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.
68 7.1 General Conclusions The objectives of this project have on the whole been achieved with the development and evaluation of the FTI system. Both the hardware and software systems of FTI are functional. The system can characterize aggregate shape, angularity, texture, surface area, and volume of a large range of aggregate sizes with high accuracy. The FTI system consists of a CCD camera, a fringe source, a parti- cle tray, and a MATLAB software package. Aggregates are placed on the adjustable-height particle tray under an angled mirror, and the CCD camera captures digital images of aggregates reflected from the angled mirror. Three images need to be cap- tured for further analysis and reconstructing of 3-D surfaces: a visual image and two images taken with lasers of wavelengths of 675 nm and 805 nm. The three images are analyzed using the MATLAB software that implements the Fourier transform method to generate a matrix of the 3-D coordinates of aggregate surfaces. The 3-D coordinate matrix is further analyzed using 2-D Fourier transform to quantify the angularity and texture of aggregates. The shape properties are directly derived from the measured three dimensions calculated from the FTI system. Based on the operations and evaluations of the FTI system, the following conclusions can be drawn: (1) The FTI system is able to characterize aggregates with a wide range of sieve sizes, from ¾ in. to #50 for individual particles. Image processing using the Fourier transform is able to generate accurate 3-D coordinates with a reso- lution of up to 35.4 µm/pixel in the x, y directions and 22 µm in the z direction. The FFT2 method can be applied to the 3-D coordinate matrix to extract angularity prop- erties of both coarse and fine aggregates as well as texture properties of coarse aggregates. The FTI analysis results also show that the angularity and texture values show different magnitudes between coarse aggregates and fine aggregates, but the relative angularity and texture rank- ings remain rational. That is to say that the angularity and texture values are aggregate size dependent. (2) The statistical analysis indicates that there is no signifi- cant difference between the FTI results and manual mea- surements of the dimensions in general. The dimension measurements are the basis for all the derived param- eters of shape, angularity, texture, surface area, and volume. This comparison actually validates the FTI meth- ods. Both ANOVA and t-test of the coarse aggregates in Set 1 demonstrate that the shape characteristics obtained through the FTI system are in good agreement with those calculated from manual measurements. The results also demonstrate the reliability of the FTI system to quanti- tatively evaluate coarse aggregates. (3) Comparison of FTI results with those from the AIMS II and UIAIA systems shows that the three aggregate imag- ing systems can analyze aggregates in Set 1 with reasonable roughness rankings, even though there are some differ- ences. Both blast furnace slag and crushed glacial gravel are shown to have higher angularity than other aggregates, followed by iron ore and copper ore. Also, rounded glacial gravel is shown by all methods to have the smoothest sur- face texture. In other words, when angularity and texture are significantly different, these systems may be sensitive enough to differentiate them; when angularity and tex- ture are not significantly different, these systems may rank them differently. (4) Since the Micro-Deval test may break particles, exposing new fractured surfaces, the FTI results of coarse aggre- gates in Set 2 presented in this report are not conclusive. This may be attributed to the small sample population of particles studied. A larger sample population of aggre- gate particles may need to be analyzed to ensure that the statistical nature of such properties is well captured. (5) Crushing has a great influence on the angularity and tex- ture for glacial gravel aggregates. Crushed glacial gravel aggregates have greater angularity and rougher surface texture than rounded glacial gravel aggregates. The FTI system has successfully captured such effects. C h a p t e r 7 Conclusions
69 (6) While the FTI system can be further enhanced through the use of different cameras and lasers, it is not realistic to use one set of hardware to reach a resolution of 1 micrometer for all aggregate sizes. It is also not realistic to image and analyze individual particles of very small sizes. (7) While it is possible to stitch several surfaces into the entire surface of a particle and calculate the volume of the par- ticle and the volumetric gradations, it is very time con- suming and not advantageous compared to other methods such as X-ray computed tomography. The capability has been developed, but no further efforts have made to calcu- late the surface area and the volume of individual particles for all size ranges. (8) Further evaluations of the FTI system should also be con- ducted via characterizing more aggregates from various sources for the robustness of the FTI system. It is essential to further test the validity and repeatability of the FFT2 analysis method. 7.2 Recommendations The original goal was to develop a system that was able to image aggregates ranging from 2 in. to #4, #4 to #200, and passing #200. While in theory the FTI approach should be able to achieve this, the commercially available components with one set of design cannot achieve it. The current FTI sys- tem is capable of imaging aggregates with the sizes passing a 1-in. sieve and retaining on a #50 sieve due to the limited view field of the CCD camera. Future research could be conducted to replace the CCD camera with one that has a larger view field to image aggregates retaining on a 1-in. sieve. Higher reso- lutions can be also achieved through appropriate selections of the components. Further evaluation of the FTI system could be conducted on more aggregates from various sources for the robustness of the FTI system and corresponding analysis method. It is essential to further test the validity and repeatability of the imaging and analysis methods. Currently the FTI system is not automatic during all oper- ations. It needs the userâs judgment to select a threshold value to identify an aggregate profile and choose reference rows to reconstruct 3-D surfaces. Further research could focus on the improvement of automatic operation. The different imaging and analysis methods for direct mea- surements of shape, angularity, and texture use different defini- tions. This prevents the development of a commonly accepted set of parameters for shape, angularity, and texture. A national joint effort to develop a consensus on defining the shape, angu- larity, and texture would be worthwhile. Once such consensus on a set of parameters has been developed, a national database on aggregate morphologi- cal characteristics could be built using the FTI system for all typical aggregates used in paving engineering. This database could provide engineers and researchers with easy access to the engineering properties of aggregates for mix design and perfor- mance assessments. Even though a lot of research has been conducted on mor- phological characteristics of aggregates, the quantitative rela- tionship between aggregate morphological characteristics and the performance of HMA and PCC is still not well under- stood. Therefore, further research could focus on developing quantitative relationships between aggregate shape, angular- ity, and texture and the performance of aggregate assembly and mixtures. In addition, the FTI system could be extended to imaging the surface of pavements for calculating rutting depth, sur- face roughness, and cracking densities, and the monitoring of pavement deformation. While it is possible to calculate the surface areas and vol- umes of individual particles, it is very time consuming. In addition, other more convenient approaches such as the XCT method are available. The team doesnât recommend work on further development of surface area and volume calculation methods for implementation at this time. Academic research along this direction may still be encouraged. 7.3 Implementation Plan and Cost Assessment 7.3.1 Implementation Plan This project developed an effective aggregate imaging technique to accurately quantify aggregate morphological characteristics such as shape, angularity, and texture. The cor- responding analysis methods are easy to interpret and imple- ment. The FTI system could be implemented for better quality control of aggregates in asphalt concrete and cement concrete mix design. The FTI system could be used for academic research or QC/QA by contractors and highway agencies. If the automa- tion of the FTI system operation were improved, the upgraded FTI system could be installed on the side of a conveyor along a production line of aggregates. In this way, the upgraded FTI system could automatically detect whether the morphological characteristics of aggregates in the conveyor are of good quality during the production process. (1) The FTI system could be used to enhance the efficiency of other imaging techniques that are labor-intensive and time consuming, and sometimes subjective. Unlike other laboratory tests, the FTI system is able to objec- tively quantify the morphological properties of either an individual aggregate or a number of aggregates at the same time. (2) The FTI system could be used to analyze the change of morphological characteristics of aggregates due to
70 polishing, and image the pavement surfaces for rough- ness and distress quantification. By upgrading the FTI system with more powerful CCD cameras and install- ing the upgraded FTI system in a vehicle, it is possible to extend its application to detect pavement roughness and cracks. These potential applications indicate the possible direction of future implementation. The following implementation steps are proposed: (1) refining the system for better automation; (2) extending NCHRP Project 4-34 into another NCHRP study or a pooled-fund study to mainly evaluate the aggregate charac- teristics of a larger number of aggregates; (3) establishing crite- ria for aggregate acceptance, mix blending, and quality control and assurance for industry; (4) developing a formal AASHTO standard; and (5) extending its application to pavement performance evaluations. 7.3.2 Cost Assessment Key products of this project include an FTI image system and the MATLAB software package. Implementation costs are estimated at a total of $24,000, made up of the following: (1) A complete setup of the FTI image system is estimated at a total cost of $20,000. (2) Training of personnel to acquire and analyze images using the MATLAB software and programs can be com- pleted in a 4-hour workshop for 10 to 20 individuals for about $4,000.
