Energy Measurement in the Brazilian SPT System

This paper presents results of the instrumentation of 373 blows from two SPT deployments performed in the Sarapuí II Test Site, located in Duque de Caxias, Rio de Janeiro. In these blows the hammer drop height, its velocity at impact, the rate of blows and also the energy transferred to the rod stem have been measured. It is therefore possible to know the loss of energy for the SPT process (and the corresponding efficiency factors), since the hammer is delivered at zero velocity up to the time the transmitted energy reaches the rod stem.


Introduction
Despite the existing problems associated with the reliability and repeatability of the Standard Penetration Test, Campanella & Sy (1994) emphasize that the SPT continues to be the most used in situ test for foundation design, evaluation of liquefaction potential and compaction control of sands and sandy silts.Many authors associate the widespread use of the test to the simplicity of the test procedure, robustness of the equipment and low operational cost (e.g., Broms & Flodin, 1988;Décourt, 1989).
One of these papers, by Schmertmann & Palacios (1979), has shown that the number of blows N varies inversely with the energy delivered to the rod stem, to N equal at least 50.After some discussions concerning the need to standardize and the choice of the proper energy to be used as a reference for the N value (e.g., Kovacs & Salomone, 1982;Robertson et al., 1983;Seed et al., 1985;Skempton, 1986), ISSMFE (1989) has established 60% of the theoretical free fall energy (or nominal potential energy) as the international reference.Therefore the corresponding N 60 is obtained as where N = measured number of blows, E = energy corresponding to N and E 60 = 60% of the international reference energy E*, E* = 474 J. Décourt (1989) and Kulhawy & Mayne (1990) have summarized the factors affecting the energy transmission from the hammer to the rods.According to Décourt (1989), the energy entering the rod stem (or enthru energy, E i ) can be obtained as E e e e E i = 1 2 3 * ( 2) where e 1 , e 2 and e 3 are efficiency (or correction) factors.The efficiency factor e 1 relates the kinetic energy just before the impact to the free fall energy and is mainly dependent on the way the hammer is lifted and released.A number of studies have been carried out on this subject (e.g., Kovacs et al., 1977Kovacs et al., , 1978;;Kovacs, 1979Kovacs, , 1980;;Kovacs & Salomone, 1982;Skempton, 1986;Tokimatsu, 1988;Décourt, 1989).
The factor e 2 is associated to the loss of energy due to the presence of the anvil (e.g., Skempton, 1986;Décourt, 1989).The efficiency factor e 3 is related to the rod length and e 3 values smaller than 1 have been proposed (e.g., Schmertmann & Palacios, 1979;Skempton, 1986) to take into account the separation between hammer and anvil for rod lengths smaller than 10 m, due to the upcoming stress wave.However, recent research (Cavalcante, 2002;Odebrecht, 2003;Daniel et al., 2005;Odebrecht et al., 2005;Danziger et al., 2006) has shown that a number of impacts may occur in a single blow, each impact being responsible for part of the energy delivered to the rod stem.Thus, e 3 should be taken as 1.The e 1 , e 2 and e 3 values are discussed below together with the corresponding values obtained herein.
The efficiency factors are related to the theoretical (or nominal) free fall energy, thus they are not the real ones.Instead, the efficiency factors are influenced by the errors associated with the non-use of the real free fall energy during the test.To the authors' knowledge, very few studies have been conducted regarding the potential energy actually used in the test (e.g., Riggs et al., 1983;Cavalcante et al., 2011), the latter only relating to the hand lifted pinweight hammer system regularly used in Brazil.However, a very experienced crew performed the SPTs in the study carried out by Cavalcante et al. (2011), and the obtained results cannot be considered typical but rather a benchmark for the best results possible to be obtained with this system.
This paper presents research to measure the potential energy of the regular Brazilian system in regular operational conditions, i.e. with a crew with regular experience.Also, the impact velocity of the hammer has been evaluated.The blow count rate was also measured, provided that there are recommendations for the rate to be used in liquefaction analysis (Seed et al., 1985).The energy reaching the rod stem has been measured and used to evaluate the efficiency factors, which have been therefore evaluated based both on the nominal free fall energy and on the measured energy.

SPT analyzer
The SPT Analyzer measures the energy transmitted to the rod stem, besides other quantities.It is composed of a data acquisition unit, instrumented rods and connection cables, as can be seen in Fig. 1.
The acquisition data unit has two channels for the force signal and other two for the acceleration signal.Its maximum sample frequency is 20 kHz.The maximum reading interval is 102.4 ms.
Rods 1 m in length have been instrumented, each with a pair of force measuring devices and a pair of accelerometers.Electric strain-gauges have been used for monitoring the force in the instrumented rods, forming a Wheatstone bridge directly fixed to the rods.Piezoelectric accelerometers, with 0.02 g resolution and capacity of 5000 g, have been used to record the rod acceleration.The accelerometers can be fixed to the rods in diametrically opposite positions and between the force sensors.The data acquisition system transforms the acceleration records into velocity upon integration with time.

High speed camera
A Casio EX-FH20 high speed camera, capable of recording up to 1000 pictures per second, was used to record the hammer drop height and impact velocity.
The images recorded by the camera were digitized and analyzed picture by picture in order to enable identification of the maximum height drop during hammer raise and the moment the hammer hit the anvil.To help determine the hammer position, an Invar ruler is positioned beside the SPT set. Figure 2 shows the system employed in the instrumentation of the SPT.

Test characteristics
Two SPT deployments have been monitored in Sarapuí II Test Site, situated at the margin of the Washington Luiz Highway, in the area of the Navy Radio Station in the municipality of Duque de Caxias/RJ.Geotechnical characteristics of the test site have been provided by Jannuzzi (2009Jannuzzi ( , 2013)).
According to Jannuzzi (2009) the soil profile in the region is formed by a very soft clay layer with a typical thickness of 7.5 m to 8.0 m, followed by minor layers of clay, sands and silts and clays once more.The water table is at ground level.
The same crew including a chief-operator and three auxiliary-operators were in charge of the two SPT borings.
An anvil with a mass of 977 g was used (see Fig. 3).It should be pointed out that although the Brazilian standard NBR-6484/2001 states that the anvil should have a mass Figure 2 -System employed in SPT energy monitoring.
ranging from 3.5 kg to 4.5 kg, anvils with a mass of around 1 kg are very often used all over Brazil.Reference must be made to, for example, Skempton 1986, Décourt 1989, Belincanta 1998and Belincanta & Cintra 1998 for the influence of the anvil mass on the energy transmitted to the rod stem.
A sisal rope was used for lifting and releasing the hammer.The pinweight hammer with a wood cushion is shown in Fig. 4. No measurement was made of the hammer mass in the present study.However the SPT company in charge of the tests has informed that the hammer mass is verified periodically and is equal to 65 kg.Measurements made in previous research (Cavalcante, 2002) indicate that this information may be considered reliable, and errors in the hammer mass may be generally considered negligible.The hammer drop height has been visually controlled, as in the usual procedure, with the aid of a mark at the pinweight hammer.
The rods employed in the tests had an external diameter of 33 mm, 3.2 kg per meter, as recommended in the Brazilian standard NBR-6484/2001.
In the first boring, named Boring 1, 141 blows have been monitored.The rod stem length (including the sampler) varied from 10.80 to 22.80 m (nominal test depths varying from 9 to 21 m).
In the second boring, named Boring 2, 232 blows have been monitored.The rod stem length (including the sampler) varied from 11.70 to 25.70 m (nominal test depths varying from 10 to 23 m).

Instrumentation results
In order to avoid significant loss of image quality, a rate of 210 pictures per second was used, corresponding to a maximum resolution of 480 x 360 pixels.The records obtained by the high speed camera have been transferred to a computer, separated picture by picture, and analyzed by AutoCAD software in a way that it would be possible to define the hammer height during drop by the action of each blow.An Invar ruler acted as a reference.
The height measured in the picture just before hammer release is defined as the hammer drop height.The hammer impact velocity has been obtained by the analysis of the hammer height picture by picture, since the instant of its release up to the imminence of impact (last picture before hammer contact with the anvil).Thus it has been possible to adjust a function that describes the relation between the height drop of the hammer and the time, according to Fig. 5.
The derivation of the hammer drop height in relation to time, when the height drop tends to zero is the impact velocity.
Different polynomial functions of two and three degrees have been tested in various blow counts that produced good agreement.The difference observed in the velocity during impact selecting one or other polynomial function has been of very low significance.The option has been then to try to adjust a second-degree polynomial function in order to simplify the numerical estimation.The hammer acceleration in time is the second derivative of the hammer drop height in time.In this way, the use of a second-degree polynomial function implies the consideration of constant hammer acceleration during the hammer release.
The hammer acceleration during its release is influenced by the gravitational force (which is approximately constant) and by friction forces.In this way, the resultant from the friction forces is considered constant during the hammer release.
Considering that the second or third degree polynomial functions did not produce significant change in the adjustments, it is reasonable to consider that the friction vs. time function, in the analyzed cases, is approximately constant.
The average values of hammer drop height (h d ), hammer velocity at impact (v i ), potential energy of hammer at release (E p ) and kinetic energy at impact (E k ) in each blow sequence of borings 1 and 2 are presented in Tables 1 and 2, respectively.The potential energy and kinetic energy at impact have been calculated as: where m = hammer mass, considered as 65 kg and g = gravity acceleration, considered as 9.81 m/s 2 .
In the three first blow sequences of Boring 2 no filming has been carried out.Furthermore, in a significant number of blows from deployments 1 and 2 (152), it has not been possible to determine the impact velocity due to problems with the video.In a smaller number of blows (102) video problems prevented the determination of hammer drop height in both deployments.
Due to errors in the SPT Analyzer operation, the rod energy in six blows from sequences 6 and 9 from Boring 2 has not been monitored.However, the hammer drop height and impact velocity of these blows have been measured.
Figures 6 and 7 show the frequency distribution of the SPT hammer drop height in borings 1 and 2, respectively.Figures 8 and 9 illustrate the percentage of blows applied in different ranges of hammer drop height in borings 1 and 2, respectively.
Figures 10 and 11 show the hammer drop height blow by blow in each sequence from borings 1 and 2, in all available cases.
Tables 1 and 2 present, for each blow sequence, the following measurements: the number of blow counts for 45 cm sampler penetration (N 45 ), the average frequency of blow count application, the working shift when the blows have been applied (see definition below), the nominal depth of the test and the length of the rod stem, the hammer drop height, the impact velocity and the energies measured, as well as energy ratios.
The average frequency of the blows has been calculated considering the interval from the initial lifting of the hammer, in the first blow of each sequence, up to the final of the hammer impact for the last blow.Figure 12 shows the hammer drop height vs. the frequency of blows.
In order to evaluate the variation in hammer drop height during the day, the working period of the boring crew was divided in four shifts, namely: first shift, from 8 and 10 AM; second shift, from 10 to 12 AM; third shift, from 14 to 16 PM and fourth shift, from 16 to 18 PM.Figure 13 shows the corresponding variation.
The energy reaching the rod stem (enthru energy) (E i ) has been calculated from Eq. 5.The values of force (F) and velocity (v) have been obtained through the measurements from the strain-gauges and accelerometers installed on the rods.Tables 1 and 2    Figure 14 shows typical force and velocity signals measured just below the anvil.Figure 15 shows typical values of energy vs. time.
Figure 16 illustrates values of E i normalized by the actual potential energy (E p ), as a function of the rod length, for borings 1 and 2.

Analysis of the results
The average hammer drop height of sequences from Boring 1 varied from 67 to 87 cm, with actual potential energy in the range 429.1 -554.0J, reaching a difference of 29%.The hammer has been lifted higher than 80 cm, or lower than 70 cm (difference higher than 5 cm from the standard value) in 64% of the blows.The average hammer drop height of the whole data from Boring 1 is 80 cm, with a standard deviation of 9 cm.
The scatter in hammer drop height values was smaller in Boring 2. The average hammer drop height varied from 69 to 87 cm, with potential energies in the range 440.8-553.0J, reaching a difference of 25%.The hammer has     been lifted to heights greater than 80 cm or lower than 70 cm in 46% of the blows.The average hammer drop height is 76 cm, with a standard deviation of 8 cm.Therefore, in spite of both borings had been performed by the same crew, using the same equipment and on the same site, under the same conditions, a difference on the average hammer drop height in Borings 1 and 2 was veri-fied (80 and 76 cm, respectively).Two other aspects observed in the tests: i) a variation in the hammer drop height in the same blow sequence; ii) a variation in the average hammer drop height from different sequences.
A tendency to increase the hammer drop height with the advance of the sequence was observed in both borings.Only in two out of 21 sequences the opposite behavior was verified.It was hypothesized that the increase in hammer drop height is caused by the fatigue of the crew, resulting in less care in the procedure, although the opposite should seem more probable.The average frequency of the blows was 19.4 and 39.7 blows per minute, in Borings 1 and 2, respectively, which is a significant difference.The smaller frequencies were observed in the longer sequences, with more than 25 blows, probably also caused by fatigue of the crew.The shorter sequences, with five or even fewer blows, presented the higher frequencies.The average frequency of blows for all sequences (Borings 1 and 2) was 27.8 blows/min with a standard deviation of 4.7 blows/min.
It is possible that high values of factor e 2 are associated to the downward movement of the rod stem during hammer blow, generating an increase in potential energy that is transferred to the rods in subsequent hammer impacts of the same blow.This occurrence, described by Odebrecht (2003), is more relevant in low resistance soils.The SPT Analyzer is capable of measuring the whole energy transferred to the rod stem, only if the process occurs before 102 ms.However, the kinetic energy is calculated in relation to the first hammer impact with the anvil, so the e 2 value can be overestimated should other impacts occur.
The results of E i /E p as a function of the rod length measured in Borings 1 and 2 are presented in Fig. 16.This figure illustrates that the energy transferred to the rod stem is not significantly affected by its length, at least in the range of lengths analyzed, from 10.80 to 25.70 m.This is corroborated by previous studies (e.g., Cavalcante, 2002;Odebrecht, 2003;Daniel et al., 2005;Danziger et al., 2008) indicating that the energy transmitted to the rod stem does not depend on its length and the e 3 factor should be considered equal to 1.00.

Conclusions
The paper presented the instrumentation results of two SPT deployments performed by the same crew, using the same procedures and equipment in the Sarapui II Experimental Test Site.The main conclusions are summarized as follows: i) Although both borings had been performed by the same crew, using the same equipment and on the same site, under the same conditions, a difference was found in the average hammer drop height in Borings 1 and 2 (80 and 76 cm, respectively).Two other aspects observed in the tests: a variation in the hammer drop height in the same blow sequence; a variation in the average hammer drop height from different sequences.ii) A tendency to increase the hammer drop height as the sequence advances was observed in both borings.The opposite behavior was verified in only two out of 21 sequences.It was hypothesized that the increase in hammer drop height is caused by the fatigue of the crew, resulting in the careless of the procedure, although the opposite should seem more probable.iii) The average frequency of blows was 19.4 and 39.7 blows per minute in Borings 1 and 2, respectively.The smaller frequencies were observed in the longer sequences, with more than 25 blows, probably caused by the fatigue of the crew.The shorter sequences, with five or even less blows, presented the higher frequencies.iv) A slight trend of higher scatter on the hammer drop height was observed when the rate of blows increased.Skempton, 1986, Décourt, 1989and Cavalcante et al., 2011).v) The work shift does not seem to have influenced the hammer drop height, since a similar scatter was observed for all work shifts.vi) The efficiency factor e 1 (E k /E*) varied in Boring 1 from 0.78 to 1.10 and in Boring 2 from 0.87 to 1.08.These values are higher than those found by Cavalcante et al. (2011) and varied in a broader range than those presented by Décourt (1989).Values of e 1 greater than 1.00 are explained by the hammer drop height above the standard value in various sequences.vii) The values of e 1 * (E k /E p ) varied in Boring 1 from 0.86 to 0.99 and in Boring 2 from 0.91 to 0.96, in a narrower range than the e 1 values.viii) The average energy measured just below the anvil, E i , varied from 318.1 J (efficiency of 67% in relation to the theoretical potential energy or nominal energy) to 493.5 J (efficiency of 103%) in Boring 1, whereas in Boring 2 varied from 403.5 J (efficiency of 84%) to 551.4 J (efficiency of 115%).The scatter in E i values is mainly due to the variation in the hammer drop height.
The obtained results indicate that even SPT performed by the same boring crew, in similar conditions, can result in N values with distinct significance.ix) When the efficiency of the energy measured just below the anvil is calculated in relation to the actual potential energy, the range in efficiency is significantly lower, varying from 74% to 93% in Boring 1 and from 85% to 96% in Boring 2. x) The efficiency factor e 2 (E i /E k ) varied in Boring 1 from 0.86 to 0.94 and in Boring 2 from 0.92 to 1.00.xi) The values of E i /E p vs. the rod length indicate that the energy transferred to the rod stem is not significantly affected by its length, and the efficiency factor e 3 should be considered as 1.00.

Figure 4 -
Figure 4 -Equipment employed in the tests in Sarapuí II Experimental Test Site.
Cavalcante et al. (2011) used polynomial functions of fourth degree to describe hammer drop height function with time.

Figure 6 -
Figure 6 -Frequency distribution of the hammer drop height from Boring 1.

Figure 7 -
Figure 7 -Frequency distribution of the hammer drop height from Boring 2.

Figure 8 -
Figure 8 -Percentage of blow counts applied in different ranges of hammer drop height from Boring 1.

Figure 9 -
Figure 9 -Percentage of blow counts applied in different ranges of hammer drop height from Boring 2.
Figure 12 -Hammer drop height vs. frequency of blows per minute.

Figure 13 -
Figure 13 -Hammer drop height vs. shift (first shift, from 8 and 10 AM; second shift, from 10 to 12 AM; third shift, from 14 to 16 PM and fourth shift, from 16 to 18 PM).

Figure 14 -
Figure 14 -Typical signals of force and velocity measured just below the anvil (Blow 21 of Sequence 9, Boring 1).

Figure 16 -
Figure 16 -Rod length vs. energy just below the anvil normalized by the actual potential energy (E i /E p ).

Figure 19 -
Figure 19 -Efficiency factor e 2 as a function of the anvil mass (adapted from Décourt, 1989 and Cavalcante et al., 2011).
3 L = length of the rod stem, including the sampler length (m); 4 h d = SPT hammer drop height; 5 A = average value; 6 SD = standard deviation; 7 v i = SPT hammer impact velocity; 8 E p = actual potential energy of the SPT hammer at the release moment; 9 E k = kinetic energy of the SPT hammer at the imminence of impact; 10 E i = energy measured just below the anvil; 11 E * = theoretical potential energy of SPT hammer from Brazilian system (478.2J).