Discussion of “A Parametric Study on Factors Affecting Ground Vibrations during Pile Driving through Finite Element Simulations” by Mo Zhang and Mingjiang Tao

Abstract

The authors applied the finite-element method (FEM) to improvecalculations of the peak particle velocity of ground vibrations gen-erated by pile driving. This method has been successfully used toresolve various static problems in geotechnical engineering. How-ever, the results are substantially different in the application ofFEM to dynamic problems for which only general tendencies canbe determined (for example, Broers and Dieterman 1992). Analyz-ing the accuracy of several prediction methods based on FEM sim-ulation, Holscher and Waarts (2003) found disappointing lowreliability in predicting vibration levels from dynamic sources.Therefore, the authors’ attempt to use FEM for assessing groundvibrations from pile installation should be welcomed.In practice, Eq. (1) of the original paper is usually used forapproximate calculations of the expected peak particle velocity(PPV) of ground vibrations at various distances from driven piles(Wiss 1981). This equation renders assessment of PPV attenuationbetween two points on the ground surface.Eq.(1)providesaveryroughassessmentofgroundvibrationsasa function of the source energy and a distance from the source.Also,Eq.(1)doesnottakeintoaccountsoilconditions,pilepenetra-tion depth, soil resistance topilepenetration, soil heterogeneity anduncertainty, and soil-structure interaction and has nothing to dowith structural vibrations, dynamic settlements, and vibrationeffectsonsensitiveequipment.Ofcourse,thisequationhasnocon-nection with prediction of ground vibrations.On the one hand, approximate calculation of expected groundvibrations and even vibration monitoring yield relative informa-tion on vibration effects on structures, and these results could beinconclusive. On the other hand, condition surveys of structuresbefore and during pile installation provide complete informationonstructuralresponsestovibrationexcitations,andthisinforma-tion can be much beneficial than vibration assessment and mea-surements. It is obvious that rough calculations of expectedgroundvibrationsandcomparisonofthosewiththevibrationlim-itsisnotvibrationriskassessment because risk managementrequires analysis of overall factors affecting structures from piledriving.TheauthorscouldapplyFEMtodevelopamethodforcalculationsof expected ground vibrations from pile driving, and they calculatedPPV ofgroundvibrationsonthebasis of the model chosenbut with-out analysis, parametric study, and further development of the appli-cation of FEM for prediction of ground vibrations generated by pileinstallation. The prediction method was not developed. The authorsmostly utilized FEM simulation for assessment of the effects of thesoil damping ratio and Young’s modulus of soil on coefficientsk and n in Eq. (1). Also, the authors studied the influence of thenumber of loading pulses on ground vibrations due to pile drivingand they received negative results. This could be expected becauseit is well known that ground vibrations attenuate between two con-secutive hammer ram impacts. In other words, powerful FEMwas employed for evaluation of the coefficients in the empiricalequation.The authors compared the results of the FEM simulations withtwoequationsderivedbyWoods(1997)andreceivedasimilartrendof them with a suitable fit to one of two lines in Fig. 2(b) of theoriginal paper. Because comparisons of the obtained results withscaled-distance equations for Soil Classes II and III from Woods(1997) were shown in the paper several times, additional informa-tion on soil classification is needed. A classification of earth mate-rials for four classes by attenuation coefficient indicating thematerial damping was proposed by Woods and Jedele (1985), andthey performed a study of attenuation of ground vibrations in SoilClasses II and III. From Woods (1997), Soil Class II is competentsoils—most sands, sandy clays, silty clays, gravel, silts, weatheredrock (can dig with shovel and 5, N , 15); Soil Class III is hardsoils—densecompacted sand,dryconsolidated clay,consolidatedglacial till, some exposed rock (cannot dig with shovel, must usepick to break up and 15 , N , 50).Foraspatialtaskofwavepropagationinthelayeredsoilmediumfrom dynamic sources, the authors used the basic two-dimensional(2D)FEMmodelforwhichsoilwaspresentedasasinglelayerwiththethicknessof50m. Suchanapproximationis nonadequatetothereal soil conditions and pile installation because the modeldoes nottake into account a number of the previously mentioned factors. Inparticular, the model used cannot reflect well-known facts thatgroundvibrationsatthesamedistancesfromthesamesourceatvar-ious directions are different and soil stratification strongly affectsvibrations at the ground surface. However, the authors receivedgood agreement with the scaled-distance equation for Soil Class II[Fig. 2(b) of the original paper]. It seems that 2D FEM model is ac-ceptableforvarioussoilconditionswithvaluesfromastandardpen-etrationtest(SPT)between5and15.Theauthorsdidnotcommenton these results. Also, it is necessary to keep in mind that pile in-stallation in Soil Class III with values from SPT between 15 and50 generates more intensive ground vibrations than pile driving insoil class II. This problem was not recognized.Before analysis of the parametric study, it is necessary to recalltwo known facts.First, as mentioned previously, Eq. (1) provides assessment ofPPV attenuation between two points on the ground surface at anydistance from the source, and according to a definition, k is PPVofgroundvibrationsatoneunitofdistancefromdrivenpiles.There-fore,forknownPPVatPoint1,Eq.(1)providescalculationofPPVatPoint2orviceversaforPPVmeasuredatPoint2,PPVatPoint1can be determined with back analysis of Eq. (1). Woods and Jedele(1985)made numerousmeasurementsofgroundvibrationsat var-ious distances from diverse dynamic sources, and they studiedground attenuation between numerous pairs of measured PPV onthe ground surface. The coefficient k as PPV of ground vibrationsdepends on the source energy and a scaled-distance from thesource.Second,experimentalstudieshaveconfirmedthatthecoefficientnchangesinthe1–2narrowrangeand,ingeneral,itappearsthatthenrangeisindependentofsoiltype,energysource,andenergylevel,