Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A laboratory study of filter compatibility with implications for coursier dam Roos, Drian Morkel 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2016_february_roos_drian.pdf [ 4.49MB ]
Metadata
JSON: 24-1.0216473.json
JSON-LD: 24-1.0216473-ld.json
RDF/XML (Pretty): 24-1.0216473-rdf.xml
RDF/JSON: 24-1.0216473-rdf.json
Turtle: 24-1.0216473-turtle.txt
N-Triples: 24-1.0216473-rdf-ntriples.txt
Original Record: 24-1.0216473-source.json
Full Text
24-1.0216473-fulltext.txt
Citation
24-1.0216473.ris

Full Text

A LABORATORY STUDY OF FILTER COMPATIBILITY WITH IMPLICATIONS FOR COURSIER DAM  by  Drian Morkel Roos  B. Eng. (Civil), University of Pretoria, 2010  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Civil Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  November 2015  © Drian Roos, 2015 ii  Abstract  Modern filter criteria are routinely used in engineering practice for the design of filters in embankment dams. Although somewhat well-developed, these criteria are based on results from a variety of non-standardised test devices and methods, and are rarely validated by means of field data or full-scale testing. Furthermore, very little work has been done towards understanding how filters built before the advent of modern filter design may be assessed.  To address this knowledge gap, the Continuing Erosion Filter (CEF) test and empirical criteria for the assessment of existing filters had been developed. Decommissioned in 2003, following a long history of sinkholes, piping and seepage-related incidents, Coursier Dam presents an excellent opportunity for study. CEF tests have been conducted on soils sampled at the dam site, to determine the material susceptibility to filter incompatibility. It is concluded that the lower core from Coursier Dam is susceptible to filter incompatibility where it is in contact with a stratum of the foundation, and that this filter incompatibility may explain the occurrence of sinkholes. The finding is supported by the results of a parametric study on soil from another dam site. Furthermore, it is found that CEF testing, in conjunction with the empirical criteria for filter assessment, provides useful insights into the phenomenon of base-filter compatibility.   iii  Preface  This thesis is original and unpublished and is the work of the author, Drian Morkel Roos. iv  Table of Contents Abstract .......................................................................................................................................... ii Preface ........................................................................................................................................... iii Table of Contents ......................................................................................................................... iv List of Tables .............................................................................................................................. viii List of Figures ............................................................................................................................... ix List of Symbols ...............................................................................................................................x List of Abbreviations ................................................................................................................... xi Acknowledgements ..................................................................................................................... xii Chapter 1: Introduction ................................................................................................................1 1.1 Preamble ............................................................................................................................... 1 1.2 Problem Statement ................................................................................................................ 3 1.3 Objectives ............................................................................................................................. 4 1.4 Outline of Thesis ................................................................................................................... 5 Chapter 2: Literature Review .......................................................................................................6 2.1 Preamble ............................................................................................................................... 6 2.2 Filter Compatibility Rules for Design................................................................................... 6 2.2.1 Base Soil Regrading ....................................................................................................... 7 2.2.2 Control Points ................................................................................................................ 7 2.3 Experimental Studies Relating to the Soil Retention Criterion .......................................... 10 2.3.1 Introduction .................................................................................................................. 10 2.3.2 Bertram (1940) ............................................................................................................. 10 2.3.3 Karpoff (1955) ............................................................................................................. 11 v  2.3.4 Lafleur (1984) .............................................................................................................. 12 2.3.5 Sherard et al. (1984a) and Sherard et al. (1984b) ........................................................ 12 2.3.6 Sherard and Dunnigan (1989) ...................................................................................... 15 2.3.7 Tomlinson and Vaid (2000) ......................................................................................... 16 2.3.8 Foster and Fell (2001) .................................................................................................. 17 2.4 Coursier Dam ...................................................................................................................... 20 Chapter 3: Apparatus and Test Program ..................................................................................30 3.1 Introduction ......................................................................................................................... 30 3.2 Test Equipment ................................................................................................................... 30 3.2.1 Permeameter Cell ......................................................................................................... 31 3.2.2 Water Supply ............................................................................................................... 32 3.2.3 Inflow Boundary .......................................................................................................... 32 3.2.4 Peripheral Base/Filter Interface ................................................................................... 33 3.2.5 Outflow Boundary ....................................................................................................... 33 3.2.6 Outflow Collection....................................................................................................... 34 3.3 Test Materials...................................................................................................................... 34 3.3.1 Field Sampling ............................................................................................................. 34 3.3.2 Lower Core .................................................................................................................. 35 3.3.3 Foundation Material ..................................................................................................... 35 3.4 Specimen Reconstitution .................................................................................................... 36 3.5 CEF Test Procedure ............................................................................................................ 37 3.5.1 Test Data ...................................................................................................................... 38 3.6 Test Program ....................................................................................................................... 39 vi  3.6.1 Coursier Dam ............................................................................................................... 39 3.6.2 Parametric Study .......................................................................................................... 39 Chapter 4: CEF Test Results and Analysis ...............................................................................47 4.1 Introduction ......................................................................................................................... 47 4.2 Coursier Dam Test Results ................................................................................................. 47 4.2.1 Test LC/U4-0.25 .......................................................................................................... 47 4.2.2 Test LC/U4-1.4W ........................................................................................................ 48 4.2.3 Test LC/U4-2.6W ........................................................................................................ 48 4.2.4 Test LC/U4-3.7 ............................................................................................................ 49 4.3 Parametric Study Test Results ............................................................................................ 49 4.4 Analysis of Coursier Dam Results ...................................................................................... 50 4.5 Analysis of Parametric Study Results ................................................................................. 51 Chapter 5: Discussion and Concluding Remarks .....................................................................55 5.1 Discussion of Results .......................................................................................................... 55 5.1.1 Coursier Dam ............................................................................................................... 55 5.1.2 CEF Erosion Category ................................................................................................. 56 5.2 Concluding Remarks ........................................................................................................... 59 5.3 Recommendations for Further Study .................................................................................. 60 References .....................................................................................................................................64 Appendix A Parametric Study ...................................................................................................... 68 A.1 Introduction .................................................................................................................... 69 A.2 Apparatus ....................................................................................................................... 69 A.3 Materials......................................................................................................................... 70 vii  A.4 Test Method ................................................................................................................... 71 A.5 Test Results .................................................................................................................... 73 A.6 Analysis of Results......................................................................................................... 78 Appendix B Calculation of Foster and Fell (2001) Erosion Boundaries ...................................... 97 B.1 Coursier Dam ................................................................................................................. 98 B.2 Parametric Study ............................................................................................................ 99  viii  List of Tables Table 2-1 Soil retention criteria used by NRCS (1994), USACE (2004), FEMA (2011) and USBR (2011) ................................................................................................................................ 22 Table 2-2: Summary of research relating to the soil retention criteria ......................................... 23 Table 2-3 Soil retention criteria proposed by Sherard and Dunnigan (1989) and Foster and Fell (2001) ............................................................................................................................................ 25 Table 2-4 Some-erosion and excessive-erosion boundaries as proposed by Foster and Fell (2001)....................................................................................................................................................... 26 Table 3-1: CEF test apparatus ....................................................................................................... 40 Table 4-1: Coursier CEF test results ............................................................................................. 52 Table 5-1: CEF test results ............................................................................................................ 61  ix  List of Figures Figure 2-1: Example filter design band with control points 1 to 7 (after NRCS, 1994) ............... 27 Figure 2-2: CEF apparatus used by Foster and Fell (2001) (after Foster and Fell, 1999) ............ 28 Figure 2-3: Contours of erosion loss for the determination of the excessive-erosion boundary for base soils with d95 > 2 mm and fines content > 35% (after Foster and Fell, 1999) ...................... 29 Figure 2-4: A cross-section of Coursier Dam (Crawford-Flett, 2014) ......................................... 29 Figure 3-1: The large permeameter configured for the CEF test .................................................. 41 Figure 3-2: Cross-sectional view of the large permeameter configured for the CEF test ............ 42 Figure 3-3: View of the west abutment with sampling locations at Coursier Dam ...................... 43 Figure 3-4: Coursier lower core material before blending ............................................................ 43 Figure 3-5: Lower core material at Coursier Dam ........................................................................ 44 Figure 3-6: Sorted Coursier foundation material before blending ................................................ 44 Figure 3-7: Reconstituted U4 foundation soils ............................................................................. 45 Figure 3-8: Typical specimen reconstitution ................................................................................ 46 Figure 4-1: Top surface of base soil after testing ......................................................................... 53 Figure 4-2: Variation of flow rate with time ................................................................................. 54 Figure 5-1: Coursier Dam test results ........................................................................................... 62 Figure 5-2: CEF test results .......................................................................................................... 63  x  List of Symbols Dx (also DFx) Sieve size corresponding to x% passing on the associated grading curve for the filter soil dx (also DBx) Sieve size corresponding to x% passing on the associated grading curve for the base soil xi  List of Abbreviations CEF  Continuing-erosion filter test EE  Excessive-erosion  NE  No-erosion NEF  No-erosion filter test NSERC Natural Sciences and Engineering Research Council SE  Some-erosion  xii  Acknowledgements I would like to thank the faculty, staff and fellow students at UBC who have supported and guided me during the course of this study.  I owe a particular debt of gratitude to Prof. R.J. Fannin for his gentle guidance and commitment towards clarity of thought. He truly exemplifies the old adage: “A scholar and a gentleman”.  My thanks also extends to my fellow students who, apart from enduring the endless barrage of noisy sieving and blending of gravel, have provided hours of stimulating conversation and differing insights into our collective interests.  I thank BC Hydro for their continued interest and financial support. They have not only been an excellent industry partner but have entertained many insightful discussions and feedback. Furthermore I offer my thanks to NSERC of Canada for providing the funding enabling me to pursue full time study.  The many new friends I have come to know in Vancouver have provided both moral support and relief from the enthralling world of academic pursuits. Lastly, I owe a great debt of gratitude to my family who have continually loved and encouraged me. Special thanks go to my loving wife without whom I would not have been able or willing to pursue this study.   1  Chapter 1: Introduction 1.1 Preamble The study of soils, and how they are affected by the presence of water, is central to the art and science of geotechnical engineering. Where one soil is adjacent to another and subject to seepage flow, it may either have its finer particles eroded and washed through, or it may filter out and retain these particles. The natural phenomenon of base-filter compatibility, or incompatibility, is one that has warranted study and has busied researchers and engineers both in the field and in the laboratory since the start of the 20th century.  Since the first filter rule, developed by Terzhagi in the 1930s (first published by Drouhin, 1936), empirical filter design criteria have been developed by researchers such as Bertram (1940), Karpoff (1955) and Sherard and Dunnigan (1989) amongst many others. Building on this body of work, the Natural Resources Conservation Service (NRCS) of the United States Department of Agriculture published a design procedure for the design of filters in 1994. This procedure has since been adopted, and slightly amended, for use by several other major agencies such as the United States Army Corps of Engineers (USACE, 2004), United States Department of the Interior Bureau of Reclamation (USBR, 2011) and the Federal Emergency Management Agency (FEMA, 2011).  These procedures form the current state of practice for modern filter design.  The use of empirical tools in design is an accepted and familiar approach and is used extensively in the field of geotechnical engineering. Care should be taken when developing these tools to ensure that they are confirmed by independent research and, where feasible, they are verified 2  using full scale testing and field performance data. Although somewhat well-developed, it is important to note several shortcomings associated with the development of modern filter design criteria. These empirical rules are based on the results of different non-standardised laboratory tests which have not all been confirmed by independent researchers. Furthermore, there is a deficit of full scale testing and field performance data, resulting in uncertainties associated with the laboratory testing, the filter design criteria, and related margins of safety.  The International Committee on Large Dams (ICOLD) has conducted extensive surveys of dam incidents (ICOLD 1974, 1983, 1995) for large dams. Foster et al. (2000) present a statistical analysis of the data gathered for earth- and rockfill dams. It is reported that of all dam failures up to 1986, a total of 48.4% may be attributed to overtopping, and 46.1% to piping and seepage related incidents. Piping and seepage related incidents encompass a variety of failure mechanisms including, internal erosion, filter incompatibility, and backwards erosion piping to name but a few. ICOLD currently maintains the most extensive international database of large dams known as the World Register of Dams (WRD), which contains information on approximately 58 266 dams internationally. Embankment dams represent 63% of the dams in the WRD and considering the advent of modern filter design, herein benchmarked to the early 1990s,  many of these dams cannot be considered to have been designed using modern methods. Considering the large number of embankment dams in operation worldwide, the need for reliable tools to assess the risk of seepage and piping related problems is clear.   3  1.2 Problem Statement Even though modern design criteria may enable the design of effective filters and drains, very little work has been done towards developing tools for assessing the risk associated with existing dams which may have filters that do not meet modern design criteria. Foster and Fell (2001) completed a study of filters that do not meet modern filter design criteria, and developed the continuing-erosion filter test, which was used along with the results of other tests to establish empirical criteria for the assessment of such filters. The continuing-erosion filter (CEF) test, as well as the empirical criteria proposed by Foster and Fell (2001), serve as a tool to assess the material susceptibility to filter incompatibility associated with existing facilities that are not compliant with modern filter design criteria.   In pursuance of dams with known seepage related problems likely associated with filter incompatibility, Coursier Dam was identified for study. Coursier Dam is an embankment dam completed in 1969 and decommissioned in 2003 following a long history of seepage and piping related problems, including the occurrence of several sinkholes (Garner et al. (2004) and Seyers (2004)). Crawford-Flett (2014) conducted a desk study to identify the likely cause of the piping and sinkholes that occurred at Coursier Dam. Modern empirical techniques were used to assess the material susceptibility for seepage induced internal erosion and the potential for filter incompatibility. It was proposed that the likely cause of piping and associated sinkholes may be attributed to filter incompatibility between the foundation and core material.  The purpose of this study is to test this hypothesis in the laboratory.  4  Coursier Dam provides an opportunity for investigation of filter incompatibility at the core-foundation interface as the likely cause of sinkholes. The intact abutments of the dam provide the opportunity for the sampling of soils that may be used to conduct a suite of CEF tests to determine the material susceptibility to filter compatibility. Furthermore the results of the tests on Coursier materials provide an opportunity for discussion of the empirical assessment criteria proposed by Foster and Fell (2001).  In order to provide additional data, and a broader base for discussion, a parametric study was conducted on available materials from another dam site. Due to contractual obligations the name of the dam is not mentioned and the test work performed on materials from this dam shall hereafter be referred to as the parametric study.  1.3 Objectives The objectives of this study with regard to filter compatibility at Coursier Dam are:  To conduct laboratory CEF testing on soils from Coursier Dam to assess the material susceptibility to filter compatibility of core-foundation soils.  To comment on whether filter incompatibility may be the likely cause of sinkholes at Coursier Dam. The objectives of this study with regard to the discussion of the Foster and Fell (2001) empirical criteria are:  To conduct a parametric laboratory study of the CEF test on materials sourced from another dam site.  To discuss the utility of the empirical filter assessment criteria proposed by Foster and Fell (2001) with regard to the results of the CEF tests performed on Coursier Dam materials and those performed as part of the parametric study. 5   1.4 Outline of Thesis The remaining sections of this study appear as follows:  Chapter 2 discusses relevant literature related to modern filter design practice and introduces Coursier Dam.  Chapter 3 discusses the laboratory apparatus used to conduct the CEF tests as well as the selection and preparation of Coursier Dam materials.  Chapter 4 presents test results and analysis of Coursier test results and provides a summary of the test results from the parametric study (further details regarding the parametric study are given in Appendix A).  Chapter 5 presents a discussion of the findings of the laboratory testing, along with concluding remarks and recommendations for further research.  Appendix A presents the parametric study with relevant details concerning the CEF tests conducted on soils from an unnamed dam site.  Appendix B presents calculations of the Foster and Fell (2001) empirical boundaries for test materials from Coursier dam and from the parametric study.    6  Chapter 2: Literature Review 2.1 Preamble Since the inception of the first filter design rule by Terzhagi (Drouhin, 1936) and its application at the Bou-Hanifia Dam, many advances have been made in the development of filter design criteria with significant contributions by Bertram (1940), Karpoff (1955), Sherard et al. (1984a,b) and Sherard and Dunnigan (1989) to name but a few. Much of their work is summarised in a design procedure published by the Natural Resources Conservation Service (NRCS) in 1994, which is taken to represent modern practice, and is reviewed in more detail in the following section.  2.2 Filter Compatibility Rules for Design The Natural Resources Conservation Service (NRCS) has published a design procedure for the design of sand and gravel filters in Part 633 of the National Engineering Handbook (NRCS, 1994) and is generally accepted in engineering practice. Following this publication several other major agencies published their own design procedures, which are very similar to that of the NRCS. These include USACE (2004), USBR (2011) and FEMA (2011).   As part of these filter design procedures, control points are used to construct a gradation envelope for filter soils based on the gradation of the base soil. Figure 2-1 shows an example filter design band with the control points marked on the plot. The following section outlines the seven control points which form part of the filter design procedure published by NRCS (1994). Each control point is described and its origins mentioned, if known. The topic of base soil 7  regrading is discussed. In addition, any differences between the NRCS (1994) design procedure and the equivalent criterion in each of the USACE (2004), USBR (2011) and FEMA (2011) design procedures are noted.  2.2.1 Base Soil Regrading The practice of regrading ‘coarser’  soils was initially proposed by (Karpoff, 1955) but only established as part of engineering practice following the work of Sherard, Dunnigan and Talbot (1984). The process of regrading involves adjusting soil gradations with particles larger than 4.75 mm such that d100 ≤ 4.75 mm. All design methods (NRCS 1994, USACE 2004, USBR 2011 and FEMA 2011) require regrading of the base soil should it contain particles greater than 4.75 mm. However, USBR (2011) and FEMA (2011) state that sands and gravel base soils (with particles larger than 4.75 mm) containing less than 15% fines that are not gap-graded and not broadly-graded, should not be regraded.  2.2.2 Control Points Control point 1 – Maximum D15 (Soil Retention)  The soil retention criterion is the most prominent of all filter criteria and has accordingly been studied in the most detail. The maximum D15 value is most commonly determined in relation to the d85 of the base soil and is usually expressed as the ratio, D15/d85. This D15/d85 retention criterion, also known as the filtering criterion, serves to ensure that the filter is able to adequately prevent movement of base soil particles through it. This ratio is determined based on categorising the base soil into one of four groups. Table 2-1 shows base soil groups as well as the soil retention criteria for the NRCS (1994) design procedure.  8   The NRCS (1994) soil retention criteria are very similar to those used by the USACE (2004), USBR (2011) and FEMA (2011). Any differences are noted in parentheses in Table 2-1. The soil retention criterion forms the basis of the rest of this study - accordingly the literature that has contributed to its development is reviewed in considerable detail in Section 2.3.   Control point 2 – Minimum D15 In order to meet permeability requirements, a minimum D15 value is prescribed for filters. The NRCS (1994) defines control point 2 with            (but not smaller than 0.1mm). This criterion was proposed by Terzhagi (Drouhin, 1936) and was investigated in several other studies, including Bertram (1940), Karpoff (1955), Sherard (1953) and USBR (1947 and 1973).  USBR (2011) filter design procedure recommends using            and USACE (2004) recommend                . FEMA (2011) has no specific criterion and cites the NRCS (1994), USBR (2011) and USACE (2011) criteria.  Control points 3 and 4 – Maximum D10 and Maximum D60 In order to prevent gap graded filters and the resulting potential for segregation, the width of the filter design envelope should be restricted. For NRCS (1994) and FEMA (2011) this is accomplished by ensuring that the coefficient of uniformity (  ) is between 2 and 6 and that the ratio of minimum and maximum D15 values (control points 1 and 2, respectively) is less than 5. These criteria ensure that a relatively uniform filter design band is attained since the Unified Soil 9  Classification System (ASTM D2487-11) classifies a gravel with       and a sand with     ,  as well-graded. The specific origin of this rule is not known.  USBR (2011) limits the width of the filter design envelope by restricting the difference between the coarse and fine limits to a maximum of 35 percent passing. USACE (2004) provides no comparable criteria.   Control point 5 and 6 – Maximum D100 and Minimum D5 Karpoff (1955) conducted a laboratory testing program to develop criteria for filters to be used for engineering design. Based on his experimental results he suggested that: “The filter material should pass the 3-in. screen for minimising particle segregation and bridging during placement. Also, filters must not have more than 5 per cent minus No. 200 particles to prevent excessive movement of fines in the filter and into drainage pipes causing clogging”. Both the NRCS (1994) and USACE (2004) prescribe to these minimum D5 and maximum D100 requirements. USBR (2011) and FEMA (2011) also recommend a minimum D5 of 0.075 mm, but limit D100 to a maximum of 51 mm.  Control point 7 – Maximum D90 Control point 7 serves to determine the maximum D90 from the minimum D10 value in order to minimise the potential for segregation during placement. A minimum D10 value is estimated from the minimum D15 value (control point 2) for a coefficient of uniformity (Cu) equal to 6 and the maximum D90 value is then read from a table. This criterion is identical for the NRCS (1994), USACE (2004), USBR (2011) and FEMA (2011) design procedures. 10  2.3 Experimental Studies Relating to the Soil Retention Criterion 2.3.1 Introduction The following section examines, in more detail, research which has contributed to the development of the soil retention or filtering criteria used for the design and assessment of filters. It is important to note that some published criteria are intended for the purposes of design, and other for the purpose of establishing the boundary of filter compatibility to be used for the assessment of existing filters. Table 2-2 contains a summary of the following section, and is useful for comparison showing which criteria are intended for use in design and which show the boundary of filter compatibility. The variety of test methods, lack of standardised testing and lack of full-scale or field testing is noted.  2.3.2 Bertram (1940) Bertram (1940) conducted a series of 12 filter compatibility tests using Ottawa sand (rounded particles) and crushed quartz sand (angular particles) to reconstitute uniform gradations of both the filter and the base materials. The tests were conducted using rigid walled lucite cylinders 51 and 102 mm in diameter, into which the soils were placed and tamped. The soil specimens were then saturated and subjected to hydraulic gradients of either 6 to 8 or 18 to 20.   Bertram concluded that, for uniform sands, a D15/d85 ratio of approximately 6 defines the limit for filter compatibility. It is noted that this conclusion is supported by the results of only two tests which yielded a D15/d85 ratio of 6.5 and were conducted at the hydraulic gradients of 18 to 20.  The remaining test data suggest that the minimum critical D15/d85 ratio of approximately 8.7 may be less conservative. Finally, Bertram proposes that the minimum D15/d85 ratios are 11  considered constant for the range of hydraulic gradients 6 to 20 and are practically independent of grain shape.   2.3.3 Karpoff (1955) Karpoff (1955) completed a total of 25 filter tests with the aim of developing criteria for the selection of suitable filter gradations in design. He tested 5 non-plastic base soils against 13 uniformly graded and 12 broadly graded filters prepared using select combinations of natural sub-rounded sand and gravel. In addition he also completed 6 tests using crushed rock filters.   The test apparatus was comprised of several transparent plastic cylinders, 203 mm in diameter and 203 mm high, bolted together with a perforated steel plate placed on the bottom to simulate a slotted pipe. Base and filter materials were placed and compacted to a thickness of 203 mm, after which the specimen was subjected to hydraulic heads ranging from 0.6 to 9.1 m.  Based on his experimental work Karpoff made the following recommendations, which form part of the NRCS (1994) criteria:  Filter material should be smaller than 75 mm to prevent segregation and bridging during placement, and should not contain more than 5% material smaller than 0.075 mm to prevent excessive fines movement in the filter and into the drainage pipes.  Grading curves of the filter and base soils should be approximately parallel in the finer range to provide stability and ensure adequate functionality of the filter arrangement.  For base soils containing particles larger than 4.75 mm ,the base soil should be analysed based on the gradation regraded to be no larger than 4.75 mm. 12   2.3.4 Lafleur (1984) Lafleur (1984) performed a total of 11 filter tests on a select combination of 3 broadly graded cohesionless till base soils and 5 sand and gravel filters of varying coarseness, in order to determine the margin of safety associated with the filter design criteria of the day. The tested soils were similar to those used extensively in the James Bay project.   Lafleur developed a flexible walled permeameter, of 150 mm diameter, capable of applying a confining stress to the base-filter specimen. The filter soil was hand placed at 70% relative density to a height of 200 mm in the rubber membrane. The base soil was compacted in a 150 mm diameter mould to a height of 150 mm at optimum moisture content to a density of 97% Standard Proctor compaction and then placed on top of the filter soil. A cell pressure of approximately 800 kPa was applied and downward seepage was imposed with hydraulic gradients of up to 8 during tests lasting from 50 to 880 hours.  Lafleur found the critical D15/d85 for soils tested in this study to be 8.4, which is in reasonable agreement with the results of Bertram (1940) who found the critical ratio to be between 6.5 and 8.7.  2.3.5 Sherard et al. (1984a) and Sherard et al. (1984b) A laboratory filter study was conducted at the Soil Mechanics Laboratory, Midwest Technical Centre, U.S. Department of Agriculture in 1981 and 1982. Sherard and his co-researchers present two papers detailing the results of this extensive test program aimed at “improving the 13  understanding of the fundamental properties and behaviour of filters”. Sherard et al. (1984a) tested uniformly graded filters of D15 between 1 and 10 mm, and Sherard et al. (1984b) tested fine grained silt and clay base soils and filters with smaller D15 values.  Sherard et al. (1984a) reported the results of 20 filter tests on relatively uniform graded filters. Uniform sand bases were tested against sub-rounded to sub-angular alluvial sand and gravel filters of varying coarseness and grading curve shape. Base-filter specimens were tested in a rigid clear plastic cylinder with a 102 mm inner diameter. Filters were compacted in a saturated condition on a vibrating table to a height varying between 127 to 178 mm, and base soils were placed and lightly tamped. 392 kPa of water pressure was applied from the municipal water supply flowing downward through the specimen.  Based on the results of their test work, and comparison with earlier research, Sherard et al. (1984a) concluded the following:  For filters with D15 larger than 1.0mm the criterion D15/d85 ≤ 5 should be used.  The shape of the particle size distribution curve of the filter and the base soil do not have to be similar to ensure filter compatibility.  Sherard et al. (1984b) carried out a total of 254 filter tests to investigate the effectiveness of a filter in sealing a preformed hole. A total of 36 base soils were tested ranging from nearly non-plastic silts to highly plastic clays and highly dispersive sodium clays. A total of 25 different filter soils were prepared from sub-rounded to rounded alluvial sands and gravels.  14  Sherard et al. (1984b) conducted two types of filter tests, namely the slot test and the slurry test. The slot test was conducted in a 102 mm diameter plastic cylinder within which a 102 mm thick lightly tamped filter soil was placed against a 165 mm thick base soil, compacted to 95% standard Proctor compaction. A preformed 13 mm wide 1.5 mm thick slot was formed in the base layer. The test was conducted subjecting the filter specimen to 392 kPa of water pressure with the cylinder laid down horizontally.  The slurry test was conducted in a similar 102 mm diameter plastic cylinder. The filter was placed 102mm thick and lightly tamped. The base soil was prepared as a slurry and poured onto the filter soil, to a height of 51 to 76 mm, after which the cylinder was closed and 392 kPa of water pressure applied. The test was conducted with the cylinder in a vertical orientation.  Both the slot and the slurry test were shown to give identical and reproducible results and are deemed to conservatively assess the capacity of the filter to seal a concentrated leak through the base soil. Using the results from their extensive laboratory study of filters, Sherard et al. (1984b) concluded the following:  For silts and clays with d85 between 0.1 and 0.5 mm the filter criterion D15/d85 ≤ 5 is conservative.  Fine grained clays, with d85 between 0.03 and 0.10 mm, may be adequately filtered with sand or gravelly sand filters with an average D15 not greater than 0.5 mm.  Fine grained silts with low cohesion and low plasticity (plotting below the A-line), and with d85 between 0.03 and 0.10 mm, may be adequately filtered by sand or gravelly sand filters with an average D15 value not greater than 0.3 mm. 15  2.3.6 Sherard and Dunnigan (1989) As an extension of the test program conducted by Sherard et al. (1984a, b), Sherard and Dunnigan (1989) conducted a series of filter tests on a wide range of base soils with the aim of identifying the characteristics of filters that are able to prevent erosion. They tested four different categories of base soils (fine silts and clays, sandy silts and clays, silty and clayey sands, and clayey and silty sands) against a range of filters.   Following the slurry and slot tests used by Sherard et al. (1984b), Sherard and Dunnigan (1989) developed a new test, namely the No Erosion Filter (NEF) test. In this test the filter material was placed in a plastic cylinder (100 mm in diameter for fine soils, and 280 mm in diameter for coarse soils) and the base soil compacted at standard proctor optimum moisture content to a thickness of 25 mm for fine soils and 100 mm for coarse soils. The test specimen was then subjected to 413 kPa of water pressure.  Sherard and Dunnigan (1989) found that the no erosion filter test (NEF) was “the best available test for evaluating critical filters located downstream of impervious cores in embankment dams”. They state that: “The conditions in the test duplicate the most severe conditions that can develop inside a dam from a concentrated erosive leak through the core discharging into a filter”. A major contribution by Sherard and Dunnigan (1989) was the grouping of base soils into four different categories based on their fines content. Using the NEF test a unique boundary D15 value for each base soil was identified. This categorisation forms part of modern design procedures as evidenced in NRCS (1994), USACE (2004), FEMA (2011) and USBR (2011) (see Section 2.2.2 16  and Table 2-1). Fines content is defined as the percentage mass of soil passing the 0.075 mm sieve.   The empirical rules for soil retention, proposed by Sherard and Dunnigan (1989) for the design of filters, are summarised below: 1. For soil group 1 (fines content between 85 and 100%), D15 ≤ 9 x d85, but not smaller than 0.2 mm. 2. For soil group 2 (fines content between 40 and 85%), D15 ≤ 0.7 mm 3. For soil group 3 (fines content between 0 and 15%), D15 ≤ 4 x d85, 4. For soil group 4 (fines content between 15 and 40%), D15 ≤ (40 - A)(4d85 - 0.7)/25 + 0.7 (where A is the percentage fines of the base soil after any regrading to the 4.75 mm sieve) Please note that the empirical rules for design, as described here, differ from a companion series of empirical rules used to delineate the onset of filter incompatibility (see Table 2-3).  2.3.7 Tomlinson and Vaid (2000) Tomlinson and Vaid (2000) performed 17 filter tests on reconstituted uniform glass bead specimens in the range, 7.3 ≤ D15/d85 ≤ 12.3, to determine the critical D15/d85 ratio considering the influence of confining pressure, filter thickness, hydraulic gradient and the formation of a self-filtration zone.  A new permeameter was developed consisting of a polished stainless steel cylinder 100 mm in diameter and 100 mm high. Filter and base soils were typically reconstituted using the water pluviation technique to heights of 37 mm and 30 mm respectively. Hydraulic gradients were 17  applied to specimens and increased either rapidly or gradually at different confining pressures of 50, 100, 200, 300 and 400 kPa.  Tomlinson and Vaid (2000) conclude that D15/d85 ratio is the most important parameter in establishing the propensity for piping erosion and associated filter incompatibility. They found that base-filter specimens with: D15/d85 < 8 were immune to piping; specimens with D15/d85 > 12 spontaneously piped; and, specimens with 8 < D15/d85 < 12 would only pipe only if a certain critical gradient was reached. These results are generally consistent with those of Bertram (1940).   2.3.8 Foster and Fell (2001) As evidenced by this literature review, much work has been completed towards an improved understanding of filter compatibility and towards developing appropriate criteria for design purposes. Foster and Fell (2001) conducted a test program with the aim of developing criteria for the assessment of filters in existing dams that are not compliant with modern design criteria. It is emphasized that the aim of their work was not to develop design criteria but rather to develop a tool for assessing any currently existing filters. Assessment criteria are intended to be devoid of any margin of safety.  Foster and Fell (2001) compiled a database of filter tests and performed statistical analysis on them. Test data included test results of conventional filter tests (Sherard et al. 1984a), base suspension tests (Kenney et al. 1984), slot tests (Sherard et al. 1984b), slurry tests (Sherard et al. 1984b), and no-erosion tests (Sherard and Dunnigan 1989) on a vast array of filter and base soils. 18  These test results were categorised based on their fines content (mass percentage passing the 0.075mm sieve) and on the mass of soil eroded from the filter. These categories are defined as:  No erosion - for cohesionless soils less than 10g of base material was eroded and the filter is deemed to have sealed (for cohesive soils no visible erosion).  Some erosion - more than 10g and less than 100g of base material was eroded and the filter is deemed to have sealed.  Excessive erosion - more than 100g of base material was eroded and the filter is deemed to have sealed.  Continuing erosion - the filter has not sealed and the base material continues to erode until the test is terminated.  As part of their study, Foster and Fell (2001) developed a new test method similar to the no-erosion filter test used by Sherard and Dunnigan (1989) called the continuing-erosion filter (CEF) test (see Figure 2-2). The device consisted of a rigid cylinder (125 or 205mm in diameter) bounded by a top and base plate. Water was supplied to the device from the municipal water supply at pressures ranging from 240 to 300kPa. The filter and base soils were placed and compacted to thicknesses of 150mm and 100mm respectively. Side material is placed at the interface between the base and filter soils near the inside wall of the cell (see Figure 2-2) to prevent preferential flow and particle movement along the inside wall of the permeameter cell. The base-filter specimen is subjected to downward seepage flow, and the water exits the filter soil through a bottom drainage layer and out of two 19mm diameter holes in the base plate. In addition a top drainage layer is placed above the base soil to mitigate the influence of point source inflow from the inlet valve. 19   A series of NEF and CEF tests were performed, at the University of New South Wales, to supplement the database so that the proposed excessive and continuing-erosion boundaries could be more precisely determined. The 8 base soils tested had fines contents ranging from 15 to 85% and were sourced from 8 different natural deposits and dam sites in Australia and New Zealand. The filters were reconstituted from rounded to sub-rounded alluvial material sieved and blended to form 14 select gradations.   Foster and Fell (2001) proposed empirical boundaries for the no-erosion, excessive-erosion and continuing-erosion categories:  The no-erosion (NE) boundary was derived considering only the results from the statistical analysis (the NEF and CEF tests results are not included here).  Similar to Sherard and Dunnigan (1989), base soils are divided into four soil groups and each group has a unique no-erosion boundary. Table 2-3 shows the Sherard and Dunnigan (1989) design criteria, the Sherard and Dunnigan (1989) filter boundaries and the Foster and Fell (2001) no-erosion boundaries. Unlike the Sherard and Dunnigan (1989) design criteria, both the Foster and Fell (2001) no-erosion boundaries and the Sherard and Dunnigan (1989) filter boundaries are intended to be devoid of any margins of safety. To this end, certain differences are observed between the design criteria and the criteria delineating the boundary of filter compatibility. As compared to Sherard and Dunnigan (1989), Foster and Fell (2001) have slightly revised soil groups 2 and 4 from 40% to 35% and renamed the groups 2A and 4A respectively. The equation of the criterion for soil group 4A (previously soil group 4) has been altered. In general the criteria proposed by Foster 20  and Fell (2001) are very similar to the filter boundaries proposed by Sherard and Dunnigan (1989) (see Table 2-3).  The continuing- and excessive-erosion (CE and EE) boundaries (Table 2-4) were derived from the results of the statistical analysis in combination with the results of the additional NEF and CEF tests conducted at the University of New South Wales. Similar to the no-erosion boundary, the empirical boundaries for excessive and continuing-erosion are intended to be devoid of a margin of safety. The excessive and continuing boundaries are a novel contribution within the topic of filter compatibility intended to be used for assessing filters not compliant with modern filter design criteria and have no historical counterparts for comparison.  2.4 Coursier Dam Following the decommissioning of Coursier Dam, in response to dam safety concerns for the cumulative impact of extensive seepage problems and recurring sinkholes, an opportunity arose for its study. Coursier Dam is a 685 m long zoned embankment water dam (see Figure 2-4) built to provide additional storage for the Walter Hardman Hydro-power scheme. It is located in the Gold Range of the Monashee Mountains, near the town of Revelstoke in British Columbia, Canada. The dam was constructed in 1963 to a height of 12m, and was raised to 19m in 1969.   As part of regular dam inspections, four sinkholes were discovered on the upstream face of the dam wall in 1992. This was followed by several investigations and related remedial works, but the discovery of two additional sinkholes in 1998 prompted the decision to decommission the dam. The decommissioning was completed in 2003, enabling Coursier Lake to drain naturally. 21  No longer in service, the remnants of the embankments and foundations of the dam represent a unique research opportunity to evaluate a number of possible hypotheses. Additional information and detail on Coursier Dam is provided by Garner et al. (2004) and Seyers (2004).  Crawford-Flett (2014) conducted a desk study of the available construction records to determine the likely cause of the sinkholes, investigating both the potential for internal instability and filter incompatibility as candidate causes. It was postulated that all observed pipes and sinkholes are associated with filter incompatibility between the lower core and the Unit 4 foundation soil (see Figure 2-4). It is the aim of this study to test this hypothesis by conducting laboratory tests using representative samples from Coursier Dam. The results of these tests may not only inform our understanding of the events that occurred at Coursier Dam but also provide an opportunity to discuss the empirical tools used to assess the uncertainty associated with filter compatibility in similar structures.            22  Table 2-1 Soil retention criteria used by NRCS (1994), USACE (2004), FEMA (2011) and USBR (2011)  Soil group Fines content by No. 200 sieve (%) Design Criterion Contributing researcher 1 85 – 100          When      is less than 0.2 mm use 0.2 mm (For dispersive soils FEMA (2011) and USBR (2011) recommend            but not less than 0.2 mm) Sherard et al. (1989) 2 40 – 85         Sherard et al. (1989) 3 15 – 39                               A - percentage base material passing the 0.075 mm sieve after any re-grading When      is less than 0.7*mm use 0.7*mm (*For dispersive soils USBR (2011) and FEMA (2011) recommend using 0.5 mm instead of 0.7 mm) Sherard et al. (1989) 4 0 – 15          (USACE (2004) state that d85 before regrading may be used. All other methods use d85 after regrading) Terzhagi (1939), Sherard and Dunnigan (1984a), Sherard et al (1989) For soils with particle sizes greater than 4.75 mm all fines content and d85 values are determined from a gradation curve re-graded to yield a maximum particle size of 4.75 mm unless specified otherwise.   23  Table 2-2: Summary of research relating to the soil retention criteria  Research Group Filter apparatus No. of tests Base soils Filter soils Test conditions Criteria type Proposed criteria Bertram (1940) Rigid walled permeameter (ϕ 51 to 102.mm) 30 Uniform Ottawa sand placed, tamped and saturated Uniform Ottawa sand placed, tamped and saturated Hydraulic gradients of either 6 to 8 or 18 to 20. Boundary of filter compatibility D15/d85 < 6 Karpoff (1955) Rigid walled permeameter (ϕ 203 mm) 25 Cohesionless soil placed and compacted 203 mm thick. Uniformly and broadly graded sand and gravel placed and compacted 203 mm thick. Hydraulic pressures ranging from 6 to 89 kPa Design D100 < 75 mm D5 > 0.075 mm Base soils with D100 > 4.75 mm should be regraded Lafleur (1984) Flexible walled permeameter (ϕ 150 mm) 11 Broadly graded till placed and compacted 150 mm thick Sand and gravel placed 200 mm thick Hydraulic gradient of up to 8 Boundary of filter compatibility D15/d85 < 8.4 Sherard et al. (1984a) Rigid walled permeameter (ϕ 102 mm) 20 Uniform sand compacted to 150 mm thick. Uniform graded sand and gravel placed and tamped 127 to 178 mm thick Hydraulic pressure of 392 kPa Design For D15 > 1 mm D15/d85 ≤ 5 24  Research Group Filter apparatus No. of tests Base soils Filter soils Test conditions Criteria type Proposed criteria Sherard et al. (1984b) Slot test Horizontal rigid walled permeameter (ϕ  102 mm) Slurry test Rigid walled permeameter (ϕ  102 mm) 254 Soils ranging from non-cohesive silts to plastic clays. Slot test compacted 165  mm thick with 13 mm wide 1.5 mm thick slot Slurry tests Prepared as a slurry and poured to a height of  51 to 76 mm. Sub-rounded to rounded alluvial sands and gravels. Placed and tamped 102 mm thick  Hydraulic pressure of 392 kPa was applied Design For silts and clays with 0.1 < d85 < 0.5 mm, D15/d85 ≤ 5.  For clays with 0.03 < d85 < 0.1 mm, D15 ≤ 0.5 mm.  For silts with 0.03 < d85 < 0.1 mm, D15 ≤ 0.3 mm. Sherard and Dunnigan (1989) No-erosion filter test Rigid walled cylinder (ϕ 100 mm or 280 mm) - Four soil groups: 1. Fine silts & clays 2.Sandy silts & clays. 3 Silty & clayey sands. 4. Clayey & sandy silts Compacted 25 or 100 mm thick with preformed hole (ϕ 1 mm or 5 to 10 mm) Sub-rounded to rounded alluvial sands and gravels. Placed and tamped 102 mm thick  Hydraulic pressure of 413 kPa was applied Design  For soil groups: 1. 85 to 100% D15  ≤ 9d85 2.  40 to 80% D15  ≤ 0.7 3. 0 to 15% D15  ≤ 74d85 4. 15 to 40% D15                   0.7+0.7  Tomlinson and Vaid (2000) Rigid walled cylinder (ϕ 100 mm) with confining stress 17 Uniform glass beads water pluviated 30 mm thick Uniform glass beads water pluviated 37 mm thick Hydraulic gradients were increased until failure. Confining pressures of 50 to 400 kPa were applied. Boundary of filter compatibility  D15/d85 < 8    25  Table 2-3 Soil retention criteria proposed by Sherard and Dunnigan (1989) and Foster and Fell (2001) Sherard and Dunnigan (1989) Foster and Fell (2001) Soil group Fines content by No. 200 sieve (%) Design Criteria Filter Boundary Soil group Fines content by No. 200 sieve (%) No-erosion filter boundary 1 ≥ 85           D15  ≤ 7d85 to 12d85 1 ≥ 85           2 40 – 85                            2A 35 – 85           3 < 15                              3 < 15           4 15 – 40                                 Intermediate between groups 2 and 3, depending on fines content 4A 15 – 35                                        For base soils with particle sizes greater than 4.75 mm the gradation curve should be regraded to yield a maximum particle size of 4.75 mm *A - percentage base material passing the 0.075 mm sieve after any regrading  26   Table 2-4 Some-erosion and excessive-erosion boundaries as proposed by Foster and Fell (2001) Base soil Filter boundary (D15) determined by test d95 Fines (1) Excessive-erosion Continuing-erosion < 0.3 mm - D15 > 9d95 D15 > 9d95 > 0.3, < 2 mm - D15 > 9d90 > 2 mm > 35% To be determined from Figure 2-3 using 0.25 g/cm2 for an average value and 1.0 g/cm2 for a coarse limit > 2 mm < 15% D15 > 9d85 > 2 mm 15 – 35%                                      For soils with particle sizes greater than 4.75 mm fines content is determined from a gradation curve regraded to yield a maximum particle size of the No.4 (4.75 mm) sieve.           27   Figure 2-1: Example filter design band with control points 1 to 7 (after NRCS, 1994) 28   Figure 2-2: CEF apparatus used by Foster and Fell (2001) (after Foster and Fell, 1999) 29   Figure 2-3: Contours of erosion loss for the determination of the excessive-erosion boundary for base soils with d95 > 2 mm and fines content > 35% (after Foster and Fell, 1999)  Figure 2-4: A cross-section of Coursier Dam (Crawford-Flett, 2014)   30  Chapter 3: Apparatus and Test Program 3.1 Introduction Crawford- Flett (2014) attributes the observed sinkholes and pipes at Coursier Dam to filter incompatibility between the lower core and unit 4 foundation soil. In order to test this hypothesis, it was proposed that laboratory tests be conducted on soil sampled from the Coursier Dam site. Furthermore, there is an opportunity to discuss the utility of the Foster and Fell (2001) empirical criteria for the assessment of filters that do not meet modern filter criteria.  As part of this study CEF testing was performed with the following objectives:  To determine whether susceptibility to filter incompatibility between the foundation and core soils at Coursier Dam may explain the occurrence of sinkholes  To discuss the empirical criteria proposed by Foster and Fell (2001) and how they are able to characterise the response of filter specimens in a CEF test procedure The test equipment, materials and the test procedure employed during the course of this study are reviewed in the section that follows.  3.2 Test Equipment This section describes the equipment used to conduct CEF tests at UBC. Any contrasts with the Foster and Fell (2001) equipment are discussed  (see Table 3-1 for summary).  31  3.2.1 Permeameter Cell  CEF tests were conducted in a large rigid walled permeameter (see Figure 3-1). The UBC large permeameter had originally been commissioned to study seepage-induced internal instability of soils. These studies are documented by Moffat (2005), Moffat and Fannin (2006) and Li (2008). For the purposes of this study, the large permeameter was configured for the CEF test.  The large permeameter, shown in Figure 3-1 and Figure 3-2, consists of a rigid transparent acrylic cylinder bounded by base and top plates at either end.  The acrylic cylinder has a wall thickness of 13 mm, is 1000 mm in length, and has an inside diameter of 279 mm. The cylinder seals against the top plate with rubber O-rings. Firm rubber strips were placed in between the base plate and the bottom end of the acrylic cylinder to leave a small gap: this gap allows water to flow freely through the bottom of the permeameter and is sufficient to prevent attenuation of flow. The base plate, top plate and acrylic cylinder are held together by six threaded stainless steel rods which are bolted in place. The top plate has a valve (the water inlet) and a port where a pressure gauge may be attached.   Inside the permeameter is a base frame on which a base-filter specimen is reconstituted. The upper and lower plates of the base frame have numerous large openings to permit water to flow freely through it and are connected by four metal rods.  Foster and Fell (2001) used 125 mm and 205 mm diameter permeameters for filters with D15 < 15 mm and D15 > 15 mm respectively, whereas the permeameter used in this study has internal diameter of 279mm. ASTM D 2434 (Standard method for permeability testing of granular soils) 32  recommends a maximum ratio of between 8 and 12 of specimen diameter to largest particle size. The largest particle size used in this study is 37.5 mm yielding a ratio of 7.4 which is within the recommended limits. Accordingly, the diameter of the UBC large permeameter is considered adequate for the tests to be performed.  3.2.2 Water Supply Water is provided by a diaphragm tank which is controlled by an air pressure regulator. The diaphragm tank provides water pressurised to 275 kPa and has a capacity of 200 l. Foster and Fell (2001) used the local municipal water supply, with pressures ranging from 240 kPa to 300 kPa, and did not control the pressure. The pressure-control system used in this study is a refinement of the system used by Foster and Fell (2001).   3.2.3 Inflow Boundary Water enters the permeameter through a valve located in the top plate (see Figure 3-1 and Figure 3-2) and proceeds to seep through a layer of gravel, 100 mm thick, which was placed on top of the test specimen (see Figure 3-2).  The gravel layer acts to mitigate any influence of the point-source inflow on the top surface of the specimen.  Experience gained in the current series of tests led to a 1.4 mm wire-mesh screen being placed at the base of the top gravel layer to eliminate the potential for any gravel particles to wash down and into an erosion-induced enlargement of the preformed hole.  The Foster and Fell (2001) CEF test device similarly incorporated a wire mesh screen (Foster, 2015).  33  3.2.4 Peripheral Base/Filter Interface  Based on past experience and successful experimentation, a polymer ring was placed on the periphery of the base-filter interface ( see Figure 3-2) to prevent any preferential movement of soil particles along the inside wall of the permeameter cell. Accordingly, no preferential migration of soil particles was observed during any of the tests.  As part of the parametric study (Appendix A) the polymer ring was used once in conjunction with modelling clay and once with a waterproofing foam sealant. Neither the modelling clay nor the foam sealant was observed to have any significant impact on the test outcome. Foster and Fell (2001) used granular side material at the interface between the base and filter soils to prevent soil erosion along the inner wall of the cylinder (see Figure 2-2).   3.2.5 Outflow Boundary The test specimen of base-filter soil rests on a stacked series of three commercially available wire-mesh screens of 1.4/6.0/6.0 mm square opening size (upper/middle/lower position in the stack, respectively).  In contrast to the array of wire mesh screens, Foster and Fell (2001) used a bottom drainage material (Figure 2-2). According to Sherard et al. (1984a) a gravel filter may act as a laboratory sieve with openings equal to 0.11D15. The range of filter soils tested in this study correspond well to the range of filters in the study by Foster and Fell (2001), with the sieve opening size used generally larger than the comparative drainage material used by Foster and Fell (2001). Therefore, the use of stacked sieves in place of the granular bottom drainage layer should have no influence on the test result.  34  The Foster and Fell (2001) apparatus (Figure 2-2) featured a base plate with two 19mmm diameter holes. As part of the parametric study (Appendix A), two tests were completed with an outlet system identical to that of Foster and Fell (2001) and no significant difference was found in the specimen response with this type of outlet boundary.  3.2.6 Outflow Collection Once seepage water has exited the filter layer it flows out through the bottom of the permeameter. The water is collected in a pail which rests on a weigh-scale (Figure 3-2). Data from the weigh scale are continuously recorded, allowing for the calculation of flow rate during the test. Once full, the pail is removed from the weigh-scale and replaced with an empty one. Seepage water is allowed to stand for a period of approximately one week so that the eroded soil particles settle out. The excess water is then decanted and the eroded soil dried and weighed.  3.3 Test Materials The materials used for CEF testing were sampled and selected such that the susceptibility to filter incompatibility could be determined for credible variations of foundation soil against a representative sample of lower core material.  The material sampling and processing are discussed in the section that follows.  3.3.1 Field Sampling In September 2014 samples were taken from the intact west abutment of Coursier Dam. Figure 3-3 depicts a view of the west abutment showing the sampling locations (LC denotes the sampling location for the lower core material and F denotes the sampling location for foundation 35  soil). The soil samples were collected in 5 gallon buckets and transported to UBC where they were prepared for processing. A total of 15 buckets of foundation soil, and 5 buckets of lower core soil was collected.  3.3.2 Lower Core The sampled lower core material from the west abutment (current study) is compared with the grading envelope defined by known gradations of lower core material from the construction records, as well as four small grab samples from Crawford Flett (2014) (Figure 3-5). The sampled soil of the current study, plots within the construction envelope (shaded blue) but it is slightly coarser between 50 and 70% passing. To produce a test gradation that is representative of the lower core material, the sample was processed by sieving and recombining (Figure 3-4) to yield a processed lower core (LC) gradation that is deemed representative given that it plots within the construction envelope and compares favourably to samples (a) and (b) (Figure 3-5) taken by Crawford-Flett (2014) from the west abutment. In addition, a sand replacement test was used to determine the density of the lower core soil on the west abutment and yielded a dry density of 1750 kg/m3.  The construction records did not contain any indications of density values for the lower core soil.  3.3.3 Foundation Material Due to practical limitations it was not possible to sample Unit 4 (U4) foundation soil from the Coursier Dam site and unit 5 (U5) foundation soil was sampled instead. Due to the identical genesis, and nearly identical particle size distributions, it was possible to reconstitute credible U4 gradations using U5 material as a parent soil. The U5 soil was meticulously processed by method 36  of dry and wet sieving and separated into different particle sizes to be weighed and recombined to produce a credible U4 gradation.  Figure 3-6 shows the sieved and sorted foundation soil ready to be blended to produce select U4 gradations.  Based on the gradation of the processed lower core (LC) soil, the corresponding Foster and Fell (2001) no-erosion, excessive-erosion and continuing-erosion boundaries were calculated (see Appendix B).  Four different gradations of Unit 4 (U4) foundation were selected for testing and are labelled according to their D15 values. Figure 3-7 shows the selected gradations along with envelope range of gradations determined using the construction records. The U4 gradations are presented in the order of testing as follows:  U4-0.25 (D15 = 0.25 mm) - Selected to be finer than the Foster and Fell (2001) no-erosion boundary.  U4-3.7 (D15 = 3.7 mm) - Selected so that to be on the Foster and Fell (2001) excessive-erosion boundary.  U4-2.6 (D15 = 2.6 mm) - Selected to be finer than U4-3.7 and within the Foster and Fell (2001) some-erosion category.  U4-1.4 (D15 = 1.4 mm) - Selected to be finer than U4-2.6 but still within the Foster and Fell (2001) some-erosion category.  3.4 Specimen Reconstitution Both lower core and unit 4 soils were blended in a dry state and moisture conditioned overnight. The filter soil (Unit 4) was moisture conditioned to a nominal value of 2.5% to limit the potential 37  for segregation. The lower core soil was moisture conditioned to a value of 17.5% to produce the targeted dry density of 1750 kg/m3.  The filter soil was placed and lightly tamped to a dry density of approximately 1900kg/m3. The base soil was placed directly in contact with the filter and a polymer ring was placed in between the filter and the base. The base soil was compacted using a Standard Proctor compaction hammer to a dry density of approximately 1785 kg/m3, which compares favourably with the measured in situ dry density of 1750 kg/m3. Compaction of the base soil around a thin polymer tube formed a hole 5mm in diameter. Figure 3-8 shows the various stages of the reconstitution of a typical base-filter specimen.  3.5 CEF Test Procedure The test method is generally consistent with the procedure reported by Foster and Fell (2001), and is summarized as follows: 1. Upon reconstitution of the test specimen, the chamber above the top gravel layer is filled with water, while permitting displaced air to escape through a valve on the top plate of the permeameter; 2. Upon complete filling of the chamber space, water pressure of approximately 275 kPa is applied through the inflow valve;  3. The response of the specimen is observed for evidence that an equilibrium condition has been reached, primarily with reference to outflow water rate and visual inspection of turbidity in the outflow water. 38  4. The filter layer is judged to have sealed at the location of the preformed hole if all of the following conditions are met: the inflow water pressure of approximately 275 kPa is constant; the outflow water rate from the filter layer has diminished to a relatively low and constant value over a period of approximately 1 hr; and the outflow water exhibits no turbidity. 5. In the event that the filter layer does not seal the preformed hole, the test is continued until the inflow water supply of 200 l is exhausted. 6. The inflow valve is closed, and the apparatus dismantled in order to conduct a post-test inspection of the test specimen; and, 7. The eroded soil is allowed to settle out in the collection pails, after which the excess water is decanted and the soil is dried and weighed. 3.5.1 Test Data The data from the weigh scale were continuously recorded using a personal computer and each test was recorded using a digital video camera. The recorded weigh scale data and the video footage were used to determine the flow rate of water through the specimen. Additional flow measurements were taken manually during each test using a beaker, of known volume, and a stop watch. These manual measurements were found particularly useful for very low flow rates, and served to verify the automated measurements.  39  3.6 Test Program 3.6.1 Coursier Dam The test program consisted of four tests. Given the experience gained in the first two tests a wire mesh was added above the base soil layer and below the overlying gravel to prevent any gravel from collapsing into the eroded hole. The test code shows the base soil of lower core (LC) against the filter soil of unit 4 (U4) followed by the D15 value of the U4 soil. The ‘W’ is included in the test code for tests where a wire mesh screen was used. The tests are presented in the order of testing as follows: 1. LC/U4-0.25  2. LC/U4-3.7 3. LC/U4-2.6W 4. LC/U4-1.4W   3.6.2 Parametric Study The tests proposed above are deemed adequate for determining material susceptibility to filter compatibility at Coursier Dam. However, in order to comment on the utility of the Foster and Fell (2001) empirical criteria more testing is required. Nine additional tests were performed on soils from another dam site. The tests are described in detail in Appendix A.      40  Table 3-1: CEF test apparatus CEF Test Setup Foster and Fell (2001) Current Study Cell Diameter (mm) 125 and 205 279 Water Supply (kPa) 240 to 300 275 Inflow Boundary Wire mesh below overlying gravel Wire mesh below overlying gravel Base/Filter Interface Granular side material Polymer ring Outflow Boundary Granular bottom drainage material Stacked wire mesh screens (1.4/6.0/6.0 mm) 2 X 19 mm outlet pipes Perforated base frame plate Outflow Collection Plastic Drums Plastic Pails      41   Figure 3-1: The large permeameter configured for the CEF test 42   Figure 3-2: Cross-sectional view of the large permeameter configured for the CEF test 43   Figure 3-3: View of the west abutment with sampling locations at Coursier Dam     Figure 3-4: Coursier lower core material before blending   44   Figure 3-5: Lower core material at Coursier Dam  Figure 3-6: Sorted Coursier foundation material before blending  45   Figure 3-7: Reconstituted U4 foundation soils  46   Figure 3-8: Typical specimen reconstitution 47  Chapter 4: CEF Test Results and Analysis 4.1 Introduction This chapter describes the results and analysis of the four continuing-erosion filter tests that have been conducted on Coursier Dam soils with emphasis on reporting observations of flow rate, turbidity and exhumation of the post-test specimen (see Figure 4-1). A summary of the test results from the parametric study is also given, for which a detailed description of individual test results and analysis is provided in Appendix A. Although part of this study, the parametric study is presented separately in Appendix A so that it may be consulted as a stand-alone body of work exploring the effects of varying base and filter soils in the CEF test.   4.2 Coursier Dam Test Results 4.2.1 Test LC/U4-0.25 The test of LC against U4-0.25 was of approximately 60 min total duration.  A very low rate of flow commenced at 20 min and continued throughout the test such that making flow measurements was impractical (Figure 4-2). The outflow was clear throughout the test.  Inspection of the top surface of the post-test specimen revealed a nominally flat surface showing the impressions formed by the overlying gravel layer. The 5 mm hole was closed (see Figure 4-1a). During exhumation of the specimen it was observed that the hole was closed over its entire depth. The absence of the wire mesh in between the LC layer and the overlying gravel did not have any effect considering that the preformed hole was closed and therefore no gravel could collapse into it. The total loss of eroded soil was 0 g (see Table 4-1).  48   4.2.2 Test LC/U4-1.4W The test of LC against U4-1.4 was of approximately 60 min total duration.  A maximum flow rate of about 0.72 l/min was measured at an elapsed time of about 6 min, which diminished rapidly to approximately 0.4 l/min at 18.5 min, after which it remained constant for the duration of the test (see Figure 4-2). Outflow water was initially turbid, and became clean after an elapsed time of about 2 min.  Inspection of the top surface of the post-test specimen revealed a nominally flat surface showing the impressions left by the wire mesh. The preformed 5 mm hole was open at the top but closed at the bottom (see Figure 4-1b). The total loss of eroded soil was 786 g or 1.29 g/cm2 (see Table 4-1).  4.2.3 Test LC/U4-2.6W The test of LC against U4-2.6, was of approximately 7 min total duration at which point the 200 l water supply was exhausted.  A maximum flow rate of about 35 l/min was measured at an elapsed time of about 0.5 min, which then stabilised to approximately 30 l/min towards the end of the test (see Figure 4-2).  Outflow water was initially turbid, and became clean after an elapsed time of about 1 min.   Inspection of the top surface of the post-test specimen revealed a nominally flat surface showing the impressions left by the wire mesh. The 5 mm diameter hole was enlarged to 50 mm widening 49  out towards the top surface (see Figure 4-1c).  The total loss of eroded soil was 1481 g or 2.42 g/cm2 (see Table 4-1).   4.2.4 Test LC/U4-3.7 The test of LC against U4-3.7 was of approximately 9 min total duration at which point the 200 l water supply was exhausted.  A maximum flow rate of about 33 l/min was measured at an elapsed time of about 3.5 min, which diminished rapidly to approximately 27 l/min at 4.5 min and 21 l/min at the end of the test (see Figure 4-2).  Outflow water was initially turbid, and became clean after an elapsed time of about 1 min.  Inspection of the top surface of the post-test specimen revealed a nominally flat surface showing the impressions formed by the overlying gravel layer. The preformed 5 mm diameter hole was enlarged to 50 mm diameter (see Figure 4-1d). The absence of a wire mesh above the base soil caused the hole to be filled by the overlying gravel over its entire depth. No core material was present in the hole and the sides of the hole were approximately vertical. The total loss of eroded soil was 1067 g or 1.75 g/cm2 (see Table 4-1).  4.3 Parametric Study Test Results CEF tests conducted as part of the parametric study were 60 min in duration (with the exception of one test which lasted 43 min). Base soils with fines content ranging from 22% to 35% were tested against filters with D15 values ranging from 3.5 mm to 12.6 mm. Flow rates varied from a maximum of 36 l/min to a minimum of 0.1 l/min. The preformed hole was closed in all of the tests except for one. In tests where the preformed hole closed, a trend of diminishing flow rate 50  over time was observed after which the flow rate remained constant for the duration of the test. Loss of eroded soil varied from 199 g to 1935 g. Complete results and further details regarding this companion parametric study are presented in Appendix A.  4.4 Analysis of Coursier Dam Results Four tests were conducted on Coursier Dam soils to determine the material susceptibility to filter incompatibility. The results show the following:  Considering the low flow rate, 0 g of eroded soil, and closing of the preformed hole in test LC/U4-0.25, the filter is deemed to have sealed indicating filter compatibility.  Considering the stabilised flow rate, 786 g of eroded soil, and closing of the preformed hole in test LC/U4-1.4W, the filter is deemed to have sealed indicating filter compatibility.  Considering the high flow rate, 1481 g of eroded soil, and no closing of the preformed hole in test LC/U4-2.6W, the filter is not deemed to have sealed indicating filter incompatibility.  Considering the high flow rate, 1067 g eroded soil, and no closing of the preformed hole in test LC/U4-3.7, the filter is not deemed to have sealed indicating filter incompatibility.  Furthermore, considering the definition of no-erosion (less than 10 g with the filter sealed), some-erosion (less than 100 g with the filter sealed), excessive-erosion (more than 100 g but with a sealed filter) and continuing-erosion (the filter did not seal) categories as defined by Foster and Fell (2001) for categorising test results (Section 2.3.8): 51   Test LC/U4-0.25 is categorised as a no-erosion (NE) response with eroded mass of 0g (Table 4-1) and considering that the filter sealed the pre-formed hole.  Test LC/U4-1.4W is categorised as an excessive-erosion (EE) response with an eroded mass of 786 g (Table 4-1) and considering that the filter sealed the pre-formed hole.  Test LC/U4-2.6W is categorised as a continuing-erosion (CE) response with an eroded mass of 1481 g (Table 4-1) and considering that the filter did not seal the pre-formed hole.  Test LC/U4-3.7 is categorised as a continuing-erosion (CE) response with an eroded mass of 1067 g (Table 4-1) and considering that the filter did not seal the pre-formed hole.  4.5 Analysis of Parametric Study Results Tests performed in the parametric study were similarly categorised according to the Foster and fell (2001) criteria and exhibited responses in the some-erosion, excessive-erosion and continuing-erosion categories (see Table A-1). A detailed and stand-alone analysis of test results is presented in Appendix A (Section A.6).52  Table 4-1: Coursier CEF test results Test LC/U4-0.25 LC/U4-1.4W LC/U4-2.6W LC/U4-3.7 Percentage fines (PF)* 50.0 50.0 50.0 50.0 d85 (mm) 0.30 0.30 0.30 0.30 d95 (mm) 0.60 0.60 0.60 0.60 NE base soil group 2A 2A 2A 2A NE boundary** (mm) 0.7 0.7 0.7 0.7 EE/CE base soil category*** Row 2 Row 2 Row 2 Row 2 EE boundary** (mm) 3.6 3.6 3.6 3.6 CE boundary** (mm) 5.4 5.4 5.4 5.4 Empirical Filter Threshold D15 < NE NE < D15 < EE NE < D15 < EE D15 ≈ EE D15 (mm) 0.25 1.4 2.6 3.7 D15/ d85 0.8 4.7 8.7 12.0 Eroded mass (g) 0 786 1481 1067 Erosion loss (g/cm2) 0 1.29 2.42 1.75 Filter Sealed (Y/N) Yes Yes No No Foster and Fell Erosion Category No-erosion Excessive-erosion Continuing-erosion Continuing-erosion *D100 < 4.75 and therefore no regrading is required **Calculations are presented in Appendix B ***After Foster and Fell (2001) (see Table 2-4)    53   Figure 4-1: Top surface of base soil after testing  54   Figure 4-2: Variation of flow rate with time               55  Chapter 5: Discussion and Concluding Remarks 5.1 Discussion of Results The results of the CEF tests are discussed in the following section with reference to the two main objectives of this study, which are:  To conduct CEF testing on Coursier Dam soils with the aim of determining whether susceptibility to filter incompatibility between the lower core and unit 4 foundation soils may account for the occurrence of sinkholes.  To discuss the utility of the Foster and Fell (2001) empirical criteria for the assessment of filters with reference to the CEF test results from Coursier Dam and the parametric study.  5.1.1 Coursier Dam It is hypothesized that filter incompatibility between the lower core and unit 4 foundation materials led to the occurrence of sinkholes at Coursier Dam. To test this hypothesis, soils were sampled and processed to produce a representative lower core (LC) material and four credible variations of the unit 4 (U4) foundation soil (see Figure 5-1). Consequently, these materials were assessed for susceptibility to filter incompatibility in a series of four CEF tests. Two of the U4 gradations, U4-0.25 and U4-1.4, sealed the preformed hole in the base soil and are deemed filter compatible, two of the U4 gradations, U4-2.6 and U4-3.7, did not seal the preformed hole and are deemed filter incompatible with the lower core soil. Accordingly, the boundary for filter compatibility for the LC gradation from Coursier Dam is 1.4 mm < D15 < 2.6 mm. This boundary lies well with the envelope of credible gradations of Unit 4 foundation soil (see Figure 5-1). The 56  test results indicate that filter incompatibility between the lower core and unit 4 foundation soil is the likely cause of the sinkholes at Coursier Dam.  The laboratory results also provide an opportunity for reflection on design practice. The NRCS (1994) design filter envelope for the LC gradation is plotted and it is shown that the  U4-0.25 gradation plots within its limits. Observing that both U4-0.25 and U4-1.4W sealed the preformed hole it follows that any soil with D15 < 0.7 mm would also seal. Therefore, for the LC gradation, the NRCS (1994) filter design envelope yields filter compatible results when subjected to evaluation by CEF testing.  5.1.2 CEF Erosion Category The empirical characterisation of base-filter specimens into the no-, some- and excessive erosion categories as proposed by Foster and Fell (2001) is evaluated with regard to the CEF test response. For the purposes of strict comparison, only the test specimens incorporating a wire mesh in between the base soil and the overlying gravel are considered, yielding a dataset of 11 tests (Table 5-1).   With reference to the no-erosion (NE) category:  The filter layer of test specimen LC/U4-0.25 (Coursier dam) yields D15 <  NE and the observed mass loss of 0 g along which occurred with sealing of the preformed hole, represents a no-erosion CEF test response that is consistent with the empirical NE categorisation.  57   The filter layer of test specimen C25/T3.5W (Parametric Study) yields D15 ≈ NE and the observed mass loss of almost 200g which occurred with sealing of the preformed hole, represents an excessive-erosion CEF test response that is inconsistent with the empirical NE categorisation.  The filter layer of test specimen C26/T3.5WP (Parametric Study) yields D15 ≈ NE and the observed mass loss of nearly 290g which occurred with sealing of the preformed hole, represents an excessive-erosion CEF test response that is inconsistent with the empirical NE categorisation.  With reference to the some-erosion (SE) category:  The filter layer of test specimen LC/U4-1.4W (Coursier dam) yields NE < D15 < EE and the observed mass loss of nearly 786 g which occurred with sealing of the preformed hole, represents an excessive-erosion CEF test response that is inconsistent with the empirical SE categorisation.  The filter layer of test specimen LC/U4-2.6W (Coursier dam) yields NE < D15 < EE and the observed mass loss of nearly 1481 g which occurred with no sealing of the preformed hole, represents a continuing-erosion CEF test response that is inconsistent with the empirical SE categorisation.  The filter layer of test specimen C23/T8.4W (Parametric Study) yields NE < D15 < EE and the observed mass loss of nearly 350 g  which occurred with sealing of the preformed hole, represents an excessive-erosion CEF test response that is inconsistent with the empirical SE categorisation.  58  With reference to the excessive-erosion (EE) category:  The filter layer of test specimen C28/T3.5(FF)WP (Parametric Study) yields D15 ≈ EE and the observed mass loss of nearly 130 g which occurred with sealing of the preformed hole, represents an excessive-erosion CEF test response that is consistent with the empirical EE categorisation.  The filter layer of test specimen C35/T8.4W (Parametric Study) yields D15 ≈ EE, and the observed mass loss of 926 g (1.5 g/cm2) which occurred with sealing of the preformed hole, represents an excessive-erosion CEF test response that is consistent with the empirical D15 categorisation.  The filter layer of test specimen C22/T8.4WP (Parametric Study) yields EE < D15 < CE, and the observed mass loss of 1135 g or 2.2 g/cm2 which occurred with sealing of the preformed hole, represents an excessive-erosion CEF test response that is consistent with the empirical EE categorisation.  The filter layer of test specimen C27/T8.4(FF)WP (Parametric Study) yields EE < D15 < CE, and the observed mass loss of 780 g or nearly 1.3 g/cm2 which occurred with sealing of the preformed hole, represents an excessive-erosion CEF test response that is consistent with the empirical EE categorisation.  The filter layer of test specimen C35/T12.6W (Parametric Study) yields EE < D15 < CE, and the observed mass loss of nearly 1935 g (3.2 /cm2) which occurred with no sealing of the preformed hole, represents a continuing-erosion CEF test response that is inconsistent with the empirical EE categorisation.  59  As evidenced in Figure 5-2 the CEF test results exhibit a general trend of increasing mass loss of eroded base soil with increasing D15/d85 ratio (base soils with d100 > 4.75 are regraded to yield d85) . The trend line in Figure 5-2 shows a consistent agreement between the Coursier Dam results and the results from the parametric study. In addition, a progression from sealing to no-sealing is observed when the D15/d85 increases from 7 to 8.7. 5.2 Concluding Remarks Regarding the tests conducted on Coursier Dam material, the following conclusions are made:  Based on the tests conducted in this study it is found that the boundary for filter compatibility for the lower core soil (LC) from Coursier Dam is, 1.4 mm < D15 < 2.6 mm and this boundary lies well within the envelope of credible gradations of Unit 4 foundation soil. Accordingly, it is concluded that filter incompatibility of the lower core where it is in contact with Unit 4 foundation stratum (see Figure 2-4) is the likely cause of sinkholes at Coursier Dam.  It is observed that the Unit 4 gradation plotting within the NRCS (1994) design filter envelope exhibited filter compatible behaviour.  The following conclusions are drawn from interpretation of the 11 CEF tests performed on specimens with a wire-mesh screen at the inflow boundary surface (see Table 5-1 and Figure 5-2):  Erosion loss in the CEF testing is found to correlate with D15/d85 ratio, showing a trend of increasing eroded mass with increasing D15/d85 ratio. 60   The no-erosion empirical categorisation was found to be in relatively poor agreement with test results, considering only one of the three base-filter specimen’s empirical categorisations was consistent with the CEF test results.  The some-erosion empirical categorisation was found to be in poor agreement with test results considering none of the three empirical categorisations was consistent with the CEF test results.  The excessive-erosion empirical categorisation was found to be in relatively good agreement with test results considering four out of the five empirical categorisations was found to be consistent with the CEF test results.  5.3 Recommendations for Further Study Although modern filter criteria may enable the routine design of filters, there are several uncertainties associated with the empirical design criteria and the data used to derive them. The following topics of study are proposed towards the development of an improved hydro-mechanical understanding of filter behaviour:  The development of a standardised base-filter test device and method. The effects of saturation, soil density, effective stress, hydraulic gradient and flow direction relative to soil fabric orientation are deemed important factors. The device should be capable of testing filters with larger soil particles to explore the practice of regrading.  The execution of field monitoring and full scale testing to assess the performance of filters in situ, allowing for comparison with laboratory testing and empirical criteria.   61  Table 5-1: CEF test results Test Empirical Categorisation D15/d85 ratio Filter Sealed (Y/N) Erosion Loss (g) CEF Response LC/U4-0.25 No-erosion (D15 < NE) 0.8 Yes 0 No-erosion C25/T3.5W No-erosion (D15 ≈ NE) 1.8 Yes 199 Excessive-erosion C26/T3.5WP No-erosion (D15 ≈ NE) 2.2 Yes 286 Excessive-erosion LC/U4-1.4W Some-erosion (NE < D15 < EE) 4.7 Yes 786 Excessive-erosion LC/U4-2.6W Some-erosion (NE < D15 < EE) 8.7 No 1481 Continuing-erosion C23/T8.4W Some-erosion (NE < D15 < EE) 4.2 Yes 353 Excessive-erosion C28/T3.5(FF)WP Excessive-erosion (D15 ≈ EE) 2.2 Yes 132 Excessive-erosion C35/T8.4W Excessive-erosion (D15 ≈ EE) 7.0 Yes 926 Excessive-erosion C22/T8.4WP Excessive-erosion (EE < D15 < CE) 4.2 Yes 1134 Excessive-erosion C27/T8.4(FF)WP Excessive-erosion (EE < D15 < CE) 5.3 Yes 780 Excessive-erosion C35/T12.6W Excessive-erosion (EE < D15 < CE) 10.5 No 1935 Continuing-erosion  62   Figure 5-1: Coursier Dam test results 63   Figure 5-2: CEF test results 64  References Bertram, G.E. 1940. An experimental investigation of protective filters. Harvard University Graduate School of Engineering, Soil Mechanics Series No. 7, Publication No. 267. Crawford-Flett, K.A. 2014. An improved hydromechnical understanding of seepage-induced instability phenomena in soil. PhD thesis, School of Civil Engineering, The Department of Civil Engineering, The University of British Columbia, Vancouver, B.C. Drouhin, M. 1936. On the contribution of permeability and seepage studies to the control of underground erosion at the Bou-Hanifia dam (in French). Trans. 2nd Int. Congress on Large Dams, Washington, D.C., Vol. 4, Annex 1, 29–53. FEMA. (2011). Filters for embankment dams - Best Practices for Design and Construction. Federal Emergency Management Agency (FEMA). Foster, M. A., and Fell, R. (1999). Assessing embankment dam filters which do not satisfy design criteria. UNICIV Rep. No. R-376, School of Civil and Environmental Engineering, University of New South Wales, Sydney, Australia. Foster, M. 2015. Personal communications. Foster, M., & Fell, R. 2001. assessing embankment dam filters that do not satisfy design criteria. Journal of Geotechnical and Geoenvironmental Engineering. ASCE, 127(5): 398-407. Foster, M., Fell, R., & Spannagle, M. 2000. The statistics of embankment dam failures and accidents. Canadian Geotechnical Journal, 37(5), 1000–1024.  65  Garner, S. J., Seyers, W. C., and Matthews, H. M. (2004). The decommissioning of Coursier dam -A case for dam safety. In Canadian Dam Association 2004 Conference, Ottawa ICOLD. 1974. Lessons from dam incidents. Complete edition. International Commission on Large Dams (ICOLD), Paris. ICOLD. 1983. Deterioration of dams and reservoirs. International Commission on Large Dams (ICOLD), Paris ICOLD. 1995. Dam failures statistical analysis. International Commission on Large Dams (ICOLD), Bulletin 99. Karpoff, K. P. 1955. The use of laboratory tests to develop design criteria for protective filters. In Proc., 58th Annual Meeting, ASTM, 55, 1183–1198. Kenney, T.C., Lau, D., and Clute, G. 1984. Filter tests on 235 mm diameter specimens of granular materials. Tech. Rep. 84-07, Department of Civil Engineering., University of Toronto, Toronto. Lafleur, J. 1984. Filter testing of broadly graded cohesionless tills. Canadian Geotechnical Journal, 21, 634-643.  Li, M. 2008. Seepage induced instability in widely graded soils. PhD thesis, School of Civil Engineering, The Department of Civil Engineering, The University of British Columbia, Vancouver, B.C. 66  Moffat, R. A. 2005. Experiments on the internal stability of widely graded cohesionless soils. PhD thesis, School of Civil Engineering, The Department of Civil Engineering, The University of British Columbia, Vancouver, B.C. Moffat, R.A. and Fannin, R. J. 2006. A large permeameter for study of internal stability in cohesionless soils, Geotechnical Testing Journal, ASTM, 29(4). NRCS. 1994. National engineering handbook. Natural Resources Conservation Service (NRCS), US Department of Agriculture, Washington, D.C. Chap. 26, Part 633. Seyers, W. C. (2004). Decommissioning the Coursier dam. 4th Canadian River Heritage Conference, Guelph, Ontario. Sherard, J. L., & Dunnigan, L. P. 1989. Critical filters for impervious soils. Journal of Geotechnical Engineering, ASCE, 115(7), 927–947.  Sherard, J. L., Dunnigan, L. P., & Talbot, J. R. 1984a. Basic properties of sand and gravel filters. Journal of Geotechnical Engineering, ASCE, 110(6), 684–700. Sherard, J. L., Dunnigan, L. P., & Talbot, J. R. 1984b. Filters for silts and clays. Journal of Geotechnical Engineering, ASCE, 110(6), 701-718.  Tomlinson, S.S. and Vaid, Y.P. 2000. Seepage forces and confining pressure effects on piping erosion. Canadian Geotechnical Journal, Vol. 37, 1-13. USACE. 2004. Filter design. In Engineer Manual EM 1110-2-1901. United States Army Corps of Engineers (USACE). Washington, D.C. Appendix B. 67  USBR. 1973. Laboratory tests on protective filters for hydraulic and static structures, Laboratory Report No. EM-132. United States Department of Interior Bureau of Reclamation (USBR), Washington, D.C. USBR. 1973. Design of small dams, Second Edition. United States Department of Interior Bureau of Reclamation (USBR), Washington, D.C. USBR. 2011. Embankment dams. In Design Standards No. 13. United States Department of Interior Bureau of Reclamation (USBR), Washington, D.C. Chapter 5.     68        Appendix A Parametric Study     69  A.1 Introduction Subsequent to the test work completed on soils from Coursier Dam, the following parametric study was conducted with the purpose of investigating soils from another dam site that plot within the same range of filter incompatibility as encountered at Coursier Dam. Accordingly base and filter soils were selected and processed such that the no-erosion and excessive-erosion boundaries could be explored using the CEF test. A.2 Apparatus The apparatus used is similar to that of the main study (see Section 3.2) but there are several variations as described below. A.2.1 Inflow Boundary  A layer of gravel, 100 mm thick, was placed on top of the test specimen (see Figure A-2).  The gravel layer acts to mitigate any influence of the point-source inflow on the top surface of the specimen (Figure A-1).  Experience gained in the current series of tests led to a 1.4 mm wire mesh screen being placed at the base of the top gravel layer so as to eliminate the potential for any gravel particles to wash down and into an erosion-induced enlargement of the preformed hole.  The CEF test device of Figure A-1 similarly incorporated a wire mesh screen (Foster, 2015). A.2.2 Outflow Boundary Two variations of outflow surface boundary were examined in the test program:  i) Flow from the specimen exits through the entire basal surface area of the filter soil just as described in the main study (see Section 3.2) 70  ii) Flow from the specimen exits through two 19mm diameter holes. This is intended to closely simulate the device used by Foster and Fell (2001) (see Figure A-1).   In both variations, there is a wire mesh screen directly below the filter layer and water flowing out of the permeameter drains through a notched opening (see Figure A-2) in an outflow box and into collection pails. Eroded soils were collected, dried and weighed to determine the total mass of eroded soil for each test. A.2.3 Preferential Particle Movement A polymer ring was placed in between the filter and base soils and in contact with the inner wall of the permeameter to prevent preferential particle movement. In two of the tests additional measures were taken to help prevent preferential particle movement. These included the use of modelling clay and polyurethane-based foam sealant. A.3 Materials A.3.1 Base Material The base material was obtained from a dam site and gradations of this material were reconstituted with fines contents ranging between 22 and 35 %, termed C22 through to C35, respectively (see Figure A-3 and Table A-1).  The base materials were intended to be representative of a typical core material from a dam with a range of particle size distributions. A.3.2 Filter Material Filter material was obtained from the same dam site and was processed and reconstituted for testing. The maximum particle size of the filter layer examined in CEF testing was limited to 71  37.5 mm, in order to satisfy the requirement of the standard method for permeability testing of granular soils (ASTM D 2434).  Five gradations of filter material, termed T3.5, T3.5(FF), T8.4, T8.4(FF), and T12.6, were examined in the program of CEF testing, yielding 3.5mm ≤ D15 ≤ 12.6mm (see  Figure A-3 and Table A-1): the (FF) sub-label identifies two gradations that were selected to closely match filter gradations in the study of Foster and Fell (2001). The filter gradations were selected to be in the same erosion categories as the filters tested for Coursier Dam. The filters ranged from the no-erosion to the beyond the excessive-erosion boundary. Furthermore the filters were selected to have similar D15/d85 values as those tested at Coursier Dam. A.4 Test Method A.4.1 Specimen Reconstitution The filter layer rests directly on the upper 1.4 mm wire mesh screen of the outflow boundary surface.  It was placed in a single loose layer, and lightly compacted to a finished thickness of 150 mm.  Experience has established the compaction effort per unit volume yielded a dry density of approximately 1 800 kg/m3.  Base material was moisture-conditioned overnight at a water content of 6.0 % before placement and compacted to a finished thickness of 100 mm.  A preformed axial hole of 5 mm diameter was created by placement and compaction around a plastic tube that was subsequently withdrawn.  Control of the compaction effort yielded a dry density between 2 020 and 2 120 kg/m3 in the last three tests (see Table A-1), a density which is characteristic of all of the CEF tests in this program. 72  A.4.2 Test Procedure The test method is generally consistent with the procedure reported by Foster and Fell (2001), with further consideration of experiences reported by Foster (2007) and additional experience gained in the prior series of three CEF tests, and is summarized in Section 3.5 of the main body of the report. A.4.3 Testing Program One test was performed without including a wire mesh screen below the top gravel layer of the inflow boundary surface, on the following combination of base and filter materials (see Figure A-3 and Figure A-4):  Test C35/T12.6: a base soil with 35% fines against a filter with D15 = 12.6 mm CEF tests were performed with inclusion of a wire mesh screen below the top gravel layer of the inflow boundary surface,   Test C35/T12.6W: a base soil with 35% fines against a filter with D15 = 12.6 mm  Test C35/T8.4W: a base soil with 35% fines against a filter with D15 = 8.4 mm   Test C23/T8.4W: a base soil with 23% fines against a filter with D15 = 8.4 mm   Test C25/T3.5W: a base soil with 25% fines against a filter with D15 = 3.5 mm  CEF tests were performed with the same wire mesh screen below the top gravel layer of the inflow boundary surface, and with inclusion of the two pipes (depicted schematically in Figure A-1) on the outflow boundary surface, on the following combinations of base and filter materials:  Test C22/T8.4WP: a base soil with 22% against a filter with D15 = 8.4 mm 73   Test C26/T3.5WP: a base soil with 26% fines against a filter with D15 = 3.5 mm  Test C28/T3.5(FF)WP: a base soil with 28% fines against a filter with D15 = 3.5 mm  Test C27/T8.4(FF)WP: a base soil with 27% fines against a filter with D15 = 8.4 mm   Permission was granted for the use of test results from a previous set of tests conducted with the same apparatus used in this study. Three tests were conducted testing the same C35 core material against filters with D15 values of 0.4 mm, 3.5 mm and 8.4 mm respectively. A.5 Test Results The test results are presented with reference to measured outflow rate, visual observations of turbidity in the outflow water, and forensic observations of the post-test specimen. A.5.1 Test C35/T12.6 The test on 35% fines Core against D15 of 12.6 mm filter was of approximately 43 min total duration.  A maximum flow rate of about 30 l/min was measured at an elapsed time of about 1 min, which diminished rapidly to approximately 10 l/min at 4 min, 1.5 l/min at 10 min, and 0.25 l/min at 20 min, after which it remained almost constant (see Figure A-5).  Outflow water was initially turbid, and became clean after an elapsed time of about 1 min.  Forensic inspection of the post-test specimen revealed the preformed 5mm hole was filled by a combination of ‘soft’ Core material (see Figure A-6) and several entrained particles of the top gravel layer that became lodged within the base layer.  The total loss of eroded soil was 1496 g or 2.45 g/cm2 (seeTable A-1).  The eroded soil is well-graded silt and sand with trace clay-size fraction (see Figure A-7).   74  A.5.2 Test C35/T12.6W The test on 35 % fines Core against D15 of 12.6 mm filter was of 6 min total duration.  A maximum flow rate of about 36l/min was measured at an elapsed time of about 2 min, which diminished to approximately 25 l/min after 6 min (see Figure A-5), at which time the inflow water supply was exhausted and the test was terminated.  The maximum flow rate is believed supply-constrained by the configuration of the permeameter inflow system, rather than by the size of the eroded hole.  Outflow water was initially turbid, and became clean after an elapsed time of about 3min.  Forensic inspection of the post-test specimen revealed the preformed 5 mm hole had enlarged to a diameter of approximately 50 mm (see Figure A-6).  The inclusion of the 1.4 mm wire mesh screen as a part of the inflow boundary surface prevented any particles of the top gravel layer from being entrained into the eroded hole.  The total loss of eroded soil was 1935 g or 3.16 g/cm2 (see Table A-1).  The eroded soil is well-graded silt and sand with trace clay-size fraction (see Figure A-8). A.5.3 Test C35/T8.4W The test on 35 % fines Core against D15 of 8.4mm filter was of 60 min total duration.  A maximum flow rate of about 28.5 l/min was measured at an elapsed time of about 1 min, which diminished rapidly to approximately 10 l/min at 4 min, 1.5 l/min at 10 min, and 0.5 l/min at 20 min, after which it remained almost constant (see Figure A-5).  Outflow water was initially turbid, and became clean after an elapsed time of about 30 seconds.  75  Forensic inspection of the post-test specimen revealed the preformed 5 mm hole was filled by ‘soft’ Core material to yield a slight conical depression in the surface profile of the base layer (see Figure A-6).  The total loss of eroded soil was 926 g or 1.51 g/cm2 (see Table A-1).  The eroded soil is well-graded silt and sand with trace clay-size fraction (see Figure A-9). A.5.4 Test C23/T8.4W The test on 23 % fines Core against D15 of 8.4 mm filter was of 60 min total duration.  Outflow water commenced at an elapsed time of nearly 6min.  A maximum flow rate of about 1.1 l/min was measured at about 15 min, which diminished to approximately 0.5 l/min at 20 min, after which it remained almost constant (see Figure A-5)  Outflow water was initially turbid, and became clean after an elapsed time of about 11 min.  Forensic inspection of the post-test specimen revealed the preformed 5 mm hole was filled by ‘soft’ Core material to yield a very slight conical depression in the surface profile of the base layer (see Figure A-6).  The total loss of eroded soil was 353 g or 0.58 g/cm2 (see Table A-1).  The eroded soil is well-graded silt and sand with some clay-size fraction (see Figure A-10). A.5.5 Test C25/T3.5W The test on 25 % fines Core against D15 of 3.5 mm filter was of 60 min total duration.  Outflow water commenced at an elapsed time of nearly 7 min.  A maximum flow rate of about 0.75  l/min was measured at an elapsed time of about 15 min, which diminished to approximately 0.5 l/min at 20min, after which it remained almost constant (see Figure A-5).  Outflow water was initially turbid, and became clean after an elapsed time of about 14 min.  76  Forensic inspection of the post-test specimen revealed the preformed 5 mm hole was filled by ‘soft’ Core material to yield a distinct conical depression in the surface profile of the base layer (see Figure A-6).  The total loss of eroded soil was 199 g or 0.3 g/cm2 (see Table A-1).  The eroded soil is well-graded clayey silt, with some sand (see Figure A-11). A.5.6 Test C22/T8.4WP The test on 22% fines Core against D15 = 8.4 mm filter was of 60 min total duration.  Outflow water commenced at an elapsed time of nearly 2 min.  A maximum flow rate of about 32 l/min was measured at an elapsed time of about 4 min, which diminished to approximately 0.4 l/min at 10 min, after which it remained almost constant (see Figure A-5).  Outflow water was initially turbid, and became clean after an elapsed time of about 6 min.  Forensic inspection of the post-test specimen revealed the preformed 5 mm hole was filled by ‘soft’ Core material to yield a distinct conical depression in the surface profile of the base layer (see Figure A-6)).  The total loss of eroded soil was 1134 g or 2.18 g/cm2 (see Table A-1).  The eroded soil comprises a nearly equal proportion of sand and fines (see Figure A-12). A.5.7 Test C26/T3.5WP The test on 26 % fines Core against D15 of  3.5 mm filter was of 60 min total duration.  Outflow water commenced immediately.  A maximum flow rate of about 3 l/min was measured at an elapsed time of about 2.5 min, which diminished to approximately 0.2 l/min at 15 min, after which it remained almost constant (see Figure A-5).  Outflow water was initially turbid, and became clean after an elapsed time of about 3 min.  77  Forensic inspection of the post-test specimen revealed the preformed 5 mm hole was filled by ‘soft’ Core material to yield a distinct conical depression in the surface profile of the base layer (see Figure A-6).  The total loss of eroded soil was 286 g or 0.47 g/cm2 (see Table A-1).  The eroded soil comprises approximately 25% sand and 75% fines (see Figure A-13). A.5.8 Test C28/T3.5(FF)WP The test on 28 % fines Core against D15 of 3.5 mm filter was of 60 min total duration.  Outflow water commenced at an elapsed time of nearly 11 min.  A maximum flow rate of about 1 l/min was measured at an elapsed time of about 21 min, which diminished to approximately 0.5 l/min at 40min, after which it remained almost constant (see Figure A-5).  Outflow water was initially turbid, and became clean after an elapsed time of about 20 min.  Forensic inspection of the post-test specimen revealed the preformed 5 mm hole was filled by ‘soft’ Core material to yield a distinct conical depression in the surface profile of the base layer (see Figure A-6).  The total loss of eroded soil was 132 g or 0.22 g/cm2 (see Table A-1).  The eroded soil comprises approximately 10 % sand and 90 % fines (see Figure A-14). A.5.9 Test C27/T8.4(FF)WP The test on 27 % fines Core against D15 of 8.4 mm filter was of 60 min total duration.  Outflow water commenced at an elapsed time of nearly 1 min.  A maximum flow rate of about 6  l/min was measured at an elapsed time of about 1.5 min, which diminished to approximately 0.1 l/min at 28 min, after which it remained almost constant (see Figure A-5).  Outflow water was initially turbid, and became clean after an elapsed time of about 2 min.  78  Forensic inspection of the post-test specimen revealed the preformed 5 mm hole was filled by ‘soft’ Core material to (see Figure A-6).  The total loss of eroded soil was 780 g or 1.28 g/cm2 (see Table A-1).  The eroded soil comprises a nearly equal proportion of sand and fines (see Figure A-15). A.6 Analysis of Results The response of test C35/T12.6 indicates the filter layer sealed the preformed hole in the base layer.  In contrast the filter layer in test C35/T12.6W did not seal the preformed hole.  The difference between the two responses is attributed to the presence of the wire mesh at the inflow surface boundary of the latter test acting to stop particles from the top gravel layer collapsing into eroded hole (see Figure A-16).    The C35 material is categorized as a NE Base Soil Group 2A (after Foster and Fell, 2001), yielding a no-erosion (NE) boundary of 0.7 mm, an excessive-erosion (EE) boundary of 7 to 11mm, and a continuing-erosion (CE) boundary of 27.9 mm.  The D15 of 12.6 mm filter yields EE < D15 < CE, and an erosion loss of nearly 3.2 g/cm2 in test C35/T12.6W that occurred with no sealing of the defect in the Core (see Table A-1 and Figure A-16 that shows the hole with its post-test wax-casting).  The test on 35 % fines Core against a filter with D15 = 8.4 mm, examined in test C35/T8.4W, yields D15 ≈ EE and an erosion loss of nearly 1.5 g/cm2 that occurred with sealing of the defect in the Core (see Table A-1 and Figure A-16).  79  The 23 % fines Core of test C23/T8.4W is categorized as a NE Base Soil Group 4A (after Foster and Fell, 2001), yielding a no-erosion (NE) boundary of 5.8 mm, an excessive-erosion (EE) boundary of 9.1 mm, and a continuing-erosion (CE) boundary of 30.6 mm.  The filter of D15 = 8.4 mm yields NE < D15 < EE, and an erosion loss of nearly 0.6 g/cm2 that occurred with sealing of the defect in the Core (see Table A-1 and Figure A-16).  The 25 % fines Core of test C25/T3.5W is categorized as a NE Base Soil Group 4A (after Foster and Fell, 2001), yielding a No Erosion (NE) boundary of 4.0 mm, an Excessive Erosion (EE) boundary of 6.3 mm, and a Continuing Erosion (CE) boundary of 30.6 mm.  The filter of D15 = 3.5 mm yields D15 ≈ NE, and an erosion loss of nearly 0.3 g/cm2 that occurred with sealing of the defect in the Core (see Table A-1 and Figure A-16).  The 22 % fines Core of test C22/T8.4WP is categorized as a NE Base Soil Group 4A (after Foster and Fell, 2001), yielding a No Erosion (NE) boundary of 4.3 mm, an Excessive Erosion (EE) boundary of 6.7 mm, and a Continuing Erosion (CE) boundary of 27.0 mm.  The filter of D15 = 8.4 mm yields EE < D15 < CE, and an erosion loss of nearly 2.2 g/cm2 that occurred with sealing of the defect in the Core (see Table A-1 and Figure A-16).  The 26 % fines Core of test C26/T3.5WP is categorized as a NE Base Soil Group 4A (after Foster and Fell, 2001), yielding a No Erosion (NE) boundary of 3.4 mm, an Excessive Erosion (EE) boundary of 5.3 mm, and a Continuing Erosion (CE) boundary of 29.7 mm.  The filter of D15 = 3.5 mm yields NE ≈ D15, and an erosion loss of nearly 0.5 g/cm2 that occurred with sealing of the defect in the Core (see Table A-1 and Figure A-16). 80   The 28 % fines Core of test C28/T3.5(FF)WP is categorized as a NE Base Soil Group 4A (after Foster and Fell, 2001), yielding a No Erosion (NE) boundary of 2.0 mm, an Excessive Erosion (EE) boundary of 3.2 mm, and a Continuing Erosion (CE) boundary of 29.7 mm.  The filter of D15 = 3.5 mm yields EE ≈ D15, and an erosion loss of nearly 0.2 g/cm2 that occurred with sealing of the defect in the Core (see Table A-1 and Figure A-16).  The 27 % fines Core of test C27/T8.4(FF)WP is categorized as a NE Base Soil Group 4A (after Foster and Fell, 2001), yielding a No Erosion (NE) boundary of 2.5 mm, an Excessive Erosion (EE) boundary of 3.9 mm, and a Continuing Erosion (CE) boundary of 28.8 mm.  The filter of D15 = 8.4 mm yields EE < D15 < CE, and an erosion loss of nearly 1.3 g/cm2 that occurred with sealing of the defect in the Core (see Table A-1 and Figure A-16).  Discussion of the results and concluding remarks are provided in the main body of the thesis (Chapter 5).  81  Table A-1: Summary of the CEF test results CEF Test C35/T12.6 C35/T12.6W C35/T8.4W C23/T8.4W C25/T3.5W C22/T8.4WP C26/T3.5WP C28/T3.5(FF)WP C27/T8.4(FF)WP Base layer          Density (kg/m3) - - - - - - 2120 2020 2080 Fines Content (FC) 35 35 35 23 25 22 26 28 27 Regraded Fines Content (FC) 41 41 41 27 30 27 30 33 32 d85 (mm) 4.75 4.75 4.75 5.30 4.75 5.30 4.80 4.80 4.7 Regraded d85 (mm) 1.2 1.2 1.2 2.0 2.0 1.4 1.6 1.6 1.6 Regraded d95 (mm) 3.1 3.1 3.1 3.4 3.4 3.0 3.3 3.3 3.2 NE Base Soil Group* 2A 2A 2A 4A 4A 4A 4A 4A 4A NE Boundary (mm) 0.7 0.7 0.7 5.8 4.0 4.3 3.4 2.0 2.5 EE/CE Base Soil Category** Row 3 Row 3 Row 3 Row 5 Row 5 Row 5 Row 5 Row 5 Row 5 EE Boundary (mm) 7 to 11 7 to 11 7 to 11 9.1 6.3 6.7 5.3 3.2 3.9 CE Boundary (mm) 27.9 27.9 27.9 30.6 30.6 27.0 29.7 29.7 28.8 Filter layer          D0 (mm) 6.68 6.68 4.75 4.75 1 4.75 1.00 1.00 4.75 D15 (mm) 12.6 12.6 8.4 8.4 3.5 8.4 3.5 3.5 8.4 Empirical Filter Threshold EE < D15 < CE EE < D15 < CE D15 ≈ EE NE < D15 < EE D15 ≈ NE EE < D15 < CE D15 ≈ NE D15 ≈ EE EE < D15 < CE Test Measurements          Max Flow Rate (l/min) 30 36 29 1 1 32 3 0.9 6.4 Eroded Mass (g) 1496 1935 926 353 199 1134 286 132 780 Erosion Loss (g/cm2) 2.5 3.2 1.5 0.6 0.3 2.18 0.5 0.2 1.3 Test Observations          Filter Sealed (Y/N) Yes No Yes Yes Yes Yes Yes Yes Yes Foster and Fell Erosion Category Excessive Continuing Excessive Excessive Excessive Excessive  Excessive Excessive Excessive *After Foster and Fell (2001): Table 4 **After Foster and Fell (2001): Table 6         82   Figure A-1: CEF test device as used by Foster and Fell (2001) (figure after Foster and Fell, 1999)   83   Figure A-2: CEF test device current study 84    Figure A-3: Gradation curves of the test materials          85     Figure A-4: Reconstitution of the CEF test specimens   86    Figure A-4 (continued)   87   Figure A-5: Variation of flow rate with time 88    Figure A-6: Post-test image of the top surface of the base soil 89   Figure A-6 (continued)  90   Figure A-7: Test C35/T12.6 gradations  Figure A-8: Test C35/T12.6W gradations 91   Figure A-9: Test C35/T8.4W gradations  Figure A-10: Test C23/T8.4W gradations 92   Figure A-11: Test C25/T3.5W gradations  Figure A-12: Test C22/T8.4WP gradations 93   Figure A-13: Test C26/T3.5WP gradations  Figure A-14: Test C28/T3.5(U)WP gradations 94   Figure A-15: Test C27/T8.4(U)WP gradations    95   Figure A-16: Post-test cross-section image of the base soil layer 96   Figure A-16 (continued) 97      Appendix B Calculation of Foster and Fell (2001) Erosion Boundaries   98  B.1 Coursier Dam B.1.1 Lower Core (LC) gradation Fines Content (FC) - 50% DB85 - 0.3 mm DB90 - 0.4 mm DB95 - 0.6 mm No Erosion Boundary Base soil group 2A DF15 ≤ 0.7 mm Excessive Erosion Boundary In Table 6 (Foster and Fell, 2001) choose Row 2                              Continuing Erosion Boundary In Table 6 (Foster and Fell, 2001) choose Row 2                                   99  B.2 Parametric Study B.2.1 C22 gradation  Regraded Fines Content (FC) – 27% Regraded DB85 – 1.4mm Regraded DB95 – 3.0mm  No Erosion Boundary Base soil group 4A                                                                                                                  2.66mm                                     Excessive Erosion Boundary In Table 6 (Foster and Fell, 2001) choose Row 5                                                         Continuing Erosion Boundary In Table 6 (Foster and Fell, 2001) choose Row 5                        100  B.2.2 C23 gradation  Regraded Fines Content (FC) – 27% Regraded DB85 – 2.0mm Regraded DB95 – 3.4mm  No Erosion Boundary Base soil group 4A                                                                                                                3.62mm                                    Excessive Erosion Boundary In Table 6 (Foster and Fell, 2001) choose Row 5                                                        Continuing Erosion Boundary In Table 6 (Foster and Fell, 2001) choose Row 5                          101  B.2.3 C25 gradation  Regraded Fines Content (FC) – 30% Regraded DB85 – 2.0mm Regraded DB95 – 3.4mm  No Erosion Boundary Base soil group 4A                                                                                                                 2.52mm                                    Excessive Erosion Boundary In Table 6 (Foster and Fell, 2001) choose Row 5                                                        Continuing Erosion Boundary In Table 6 (Foster and Fell, 2001) choose Row 5                           102  B.2.4 C26 gradation  Regraded Fines Content (FC) – 30% Regraded DB85 – 1.6mm Regraded DB95 – 3.3mm  No Erosion Boundary Base soil group 4A                                                                                                                   2.13mm                                     Excessive Erosion Boundary In Table 6 (Foster and Fell, 2001) choose Row 5                                                         Continuing Erosion Boundary In Table 6 (Foster and Fell, 2001) choose Row 5                          103  B.2.5 C27 gradation  Regraded Fines Content (FC) – 32% Regraded DB85 – 1.6mm Regraded DB95 – 3.2mm  No Erosion Boundary Base soil group 4A                                                                                                                   1.56mm                                    Excessive Erosion Boundary In Table 6 (Foster and Fell, 2001) choose Row 5                                                         Continuing Erosion Boundary In Table 6 (Foster and Fell, 2001) choose Row 5                          104  B.2.6 C28 gradation  Regraded Fines Content (FC) – 33% Regraded DB85 – 1.6mm Regraded DB95 – 3.3mm  No Erosion Boundary Base soil group 4A                                                                                                                   1.27mm                                    Excessive Erosion Boundary In Table 6 (Foster and Fell, 2001) choose Row 5                                                         Continuing Erosion Boundary In Table 6 (Foster and Fell, 2001) choose Row 5                          105  B.2.7 C35 gradation  Regraded Fines Content (FC) – 41% Regraded DB85 – 1.2mm Regraded DB95 – 3.1mm No Erosion Boundary Base soil group 2A (35% < FC < 85%) DF15 ≤ 0.7mm  Excessive Erosion Boundary In Table 6 (Foster and Fell, 2001) choose Row 3 (DB95 > 2 and FC > 35%) Percentage Fine Medium Sand = 85 – 41 = 44 (from regraded curve) Read Average value of 7mm or coarse limit of 11mm from Figure 9 (Foster and Fell, 2001) Continuing Erosion Boundary                            

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0216473/manifest

Comment

Related Items