Spatial prediction model of soil bearing capacity using artificial neural networks, case study: Shahr-Azershahr

Number of pages: 100 File Format: Not Specified File Code: 29415
Year: Not Specified University Degree: Not Specified Category: Civil Engineering
  • Part of the Content
  • Contents & Resources
  • Summary of Spatial prediction model of soil bearing capacity using artificial neural networks, case study: Shahr-Azershahr

    Dissertation for M.Sc»

    Treatment: Soil and Soil

    February 2013

     

    Abstract:

    Permissible bearing capacity or safe bearing capacity is a virtual pressure that provides a safe range against collapse due to shear failure, and usually bearing capacity Allowed is a fraction of the final net bearing capacity, which is the maximum compressive stress that the soil can bear. With this description, the main basis for the construction of any building is the accurate determination of the bearing capacity of the soil and it must be determined precisely. Considering the time-consuming and costly conventional methods (in situ and laboratory) to determine the bearing capacity of the soil, in this research we will show that with the help of artificial neural networks, the bearing capacity of the soil can be predicted to an acceptable and reliable level. In order to reach the best solution, we have analyzed three artificial neural network models: back propagation 1, back layer 2 and cascade correlation 3. Based on the obtained results, the best network, i.e. the cascade correlation artificial neural network model, is selected and suggested for spatial prediction of soil bearing capacity in the study area. Keywords: soil bearing capacity, artificial neural network, cascade correlation. Chapter 1. General. 1-1 Introduction. Soil is one of the oldest and most complex engineering materials. Our ancestors used soil as building material to build tombs, flood protection and shelters. In Western civilization, from the Romans  They have mentioned the importance of soils in the stability of structures. Roman engineers, especially Vitruvius (Vitruvius), who served in a century BC, paid a lot of attention to the types of soils (sand, gravel, etc.) and the design and construction of solid foundations. At that time, there was no theoretical basis for design, and experience from trial and error was sufficient. [1]

    Columb (1773) is known as the first person who used mechanics to solve soil problems. Since the beginning of the 20th century, with the rapid growth of cities, industry and trade, the emergence of various construction systems such as skyscrapers, large public buildings, dams for electricity generation and reservoirs for water and irrigation, tunnels, roads and railways, port equipment, bridges, airports and runways, mines, hospitals, health systems, drainage systems and towers for communication systems become necessary.

    2

    have economic needs, now new questions about soils have been raised. For example, what is the state of stress in a soil mass? How to design a safe and economical foundation? How much will a building sit? And how is the stability of structures built on a soil or inside it? To answer these questions, special methods were needed and as a result, soil mechanics was born. Karl Terzaghi (1883-1963) is the undeniable father of soil mechanics. The publication of his book called " Erdbaumechanik " In 1925, he laid the foundation of soil mechanics and revealed the importance of soil in engineering activities. Soil mechanics, which is also called geotechnics or geomechanics, is the application of engineering mechanics in solving problems that deal with soil as a foundation and building materials. Engineering mechanics is used to understand and interpret the properties, behavior and performance of soils. [1]

    Soil mechanics is a subset of geotechnical engineering and includes the application of soil mechanics, geology and hydraulics to analyze and design geotechnical systems such as dams, earthworks, tunnels, canals, waterways, bridges, roads, buildings and solid waste burial systems. In every application of soil mechanics, there is uncertainty due to changes in soils, changes in their layers, their compositions and engineering properties. Therefore, engineering mechanics can only give part of the answers to soil problems. Experience and approximate calculations are very essential for the successful application of soil mechanics in practical problems. [1]

    Sustainability and economy are two basic beliefs of engineering design. In geotechnical engineering, the uncertainty of soil behavior, the uncertainty of applied loads and unusual cases in natural forces, lead us to choose from complex analyzes to simple analyzes or approximate methods. [1]

    3

    Loads from a building are transferred to the soil through the foundation. The foundation itself is a structure that is often made of concrete, steel or wood.The foundation must have the following two conditions for stability:

    The foundation must not collapse or become unstable under any imaginable loading.

    The leakage of the structure must be within the permissible limits.

    Usually, they use the limit equilibrium method to find answers to various problems, including the bearing capacity of foundations, the stability of retaining walls and slopes. [1]

    1-2 statement of the problem

      One of the important issues in geotechnical engineering is determining the bearing capacity of foundations for different conditions of the layers below the foundation.  The bearing capacity of the soil is the average contact stress between the soil and the foundation, which leads to the shear failure of the soil. Allowable bearing stress is the bearing capacity value, which is reduced by the safety factor. Sometimes, in places with soft soil, the soil under the foundation can have many settlements without real shear failure. In some cases, the allowable bearing stress is calculated according to the maximum allowable settlement. [1]

      Determining the bearing capacity of the soil under the foundations has long been of interest to researchers and designers in the geotechnical field, and for this reason, its design is not considered a new problem, but the use of new computational methods and the testing of proposed models for the soil and the advancement of computers bring forth new perspectives in the field of this issue, which justify new efforts in this field. 4

      The final bearing capacity of the foundation is a function of the shear strength of the soil, which has been estimated by Terzaghi, Meyerhoff, Vesik and others using different methods. Recently, to avoid spending a lot of time and money and performing multiple experiments, there has been an increasing trend towards computer tools that are similar to biological (pseudo-biological) systems.

      One of the most basic problems of soil mechanics is to calculate the reliability factor against the final failure of a soil mass, which in foundation engineering leads to the determination of the final bearing capacity of the foundation. Based on the theorems of limit states, the exact answer is obtained when the answer of the upper limit and the lower limit are the same. To stabilize the lower limit, a correct stress field that does not violate the yield criterion at any point of the problem area is considered, and based on that, the final capacity is calculated. However, in order to find the upper limit, an acceptable mechanism for rupture is assumed and the corresponding answer is determined using the equality of work of internal and external forces. The permissible bearing capacity or the safe bearing capacity is a virtual pressure that provides a safety margin against collapse due to shear rupture for the building and usually the permissible bearing capacity is a fraction of the final net bearing capacity, which is also the maximum pressure that the soil can withstand in addition to the pressure from the upper layers [1]. With this description, the main basis for the construction of the building is the accurate determination of the bearing capacity of the soil and the type and amount of materials, as well as the way to continue the work. to be specified for the construction of the building. 1-4 Research Objectives

      Considering the importance of soil bearing capacity in the construction of any structure, its precise determination is inevitable and the main goal of the above research is to predict the location of soil bearing capacity with the help of artificial neural networks. And the purpose of using artificial neural networks to determine the carrying capacity can be to reduce costs and reduce the time required to accurately determine the carrying capacity.

    1-5 main questions

    Is it possible to determine the bearing capacity of the soil in a specific location using artificial neural networks?

    To what extent can we trust the answers obtained from artificial neural networks in determining the bearing capacity of the soil?

    1-6 research hypotheses

    Spatial prediction of the bearing capacity of the soil with the help of artificial neural networks is possible with high accuracy.

  • Contents & References of Spatial prediction model of soil bearing capacity using artificial neural networks, case study: Shahr-Azershahr

    Chapter One: Generalities

    Introduction .. 2

    Statement of the problem .. 4

    The necessity and importance of the subject 5

    Research objective .. 6

    Main question .. 6

    Research assumptions .. 6

    Variables .. 7

    Research method .. 10

    Chapter Two: Theoretical foundations and research background

    2-1 Introduction .. 12

    2-2 History .. 12

    Chapter Three: Methods and Materials

    3-1 Introduction .. 18

    3-2 Introducing the methodology and method of doing the work. 19

    3-3 artificial neural networks. 20

    3-4 Mathematical Aspects .. 23

    3-5 Network Learning .. 24

    A

    Table of Contents

    Page Title

    3-6 Post Release .. 26

    3-6-1 Post Release Algorithm. 27

    3-7 Graded Shuffle Algorithms. 33

    3-8 base radius function .. 34

    3-9 cascade correlation algorithm. 36

    3-10 recurrent artificial neural networks. 38

    3-11 Self-organizing feature maps. 39

    3-12 important aspects of artificial neural network modeling. 41

    3-12-1 Selection of input and output variables. 41

    3-12-2 Data collection and processing. 42

    3-12-3 Artificial neural network design. 43

    3-12-4 Training and mutual training. 45

    3-12-5 Validation of the model. . 47

    3-13 Some other problems .. 47

    3-14 Strengths and limitations .. 48

    3-15 Some capabilities of neural networks in civil engineering. 50

    B

    Table of Contents

    Page Title

    3-16 Application of artificial neural networks in the optimization of structures. 50

    3-17 Introduction of Matlab software. 51

    3-18 modeling steps .. 53

    Chapter 4: Results of artificial neural network model for spatial prediction of soil bearing capacity

    4-1 Introduction .. 56

    4-2 Application of artificial neural network in civil engineering. 56

    4-2-1 Application of artificial neural networks in the optimization of structures. 57

    4-3 The scope of the study .. 57

    4-4 The modeling process .. 60

    4-4-1 Used parameters. 61

    4-4-2 Data sorting. 62

    4-4-3 Specifications of artificial neural network model for spatial prediction of soil bearing capacity. 64

    4-4-4 Evaluation of models .. 66

    4-4-4-1 Model construction. 66

    4-4-4-2 FFBP post release network. 66

    4-4-4-3 LRN return layer network. 70

    P

    Table of Contents

    Page Title

    4-4-4-4 CFBP cascade correlation network. 72

    Chapter Five: Discussion and Conclusion

    5-1 Evaluation of artificial neural networks to predict soil bearing capacity. 78

    5-2 Conclusion .. 81

    5-3 Suggestions .. 82

    5-4 References .. 83

Spatial prediction model of soil bearing capacity using artificial neural networks, case study: Shahr-Azershahr