JKAU:Eng. Sci., Vol. 2, pp. 65-79 (1410 A.H. 11990 A.D.)An Investigation of the Influences of Particle Size,Size Gradation and Particle Shape on the Shear Strengthand Packing Behavior of Beach Sands in Saudi ArabiaA YAD A. ALZAYDICivil Engineering Department, Faculty of Engineering,King Abdulaziz University, Jeddah, Saudi Arabia.ABSTRACT. This investigation is a study of the shear strength and packingbehavior of Dammam and Abhur Beach Sands in Saudi Arabia in terms ofthe geometric particle characteristics of size and shape. A large number ofuniformsize sands and a limited number of improvedgradations arestudied. Particle size is defined by mechanical sieving and particle shape isquantified in terms of sphericity and roundness. Surface texture is recog-nized to be a significant variable as is particle mineralogy. Packing behavioris studied with respect to the maximum and minimum limiting void ratiosand void ratio spreads. Shear resistance is investigated in terms of the angleof internal friction. Results indicate that both packing behavior and shearbehavior are influenced significantly by particle roundness and surface tex-ture. The uniform sands tested were normalized into a linear strength-pack-ing-shape relationship at a given relative density. Chal).ges in grading affectpacking behavior more than strength behavior and the normalizations arerather sensitive to changes in void ratio spread. Normalizations appear sig-nificant in that they include surface texture variations and afford a possibletechnique for a more comprehensive approach to sand behavioral formula-tion in terms of the variables of particle size, size gradation, shape, surfacetexture, mineralogy and relative density_IntroductionIn the field of geotechnical engineering, the behavior of sands is dependent on quan-tities and characteristics such as relative density, confinement, particle size, particlesize distribution, particle shape, surface texture, mineralogy, and type of test con-ducted to determine sand response[I-41.Given a particular type oftest in which; ide-ally, relative density and confinement can be controlled, the behavior of a given sandwill then dependon the above sand characteristics. If it were possible to evaluate6566these characteristics for a given sand, it is conceivable that its behavior could be pre-dicted if sufficient test resulls in terms of all quantities and characteristics were avail-ableto eSlablishfunctional relationships. Thus. inorder to differcnliale onesandfrom another. a system of designating particle size, panicle size distribution. particleshape and surface texture is necessary.Panicle size and size distribution are reproducible by the common sieve analysis.However. a close look al any of the specific groups obtained by sieve analysis will re-veal thai no twoparticles arc exactly alike in shape or surface texturels.6l. Hence.shape and texture can playa major role in the behavior of particulate materials.Waddell7.81 showed that particle shape essentially consisted of two distinct charac-teristics. sphericity androundness. Sphericity. after Krumbeinl'll. is theratio of thevolume of the panicle to the volume of the smallest circumscribing sphere. Perfectsphericity (perfect circularity in two dimensional views) gives a ratio of 1.0. Round-ness is the ratio of the curvature of the corners and edges of the particle to the aver-age curvature of the particle. A particle which gives a perfect circle as a projectionhas a roundness of 1.0. It will be shown thai any given sieve portion of any sand doesnot have singular values of sphericity androundnessbut rather has distributions ofeach. Therefore, in order to study sands realistically. the behavioral variations of dif-ferent size ranges must be studied first and compared in terms of particles shape andtexture to formulate sand behavior in terms of these characteristics.Surface texture of sand grains is an extremely difficult propeny to formulate. if notanimpossible one. Themagnitudes of the surfacefeaturesallributableto surfacetcxture are subordinate in size to those features included in particle shapedetermina-tion. For agiven particle size. size distribution and particle shape. surface texture canbe defined only in terms of its effects on sand behavior.Another property of sands is the mineralogy. It is important since (I) absolute tle-nsity in terms of specific gravity depends on the relative contributions of the variousmincmls comprising the sand particles. (2) tendencies of different minerals to cleaveand/or fracture in different ways and the geological history of each particle influencethe sh:lpeandtextural featuresofthe particles, and(3) grain-to-grainfrictionalphenomena arc influenced by the Iypes and relative amounts of the various mineralsinvolved.Itis intended to determine. as far aspossible, the relative order of magnitude ofthe influences of particle size, particle size gradation, particle shape and surface tex-tureonthebehavioral response of sand, to determinethesignificance of eachofthese characteristics on suchresponse. to seekpossible correlations among them,and to establish the feasibility of such an approach to sand behavior. This investigation is not concerned with information on panjcle formation, its interest is only in thegeometries and surface features of panicles as they exist at present in nature. How-ever. the study will assist the reader in understanding the complexity of any attemptto accurately describe a grain of sand.An o{tlte /npuenuso{Size...Laboratory Investigation67Sand PropertiesThe materials studied in this investigation were derived from three sands obtainedfrom Ottawa (USA). Dammam Beach on the eastern coast. and Abhur Beach on thewestern coast of theKingdom of Saudi Arabia. Sands were sieved into several un-iform sizes with a uniformity coefficient. Cu' of 1.2.Mechanical sieving. though artificial in nature, isan adequate classification test fordefining uniformparticle size ranges anduniform size gradation in terms of theun-iformity coefficient. however this coefficient is generally inadequate for describingparticle size gradations. The range of particle size investigated extended fromO.84mm (No. 20)to O.lOmm (No. 140). plus two special mixesI and 2 as showninTableI.TABl.EI. MaterlalsinvestigatedSinofMeanSh.Valucs. Void Ralio Pol'O:lit)' FllClion>. -

u.s. Sieve

Spoof.H"So,(11m)""'.RoundGnll'it),... ...'"cil)'- '-"I"1'1 I"""'" lJJ.".,.."I.'"0.768 0..162,.""31.6 11.8'"379 1.13545 500-355

2.66 0.821 00478 0.313 a5.1 32.3 12.8 32.9 1.1.s.'W ,.0.52 2.66 ,....0.5.18'.J47.0 125 31.7 I.l10-100 212150 I.'"0.915 0.537 0.3711 aU

12.9 32.3 39.2 1.2IlkI-JoIO I.'"o.m0.5350.""aU 12.9 32.5 39.0 1.1".".,.."D.M I.W 0.",0.'""00 .10.1 )2.1 I.lnAS l0.251.2tmax1.0@0.8II)-,{tmin0-""0"r-0 0.8a:@ 0.4U. S. Sieve Size140 10070 60 45 35 20o 100 200 300 400 500 800 700 800 900MeanPartie IeDiameter, p mFIG. 3. Mean particle diameter versus void ratio and void ratio spread.An Investigation of the Influences of Particle Size... 730.45CIl 0.40

0.40.70 2 ~ __----7=----- ---l.. -.J0.4 0.5 06Roundnp.ss, RFIG 4. Void ratio and void -atio spread versus roundness.Figures 5and 6are examples prepared usingelectron scanning microscopytechnique to manifest the effect of surface texture on sand behavior. The figures areself-explanatory and indicate the smoother surface texture of Dammam Beach sandcompared to Abhur Beach sand. A comparison between these photographs and val-ues in Table 1 indicates that smoother surface texture results in decreased packinglimits, decreased void ratio spread and decreased shear resistance at a given relativedensity.Fl(l 5. Smuothersurface leKluTc or (35-45) D,lmm,lm Belich Slind in A.B_ and C. compared 10 the surfacelexture of (35-45) Abhur Beach sand in D.E. andF. Scale is inj.l.m.Flo 6. Smoother surface texlUre or (70-100) Dammam Ikach sand inA,B, and C, compared 10 the sur-face textureof(70-IOO) Abhur Beach sand in D.E, and F. Scale is inI/om.76 Ayad A. AlzaydiStrength behavior in terms of the angle of internal friction increases significantlywiththeincreaseof particleroundnessandsizeforDammamandAbhur Beachsands as shown in Figures 2and 7. However, the influences of particle size on Ottawasand strength response were indeterminate either because they are small comparedto the influences of particle shape and were masked by scatter or because the rangesof particle size investigated werelimited(whichisdoubtful since the sands testedranged fromfinetomedium coarse according to M.LT. classification). Improvedparticle size gradationrepresentedby Mixes1 and 2 in Table1 results in reducedpacking limits and reduced void ratio spreads. Conversely, its influence on strengthbehavior is minor as compared to roundness and surface texture.____ :"oose 3amclesUniformSie'VeSize, C.J .:!I.2........ )t1owo SandDommoma.oenSana'-,:) Abhur 3eacnSand__Densey..1if / ,L _-?,-JIl/ I ---,lJ .J, " ::. ) :5 :) : LI

a::.!! 40 r'0"c.c:-IrJL.) i..?.r.--=J\ -.j./,...:.11., ....)\.i. j'.& ....-:..J..li) .QJ...l1U..l.:J1 j' , U..d\ j' 4.l2::l1 ol,L...-l1 ! . i)..\:... '11 j' y;.:J1 J>-JI...,......;i..?.r.--=J1 -.j.r:J1 ...:.l \c5 .1...Lo. .jj' ....)\.i.) , w\S.\jJl)L..:..:;I)wU:.Ir"J1. J>-I..ul'11 1)(; ;t:o1.+.15i..?.r.--=J1-.j ....)\.i. .:.>fJs- C:L..:JI..LibJl .......,.....,al 4.l2::l1JL.)I .i..J;\.:... -:..J. ,)t Js- c:U1 ...:.J, \c5) . ;.;l;SJ\.....)L:J.I .:..'" ;.,.la.>