Estimating compaction characteristics of engineering fill materials based on soil index parameters using artificial neural networks
2012
0 views
0 downloads
Advisor: Doç. Dr. Gürkan Özden
Abstract (EN)
Maximum dry unit weight and optimum water content are the two compaction characteristics obtained by means of Proctor test in the laboratory. Earth structures often need large quantities of soil and sometimes it may be difficult to obtain the desired amount from a single borrow area resulting in the determination of compaction characteristics of several types of soils. It may be preferable to determine compaction characteristics by means of prediction models instead of performing whole compaction test.This dissertation presents artificial neural networks prediction models for estimating compaction characteristics of coarse and fine-grained soils. A total number of two hundred soil mixtures are prepared by blending different amount of gravel-sand-clay soil materials. These soil mixtures are compacted under Standard and Modified Proctor energy levels by means of an automatic soil compactor machine. The trained networks are able to accurately estimate compaction characteristics of soil mixtures. It is found that data clustering and division methodologies have a considerable effect on performance of developed networks. Multiple linear regression models are also developed.The effect of fine content on compaction behavior is observed during the testing program. This behavior is also captured by means of computed tomography imaging technique. It is found that both maximum dry unit weight and optimum water content of the compacted soil mixtures exhibit a transitional behavior nearby a transition fine content. Prior to the transition zone, soils behave like a sand or gravel and beyond this zone fine matrix governs soil behavior. The compaction curve obtained using automatic soil compactor is also compared with that obtained by hand. It is found that the automatic soil compactor has a remarkable effect on compaction characteristics of soils compacted under Modified Proctor energy level. Regarding the Standard Proctor compaction curves, there is a slight difference between compaction characteristics of soils.
Author
Dr. Fatih Işık
How to Cite
Fatih Işık (Doctorate thesis). Estimating compaction characteristics of engineering fill materials based on soil index parameters using artificial neural networks, 2012, Dokuz Eylül University.
License
Tüm Hakları Saklıdır
This work is shared under the specified license terms.
More theses from Dokuz Eylül University
- Analysis of speech clarity parameters in open plans offices(2021)
- AFAD gönüllülük sisteminin etkin müdahale açısından analiz(2020)
- The musical analysis of W. A. Mozart, J. N. Hummel and C. M. Von Weber' s bassoon concertos(2006)
- Examination of martian habitats from the viewpoint ofstructure(2022)
- Environmental graphic design and public installation in the context of 21st century postmodernism(2022)
- Muscula skeletal injuries of the professional dancers(2006)
