and the quality of the milled surface is visually inspected. The engineer also checks the slope and cross slope using a straight edge to ensure the milled surface aligns with the design, as shown in Figure (b). The micro-milling should remove the entire OGFC layer to ensure no residuals are left on the micro-milled surface. On the I- project, originally a micro-milling of / in. was specified in the contract. However, this depth could not guarantee removal of the entire OGFC layer because some areas (close bridge) have OGFC deeper than / in. A change order was issued to use variable milling depths to enable removal of the entire OGFC residuals. Based on the actual milling situation, this has been subsequently adjusted in the revised provision, using a variable milling depth to completely remove the residuals instead of using a fixed milling depth. The speed of milling machine was closely observed on the I- project for studying the association between the speed and micro-milled surface texture. The milling machine was operated at approximately 0 feet per minute on the I- project and a speed lower than feet per minute was recommended. A higher speed (approximately feet per minute) using a high RPM was used to compensate for the original low speed, and the higher speed was used on the I- project. Instead of the milling machine speed alone, the combination of the milling machine speed, the cutting drum speed, the teeth pattern, and the underlying material need to be adjusted to ensure the final milled surface texture meets the RVD requirement. On the I- project, the milled materials were on the teeth and resulted in rough milled surface. A soap solution was added to the water sprayed on the front of the milling drum to clean the teeth. In addition, the drum teeth needed to be replaced frequently to maintain a smooth micro-milled surface. Based on the contractor's experience on the I- project, the micromilling operation has less tolerance for worn-out teeth. A few worn-out teeth will quickly develop a visible pattern that is a problem on the micro-milled surface. Therefore, the contractor developed a routine to check and replace worn-out teeth daily. (a) Micro-milling (b) Checking the micro-milled surface FIGURE Micro-milling operation. Cleaning a micro-milled surface

9 Tsai 0 A clean milled surface that is free of loose aggregates, dirt, and dust, is essential to ensure adequate bonding of the binder to the existing base layer. Therefore, a cleaning process immediately followed the micro-milling operation to remove loose aggregates and dust from the micro-milled surface. The contractor swept the surface dust using a power broomer; the dust was loaded into dump trucks using a front end loader. Compared to cleaning the conventional milled surface, the broom has to move slower to clean the micro-milled surface because the particles from the micro-milling operation are small. The cleaned surface was inspected visually because there is no clear test method available to measure it. The micromilled surface was very smooth, having a HCS IRI less than mm/km (~ 0 in/mile), which is the requirement for newly paved pavement surface. Because the micro-milled surface can provide a smooth ride, GDOT allows the micro-milled surface to be opened to traffic for days (later extended to days) before being paved, if needed. This provides the flexibility to the construction schedule, especially for the nighttime construction on interstate highways. It is notices that the micro-milled surface can be cleaned after exposed to traffic. The pockets or holes on the milled surface can still be partially filled with milling dust even after the surface is swept clean. Traffic can blow out the dust trapped in the holes and, thus, expose the clean surface. This was observed on the sections on I- and I- that were exposed to traffic, as shown in Figures a and b. However, the ridges of the milled surface textures can be smoothed out because of heavy wheel loading and tire pressure from the highway traffic and, thus, reduce the ridge-to-valley depth. Such an effect could be more obvious in the summer when the high daytime temperatures could soften the milled asphalt pavement surfaces and, thus, reduce the ridge-to-valley depth. Therefore, the days the micromilled surface is allowed to be exposed to traffic should be limited. 0 FIGURE Cleaning a micro-milled surface. Quality Control on Micro-milled Surface GDOT s laser road profiler was retrofitted for measuring RVDs using the longitudinal profiles to effectively conduct quality control on micro-milled surface texture. On the I- project,

10 Tsai GDOT s laser road profiler was operated at, approximately, 0 mph to measure the RVD and IRI after the micro-milled surface was cleaned each day. The 0.-mile RVD and IRI were reported on the site for quality acceptance in accordance with the Special Provision. The testing results provided in-time feedback to the quality of the micro-milled surface. If any of the requirements was not met, corrective action was taken before performing more micro-milling operations. Figure shows the laser road profiler and the data collected for measuring the RVD. A mean RVD was tentatively selected for compliance of the.-mm RVD on I-. Tsai et. al. (0) conducted a study to characterize micro-milled pavement texture using RVD collected on I-. In addition, Tsai et. al. (0) has explored the alternative method to have a full coverage measurement of micro-milled surface texture using D line laser imaging technology. The average RVD and HCS IRI on the I- project are presented in Table. 0 0 (a) GDOT s Laser Road Profiler (b) Reporting RVD on site FIGURE Measured RVD using GDOT s laser road profiler. TABLE Summary of RVD and HCS IRI on I- HCS IRI SB SB SB NB NB NB (mm/km) Average 0 Maximum Minimum Paving Operation After the micro-milled surface was cleaned, a tack coat was applied, and an OGFC layer was placed on top of it. GDOT requires a PEM layer to be laid when the ambient temperature is above degrees Fahrenheit. Because the micro-milled surface is smoother than a conventionalmilled surface, bonding strength tests were conducted to test the bond strength between the PEM overlay and the micro-milled surface for different tack coat application rates, 0.0 gal/yd and 0.0 gal/yd. The tests were conducted on four test sections on the I- project, and the results from the experimental tests showed that using 0.0 gal/yd tack rate produced noticeably higher bond strength between the micro-milled surface and the PEM overlay than a 0.0 gal/yd tack rate. Therefore, the tack coat rate of approximately 0.0 gal/yd was suggested micro-milled surfaces for PEM overlay paving. Following the tack coat application, the PEM was placed on top of it and compacted using a rubber roller, as shown in Figure.

11 Tsai (a) Application of tack coat (b) Compaction operation 0 (c) Newly paved PEM (d) A close look of the PEM FIGURE Paving operation and paved PED surface. Acceptance Testing GDOT s laser road profiler was, again, used for acceptance testing on the newly paved pavements. The HCS IRI (Half Car Simulated International Roughness Index) values constitute the basis for acceptance or rejection of pavement ride quality. The indices for the smoothness of the newly paved surface measured must meet a target value of mm/km and not exceed the correction index of 00 mm/km. When testing an asphaltic concrete pavement, if the data shows that a section has failed, the profiler operator shall repeat the test two more times. If two out of the three sets of test data are equal to or below the specification requirement target, then the section is accepted as passed. GDOT S SPECIFICATION MICRO-MILLING GDOT Specification Section (), Mill Asphalt Concrete Pavement, was developed to ensure the quality of the micro-milled pavement surface texture. An indicator, Ridge-to-Valley- Depth (RVD), was developed through research studies to control the quality of the micro-milled surface texture. The RVD is defined as the difference in height between the ridge (highest) and valley (lowest) points within the base length of 0 mm, which is consistent with the length specified in the ASTM E (). The RVD and smoothness are used as the indicators for monitoring the micro-milled pavement surface texture. The RVD and IRI were measured using

12 Tsai the laser road profiler, which was operated at, approximately, 0 mph, and reported for every 0. mile. The following are the requirements for the micro-milled pavement surface texture:. Any areas exceeding / in (. mm) between the ridge and valley of the mat surface or failing to meet pavement surface acceptance testing using the Laser Road Profiler shall require the underlying layer to be removed and replaced with material as directed by the Engineer at no additional cost to the Department. All corrective work shall be performed in a minimum 00-ft section.. The indices for the smoothness of the milled surface measured must meet a target value of mm/km and not exceed the correction index of 00 mm/km.. Ensure the cross slope is uniform and no depressions or slope misalignments greater than / in per ft ( mm in. m) exist when the slope is tested with a straightedge placed perpendicular to the center line. Lai et. al. (0) discussed the parameters ( percentile and mean) tentatively used on I- and I- to represent the. mm RVD threshold value and concluded the selection of an appropriate parameter should be further studied based on the long-term performance of the thin overlay (e.g., PEM layer). CONCLUSIONS AND RECOMMENDATIONS A new, sustainable pavement preservation alternative using micro-milling in conjunction with a thin overlay has been developed to cost-effectively replace only the top, worn-out surface layer (e.g., OGFC) while the underlying layer is still in good condition. This new pavement preservation method can reduce milled and recycled materials by minimizing the disturbance and damage of the layer underneath the worn-out surface layer, minimize traffic interference by opening to the traffic right after milling, improve the construction scheduling flexibility, and save money. The challenge for successfully implementing such a preservation alternative is to have smooth micro-milled surface to ensure good bonding between the OGFC or PEM layer and the milled surface. Therefore, processes, including pavement surface preparation using deep patch, quality control on the micro-milled surface, and a new performance indicator (RVD) were developed for implementing the new method on two full-scale projects, on I- near Perry in 00 and on I- near Savannah in 0. The new method has an estimated savings of $. million and $. million on the I- and I- projects, respectively, and the pavements are still in good condition. Further research is recommended as follows: The pavement surface preparation can be further studied to include detailed treatment methods needed (e.g., crack sealing or filling with different types of materials) for different prior pavement conditions, especially for low severity of cracks. The selection of an appropriate parameter (e.g., mean, percentile, or percentile) for compliance with the. mm RVD requirement should be further studied based on the longterm performance of the PEM (or OGFC) placed on the micro-milled surface. A life-cycle cost analysis needs to be conducted to compare the cost-effectiveness of overlay with micro-milling and conventional milling using test sites with similar traffic and pavement design. Long-term performance monitoring on the I- and I- projects is needed to evaluate the actual performance of the new pavement preservation method. Sensing technology can be applied for a comprehensive, full-coverage pavement condition evaluation, especially for

13 Tsai studying the change of pavement texture and detecting early stage raveling, which are, otherwise, difficult to observe. It is recommended to establish a national standard and specification on this new pavement preservation technology that is cost effective and sustainable, to provide alternative pavement preservation be developed. ACKNOWLEDGEMENTS The author would like to acknowledge the assistant of Mr. Will Murphy (District Preconstruction Engineer), Mr. David Jared (Chief, Research and Development Branch), Mr. Peter Wu (Bureau Chief, Technical Assistance Bureau), and Ms. Sheila Hines (State Bituminous Construction Engineer ) of the Georgia Department of Transportation. In addition, the author would like to thank Dr. James Lai for his inputs. REFERENCES. Tsai, Y., Y. Wu, and E. Pitts. Improving GDOT s Annual Preventive Maintenance Using A Collaborative Decision Support System. In th international conference on managing pavement assets, Calgary, Canada, 00.. Lai, J., M. Bruce, D. M. Jared, P. Y. Wu, and S. Hines. Georgia s Evaluation of Surface Texture, Interface Characteristics, and Smoothness Profile of Micro-milled Surface. Presented at th Annual Meeting of the Transportation Research Board, Washington, D.C., 00.. Lai, James S. Assessing Techniques and Performance of Thin PEM Overlay on Micro-milled Surfaces. Final Report submitted to the Office of Materials and Research, Georgia Department of Transportation, 0.. Lai, J., M. Bruce, D. M. Jared, P. Y. Wu, and S. Hines. Georgia s Pavement Preservation with Micro-milling A Follow Up Study. Submitted to th Annual Meeting of the Transportation Research, Washington, D.C., 0.. Mokarem, D. W. Hot-Mix Asphalt Placement: Virginia s Move to a two-inch Drop-Off. Report 0-R. Virginia Transportation Research Council, Richmond, Feb Brown, D. Fine-Mill Pavements for Smooth Thin Overlays. Pavement Preservation Journal, Spring 0.. GDOT. Pavement Condition Evaluation System. 00. Georgia Department of Transportation, Atlanta, GA.. Tsai, Y., Wu, Y., Lai, J., Geary, G. (0) Characterizing Micro-milled Pavement Textures Using RVD for Super-thin Resurfacing on I- Using A Road Profiler, Journal of The Transportation Research Record, No.0, pp.-0.. Tsai, Y., Wu, Y., Lewis, Z. (0) A Full-Lane Coverage Micro-Milling Pavement Surface Quality Control Using Emerging D Line Laser Imaging Technology, ASCE Journal of Transportation Engineering.. GDOT. GDOT Item Summary Mean. Georgia Department of Transportation, Atlanta, GA.. GDOT. GDOT Special Provision Section Mill Asphaltic Concrete Pavement (Micro- Mill). Georgia Department of Transportation, Atlanta, GA.. American Society of Testing and Materials. ASTM Designation E 0: Standard Practice for Calculating Pavement Macrotexture Mean Profile Depth (Reapproved 00). Annual Book of ASTM Standards, Vol..0. ASTM International, West Conshohocken, PA, 00, pp -.