Konsolidasyon Özelliklerinin Arttırılmış Veri Seti İle İstatistiksel Analizi

Abstract

Konsolidasyon parametrelerinin doğru ve süratli bir şekilde tahmin edilebilmesi konusu fizibilite çalışmaları aşamasında çok önemlidir. Doğruya en yakın tahminler yapabilecek bağıntıların aranması bilim adamlarına daima cazip gelmiştir. Bu tür bağıntıların tahmin gücünün artması kullanılan veri setindeki gözlem sayısının artması ile doğru orantılıdır. Bu çalışmada, mevcut 299 adet konsolidasyon deney verilerine 261 adet konsolidasyon deney verileri ilave edilerek elde edilen arttırılmış 560 adet konsolidasyon deney verileri kullanılarak, konsolidasyon parametreleri ile zemin indeks özellikleri arasında geliştirilmiş mevcut bağıntılardan daha güçlü istatistiksel ilişkiler elde edilmesi amaçlanmıştır. Bu çalışma kapsamında tamamı İ.T.Ü. zemin mekaniği laboratuvarında yürütülmüş konsolidasyon deneylerini içeren iki ayrı veri seti ve bu iki setin birbirine eklenmesi ile elde edilen üçüncü bir veri seti kullanılmıştır. Her bir set; killer, siltler, aşırı konsolide zeminler, normal konsolide zeminler, yüksek plastisiteli zeminler, düşük plastisiteli zeminler, yüksek plastisiteli killer, düşük plastisiteli killer, yüksek plastisiteli siltler ve düşük plastisiteli siltler olmak üzere 10 alt gruba ayrılmıştır. 1959-1982 tarihleri arasında yapılan 299 adet konsolidasyon deneyi ile elde edilen veri setini kullanarak yapılan çalışmaları geliştirmek amacıyla 1993-2010 yılları arasında yapılan 261 adet konsolidasyon deneyini içeren yeni bir veri seti oluşturulmuş ve eski veri seti ile toplamından oluşan 560 adet veriye ait sonuçlar istatistiksel yöntemlerle incelenmiş; sıkışma indisi, kabarma indisi ve hacimsel sıkışma katsayısı parametrelerinin indeks özellikleri ile olan korelasyonları tek değişkenli lineer bağıntılar halinde sunulmuştur. Zemin mekaniği literatüründe, ön tasarımda ve fizibilite çalışmaları sırasında kullanılmak amacıyla, indeks özellikleri ile sıkışma indisi arasında geliştirilmiş birçok bağıntı bulunmaktadır. Sıkışma indisi, konsolidasyon oturmasının belirlenmesinde kullanılan bir parametredir. Yani temel sistemi ve maliyetin belirlenmesi açısından büyük önem taşımaktadır. Bu bağıntılar sayesinde ön keşif esnasında, temel tipine karar vermek amacıyla 15 gün sürecek konsolidasyon deneyi sonucunu beklemek yerine pratik olarak hesap yapıp ön tasarım ve maliyet belirlenebilmekte ve zamandan kazanım sağlanmaktadır. Oturma hesaplarında kullanılan, sıkışma indisini belirlemeye yönelik literatürde bulunan bu bağıntıların, sıkışma indisini ne kadarlık bir hata payıyla belirleyebildiği, bu çalışmada kullanılan 560 adet konsolidasyon deneyi sonucunda hesaplanan sıkışma indisi değerleri ile sınanmıştır. Literatürde mevcut olan bu bağıntılar belirli zemin tipleri için özelleştirilmiş olarak bulunmaktadır. Bu çalışmada, veri setindeki zemin tiplerine uygun olan bağıntılarla hesaplanan sıkışma indisi değerleri ile laboratuvar deneyleri sonucu elde edilmiş sıkışma indisi değerleri aynı grafik üzerinde gösterilip, parametre belirleyen eşitliklerin sınanmasında kullanılan RMSE (Root Mean Square Error) yöntemi ile bağıntının parametreyi ne kadarlık bir hata payıyla belirlediği araştırılmıştır. Bu sınamalar neticesinde ilk öne sürülen ve günümüze kadar geçerliliğini sürdüren bazı bağıntıların yüksek hata paylarına sahip olduğu, yakın zamanda yapılan çalışmalarda elde edilenlerden ise çok güçlü bağıntıların mevcut olduğu sonucuna varılmıştır.

The soil investigation has great importance in civil engineering before the construction of building. In scope of the soil investigations, the foundation system of building is decided through performing the field and laboratory tests. There are two main factors to be considered in the design of foundation. The foundation must handle the applied loads without any collapse whereas the settlements to be occurred in the soil must stand below the allowed limits. The settlement is one of the most important problems that might occur on the soil. The loads transferred by the foundations lead the compaction of soil layers. Depending to the applied load, the soil layers are exposed to settlements. The settlement occurred as soon as load applied is named as the immediate settlement. On the other hand, the settlements, which continue long time, are named as primary and secondary consolidation settlements. Total settlements must remain the allowable limits following the construction of buildings. The consolidation settlement refers to the percolation of pore water within the soil by the applied load whereas the secondary settlements are the plastic deformations occurring after the extinction of pore water pressure. The settlement of soils is determined with the consolidation tests conducted in the laboratory. By the consolidation test, the parameters required for the settlement calculations in the field are determined and the settlement amount to be occurred in the soil is calculated by using these parameters whereas the foundation design is prepared in accordance with the stated data. The consolidation test is a test taking long time. The preparation and carry out of test together with the calculation of parameters are performed in such a period of 15 days. This case causes a considerable loss of time in the civil engineering. The correlations enabling us to acquire the consolidation parameters in a certain approximation and a short while have great importance in the civil engineering applications. In the consolidation tests, the loadings are applied to the sample gradually in order to calculate the parameters of various load phases. Because of the fact that this condition delays the settlement calculations and preliminary design phase, the consolidation parameters are determined by the empirical equations. Instead of waiting the calculation results of consolidation test taking 15 days in order to use in feasibility studies, the index tests enabling us to acquire the parameters easily have been conducted whereas a foundation system complying with the soil profile and the structure loads has been anticipated in a short while by using the correlations existed in literature and the construction method and costs of this preferred foundation system have been determined. There are correlations developed for various data sets existed in literature that might be used in calculation of settlement levels as faster, more practical and closer to the real value in the preliminary design phase. The stated correlations exist as developed for the specific local soils or as customized for both all soils and the sub-groups of soils, which are clays, silts, normally consolidated soil, over consolidated soil, low plasticity soil and high plasticity soil. In this part, the question regarding the calculation of settlements without the over consolidation ratio on an over consolidated soil might flash. However, the feasibility studies have the characteristic of anticipating the design by means of standing in the approximate and reliable side. Therefore, using the correlations developed for the normally consolidated soils as well as the over consolidated soils shall provide the settlement values enabling us to take decisions through standing in the reliable side. During the calculation of consolidation settlement together with the correlations existing in literature, regardless that it seems a disadvantage to lack knowledge on normally or over consolidated status of the soil before the conduct of consolidation test; the correlations determining the settlement amount by considering every soil as normally consolidated actually indicate that we shall stand in the reliable side within calculations. In this study, the relations between the index properties of the soil and the consolidation parameters required to be known in foundation design have been analyzed whereas it has been aimed to develop substantial statistical correlations explaining these relations. The representation capacity of correlations acquired in the regression analyses stands in direct proportion to the data set consisting of numerous observations. The aim of this study is to examine existing relationships as well as to find out statistical correlations between consolidation parameters and soil index properties based on an increased data set to 560 observations. This study has been performed in order to provide the relations between the major data sets and the parameters by increasing the number of observations with the aim to enhance the previously conducted studies (Güneş, 1986; Kılıç, 2007) whereas the substantial correlations based on high amount of observations have been developed. Two separate data sets, and a third data set which was obtained by adding these two data sets together, were used in the scope of this study which involved consolidation tests that were entirely conducted at ITU soil mechanics laboratory. Each set was divided into 10 subgroups as clays, silts, over-consolidated soils, normally consolidated soils, high plasticity soils, low plasticity soils, high plasticity clays, low plasticity clays, high plasticity silts, and low plasticity silts groups. In order to enhance the studies which were carried out by using the data set that was obtained from 299 consolidation tests which were conducted between 1959 and 1982, a new data set was formed which involved 261 consolidation tests that were conducted between 1993 and 2010, and results of 560 data, which is composed of the old data set and its cumulative, were examined by statistical method; and correlations of compression index, swell index, coefficient of volume compressibility and compression modulus parameters with index properties are presented by single variable linear relations. Many relations have been developed between index properties and compression index in the soil mechanics literature in order to use them in preliminary design. Compression index is a parameter that uses to determine the settlement of consolidation. That is of great importance for determination of foundation system and cost. Thanks to these relations can be made practical calculations in short time. The extent of error margin in determination of the compression index of these relations, which are used in settlement calculations and are available in the literature aiming to determine the compression index, is tested by means of compression index values which are calculated as a result of 560 consolidation tests that are used in this study. The relations which are found in the literature are available as customized for certain soil types. In this study, the compression index values that are calculated by relations compatible with the soil types involved in the data set and the compression index values obtained by means of laboratory tests are displayed on the same graph and the extent of the error margin of the relation in determining the parameter is investigated by means of the RMSE (Root Mean-Square Error) method which is used to test the parameter determining equations. In consequence of these tests, it is concluded that certain relations, which were initially proposed and maintained their validity until today, involved high error margins while the relations that were obtained from recent studies included very strong relations. The statistical data analysis software of SPSS has been used for the determination of significant relations. The results of consolidation tests have been analyzed with statistical methods whereas the correlations of compression index, the swell index and coefficient of volume compressibility with the index properties are provided in single variable linear relations.