Kohezyonlu zeminlerin dinamik davranışına boşluk suyu basıncının etkisi

Abstract

Ö Z E T Nükleer reaktörler gibi önemli yapıların deprem yüklerine karşı korunabilmeleri ve açık deniz yapılarının dalga yüklerinden zarar görmemeleri, zeminlerin dinamik yükler al tındaki davranışının gerçeğe yakın bir şekilde tahmin edilebilmesi ile mümkündür. Yüksek su muhtevalarından başlayarak konsolide edilen Arnavutköy Kaolini üzerinde, şartları dikkatle kontrol edilen 58 konsolidasyonlu drenaj sız dinamik deney ve 9 konsolidasyonlu drenaj sız statik deney ile 6 adet ödometre dene yi ve standart zemin sınıflandırma deneyleri yapılmıştır. Deney sonuçlarından elde edilen sonuçlar özetle şöyledir: 1) Statik deneylerden elde edilen Kritik Durum Eğrisi ile dinamik yüklemeyi takip eden statik kesme deneylerinden elde edilen Kritik Durum Eğrisi büyük bir uygunluk göstermektedir. 2) Dinamik deneyler sırasında biriken boşluk suyu basıncı, uygulanan sabit ve dinamik yükler cinsinden ifade edilebilir. 3) Dinamik yükleme sırasında biriken boşluk suyu basıncı değerinin artması ile numunenin dinamik yüklemeden sonraki statik kesmede kırıldığı deviatör gerilmenin değeri azal maktadır. 4) Dinamik yüklemeden sonra statik olarak kesilen numuneler aşırı konsolide killere benzer davranış göstermektedirler.II 5) Dinamik yükleme sırasında oluşan plastik deformasyonlar, aynı deney sırasında biriken boşluk suyu basıncı değerinden bulunabilmektedir. 6) Tekrarlı Ödometre deneyleri sonucunda zeminin boşluk oranının statik yüklemeye kıyasla daha küçük bir değere düştüğü gözlenmiştir.

III S ÜMMARY In the last two decades the importance of dynamic loading increased due to several reasons. Erection of nuclear po wer plants, the failure of which can cause hazardous results and the building of off-shore structures with unusually high repeated loadings can be counted as the two main reasons. Since the dynamic testing has not been completly standardized yet, it is difficult to compare the results of different researchers. Because of the cumulative nature of behaviour which governs dynamic properties, also errors add up to cause a larger scatter in test results in comparison with those obtained from static tests. There exists a considerable body of data on the behaviour of sands under cyclic loading conditions and engineer ig theories have been developed for particular classes of problems. Recently data for the behaviour of clays under cyclic loading have been obtained. The theories developed for static loading in the early years of soil mechanics treated different aspects of soil behaviour with different tests, formulas and theories. In the year 1958 Roscoe, Schof ield and Wroth brought a new understanding to the relation of water content, shear stress and mean effective stress and explained many aspects of existing theories with just one single theory. The development of theories on dynamic loading seems to follow a similar pattern. First experiments in this new field aimed to bring solu tions to immediate problems that appeared as mentioned above. Then, the rapidly increasing test results made a better understanding of dynamic behaviour possible. So the latest research works in this field try to explain the behaviour of soils with a single fatigue parameter and mostly with modified Critical State theories.IV Tests in this research are conducted on a remoulded, saturated cohesive soil which has highly uniform charac teristics (Arnavutköy kaolini). The liquid limit of this soil is w^ `= %65 and the plastic limit is Wp = %30. A batch of material is mixed at a water content twice the liquid limit and consolidated to 100 kN/nr in a Rowe consolidation cell. Triaxial specimens were obtained from it and consolidated isotropically to pc= 200, p^ 300 and ?p.ç^'400 kN/m2. All tests have been conducted on rormally consolidated samples. Dynamic load is applied to the system with a pneumatic actuater. The air pressure that runs the actuater is con t role d with a solenoid valve. The solenoid valve is commanded by a Servo System Modul. The S SM takes the chosen signal from the function generator, conditions it to a desired amplitude and adds a static potential to it. Then it compares this signal with the Feedback signal coming from the load-cell, and sends a command signal to the solenoid valve to apply the necessary pressure. The func tion generator can generate sinusoidal, triangle and square shaped waves in frequencies ranging from 0.1 Hz to 1.1 kHz. The S SM makes it possible to apply any level of sustained and dynamic load to the specimen. The cell pressure on the load-cell can be balanced with the aid of air pressure and a pressure regulator. To prevent static friction a Dither wave with very high frequency and low amplitude is added to the command signal. This reduces the friction coefficient from its static to its dynamic value and allows the smooth application of the periodic function. Since the load-cell is placed beneath the lower pedestal, the friction at the loading piston doesn't effect the measured vertical load at all. Displacement is measured with a LVDT. The core of the LVDT is directly connected to the actuater rod. Pore pressure is measured with a Pore Water Pressure Transducer. During dynamic testing all data is recorded with recorders. During the static shearing, which follows the dynamic loading a printer has been used. The load increments during these tests are chosen as 20 kN/m2 at the early stages of the tests and 10 kN/m2 near failure.The sustained + dynamic loading is applied to the specimen as long as the residual pore water pressure increased. After the residual pore water pressure comes to an equilibrium state, the specimen is sheared to failure statically. At every load level during static shearing the cease of pore water pressure increase is waited for, before moving to a higher stress level. Failure is defined at the dynamic tests as well as at the static tests after dynamic loading as the stress which causes the specimen to fail along a smooth failure surface. The projection of the results on the p-q plane of the static tests, which are run for comparison, are given in Fig. 4.1.. For the specimens isotropically consolidated at 200 kN/m2 the mean deviator stress at failure is found to be qf = 128 kN/m2 and the results of three tests give the mean effective stress at failure as pf= 162. For the specimens consolidated at 300 kN/m2 qf and p£ are 194 kN/m2 and 247 kN/m2 respectively. For the specimens consolidated at 400 kN/m2 the mean values of stresses at failure can be given as qj=243 kN/m2 and pf= 324 kN/m2. From those values the soil parameter M is found to be M- 0.775. The projection of the Critical State Line on the v - p plane is shown in Fig. 4.2. The mean water content of the specimens consolidated at 200 kN/m2 is 38.1 % ; at 300 kN/m2 consolidation pressure the mean water content is 35.5 % and at 400 kN/m2 consolidation pressure 33.4 %. The residual pore water pressure that builds up during dynamic loading is chosen to be the main parameter that effects the behaviour of dynamically loaded clay and will be denoted in this work by u^^. The sustained load-dynamic load combinations have been varied to contain all possible stress conditions. The results of the dynamic tests are given in Table 4.1, In this Table among u^^n and e<ün values, also the pore water pressure increase during static shear after dynamic load application (Auf) and the extra deformation observed tillVI failure are shown. The water contents of specimens measured at the end of the tests are also given in this Table. The second of values don't exist at tests where the specimens failed under dynamic loads. Many preparatory tests have been conducted and the test method and equipment have been improved before the main test series have been started. After this, every test con ducted is used for evaluation. So the danger of overseeing some results, because of unproper selection is tried to be avoided. It has been observed that the Critical State Line does not change after the dynamic load application Fig. 4.4.. The effect of dynamic loading can be seen on the stress paths of static shear following dynamic loading (Fig. 4.5, 4.6 and 4.7). It can be seen that with increasing residual value of the pore water pressure during dynamic loading, the initially normally consolidated clay behaves like an overconsolidated clay. Although the specimen fails when it reaches the critical value of q/p, it has been observed that the intersection.of the stress path with the Critical State Line happens at lower values, if compared to the static tests, and this reduction increases with increasing pore water pressure that builds up during dynamic loading (Fig. 4.8, 4.9 and 4.10). This result can be interpreted as the softening of the failure surface during dynamic loading. Since there is an inner relationship between pore pressure increase and plastic deformation, the final plastic strains and pore water pressures of specimens after dynamic loading have been compared and a good relation has been found (Fig. 4.11, 4.12 and 4.13). The above assumption of the overconsolidated-like behaviour of dynamically loaded soils has to show its effect on the pore wa ter pressure increase during static shear. As it is well known,VII overconsolidated specimens develop less pore water pressure during static shear than normally consolidated clays during static shear. A similar relation is found in our tests (Fig. 4.14, 4.15 and 4.16). Similar relation between pore pressure increase and axial strain during static loading which follows the dynamic loading has been seen to be valid (Fig. 4.17, 4.18 and 4.19). Since the residual pore water pressure that builds up during dynamic loading has been chosen as the main parameter, we have to be able to predict the increase of pore water from the loads we apply. In a three coordinated system the sustained load is placed on x^ axis, the dynamic load is placed on ^ axis and the Y axis is chosen as to be the residual pore water pressure increase u^^. The estimated surface through the test points can be given as Y - bQ ?». b-^x.. + box2* Tne b » bl and b? coefficients are determined to provide the minimum sum of squares of differences between the observed Y' s and this linear combination of the x values. When these computations are made, the surface with the minimum sum of squares of differences between the observed and calculated Y values can be found. The contours for fixed values öf u,. /p are dm *c seen in Fig. 4.20. This, surface is limited from its right side with the locus of points which failed during dynamic loading. This curve has the form: Sustained stress Dynamic stress 2= 2.12() ~ p P *c rc Dynamic stress -2.25 ()*0.60 Pc From the odeometer tests it can be seen that the increased intensity of dynamic loading, causes increased densifica- tion of the soil when compared with the virgin consolida tion line. The dens if ication per cycle tends to decreaseVIII with increasing application number. The final results drawn can be summarized as follows. 1) 58 dynamic, 9 static and 6 odeometer tests have been conducted on samples consolidated from a water content of twice the liquid limit to different consolidation pressures. 2) The Critical State parameters coincide for the results before and after dynamic loading. The parameters are M =0.775 X = -0.186 T = 2.96 3) The residual pore water pressure at the end of dynamic loading can be calculated from applied loads by the following equat ion : Where u,., Sustained stress and Dynamic stress are given in kN/nr 4) The deviator stress at failure depends on the amount of din q£ = 0.63 p - 1.30 u,. nf ? rc dm In this equation q, p and u,. are given in M/m~ 5) There is a relation between the pore pressure that builds up during dynamic loading and the extra pore pressure increase during the subsequent shearing which canIX be seen as a measure of overconsolidation: An = 0.42 p - 0.81 u,. f rc din 2 Where Au,., p and u,. are given in kN/m. x c din 6) The final plastic deformations caused by dynamic loading can be predicted from the residual pore water pressure increase: u,. == 0.31 p log e,. ?? 42 ? dm *c & dm Similar results exist between the extra increase at failure after dynamic loading and extra strain : Au_ =0.34 p log £ ? - 9.7 f c ° f Also in 'the above equations u., p and Au_ are given., »_ / Z din c t m kN/m. 7) The odeometer tests show that with repeated loads the soil tends to intensify.