ABSTRACT
This study adopted experimental design to investigate the suitability of abandoned solid waste site soil (ASWSS) as a foundation material for building construction. Measurements of geotechnical properties of stratified random soil samples of ASWSS and adjoining natural soil (NS) at depths 1.5, 2.0, 2.5, 3.0 and 3.5 m were obtained from six test points in Kaduna, Nigeria. The soil samples were subjected to sieve analysis, Atterberg limits (liquid limit, plastic limit, plasticity index and shrinkage limit), compaction, consolidation, triaxial, specific gravity tests as well as chemical characterization. Data treatment of ASWSS was carried out by applying 15% upper trim and 15% lower extended mean. These were done to forestall the effects of ‘reinforced earth scenario’ (unusual high strength spots caused by mix matrices of soil and fibrous materials) and unnoticed randomly distributed weak spots. Design data were evaluated in accordance with the provision of European code (Eurocode 7). The responses of ASWSS and NS to loadings were investigated by carrying out spread foundation designs on both of them using the same loading and geometric conditions. The two sets of designs were subjected to safety measurements by first order reliability method and Monte Carlo simulation respectively. The comparative reliability of ASWSS and NS with respect to structural loading was obtained in forms of reliability index and probability of failure. Significant differences in the geotechnical properties of ASWSS and NS were observed. The liquid and plastic limits of ASWSS fell in the ranges of 28 – 32% and 25 – 37% respectively. The angles of internal resistance ranged from 7 – 15º for ASWSS and 8 º – 17º for NS. Clay and silt accounted for up to 90% of ASWSS in some cases while as low as 9 kN/m2 cohesion was recorded. The composition of organic matters in ASWSS was found to be in the range of 2.1 – 5% while that of calcium/magnesium ranged between 106 mg/kg and 1000 mg/kg. Corrosive agents of sulphate and carbonate were found in the ranges of 235 – 903 mg/kg and 20 – 50 mg/kg respectively. The main mineral composition was quartz (silicon oxide), rutile (titanium oxide) and stolzite (lead tungsten oxide). Design values of cohesion, angle of internal resistance and unit weight of soil were obtained in the ranges of 9.5 – 12 kN/m2, 7 – 20º and 12.9 – 14.3 kN/m3 respectively for ASWSS. The safety of foundation designs on ASWSS and NS was obtained in terms of reliability index and probability of failure. Despite the record of small probability of failure of 0.00013, corresponding to reliability index of 3.75, there were few cases of zero reliability indices corresponding to probability of failure of 0.5 on ASWSS. These values placed ASWSS in the category of ‘hazardous to high’ safety index in the standard performance classification formats. Sulphate resistant cement, large reinforced concrete basement or foundations covering large areas and a minimum foundation depth of 2.0 m are recommended for all structural foundations built on abandoned solid waste sites.
TABLE OF CONTENTS
Approval i
Approval ii
Dedication iii
Acknowledgments iv
Table of Contents v
List of Tables ix
List of Figures xi
List of Plates xiii
List of Symbols/Abbreviations and Units xiv
Abstract xvii
CHAPTER ONE INTRODUCTION 1
1.1 Background 3
1.2 Statement of problem 7
1.3 Aim of the study 10
1.4. Objectives of study 10
1.5 Significance of the study 10
1.6 Delimitation of the study 11
1.7 Limitations of the study 12
1.8 Research questions 12
1.9 Organization of this study 13
CHAPTER TWO LITERATURE REVIEW 15
- Properties of solid waste soil 15
2.1.1 Classification of municipal solid waste 16
2.1.2 Compression characteristics of ASWSS 18
2.1.3 Shear Strength Characteristics of ASWSS 20
2.2 ASWSS as foundation soil 21
2.2.1 Spatial variation in ASWSS 24
2.2.2 Characteristic and representative values 25
2.3 Safety analysis of ASWSS 28
2.3.1 Risk and reliability analysis 30
2.4 Classification of geotechnical category 35
CHAPTER THREE PLAN AND METHODOLOGY 39
3.1 Location of study area 40
3.2 Design of study 46
3.3 Population and sampling technique 47
3.4 Instruments for data collection and administration 49
3.5 Procedures and methods of data analysis 50
3.5.1 Determination of engineering properties 50
3.5.2 Design values of soil data 51
3.5.3 Evaluation of reliability index and probability of failure 53
3.5.4 Reliability analysis using Hasofer-Lind approach (FORM) 53
3.5.5 Solution of reliability index equation 57
3.5.6 Reliability analysis of Monte Carlo Simulation (MCS) 60
3.5.7 Regression Analysis of ASWSS and Natural Soil properties 64
CHAPTER FOUR RESULTS AND DISCUSSION 68
4.1 Graphical representation of test result 69
4.1.1 Atterberg limits test results 69
4.1.2 Compaction test results 82
4.1.3 Consolidation and triaxial tests results 89
4.1.4 Specific gravity and sieve analysis results 102
4.2 Chemical analysis of ASWSS and NS 109
4.3 Organic matter content 135
4.4 X – ray diffraction test 139
4.5 Student’s t –test of triaxial test results 141
4.6 T – test atterberg limits, compaction and specific gravity test results 161
4.7 Evaluation of design data 184
4.7.1 Trimmed upper and extended lower mean (ASWSS) 190
4.7.2 Design values of soil properties and load effects 194
4.8 Determination of foundation width 197
4.9 Estimation of settlement value 202
4.10 Reliability of foundation design by Monte Carlo Simulation 203
4.11 Reliability of foundation design by first order reliability method (FORM) 208
4.12 Influence of foundation dimensions on the reliability of foundation design 213
4.13 Classification of ‘expected performance’ of foundation design 220
4.14 Geotechnical properties of ASWSS and NS 224
4.15 Design values 226
4.16 Safety indices of ASWSS and NS 227
4.17 Performance classification of ASWS S 228
CHAPTER FIVE CONCLUSION AND RECOMMENDATIONS 230
5.1 Summary 230
5.2 Conclusion 231
5.3 Recommendations 233
References 236
Appendix A: Safety Analysis using first order reliability method FORM 5 248
Appendix B: Determination of foundation width using MATLAB 250
Appendix C: Results of laboratory determination of geotechnical properties of ASWSS and NS 251
Appendix D: Application of t –table in the computation of t values 261
CHAPTER ONE
INTRODUCTION
In most geotechnical engineering work concerning Solid Waste Sites, efforts seemed to have been directed mainly towards discovering the mechanical properties of solid waste site soil (SWSS) in order to determine the safe and reliable landfill inclination slopes and consequently landfill capacity. In other words, landfill structure is the main focus and its safe design the concern. However, the high cost and quest for municipal space for infrastructural development invariably calls for judicious management and use of available ones by optimizing the benefits derivable from all lands including old landfills. A higher scale of this phenomenon is clearly and currently observed in most areas with the resultant economy of space almost over-emphasized in the affected wards. Establishing the geotechnical properties of SWSS is primary to the use of landfill for any engineering purpose though geotechnical investigation results by their very terms do not portray absolute conclusion of the elements they speak of, especially in their precise order. Spatial and geotechnical uncertainties exist significantly in different forms.
In addition to the wide ranges of approximations attending the practice, assessment of SWSS properties is categorically fraught with uncertainties in obtaining representative samples, time –dependent variations in soil characteristics and different or almost incompatible reactions of layers of SWSS to the applied stress values as a result of the heterogeneity of waste composition. These uncertainties are either compensated for, using relevant probabilistic and statistical theories or designs and engineering judgments are made based on incomplete geotechnical information and traditional factor of safety which by nature lacks logical competence to address the inconsistency posed by these wide ranges of uncertainties. Structural and mechanical engineering practices which deal with specified material geometries and qualities have their uncertainties arising from the prediction of tolerances to which the structural members may be built and also the stresses and environmental conditions to which they may be exposed (Phoon, 2008). However the practice is different in geotechnical engineering. Geological materials are investigated in their natural state and their conditions are, of necessity, inferred from measurements carried out on limited sample sizes (Baecher and Christian, 2003). The uncertaintities therefore, arise from the accuracy and completeness with which the geotechnical properties are discovered and the prediction of the mobilized resistance of soil and rock materials. Reliability-based design incorporates interalia, the principles of probability, statistics and other mathematical solutions to give expressions and make allowances for uncertain elements in the use of evaluated soil properties for foundation design.
Due to social and economic reasons, strong preference has become an important element in the settlement pattern of most of the world cities. This has created an informal polarization of settlement and uneven distribution of population among the various wards that make up the townships. Undoubtedly, waste generation has followed the same pattern.
A study of this nature therefore, is most desirable at a time when pressures on land acquisition, coupled with lack of strict regulation, have driven individuals, public and private outfits into indiscriminate use of abandoned solid waste site soil (ASWSS) for various purposes. This study comes in to equip the public with the knowledge of the varying risk levels involved and the required geotechnical procedures to adopt in the use of ASWSS for different developments.
1.1 Background
The first rational system of reference for the classification of geological materials behaviour and interpretation of observations/experience developed out of a scientific approach launched by Karl Terzaghi (1883-1963) in 1925 to study varying responses of soil and rock under differently specified stress characterization using the knowledge of physical science and engineering mechanics (Baecher and Christian 2003). From here geotechnical engineering took off and went through series of technological refinements to arrive at the present geotechnical reliability which is the integration and extension of the works of Freudenthal (1947), Purgsley (1955) and Cornell (1969).
These pioneers of geotechnical engineering warned that the results of laboratory tests, their own observation/assertions or anybody’s else do not advance conclusive narrative, since applying finite efforts to discover the state of an engineering site as laid down by nature obviously involves a number of unpleasant approximations and uncertainties which must be quantified and compensated for using reliability based methods. In practical terms, reliability deals with the relationship between the loads a system has to carry and its ability to carry those loads (Baecher and Christian 2003). The interaction between the load and resistance becomes uncertain if the quantitative evaluation of the load and resistance variables bears any uncertain elements. The widest and simplest expression of reliability is in the form of reliability index and probability of failure which may be related mathematically.
There has been an intensive search for a design model that has sufficient probabilistic and statistical robustness to address soil properties and model variability in geotechnical engineering until 1978 when load and resistance factor design (LRFD) method was discovered in the proposal submitted by Ravindra and Galambos (1978). Their submission which received approval for publication in the first edition of load and resistance factor design manual for steel construction, published in 1986, formed the basis for the development of a safety control format for steel structures in United States (US) codes. The code clearly defines the material capacity as the resistance and the aggregate stress to be imposed on the structure as the load.
This period was actually preceded by the period of implicit consensus and understanding that the traditional methods (allowable and working stress designs) were not capable of meeting the technical challenges posed by the model and soil properties uncertainties. According to Phoon (2008), ‘LRFD is used in a loose way to encompass methods that require all limit states to be checked using a specific multiple-factor format involving load and resistance factor’. It started as partial factor design approach (DA) in Europe; limit state design (LSD) in Canada and LRFD in United States, where practical application of the model and development of its code have been recorded. It is
noteworthy to say that the European design approach (DA) has undergone tremendous reliability-based refinement both in code calibration and design modeling that resulted in the recent design standard called Eurocode 7.
The National Research Council (2006) report acknowledged the fact that the inherent and unavoidable uncertainties resulting from soil properties measurements and model imprecision, and how they affect design decisions, need to be assessed by modern and improved methods, simplified enough to attract world-wide acceptance. The very attempt to evaluate the reliability of a system is an acceptance of the fact that it is unrealistic to attain absolute reliability if uncertain elements have been identified in the system and thus making probabilistic analysis imperative.
Three philosophical issues here may be identified: the readiness of geotechnical engineering community to redirect the mind set towards a reliability based design format that has a good portion of its concept based on probabilistic analysis; the need to reduce the mathematical complexity in reliability-based design (RBD) to a simplified model that can be handled by non-specialist in numerical and statistical analysis and lastly the reliability based calibration of comprehensive multiple factor formats that capture the variability in the sources of uncertainties (Robert et al, 2008).
RBD was introduced to civil engineering in form of structural reliability theory by Freudenthal (1947) and Pugsley (1955). Like any other new concept, it was developed to improve the management of failure tendencies by carrying out design based on certain criteria that consistently reduce the probability of failure to its acceptable minimum. However, the mathematical rigour involved in the application of the theory even in simple designs, made the concept unpopular. Several attempts were made towards simplifying RBD theory, but the most popular was the one by Cornell (1969) where he reformulated Gaussian equation into a model requiring just the second moment statistical descriptors (mean and covariance) of uncertain material parameters to evaluate the reliability index equation.
At this stage, however, the solution of Cornell’s equation was still found inconsistent when the performance function of factor of safety was replaced by that of margin of safety, though their limit state equations were mechanically equivalent. Hasofer and Lind (1974) addressed this problem by proposing an equation of uncertain elements containing dimensionless variables whose mean value and standard deviation are zero and unity respectively. They redefined reliability index whose geometric interpretation is the linear displacement between the closest point on the failure surface and the point defined by the expected values of the variables
The probability characteristics exhibited by inherent spatial variability in ASWSS properties make it exceptionally suitable for probability-based reliability treatment. It is clear that of all the sources of uncertainties, the natural random soil heterogeneity appears to have the worst effect on the failure mechanism of structural foundation soil. This fact and the deviation of the actual failure surface from its theoretical domain may be dramatic for ASWSS.
The earlier solution of reliability equations ‘postulates an existence of average response that depends on the average values of the soil properties’ (Phoon et al. 2003). This average response is assumed to be characteristically identical with that observed from a corresponding homogenous field having the same properties as the average properties of randomly heterogeneous soil. However recent works have revealed the possibility of the deviation of actual failure surface from its theoretically evaluated domain to a weaker part of material formation and thereby rendering the evaluated average strength of soil material higher than the actual mobilized strength. This has a serious consequence on design.
It is understandable that a comprehensive description of a random geological formation that mimics the exact spatial heterogeneity of its materials is not realistic. The progress now, therefore, is the evolution of models and standards like Eurocode 7, first and second order reliability methods (FORM and SORM), Monte Carlo simulation (MCS), cross correlation structure (CCS), cross-spectral density matrix (CSDM) etcetera, that reduce approximations and give description of random fields more accurately (Baecher and Christian, 2003 and Hema and Emil, 2015)..
1.2 Statement of Problem
The management of solid waste, though not totally neglected, has witnessed several but failed attempts to make it worthwhile especially in most part of the developing world. In the absence of engineered repositories, relevant SWSS management skills and controlled disposal points, waste is indiscriminately disposed at open dumps situated at low lying areas or undeveloped and unused land masses usually not in close proximity to dwelling places (Ramaiah, et al., 2010). However, municipal expansion and proliferation of social and economic activities soon render such dump sites an environmental misnomer with subsequent abandonment of the use and re-allocation of it for infrastructural development ultimately. Not enough is seen in the treatment of SWSS by composting, and the high water content makes incineration critically seasonal.
Where professionals are involved, thorough investigation of ASWSS is often recommended with the result revealing most of the time the anticipated weakness in soil data. Upon this weakness has been based the argument against the use of ASWSS and in favour of the search for alternative sites most of the time. Traditionally, the empirical expression of these risk levels (weakness) coherently and numerically is what has been absent. This is the solution provided by the recent development in geotechnical reliability and with the risk level of ASWSS explicitly, though not unanswerably, presented by a tested and approved risk controlling model, ASWSS designs and decision making are made simple. This study was intended to make its impact in this direction.
The use of ASWSS for development, whether the preference of the developers or not, is a widely known practice despite the structural foundation failures recorded in the practice in the recent times. Developments on ASWSS require more than adherence to the mandatory provisions of building codes. The design details should be the product of tested and approved risk controlling techniques like reliability – based method. Many catastrophic and fatal failures of landfill structures and waste dumps were recorded between 1997 and 2005 world-wide, resulting into the death of over 600 people and mobilization and redistribution of over 1.5 million m3 of waste (Gandolla et al., 1979; Eid et al., 2000; Blight, 2004 and Merry et al., 2005). The environmental damage caused by these catastrophes, most of which were reported to have occurred in developing nations, was almost irreparable (Blight, 2008). One of the most recent and surprising cases was the 2011 failure of a students workshop cited on an ASWSS in a tertiary institution having a notable civil Engineering department in Nigeria.
In the face of the current scarcity of municipal land, old and abandoned Solid Waste Sites may not be allowed to waste without development. On the other hand, if failures are recorded despite claims of adequate site investigation and propriety of designs, it means something has to be done to acknowledge and accept the peculiarities of SWSS that require relevant statistical and reliability treatment of its soil data to make it a fair representation of both the tested sample and untested mass in the parent population. Until this is done, there will continue to be the tendency of either the design of structural foundation members below its failure level or elusive selection of factor of safety in a defensible and uneconomical manner.
Results obtained from modern methods (especially Monte Carlo simulation) have revealed worrisome discrepancies between the average response of spatially variable soils and the response of corresponding homogenous soil (Phoon et al 2003). For instance, Nobahar and Popescu (2000) and Griffiths et al. (2002) discovered up to 30% decrease in the mean value of bearing capacity of spatially variable soils having coefficient of variation of 50%, compared with the bearing capacity of corresponding homogenous soil with the same average soil properties.
An increase of 12% in the average settlement of spatially variable soil having coefficient of variation of 42% was equally discovered by Paice et al. (1996) over the settlement of corresponding homogenous geological formation having equivalent mean soil properties. The summary results of the work of Popescu et al. (1997) projected up to 20% increase in pore-water pressure for a non-homogenous soil deposit having coefficient of variation of 40%, over that of corresponding homogenous soil deposit with equivalent mean soil properties.
The technical issues raised by these uncertainties constituted the target of some of the recent works. It is obvious, however, that not all the modern methods have what it takes to address the effect of these discrepancies in the use of soil data for design and analysis. While appreciating the efforts of these researchers and those mentioned earlier, one question remains unanswered and that is ‘how can the combined efforts of these researchers be harnessed to solve the obvious problem of variability in geotechnical reliability?’’ This is part of the focus of this study.
1.3 Aim of the Study
The aim of the study was to explore the contrasting responses of ASWSS and adjoining NS to structural loading so as to establish the peculiar geotechnical characteristics of ASWSS.
1.4 Objectives of the Study
The specific objectives were
- To obtain the properties of ASWSS and adjourning Natural Soil (NS) for comparison with those reported in similar and recent works of other areas.
- To obtain design values of soil properties using Eurocode 7 and reliability methods for both ASWSS and adjourning natural ground.
- To evaluate the reliability indices and probabilities of failure of foundation designs for both ASWSS and adjourning natural soil in all cases using reliability based computer program (FORM5).
- To categorize, from the results of ii and iii, the safety indices of ASWSS foundation designs and the proportion of deviation from those of the natural adjourning soil and make recommendations regarding the use of ASWSS for developments.
1.5 Significance of the Study
Despite large scale investigation of random fields for representative values, shear strength characteristics and values so far reported in literature fall in an amazingly wide range as a result of SWSS field variation in its geotechnical character. The design engineer will therefore be in dilemma as to which value to adopt. Lack of the application of modern and effective probability-based reliability methods in establishing the engineering behaviour of ASWS is partly responsible for the wrong selection of design data used in the prediction of failure zones that are higher than the actual field values.
The design process of this study is illustrative and its results the direct products of reliability-based handling of ASWSS and thus may be reliably applied in judgments and decision making concerning the loads to be imposed on ASWSS sites. This will also save the community, developers and a relevant professional body further loses from ASWSS structural foundation failures and the embarrassment of being associated with such failures in the practice.
1.6 Delimitation of the Study
The application of only the geotechnical aspect of foundation design based on Eurocode 7, Monte Carlo simulation and first order reliability method defined the analytical boundary of the study. The structural component of foundation design which requires second order reliability method (SORM) was not included.
A total of six abandoned solid waste sites (ASWSS), two in Kaduna North; two in Kaduna South and two in the land between, were selected for study. This is due to the facts of literature that have shown appreciable similarity among the results of studies conducted on ASWSS. A depth range of 1.5 to 3.5m were selected for material sampling and a scheme of geotechnical investigation was designed to include the conduct/determination of: unit weight of soil, water content, direct shear box tests, atterberg limits, triaxial compression tests, grain-size distribution, consolidation and compaction test. Direct and indirect application of some of these measurements were made in relevant models while others were used in comparative assessment of the states of ASWSS and NS.
Bearing capacity measurement is conspicuously absent from the list, and this is due to the fact that there was no intention of considering stress imposition on ASWSS soil by highway structures, though that may be a possible dimension of consideration. Furthermore, special attention is often given to heterogeneous soil outcrop and exceptionally weak formation in both design and construction of highways.
1.7 Limitations of the Study
An overall or system reliability (SYSREL) evaluation requires the analysis of both the geotechnical and structural components of foundation structure. A higher order reliability method, however, is required for the structural analysis and this was not included in this study because of lack of requisite knowledge of the author in structural reliability theory and practice.
1.8 Research Questions
To provide a direction and focus to the execution of this study the following research questions are generated
- What are the differences between the engineering properties of ASWSS and those of the adjourning natural soil and how do they compare with those reported in literature?
- What procedures are employed to arrive at the appropriate design values of soil data?
- How do the reliability indices and probabilities of failure of ASWSS compare with those of corresponding natural soil in each case of design method?
- What are the tolerances and levels of safety indices of ASWSS on the standard expected performance classification table?
1.9 Organization of this Study
This study is divided into five chapters, the first of which presents the problem and current situation of ASWSS and advances some modern techniques and approaches in design and statistical treatment of data to address it. It highlights the evolution and improvement trend of these approaches and generates their application objectives to achieve the overall purpose of the study within the defined measurement and analytical ambits. It is concluded with a note on the potential weakness of the study, scholarly limitations of the author, benefits and beneficiaries of the results of the study. The second chapter describes mainly the published principles and practical approaches to a realistic characterization of ASWSS and corresponding design methodologies that take into account the randomly variable distribution of its soil character. In this section, the principles of reliability-based design, their theoretical basis and validity and the feasibility of their practical application are well discussed.
The third chapter relates to the plan for the execution of the study and contains briefs under the following subheadings; design, subjects, instrumentation and procedure/methods of data analysis. The fourth chapter presents a procedural model of ASWSS data statistics and employs first order and Monte Carlo Simulation reliability to predict the expected responses of ASWSS and its corresponding NS under differently specified stress characterization. The fifth chapter is a brief that summarizes the salient features of the study including key findings from modern design and analytical efforts to discover the true state of ASWSS. It is concluded with relevant proposals on the possible geotechnical and statistical solutions to the problems associated with the use of ASWSS for developments.
Do you need help? Talk to us right now: (+234) 08060082010, 08107932631 (Call/WhatsApp). Email: [email protected].IF YOU CAN'T FIND YOUR TOPIC, CLICK HERE TO HIRE A WRITER»
