Preparations for the load tests

Preparations for the load tests at Vatthammar were started in September 199 5. The four holes for the expander bodies were drilled by screw auger down to 10 metres depth and the holes stood open without any internal support. The expander bodies were lowered into the holes , connected to the tie rods and expanded by injection of cement grout under high pressure.

A pit was then excavated. Because of the stiffness of the soil, it was uncertain whether failure would be obtained even for the smallest plate if it was installed in an ordinary back-filled pit. The pit was therefore made fairly large and with such dimensions that a possible bearing capacity failure for the smallest plates would not be affected by the soil at the sides of the pit. The excavation was made 1.35 metre deep with base dimensions of approximately 11 x 4.5 metres. The slopes were given an inclination of about 1: 1.5 and the outer dimensions of the excavation were 14.5 x 8 metres.

This part of the excavation work was carried out with an excavator. The work was conducted very carefully with only a small deepening of the pit in each digging operation. The ground was very hard but, because of the pronounced layering, the soil could be peeled off layer by layer without any obvious disturbance of the

SGI Report No 54

126

Pressure, kPa

Results of pressuremeter tests of the Menard type in the test field at Vatthammar.

Investigations and load tets in silty soils 127

underlying soil. Two open containers were placed at the middle of the long sides and a few metres away from the excavation. These were filled with excavated soil.

Smaller pits for the test plates were then dug manually to a depth of 0.15 metre below the bottom of the large pit. Also this operation was performed very carefully by peeling off layer after layer. The outer dimensions of the small pits were made only marginally larger than the plates to allow insertion of the prefabricated moulds.

After insertion of the moulds, the instrumentation and reinforcement were put in place. The instrumentation below the plates was in this case confined to three settlement gauges under each plate. As in the previous test field, the settlement gauges were placed under the characteristic points at three corners of the plates.

In this case, however, the settlement gauges for all three plates were placed at 0.25, 2.0 and 4.0 metres below the bottom of the plate. This was done in order to obtain a better picture of the distribution of settlements versus normalised depth when all results were combined. The gauges were of the same type as in the previous test field, Fig. 4.2.24.

No attempt was made to measure the possible pore pressure generation in this non­

saturated soil. This was mainly due to practical problems because no suitable equipment for such measurements exists. BAT piezometers could not be readily used because of the need for a stiff tube in which the reading transducer can be lowered and a possible use of retractable tensiometers would require a number of insertions to be made in the stiff soil into which nothing could be pushed without using a jack or a drill rig. The risk of significant disturbance was therefore judged to be greater than the possible benefit of such measurements.

A few days after the concrete had been poured, the moulds were removed and the 0.15 metre deep slots at the sides of the plates were filled with soil.

The ground surfaces at the ground anchors were scraped off and levelled and stacks of excavator mats were placed as supports for the long railway beams. These were put in place on top of the stacks, bolted together and aligned with the centre line of the plates. The short beams were placed at the ends and connected to the tie rods, which were pre-stressed and the reaction system was fixed. The long beams were further fixed in position sideways by tightening straps fastened in the soil-filled containers to the centre of the beams, Fig 4 .3.13.

A plan of the test set-up is shown in Fig. 4.3.14.

SGI Report No 54

128

Smoothed side of the excavation showing the layering of the soil.

Assembly of the reaction system and injection of the ground anchors.

Fig. 4.3.13 Preparations for the load tests.

Investigations and load tets in silty soils 129

... w 0 Vl ~ ;;o (1) -.::, 0 ;:; z 0 u, .I>.

-f -·-·-

A Plan plates, for the

Ground anchors with expander bodies ---...

G . .

---~--~ -----·-·-·-·-·-·t===i·-·-·-·-··-·-13J.-·-·-·-·-·-·-El-·-0

---,A

1'---i-1--y---i,r--~1--1 Instrumentation Settlement gauge 0.25 m Settlement gauge 2.0 m Settlement gauge 4.0 m 0 5 10 (m) Fig. 4.3.14 Configuration of the excavation, test instrumentation and reaction system load tests at Vatthammar. Section A-A

The test pit was protected by tarpaulins using the long reaction beams as a ridge.

Ditches were also dug around the excavation in order to lead away any water from precipitation.

Any excess pore pressures were expected to dissipate rapidly. Normal Swedish practice for load tests in free draining soils is to apply each load step for up to 8 minutes. In this case, a duration of 16 minutes was chosen in order to be on the safe side. The relatively short duration of the load steps enabled the use of a manual load regulation system while the automatic electronic system was occupied in another project. Apart from this, the loading system was basically the same as for the tests at Vagverket. It consisted of a hydraulic pump with regulation valves, in this case operated manually, a hydraulic jack and a load cell. The circular steel plate for load distribution was placed directly on the top surfaces of the two smaller plates and on the larger plate on top of the larger square load distribution plate. The hydraulic jack used in this field had a longer stroke which eliminated the need for extra gap filling plates.

The measurement system consisted of a data acquisition and display system, the load cell and seven displacement transducers. The displacement transducers were fixed on the measurement ladder, which in turn was fixed by screw anchors in the soil well outside the expected zone of influence from the load tests. The displace­

ment transducers measured the vertical movement of the settlement gauges under the plate and of the four corners of the plate. The measurement ladder was supplied with a vertical scale and was continuously levelled by a precision instrument during the tests. This instrument had to be placed inside the excavation pit and was therefore also regularly checked against a fixed point further out from the test area.

Attempts to estimate the failure load by different methods yielded highly scattered results and it was decided to start the test series very carefully with small load steps on the smallest plate and to let the results govern the subsequent process.

The load tests were performed at the beginning of October 1995. No problems were encountered with the performance of the tests or the measurements. The weather, however, was miserable and caused some practical problems, and may also to some extent have affected some of the results. During the first test, the weather was windy and cloudy but dry. The following night, an autumn storm broke out with heavy rain and strong winds. The next morning, the tarpaulins were found to be torn to pieces and a considerable amount of water had entered the excavation. However, most of it had been drained away through the holes from the previous testing and sampling operations which still stood open and acted as

Investigations and load tets in silty soils 131

gutters leading down to deeper strata. Most of the remaining water was removed the same way and new tarpaulins were put in place, Fig 4.3.15. The weather then successively cleared up and the strong wind dried up the area. Nevertheless, it must be expected that the water content in the soil was somewhat higher during the second and third tests, i.e. on the 1 xl and 2 x 2 metre plates, than during the investigations and the first load test.

The pore pressures in the ground outside the load test area were measured in the previously installed piezometers at 2.5, 5.0 and 7.5 metres depth. However, at this stage the measured suctions were relatively low and it must be expected that diffused air had entered the piezometers. The following investigations conducted by Oberg (1997) shortly afterwards showed much higher matric suctions corre­

sponding roughly to the initially measured values. The pore pressure readings during the load tests are therefore not considered to be relevant. The pore pressure situation in the ground during the tests may be assumed to have corresponded to the normal situation as outlined in Fig. 4.3.9, except for the last two tests when the suction in the ground below the plates may be assumed to have been smaller at the start of the tests.

SGI Report No 54

132

Load test on 0.5 x 0.5 m plate in progress.

Clearing up after the storm.

Load test on I x I m plate in progress.

Fig. 4.3.1 5 Load tests at Vattham­

mar in progress.

Investigations and load tets in silty soils 133

---4.3.4 Results of the load tests

• 0.5 x 0.5 metre plate

The load test on the 0.5 x 0.5 metre plate started with small load steps of 10 kN.

Each step was applied for 16 minutes. The shapes of the settlement - log time curves connected to a straight line relation within 3 minutes and the application time was therefore considered to be sufficient, Fig 4.3.16. The loading was continued in small steps up to 290 kN. An indication of a change in behaviour in terms of increased creep rate and settlements had been observed at a load of 140 kN, but instead of continuously accelerating creep rates and settlements, the load-settlement relation only changed into a new, almost straight line, Fig 4 .3.17.

Since no sign of a failure could be observed, the load steps were doubled to 20 kN and later, at a load of 390 kN, increased to 30 kN. The test was terminated at a load of 420 kN, not because of signs of imminent failure but because the plate, which by then had settled almost 120 mm straight down, started to tilt, Fig. 4.3.18. No signs of heave or settlements could be observed outside the plate, but a pattern of fine radial cracks had developed, which spread and widened as the test proceeded, Fig 4.3.19. Time, s

Load, kN

0 50 100 150 200 250 300 350 400 450

0

··-- _···

-2

···-· ^-.. .

4

··--···· ^•••

en 0

~ E ⁶ .

._

~

8

c. ^Cl>

·- ^{. ._}

f!

-~

(.) 10

•

12 14

•

Fig. 4.3.17 Creep rate versus load for the load test on the 0.5 x 0.5 metre plate at Vatthammar.

Load, kN

0 50 100 150 200 250 300 350 400 450

0

- -Corner 1

20

--Corner2

40 ^--Corner3

E _{_ _ Corner4}

E

.J 60 _ _Average settlement

C:

Cl)

E Cl) 80

=

Cl) U) 100

120 140

Fig. 4.3.18 Settlements of the four corners of the 0.5 x 0.5 metre plate at Vatthammar.

Investigations and load tets in silty soils 135

Fig. 4.3.19 Pattern of cracks in the ground around the 0.5 x 0.5 metre plate at Vatthammar at the end of the load test.

The measurements of the settlement gauges showed that the settlements started at the top just below the plate and then spread downwards with increasing load. The upper gauge was placed 0.25 metres below the plate, which corresponds to 0.Sb, and the next gauge was placed 2.0 m below the plate (or 4b). Roughly speaking, about half ofthe settlements occurred down to 0.Sb and the other half further down, but the exact relation changed continuously as the test proceeded, Fig 4.3.20.

Load, kN

0 50 100 150 200 250 300 350 400 450

0

- Average plate

20 settlement

E E

..,.

40 ^- Gauge 0.25 m

C below plate

Cl)

E - Gauge 2.0 m below

Cl) 60

"

^plate

"'

Q.

.,

-Gauge 4.0 m below

=c 80 ^plate

«i -~ 100 t:: Cl)

> 120

Fig. 4.3.20 Vertical displacement of settlement gauges in the load test on the 0.5 x 0.5 metre plate at Vatthammar.

SGI Report No 54 140

136

The evaluation of the creep rate enabled an extrapolation of the measured settlements to a corresponding settlement after 10 years. Both load-settlement curves show a break in the relation at a load of about 140 kN. Instead of a continued accelerated bending-off downwards at further loading, the curve enters a new, almost straight load - settlement relation with only a slight curvature downwards.

The next knee in the curve at about 300 kN is related to a doubling of the load steps at this load, Fig 4.3.21.

Load,kN

0 50 100 150 200 250 300 350 400 450

oru-..~Nijl~;;;:::::=---20 40 E 60 E -· _C: 80

~100

~ 120

Ji

¹⁴⁰

160

---0-Measured 16-min settlement

- e - Extrapolated 10-year settlement

•

. ' ^,. _{' ,.}

180 200

Fig. 4.3.21 Measured and extrapolated load - settlement curves from the load test on the 0.5 x 0.5 metre plate at Vatthammar.

The parameter c describing the relation between settlement and creep rate was almost constant at 0.07 evaluated with respect to total settlements. Evaluated with respect to the settlement in the individual load steps, it rose from 0.11 at low load to about 0.18 at the change in behaviour at a load of 140 kN. It then remained constant for higher loads. As before, the chosen reference time is 1 year, Fig. 4.3.22.

Investigations and load tets in silty soils 137

••••••••• • • • ^• •

Fig. 4.3.22 Creep rate in relation to settlements measured in the load test on the 0.5 x 0.5 metre plate at Vatthammar.

1111 J x 1 metre plate

The load test on the 1 x 1 metre plate was performed in steps of 40 kN up to a maximum load of 800 kN. All load steps had a duration of 16 minutes except for the steps in an unload - reload cycle from 400 kN to 60 kN and back, for which the duration for each step was only one or two minutes. The plate settled with a small diagonal tilt and, by the end of the test, the corners of the plate had settled between 24 and 30 mm, Fig. 4.3.23.

As in the previous test, the settlement - log time curves formed straight lines after less than 3 minutes, Fig. 4.3.24.

The evaluated creep rates showed a significant change in behaviour at about 440 kN load, Fig. 4 .3 .25. However, they did not start to accelerate for higher loads but only formed a new, almost straight line relation with the applied load.

The measured load - average settlement curve and the curve for 10-year settle­

ments extrapolated by using the measured creep rates show a fairly elastic behaviour up to a load of about 440 kN. The curves then start to bend off downwards more markedly but the curvatures are smoothly rounded and a bearing capacity failure is far distant at the end of the test at 800 kN load, Fig. 4.3.26.

SGI Report No 54

138

-- _---

Settlement of the four corners in the load test on the I x I metre plate at Vatthammar.

Investigations and load tets in silty soils 30

139

Load, kN

Fig. 4.3.26 Measured and extrapolated settlements in the load test on the I x I metre plate at Vatthammar.

SGI Report No 54 800

140

The settlement gauges under the plate were placed at 0.25, 2.0 and 4.0 m below the plate, which in this case corresponded to 0.25b, 2b and 4b. The measurements showed that by the end of the test about 20 % of the settlements were due to compression of the soil down to 0.25b and the remainder had occurred between 0.25b and 2b, Fig 4.3.27.

Load,kN

0 100 200 300 400 500 600 700 800

0

5

- Gauge 4.0 m below plate

E 10 _- Gauge 2.0 m below plate

E

..,;- - Gauge 0.25 m below plate

C: a, 15 - Average plate settlement E

i

a,

"'

²⁰

25

30

Fig. 4.3.27 Vertical displacement of the settlement gauges in the load test on the I x I metre plate at Vatthammar.

The parameter c describing the time dependence of the settlements showed a fairly stable value around 0.07 when referring to total settlements, being somewhat smaller in the "elastic" load range. When referring to settlements in the separate load steps, the c -parameter increased from 0.07 to about 0.18 at the yield stress, whereupon it remained fairly stable, Fig. 4.3.28.

The results of the load test on this plate were in principle very similar to the results from the previous test. However, the average stress at yield, 440 kN/m², was in this case about 20 % lower than in the previous test, in spite of the fact that most methods of predicting bearing capacity would predict an equal or higher stress for the larger plate. This may be related to the rainstorm preceding the test.

Investigations and load tets in silty soils 141

••

Fig. 4.3.28 Creep rate in relation to settlements measured in the load test on the lxl metre plate at Vatthammar.

1111 2 x 2 metre plate

The load test on the 2 x 2 metre plate was performed in steps of 50 kN up to the maximum load of 800 kN. One unloading-reloading cycle was performed from 450 kN to 100 kN and back. The load steps were applied for 16 minutes except for the unload-reload cycle, in which they were only applied for 1 to 2 minutes. Some of the steps at the unloading after the test were also maintained for a somewhat longer time and showed that the rebound was also somewhat time dependent. The settlements at the end of the test amounted to only about 2.4 mm and the measurements in the unloading - reloading cycle and at the final unloading showed that they were essentially purely elastic, Fig 4.3.29.

All significant settlements occurred within at depth of 2b (4 metres) below the plate. About 12 % occurred due to compression of the soil down to 0.125b, 75 % between 0.125b and lb and 13 % between lb and 2b, Fig. 4.3.30.

The settlement - log time curves showed the same pattern as in the previous tests, although the settlements and creep rates were so small that the resolution and accuracy of the measuring system affected the results. The evaluated creep rates versus load formed a fairly straight line relation without any significant break which could be interpreted as a yield stress, Figs. 4.3.31 and 32.

SGI Report No 54

142

0 100 200 300

0.5

E E

. ; ' C:

CII E E CII V) CII

1.5

2

2.5

3

0 100 200 300

0.5

E E

~ CII

E CII

~

V)

1.5

2

2.5

Load, kN

400 500 600 700 800 900

- Comer1

- Comer2

- Comer3

- Comer4

Load, kN

400 500 600 700 800 900

Fig. 4.3.29 a) Settlements of the four corners in the loa.' test on the 2 x 2 metre plate at Vatthammar.

b) Average settlement of the 2 x 2 metre plate at Vatthammar.

Investigations and load tets in silty soils 143

Load,kN

0 100 200 300 400 500 600 700 800 900

0

0.4

E 0.8 E C: 41 1.2

~ E ai 1/) 1.6

- -Gauge 4.0 m below plate 2 - -Gauge 2.0 m below plate --Gauge 0.25 m below plate - -Awrage plate settlement 2.4

Fig. 4.3.30 Vertical displacement of the settlement gauges under the 2 ^x 2 metre plate at Vatthammar.

SGI Report No 54

144

0

Settlement versus time in the load steps in the test on the 2 x 2 metre plate at Vatthammar.

Investigations and load tets in silty soils 145

Load, kN

The measured and extrapolated load - settlement curves also showed fairly linear relations in accordance with an almost fully elastic behaviour, Fig. 4.3.33.

The evaluated values of the parameter c shows that when the creep rate is refeITed to total settlements in the elastic range, the cvalue is low and only about 0.02 -0.03. When it is refeITed to settlements in the individual load steps, the c-parameter increases linearly with load from 0.06 to 0.14 at a load of 800 kN, Fig 4.3.34.

Assuming that it would adopt the same pattern and the same value at yield as in the previous tests, this points to a yield load around 1200 kN. Also this would correspond to a lower stress level than for the previous plate, which could be related to the later observed effect of soaking of the ground surface spreading to greater depths and soil volumes with time.

1111 Distribution of settlements with depth

The combined results of the three load tests yield a consistent picture of the distribution of settlements with depth. Since the settlement gauges were placed at the same depths for all plates, there was a fairly even distribution of gauges with- I

depth normalised against the width of the plates when the results are combined.

This combination of the results is not quite unambiguous because the distribution of settlements in a single test changes somewhat with the applied load and the soil

This combination of the results is not quite unambiguous because the distribution of settlements in a single test changes somewhat with the applied load and the soil

In document LARSSON ROLF (Page 128-151)