| Popis: |
ÖZET Bu çalışma, Büyük Sürmeli Oteli (Turotel) Yapı Çukurundaki iki tür derin kazı destekleme sisteminin yatay hareketlerinin belirlenmesini, karşılaştırılmasını ve yorumlanmasını içerir. Bilindiği üzere 6 m 'den daha derin olan kazılar, gerek kazı çevresindeki yolların ve yapıların, gerekse inşa edilmekte olan temel çukurunda çalışanların güvenliğini sağlamak üzere desteklenirler. Bir kazıyı desteklemek için kullanılabilecek elemanlar; ankrajlar, yatay destekler, kazıklı perdeler, betonarme perdeler olmak üzere sıralanabilir. Bu elemanlar kullanılarak derin kazılar, zeminin özelliklerine ve sistemi etkileyecek yüklere göre bir destekleme sistemiyle desteklenirler. Çalışmada ele alınan destekleme sistemleri; kuyu tipi ankrajlı perdeler ve mini kazıklı ankrajlı perdelerdir. Bu iki tür destekleme sisteminin yatay hareketlerini belirle mek üzere, bir kontrol ağı kurulmuş ve her kazı kademesin de periyod ölçüleri yapılmıştır. Toplam dört periyod ölç mesine dayanılarak yapılan deformasyon analizi sonucunda, kuyu tipi ankrajlı perde üzerinde, 6; mini kazıklı ankrajlı perde üzerinde, 14 obje noktasının yatay hareketleri belirlenmiştir. Deformasyon analizinde temel amaç, mümkün olduğu kadar varsayımlardan kaçınmak ve kaçınılamayacak olanları matematik-istatistik yöntemlerle test etmektir. Bu nedenle geçerli bir fonksiyonel model kurulduktan sonra, ölçü presizyonlarının, dengeleme sonuçlarına etkilerini gösteren stokastik model kurulmuş ve geçerliliği test edilmiş tir. Ayrıca, aynı ölçü kümesine ait olmayan uyuşumsuz ölçüler, ^-testiyle belirlenerek ayıklanmıştır. Periyod ölçüleri önce tek tek, sonra ikişerli gruplar halinde tüm den serbest dengelenmiş; tümden dengelemelerde, hareket ettiği test edilerek belirlenen sabit noktalar, obje noktaları gibi düşünülmüştür. Tümden dengeleme sonucunda, her obje noktasının ti ve tj zamanlarına ait koordinatları ele alınarak, noktaların hareket edip etmedikleri analitik yöntemle belirlenmiştir. Söz konusu bütün bu işlemlerin teorik ve pratik aşamaları, tez içerisinde ele alınmaktadır. SUMMARY DETERMINATION OF HORIZONTAL DEFORMATION ON THE SUPPORTING SYSTEMS OF DEEP EXCAVATIONS This study contains the detection of horizontal movements in two types of support wall systems. These systems are the anchor supported minipile wall and the anchor supported caisson wall. Excavations deeper than six meters are supported to secure workers, neighbouring buildings and roads by se veral supporting systems. In selecting and projecting of supporting system, some subjects should be taken into consideration. - The period of the excavation, - Geologic and topographic conditions of the ground and rock formations, - The situation of underground water, - Positions and heights of the buildings, - The computation of horizontal wall movements and the estimation of the supporting system dimensions. The changes of shapes and positions on a structure are usually called, `def ormation`. Measurements made in or der to detect these changes are named, `deformation measure ments. ` The deformation analysis can be examined in three different models. These are dynamic, kinematic and static models. Static model is the most applied geodetic model. The deformation of a point can be determined by physical or geodetic measurement methods. The geodetic mea surement. method is classified into two groups: detections of vertical and horizontal deformations the vertical defor mation of a point can be determined by hydrostatic, trigonometric or geometric level measurement methods. In determinations of horizontal movements, the most advanced method is micro geodetic net method. A micro geo detic net consists of control (datum) and deformation points. Although deformation points are established on the characteristic places of the structure control points are located out of the area where deformation is expected to take place. The basic principle of the analysis is to keep away from hypotheses. Therefore, the horizontal control net is adjusted as a `free net` to avoid hypotheses concer ning the datum parameters of the geodetic net. However, VIunavoidable hypotheses are tested by mathematical-statisti cal analysis methods. These are; ment, ment, - The test of the functional model of the adjust- - The test of the stochastic -model of the adjust- - The detection of the outliers, - The test of stabilities of control points, - The test of the movements of deformation points. 2 G6, a priori variance found by experimental ways has to be harmonic with (5q2,a posteriori variance which was obtained after the adjustment. The null hypothesis is, T =Ç°1- or T = Co: 2-2 2 Av Cb C752 This value is compared with F f,. which is found from F-distribution, s' 1' 2 S= l- F X2 tl-0^,f1,f2, the null hypothesis is rejected. In other word, the weights of the observations were not selected correctly. The rates of the weights should be changed. This test is called,`global test`. A large number of measurements increases the risk of errors in instrument reading, data writing etc. There fore a check should be made in order to detect these gross errors. The gross errors are called, `outliers`. Outliers are detected during the computation and they are omitted from the observation groups. Residuals can be tested according to BAARDA, POPE and HECK for the detections of outliers.. The cof actor matrix of residuals is, vilQ = Q.. - A Q A - vv -11 - -xx - The test value is, 'V Co V vivi This value is compared with the criterion obtained from the normal distribution. If, i vi i T =. > N 4 / fa H l-CX/2 V1V1 the observation L. is an outlier. Here,Cg is a priori variance. This test which is called, `data snooping` is suggested by Baarda. If Go2 a priori variance is not known exactly,^2 a posteriori variance can be used to test, rp = x > G7 0o V uvivi The test value, T5 belongs to % -distribution with the degree of freedom, f. If inequation is realized, the observation Lj_ is an outlier. This test called,`*?! -test` is suggested by Pope in 1976. If the model which does not contain the observati on, L^ is adjusted, an other a posteriori variance Çr2 is obtained. When this varience is used, the test value is. T6 =-' ~ 0±/J Qvivi This value belongs to t-distribution with the deg ree of freedom, f-1. T = > t 6 A^//7^ > f `1/1-^/2 ^iV vivi vniIf this inequation is realized, the observation, L., is an outlier. The test called, `t-test` is suggested by Heck. follows, The error equations are given in matrix form as V = A x - 1 using the notation, N = A'P A and n = A' Pi unknown matrix, x, is, x = N. n but, the inverse of the coefficients matrix of normal equ ations is not calculated in a free network. Therefore the Moore-Penrose inverse, N+ / is used to solve unknows,x. Although they give the same solution, there are different methods for the calculation of Moore-Penrose in verse. The solution of this inverse can be obtained by adding condition equations. These condition equations are established for two translations and one rotation. (If there is not any distance measurement in a net, one conditi on equation is also established for the scale variation.) IXi=o I(Xn.y. - Yn. x.) =0 i=l,2,...,n Ori Oil These condition equations form a coefficient matrix, G, G' = 1 0 n 0 1 x n When the coefficient matrix, G, is combined with the coefficient matrix of the normal equations, Moore-Pen rose inverse can be assumed as, IXN N G G1 0 -1-1 Free net adjustment satisfies V P V = min and x'x = min. After each period is adjusted separately as a free net, control points whose positions were changed are determi ned by a test method. These control points are considered as deformation points. Then, all the observations in two periods are adjusted together as a free net by taking cont rol (datum) points which are defined to be stable. The difference vector can be calculated from two estimated coordinates pair of the point, Pj_,in combined adjustment. d. = -i ı2 il Y. ` - Y., i2 xl d. XI *. yx d! - i The statistical test of the difference vector is, di Sdi 2^o o F. ` _ the null hypothesis, H :E(d.)=0, 0. &* t £ i X r O X is rejected. Therefore it is called that the point, P., significantly had moved between period t i and period t. After all computations, it can be said that the anchor supported caisson wall is more rigid than the other wall. It was determined that the anchor supported minipile wall had horizontaly moved approximately 1,5 cm. The detections of wall movements is so important to prevent some damages. Therefore, in this study, an investigation was made about the horizontal movements of support systems in a weak and heterogeneous rock formation. XI 97 |
