Geogrid-donatılı kum zemine oturan temellerin taşıma kapasitesi

Abstract

Donatılı zeminler, son yıllarda değişik geoteknik uygulamalarında yaygın olarak kullanılmaktadır. Sınırlı sayıda da olsa yapılan çalışmalar, taşıma kapasitesini artırmak amacıyla temel zemini içerisine çekmeye dayanıklı ve zemin ile arasında yeterli sürtünmeye sahip donatılar yerleştirilmesinin, konvansiyonal yöntemlere oranla daha hızlı, daha efektif ve daha ekonomik alternatif çözümler oluşturduğunu ortaya koymuştur. Dolayısıyla, donatılı zeminlerin yapı temelleri altında kullanılması, geoteknik mühendisliğinde potansiyeli olan daha yeni bir diğer donatılı zemin uygulamasıdır. Bu çalışmada, geogrid-donatılı kum zemine oturan temellerin taşıma kapasitesi araştırılmıştır. Donatı konfigürasyonu ve donatı rijidliği parametre alınarak bir seri laboratuvar model deneyi ve sonlu elemanlar analizleri yapılmıştır. Bu deney ve analizlerden aşağıda belirtilen genel sonuçlar elde edilmiştir; 1. Temel zemini içerisine belli derinliklere donatı yerleştirilmesi ile taşıma kapasitesinde çok önemli artışlar olmaktadır. Bununla birlikte, donatılı veya donatışız zemine oturan temellerin göçme anındaki oturma değerleri arasında önemli bir fark bulunmamaktadır. 2. Tek tabaka donatılı kumlarda birinci donatı tabakası derinliğinin optimum bir değerinde taşıma kapasitesi maksimum olmaktadır. Çok tabaka donatılı kumlarda ise, birinci donatı tabakasının temel tabanından olan uzaklığı azaldıkça taşıma kapasitesi büyümektedir. 3. Yatay olarak yerleştirilen donatı tabakaları arasındaki düşey uzaklığın optimum bir değerinde taşıma kapasitesi en büyük değerine ulaşmaktadır. 4. Temel tabanından itibaren belli bir zon içerisinde kalan donatı tabakalarının sayısı ile taşıma kapasitesi önemli artışlar göstermektedir. Bu kritik derinlikten sonraki donatı tabakalarının taşıma kapasitesine etkisi önemli derecede değildir. 5. Donatı boyutunun artması ile taşıma kapasitesi de artmaktadır. Ancak, temel kenarlarından itibaren belli bir uzaklıktaki donatı boyutunun taşıma kapasitesine çok önemli bir etkisi yoktur. 6. Belli bir değerden sonra donatı rijidliğinin artması ile taşıma kapasitesinde gözlenen artış sınırlı kalmaktadır.

Reinforced soil can be basically defined as a composite material obtained by incorporating reinforcement elements in the form of fibers, bars, strips or sheets inside a soil mass. Reinforcement elements have to possess enough tensile strength and capacity to bond with the soil through interfacial friction in order to allow the soil-reinforcement system to act as a composite material. The concept of reinforcing an earth fill using natural materials such as rope fibers or bamboo strips for road, house, or retaining wall construction originated in ancient times. However, the first systematic and scientific study on reinforced soil can be credited to Vidal (1968). Since then, an increasing number of analytical and experimental studies on the subject have been conducted by several researchers. The reinforcement elements used in practice can be manufactured from both metal and polymer raw materials. Corrosion is a major problem for metal reinforcements. The viable alternative reinforcements are "geosynthetics" produced from polymer raw materials. Geosynthetic reinforcements can be classified into two main groups according to their production methods: (1) "geotextiles," e.g., woven, non-woven or knitted fabrics, and (2) "geotextile related products," e.g., geogrids, geonets or geomats (John, 1987). In general, geosynthetics are more easily handled and constructed, and more resistant to corrosion and bacterial action than many conventional materials including metals. Besides serving as reinforcement, geosynthetics can serve many other functions such as separation, drainage, and filtration. However, the effects of ultraviolet light on geosynthetics are generally more severe. Also, the long-term behavior of geosynthetics is still not fully understood. Woven or non-woven geotextiles and geogrids are the most common geosynthetic reinforcement elements used in practice. Woven geotextiles are usually produced from continuous monofilament fibers or slit-film strips. Non-woven geotextiles can be manufactured from continuous fibers laid down in a random pattern, then jointed together by various mechanical, thermal or chemical processes. Geogrids Vll are produced from a sheet of polymer (polyethylene or polypropylene) that is punched with closely spaced holes in a uniform pattern, then elongated uniaxially or biaxially (Koerner and Welsh, 1980). The behaviors of geotextiles and geogrids are generally similar in reinforced soils except for the mechanism of soil-reinforcement interaction. Jewell et al. (1984) and Milhgan and Palmeira (1987) pointed out that the mechanism for mobilization of factional resistance in geogrids was different from that of geotextiles. In geogrid-reinforced soils, the transverse members of the geogrid together with the longitudinal ribs frictionalry interact with soil, mobilizing interfacial shear strength opposing the lateral flow of sou. The mode of factional interaction is not just the interfacial one, but also includes the passive resistance mobilized by the bearing of soil particles against the lateral (transverse) elements. This study indicated that geogrids generally offer a higher shearing resistance than geotextiles. Also, geogrids generally exhibit higher tensile strength, higher stiffness, and lower creep than geotextiles. However, the cost of geogrids is usually higher. Recent advances in technologies related to soil reinforcement with either metal or polymer tensile inclusions have led to an extensive use of reinforced soils in geotechnical engineering applications all over the world, including Turkey. Reinforced soils have been used in many applications such as embankments over soft ground, earth retaining walls, bridge abutments, foundation mats, slope stabilization, contaminants dikes, and bulk storage to improve structural performance. Reinforced soil applications facilitate ease and speed of construction and can also reduce costs. In comparison with other applications of geosynthetic-reinforced soil, especially pertaining to geosynthetic-reinforced sou embankments or retaining walls, relatively less emphasis has been placed on reinforced soil beds (i.e., use of geosynthetics for reinforcing soil foundations). Binquet and Lee (1975a and b) were among the first to report a systematic study on bearing capacity of reinforced soil beds. Since then, a number of analytical and experimental studies on the subject have been conducted by several researchers. Even though these studies clearly indicated that reinforcement inclusions could increase the bearing capacity very significantly, there has not been a general consensus regarding the effects of reinforcement configuration on the bearing capacity of reinforced soil foundations. This study was undertaken to investigate the effects of reinforcement configuration and reinforcement stiffness on the bearing capacity of rectangular footings on geogrid-reinforced sand by performing laboratory model tests as well as finite element analyses. Chapter 1 presents an introduction to this study. In this chapter, the problem was stated, and both the research objectives and the methods of investigation were presented. Chapter 2 provides a detailed literature review on the experimental studies conducted by several researchers to investigate the bearing capacity of reinforced soil foundations. In most of these experiments, strip footings were employed, despite the fact that rectangular and square footings are far more common in viii practice. Also, geotextiles, metals (strips, sheets or bars), or fibers (natural or man- made) were used as reinforcement in most of these studies. Guido et ai (1986) conducted laboratory model tests on reinforced soil foundations and compared the performance of geotextile and geogrid as soil reinforcement. They indicated that the geogrid reinforcement was more effective than the geotextile from the standpoint of improving the bearing capacity of footings on reinforced sand. The discussions on the available literature point out the fact that further studies are needed to better understand the behavior of reinforced soil foundations. In Chapter 3, the analytical studies carried out by several researchers to predict the behavior of reinforced soil foundations were presented and discussed. Most of the theories were developed for strip foundations using the limit equilibrium method. The classical solutions based on the theory of elasticity were employed in most of these theories. Since some of the assumptions utilized in these theories are not very realistic and the nature of the problem is very complicated, significant discrepancies between the theoretical and observed experimental results often exist. A limited number of analytical studies on the subject were conducted by the finite element method. These studies indicated that a proper finite element model could be adopted for the analysis of reinforced soil foundations. The experimental study conducted as a part of this research endeavor was presented in Chapter 4. The set-up of the tests, the test procedure, and the material properties of sou and reinforcement used in the tests were explained. The test results were discussed and compared with those published in the literature. The experimental study involved performing more than one hundred laboratory loading tests. Some of the tests were repeated as many as four times to assure the accuracy of the measured results. The parameters that have been investigated are: (1) the depth to the first layer of reinforcement (u), (2) the vertical spacing of reinforcement layers (ü), (3) the number of reinforcement layers (N), and (4) the width of reinforcement sheet (Br). The model loading tests were conducted inside a cubical steel tank of 70 cm by 70 cm in plane and 100 cm in depth. The sand was emplaced in lifts inside the tank and compacted by a vibration hammer to a relative density of Dr = 70% to 73%. The sheets of reinforcement were placed horizontally at the predefined elevations. Upon reaching a prescribed height in the test tank, the top surface was levelled and the footing was placed at the center of the top surface. The test footing was a rectangular steel plate of 12.5 mm thick, 101.5 mm wide and 127 mm long. The bottom surface of the footing was smooth. The footing was loaded at a constant rate until an ultimate bearing state was reached. Static vertical loads were applied using a 100-kN capacity electrical ¦hydraulic pump. Loads transferred from the pump to a hydraulic jack were carefully recorded by a proving ring installed between the jack and the test footing. Settlements of the footing were measured using two dial gauges situated in diagonal directions. IX A clean, oven-dried, uniform quartz river sand (Yalıköy sand) was used in the tests. The reinforcement used in the tests was a uniaxial polypropylene geogrid, Terragrid GS 1000 (produced in Turkey). To compare load-settlement behavior of reinforced and unreinforced sands, a dimensionless term BCR (Bearing Capacity Ratio) is defined as: BCR = q/q0 (1) in which, q0 is the ultimate bearing capacity of unreinforced sand, and q is the ultimate bearing capacity of reinforced sand. The test results indicated that the ultimate bearing capacity of reinforced sand could be four times as high as that of unreinforced sand. However, the settlement at failure was not affected significantly by the geogrid reinforcement. The ultimate bearing capacity was mobilized at a settlement of approximately 3% to 5% of the footing width (i.e., at settlement ratio «? 3% to 5%) for all the unreinforced and reinforced sands, while the BCR varied from about 1 to 4. Also, the test results indicated that geogrid-reinforced sand could exhibit a failure mode which resembles typical punching shear failure. The effect of depth ratio on the BCR in single-layer reinforced sands (i.e., N = 1) was found to be different from that in multi-layer reinforced sands (ie., N > 1). The depth ratio is defined herein as the ratio between u and B (i.e., u/B). In single-layer reinforced sand, the tests indicated that there was an optimum value of depth ratio at which the BCR was the highest. This optimum depth ratio is around 0.25 and appears to be independent of reinforcement size. In multi-layer reinforced sands, the BCR simply decreased with the increasing of the depth ratio. The tests in the multi-layer reinforced sands indicated that there was no optimum value for the depth ratio. In both single-layer and multi-layer reinforced sands, the BCR values approached a constant when the depth ratio was greater than about 1.0. The tests in the multi-layer reinforced sands also indicated that there was an optimum value for the vertical spacing of horizontally placed reinforcement layers. The optimum vertical spacing at which the BCR is the highest appears to be around 0.15B. The BCR was found to increase with the increasing number of reinforcement layers within a depth of 3B below the footing base. However, the rate of increase in BCR was less significant beyond a depth of 1.5B. In other words, placing geogrid reinforcement beyond an effective depth of l.SB would not significantly increase the bearing capacity. The tests revealed that the BCR generally increased slightly with increasing reinforcement size (Br). The increase was more pronounced for reinforcement size ratio (Br/B) up to approximately 4.5, beyond which the BCR remained more or x less constant. In Chapter 5, the finite element computer program (DACSAR) employed in this study was introduced. The element types and material models included in the program were presented along with a detailed description of the modified Duncan hyperbolic soil model. The finite element computer program, DACSAR (Deformation Analysis Considering Stress Anisotropy and Reorientation), was originally developed by Iizuka and Ohta (1987) at the University of Kyoto in Japan. The computer program has been validated through numerous laboratory element tests, model tests, full-scale loading tests, and field tests. The modified Duncan hyperbolic soil model included in the program was also verified by three triaxial CD tests performed on Yahköy sand samples. Chou (1992) conducted a comparative study of four finite element programs: SSCOMP, CRISP, CON2D-86, and DACSAR. DACSAR was judged to be a more appropriate finite element program for the simulation of reinforced sou structures. In Chapter 5, the finite element analyses and their results were also presented. Besides the parameters of reinforcement configuration examined in the tests, the stiffness of reinforcement was also taken as a parameter in the analyses. The analyses were conducted under axi-symmetric conditions with the soil, reinforcement, loading and boundary conditions mimicking those of the experimental tests. Singh (1988), based on his experimental results, suggested that the effect of shape factors of footings on the bearing capacity of reinforced soils was insignificant. In the analyses, the rectangular footing (B = 101.5 mm, L = 127 mm) was treated as an equivalent circular plate (D = 130 mm) of the same footing area. The geogrid reinforcement and the footing were represented by a series of discrete axi-symmetric shell elements, while the sand represented by an assembly of axi-symmetric quadrilateral and triangular elements. Vertical loads were applied sequentially in equal increments of 25 kPa. Each load increment was divided into five steps to achieve better computational accuracy. The geogrid and footing were simulated as linear elastic materials. The stress- strain-strength behavior of the sand was simulated by the modified Duncan hyperbolic model (Duncan et al, 1980). The values of the soil model parameters were deduced from the results of the CD triaxial tests performed on Yahköy sand samples having the same density as that of the model loading tests. In the discussion of the results of analyses, the BCR defined at a given settlement was used. In other words, instead of the ultimate bearing capacities, q and q0 in Equation 1 were considered the average contact pressures at a given settlement in reinforced and unreinforced sand, respectively. Similar to the experimental findings, the analyses indicated that the effect of the XX depth ratio (u/D) on the BCR in single-layer reinforced sands was different from that in multi-layer reinforced sands. There was an optimum depth ratio at which the BCR was highest in single-layer reinforced sands. However, the optimum depth ratio in the single-layer reinforced sands was found to be somewhat higher than 0.25 when the settlement ratio (s/D) was higher than 6%. On the other hand, the BCR corresponding to any settlement ratio simply decreased with the increasing depth ratio in multi-layer reinforced sands. There appeared to be an optimum range of values for the vertical spacing of horizontally placed reinforcement layers in multi-layer reinforced sands. Even though the analyses did not show a clearly defined optimum value for vertical spacing, the maximum BCR values occurred between z/D = 0.2 and z/D = 0.4, depending on the number of reinforcement layers. The analyses also indicated that the BCR values increased significantly with reinforcement size and reinforcement layer number within a certain effective zone. The extent of the effective zone was found to He approximately within 1.5D from both the base and edges of the footing. In addition, the analyses indicated that increasing reinforcement stiffness beyond a certain effective value would not bring about a further increase in the bearing capacity. The axial stiffness per unit width of reinforcement was defined herein as the product of E (modulus of elasticity) and t (thickness of reinforcement). For the conditions investigated in the analyses, the effective value of reinforcement stiffness was 1000 kN/m. Chapter 6 presents a comparison of the experimental and analytical results. In both unreinforced and reinforced sands, the findings obtained from the tests and analyses are compared and discussed. The load-settlement curves obtained from both the experimental test and finite element analysis for unreinforced sand exhibited a failure mode which resembles typical general shear failure. The ultimate bearing capacity for the unreinforced sand obtained from the test (quU = 316 kPa) and the analysis (q"te = 311 kPa) are in very close agreement. However, the settlements obtained from the analysis at different contact pressures were much higher than those from the test. The settlement ratio for the unreinforced sand at failure was approximately 3% (i.e., settlement/footing width) in the test, and 16% (ie., settlement/footing diameter) in the analysis. This was attributed to the fact that compaction-induced stresses were not accounted for in the analysis. It is well known that prior stress history, which may influence stress-strain behavior of a sand very significantly, has no discernable effect on the shear strength (Lade and Duncan, 1976; Lambrechts and Leonards, 1978). In reinforced sands, the BCR values obtained from the tests and analyses at a settlement ratio of 2% were compared for varying reinforcement configurations. It can be concluded that the experimental and analytical results are generally in very good agreement. XIX Chapter 7 summarizes the main findings of this investigation. It should be noted that since the influence of foundation size and the scale effects on the bearing capacity of reinforced soil foundations have not been investigated fully, the behavior of actual foundations is not yet well known. Hence, further studies are still needed to establish more accurate design criteria for reinforced soil foundations.