Skip to contentUnited States Department of Transportation - Federal Highway Administration FHWA Home
Research Home
Report
This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-RD-99-194
Date: June 2000

Development and Field Testing of Multiple Deployment Model Pile (MDMP)

CHAPTER 3. THE MULTIPLE DEPLOYMENT MODEL PILE (MDMP)

3.1 General Description

The Multiple Deployment Model Pile (MDMP) is an in situ soil testing device very similar to the 7.26-cm (3-in) model pile previously described in section 2.6. The MDMP is composed of a series of modular sensors that are screwed together in any desired configuration. The model pile is capable of measuring axial loads, pore water pressure, total radial stresses, local displacement, and pile acceleration. A typical configuration of the modular MDMP is shown in Figure 17. In summary, the MDMP instrumentation includes three load cells, three accelerometers, a displacement transducer, a pore pressure transducer, and a total pressure cell.

Two load cells are positioned in a series with a total stress cell and pore pressure cell (located in transducer housing) centered between the load cells. The friction sleeve between the two load cells has a surface area of 2000 cm2 (310 in2). By subtracting the measured loads of each load cell and using the surface area, the friction along the friction sleeve can be calculated. An additional load cell is located near the tip of the pile. This additional load cell offers another measurement of friction along a greater length of the pile, as well as a measurement of tip resistance. A slip joint is located 16.7 radii from the total stress cell and pore pressure cell. The slip joint utilizes a direct current-linear variable displacement transducer (DC-LVDT) to measure up to 5 cm (2 in) of local displacement. During load tests, the local displacement measured by the DC-LVDT and the load cells should yield a load-displacement curve that is independent of any slack and compression of the drill rods. The friction sleeve can be made of different materials of various surface finishes (different roughness), allowing the examination of surface roughness effects on the frictional pile resistance. Pile acceleration is measured in the model pile utilizing high-impact accelerometers. The accelerometers are mounted inside the model pile at the load cell locations, thereby allowing for force and velocity records at the same location and minimizing uncertainties in the acceleration records due to drill rod connections. Two sets of load cells of different capacities were designed for use with the MDMP in a variety of subsurface conditions. The MDMP can be used to model large displacement piles by using a closed-ended tip. Also, small displacement piles can be modeled by using an open-ended tip.

3.2 Requirements

The Multiple Deployment Model Pile (MDMP) must be able to record the following measurements during driving, static load testing, and restrikes:

  • Axial loads at multiple locations along the pile (static and dynamic).

  • Pore pressures (static and dynamic).

  • Tip resistance (static and dynamic).

  • Total radial stresses (static).

  • Local displacement (static).

  • Accelerations (dynamic).

Figure 17. View Alternative Text

Figure 17. Typical Configuration of the Modular MDMP.

Total capacity, load transfer, and time-dependent information can be determined from these measurements. Initially, the MDMP tests were planned to be conducted in medium to soft Boston Blue clay deposits in the eastern Massachusetts area. Additional requirements were included so that the MDMP could be deployed in stiffer Boston Blue clay, glacial till, and/or dense sands.

The major objective of the MDMP is to simulate the installation and stress history that full-scale piles experience. To achieve this, the MDMP must be designed and constructed rugged enough to withstand driving stresses and, more importantly, the instrumentation must maintain the required standard of accuracy throughout the testing sequence. Measurements need to be recorded during the three stages of pile history - installation, stabilization (equilibration), and static loading conditions. Total radial pressures and pore pressures are planned to be continuously monitored during all stages. Axial strains and accelerations need to be monitored during driving to provide for the dynamic prediction of the pile capacity. The MDMP can therefore be restruck to assess the gain of capacity with time. Axial strains need also to be recorded during static load testing. The MDMP has several different tip configurations (see Figure 18), including an open-ended and a closed-ended cutting shoe that simulate small displacement and large displacement piles, respectively. The cutting shoe acts as an anchor during load tests by providing a reference point for local displacement measurements in the slip joint.

To determine the required ranges of measurements for the various instrumentation, a "typical" soil profile was established (see Figure 19). This typical soil profile is based on subsurface conditions found in the Boston area and consists of the following layers (from the surface downward):

  • 6 to 9 m (20 to 30 ft) of fill, organic material, silty sand, sand and gravel, and stiff clay (OCR Approximately 8).

  • 9 to 12 m (30 to 40 ft) of medium Boston Blue clay (BBC) (OCR 1.5 to 7, decreasing with depth), with an Su of about 0.4 to 0.6 tsf.

  • 15 to 38 m (50 to 125 ft) of normally consolidated soft BBC (OCR Approximately 1 to 1.5), with an Su Approximately 0.22 Sigma'vo.

This subsurface profile was used to calculate the expected conditions that the model pile will be subjected to during installation and testing. In addition, the MDMP was designed to allow for testing in stiff BBC, glacial till, and dense sand. A soil profile consisting of dense sand as shown in Figure 20 was used to represent these more difficult driving conditions.

3.3 Analysis of the MDMP Loading Conditions

3.3.1 Overview

The loads (soil resistance) that the MDMP is expected to be subjected to will vary depending on the installation mode, soil type, and pile geometry. Both dynamic (driving conditions) and static (static load test conditions) analyses were conducted to evaluate the MDMP's condition under the expected loads. The MDMP load cells will be subjected to a large range of axial loads due to testing in a variety of soil profiles. Soft BBC was used to represent the lower soil resistance to be measured by the load cells. Dense sand and/or glacial till was selected to represent the upper load measurements. These two limiting cases were used in both the dynamic and static analyses outlined in the following sections. Appendices A and B detail the calculations carried out for the static and dynamic analyses, respectively.


Figure 18. View Alternative Text

Figure 18. Tip Configurations of the MDMP.



Figure 19. View Alternative Text


Figure 19. Typical Soil Profile for the Boston Area.

3.3.2 Static Capacity Analysis

Based on the review of existing model piles in Chapter 2, the MDMP was assumed to be 76.2 mm (3 in) in diameter (O.D.), with a 9.525-mm (3/8-in) wall thickness. The total length was assumed to be 4.88 m (16 ft), with a closed-ended configuration. Several empirical methods were employed to evaluate the loads under static conditions for both the soft BBC and dense sand cases. These loads were calculated at depths ranging from 12.2 to 33.5 m (40 to 110 ft).

The typical profile presented in Figure 19 was chosen to represent the soft BBC profile. The Alpha (Tomlinson, 1971) and Gamma (Vijayvergiya, 1972) methods were used for the determination of the skin friction, while traditional and CPT (de Ruiter, 1975; Toolan and Fox, 1977; and de Ruiter and Beringen, 1979) methods were used for the tip resistance. Appendix A outlines the details of the static analyses. Table 2 summarizes the calculated skin resistance, tip resistance, and total resistance for the range of depths indicated. The values presented in Table 2 are the average values from the various methods used. The total resistance acting on the 4.88-m (16-ft) MDMP section ranges from approximately 56 to 82 kN (12.6 to 18.4 kips). Therefore, the MDMP is expected to experience loads of around 45 to 90 kN (5 to 10 tons) in the soft BBC.

Figure 20. View Alternate text

Figure 20. Hypothetical Soil Profile of Dense Sand.

Table 2. MDMP Static Load Resistance in Soft BBC (Lower Limiting Case).

Depth @ Tip (m)

Skin Resistance (kN)

Point Resistance (kN)

Total Resistance (kN)

12.2

53.7

2.6

56.3

15.2

55.2

2.6

57.7

18.3

57.7

2.6

60.3

21.3

59.2

2.6

61.8

24.4

60.5

2.9

63.4

27.4

64.8

3.2

68.0

30.5

72.1

3.6

75.6

33.5

78.3

3.9

82.2


Figure 20 presents the dense sand profile for the evaluation of the upper soil resistance limit. Meyerhof's (1951, 1976), Vesic's simplified (1975), Vesic's advanced (1977), and the American Petroleum Institution (API) (1984) methods were used for the determination of the tip resistance. The traditional, McClelland (1972), and Bhushun (1982) methods were used for the determination of skin friction. Appendix A outlines the details of the static analyses for this case. Table 3 summarizes the calculated skin resistance, tip resistance, and total resistance for the range of depths indicated. The values presented in Table 3 are the average values from the various methods used. The total resistance acting on the 4.88-m (16-ft) MDMP section ranges from approximately 186 to 465 kN (38 to 95 kips). Note that these values are conservative (for the purpose of the upper load evaluation) because they do not employ the critical depth adjustment for both resistance components - tip and skin. Based on these results, the MDMP may be expected to experience loads of around 185 to 465 kN (20 to 50 tons) in dense sands.

3.3.3 Dynamic Analysis

To evaluate the dynamic loads and accelerations during driving, wave equation analyses using the software program GRLWEAP (Goble, et al., 1995) were performed. The use of several hammers was investigated, including 0.62-kN, 1.33-kN, and 2.22-kN (140-lb, 300-lb, and 500-lb) drop hammers and a Delmag D-5 diesel hammer. A 0.762-m (2.5-ft) stroke was used for the drop hammers. The 2.22-kN (500-lb) drop hammer and the diesel hammer were included because they represent the most difficult possible driving conditions. It is more likely, however, that the 0.62- and 1.33-kN (140- and 300-lb) hammers will actually be used.

The drop hammer driving system used in the wave equation analyses is illustrated in Figure 21. Appendix B presents the details of the dynamic analysis related to this system. Since the MDMP will be advanced using conventional wash and drive drilling methods, N-rods with a 60.325-mm (2.375-in) O.D. and 4.763-mm (3/16-in) wall thickness were modeled to connect with the MDMP. No hammer cushion is shown, which is typically for the standard penetration test (SPT). The MDMP was assumed to be 4.57 m (15 ft) long and 76.2 mm (3 in) in diameter (O.D.), with a 9.525-mm (3/8-in) wall thickness. The tip configuration was assumed to be closed-ended.

Table 3. MDMP Static Load Resistance in Dense Sand (Upper Limiting Case).

Depth @ Tip (m)

Skin Resistance (kN)

Point Resistance (kN)

Total Resistance (kN)

12.2

127.2

59.6

186.8

15.2

159.2

59.6

218.8

18.3

190.4

68.5

258.9

21.3

221.5

78.3

299.8

24.4

253.5

88.1

341.6

27.4

284.7

97.9

382.5

30.5

315.8

107.6

423.4

33.5

346.9

118.3

465.3

Figure 21. View Alternative Text

Figure 21. Drop Hammer Configuration Modeled in the Wave Equation Analyses.

The minimum and maximum driving stresses that the N-rods and the MDMP are expected to be subjected to were determined based on the static soil resistances presented in section 3.3.2. The lower soil resistance of 0.44 kN (0.1 kips) was used to represent easy driving conditions in soft clay, and the higher soil resistance, which varied between 44.5 and 445 kN (10 and 100 kips), was used to represent hard driving conditions in dense sand. Table 4 summarizes the maximum and minimum driving stresses in the N rods and the MDMP for each hammer type. These stresses were evaluated using three penetration lengths for each of the soil conditions. Since the MDMP will be installed at the bottom of a cased hole, soil resistance was modeled only along the shaft and tip of the model pile. The maximum static soil resistance that could be overcome during driving varied with the hammer size and pile length. The static soil resistance increased as the hammer size increased (i.e., increase in energy) and/or the pile length decreased. Static soil resistances developed for blow counts greater than 79 blows per 10 cm (240 blows per foot) were considered unrealistic and were not included.

Table 4. Dynamic Loads and Accelerations in the MDMP During Easy or Hard Driving.

Hammer Type

Pile Penetration (meter)

Relative Driving Resist.

Total Static Soil Resist. (kN)

Blow Count (blows/10cm)

Max. Compr. Stress (MPa) N Rods/MDMP

Max. Tensile Stress (MPa) N Rods/MDMP

Max. MDMP Acceleration (g's)

0.623 kN Drop
0.762 m
Stroke

7.6

Easy

0.4

1.0

175.1/72.4

84.1/46.9

1200

7.6

Hard

53.4*

39*

175.8*/72.4*

53.8*/44.1*

1100

38.1

Easy

0.4

2.0

174.4/71.7

95.8/40.0

1000

38.1

Hard

44.5

43.0

175.1/72.4

68.3/31.0

900

114.3

Easy

0.4

2.0

167.5/69.0

89.6/33.8

500

114.3

Hard

44.5

43.3

168.2/69.6

57.2/23.4

500

1.334 kN
Drop
0.762 m
Stroke

7.6

Easy

0.4

0.7

224.1/92.4

95.8/58.6

1600

7.6

Hard

137.9

52.8

224.8/92.4

44.8/23.4

1400

38.1

Easy

0.4

1.3

225.5/93.1

106.9/53.1

1300

38.1

Hard

120.1

>79

226.8/93.8

104.8/42.7

1100

114.3

Easy

0.4

2.0

211.7/87.6

102.0/42.7

700

114.3

Hard

120.1

>79

214.4/88.9

46.2/15.9

500

2.224 kN
Drop
0.762 m
Stroke

7.6

Easy

0.4

0.7

243.4/100.0

145.5/81.4

1900

7.6

Hard

222.4

>79

318.5/135.8

30.3/17.2

1600

38.1

Easy

0.4

1.0

244.8/100.7

91.7/57.2

1400

38.1

Hard

124.5

40.0

246.2/101.4

125.5/51.7

1200

114.3

Easy

0.4

1.6

232.4/95.8

112.4/46.9

800

114.3

Hard

124.5

42.7

235.8/102.0

51.0/13.8

600

D-5
Diesel

7.6

Easy

44.5

1.0

151.0/57.2

0/0

100

7.6

Hard

444.8

23.6

653.6/268.2

37.9/15.9

200

38.1

Easy

44.5

1.0

153.8/47.6

0/0

100

38.1

Hard

231.3

44.0

319.9/131.7

0/0

100

* Interpolated from the dynamic analyses graphical results.

The allowable driving stresses (tensile and compressive) for steel are 0.9f'y, which equals 223.4 MPa (32.4 ksi) based on grade 36 steel. As expected, the worst compressive stresses (653.6 MPa) and tensile stresses (145.5 MPa) were created by the D-5 diesel and 2.22-kN (500-lb) drop hammers, respectively. These stresses were calculated in the drill rods at the connection with the MDMP, where the change in cross-section (increased impedance) creates larger compressive stresses as a result of reflections in the stress wave. Based on the 2.22-kN (500-lb) hammer and, especially, the D-5 hammer simulations, the smaller cross-sectional area of the N-rods is expected to be damaged before the MDMP. If the larger hammers are used, the cross-sectional area of the drill rods may have to be increased to a larger size in order to accommodate the higher driving stresses. Alternatively, the use of cushions and a reduction of the stroke can be employed.

Discounting the D-5 diesel hammer driving simulation, the maximum compressive stress that occurred in the MDMP was 135.8 MPa (19.7 ksi) using the 2.22-kN (500-lb) drop hammer. The maximum tensile stress that occurred in the MDMP was 81.4 MPa (11.8 ksi) using the 2.22-kN (500-lb) drop hammer. The compressive stresses typically occurred at the connection of the MDMP with the drill rods (at the top of the MDMP). In general, the maximum tensile stresses were encountered toward the middle of the MDMP, although under the harder driving conditions and longer pile lengths, they were observed at the top of the MDMP. It is highly unlikely that the D-5 diesel hammer will be used (especially together with N-rods) and, therefore, excluding the D-5 simulation results, all driving stresses in the MDMP remain within the allowable stress level.

The accelerations presented in Table 4 were determined using 152.4-mm (6-in) segment lengths, except for the 114.3-m (375-ft) pile length, which was modeled using 304.8-mm (12-in) segment lengths. Based on Rausche (1995), smaller increments are required to properly analyze SPT driving systems. By increasing the number of pile segments, the stress wave can be more clearly defined as it propagates down the pile. This is especially important for uncushioned steel on steel impacts, where the impact stress signal is a high peak of short duration.

The maximum range of accelerations expected in the MDMP vary between 1500 g's and 2000 g's for shorter pile lengths (7.6 m) driven with the 1.334- and 2.224-kN (300- and 500-lb) drop hammers. The lower accelerations may be around 500 g's for the longer pile lengths (114.3 m) driven with the 0.623-kN (140-lb) drop hammer. These accelerations are approximately two to five times higher than accelerations observed during the driving of full-scale piles. Lower accelerations (100 g's to 200 g's) are obtained for the diesel hammer analysis due to the different mode in which the hammer impacts the pile.

3.3.4 Summary of Load Requirements

Table 5 summarizes the maximum loads in the MDMP obtained from the static and dynamic analyses. The indicated loads are based on a cross-sectional area of 19.94 cm2 (3.09 in2). Both analyses are based on conservative assumptions to ensure that the upper limits have been identified. The different values of static capacity, 89 kN to 463 kN (20 kips to 100 kips), suggest that two separate load cell systems may be required in the MDMP so that accurate measurements can be obtained under the two soil conditions. The dynamic capacity was determined from the product of the stress and the area of the MDMP. Based on the 1.334-kN (300-lb) drop hammer in the soft to medium clay, the dynamic capacity in tension and compression are 117 kN and 186 kN (24 kips and 51 kips), respectively. The dynamic load cell requirements for the soft to medium clays, therefore, can be rounded to those indicated under "Design Requirements for MDMP " in Table 5. The dynamic capacity values for soft to medium clay (162 kN tension and 201 kN compression) are within the load cell overload range of 2.5 times the static capacity. The dynamic design requirements for the dense sand (225 kN tension and 550 kN compression) are well within the 250% overload range, even with the use of the larger hammers.

Table 5. Summary of Load Cell Capacity Requirements.

Soil Condition

Static Capacity (kN)

Dynamic Capacity

Design Requirements for MDMP

Maximum Tensile (kN)

Maximum Compressive (kN)

Static kN)

Dynamic Tensile/Compressive (kN)

Soft to Medium Clay

89

162

201

89

225 / 225

Dense Sand/Hard Clay

463

103

535

445

225 / 550

3.4 Specifications for Instrumentation and Mechanical Parts

3.4.1 General

The MDMP was based on a modification of the 7.62-cm (3-in) Model Pile originally developed by Bogess et al., 1983. The specifications for the load cells, accelerometers, pore pressure transducer, total pressure cell, connector housing, slip joint, tip segment, and loading frame are described below. The description of these components and other MDMP components is provided in section 3.5.

3.4.2 Accelerometers

Accelerometers are mounted in the model pile to provide accurate records of pile acceleration that are independent of drill rods and hammer. The accelerometers must be capable of measuring accelerations up to 2000 g's. The accelerometers are installed at the load cell locations in the interior of the model pile and are securely attached to the pile. The accelerometers will be monitored using the PDA (Pile-Driving Analyzer). The PDA is capable of monitoring up to four accelerometers (two piezoelectric and two piezoresistive). As such, two of the MDMP accelerometers (top and bottom load cells) will be piezoelectric and one will be piezoresistive (middle load cell), with the fourth accelerometer being mounted at the top of the drill rods.

3.4.3 Load Cells

(a) Top and Middle Load Cells

Two load cells are required to measure the friction along a section of the pile. Since a variety of soil profiles will be tested, two sets of load cells are required to measure the skin friction. Refer to section 3.3.4 for the required load ranges in the soft to medium BBC clay and dense sand. There is 2000 cm2 (310 in2) of surface area between the two load cells to measure skin friction and load transfer.

(b) Tip Load Cell

The load cell at the tip was included to measure end-bearing capacity during compression tests and evaluate the friction along the lower pile segment. The tip load cell may be used to correlate MDMP results to the more conventional CPT. The load at the tip also provides a performance check of the slip joint, ensuring that no load is being transferred through the slip joint during tension load tests. During compression load testing, the tip load cell must be able to measure the tip resistance when the slip joint is fully compressed. During installation, the tip load cell will continuously measure the tip resistance.

3.4.4 Pore Pressure Transducer

The pore pressure transducer must be able to measure pore water pressure during initial driving, restrike, and static load tests. The pore pressure transducer must to be able to measure the excess pore pressure dissipation continuously for several days after driving. When driving in dense overconsolidated clays or silts, the pore pressure transducer must also be able to measure negative excess pore pressures. Most importantly, the pressure transducer must physically be able to withstand stresses during driving. The increase in pore pressure due to driving can be expected to be 2.29 times the vertical effective stress (Paikowsky et al., 1995). Based on the typical soil profile in the Boston area (maximum test depth at 33.5 m (110 ft)), the predicted pore pressure immediately after driving is 1070 kPa (155 psi).

Two porous filters must be located 180° apart and mounted flush to the pile wall, maintaining the radius of the MDMP so that no local discontinuities are present. The porous filters must be permeable to enable quick response to pressure changes, but must also be of sufficiently low permeability to maintain saturation while the model pile is driven through unsaturated material. In addition, the porous filters must be durable for use in hard driving conditions, easily replaceable, and compatible with other materials utilized in the model pile to prevent corrosive effects. Lastly, a method to saturate and de-air the porous filters is required.

3.4.5 Total Pressure Cell

The total pressure cell must be able to measure total pressure during the entire duration of the MDMP test sequence. The total pressure cell, like the pore pressure transducer, must be able to physically withstand the stresses during driving. The loading caps of the total pressure cell must also maintain the radius of the MDMP so that no local discontinuities are present. Considering the typical soil profile in the Boston area (maximum test depth at 33.5 m (110 ft)), the predicted total pressure is 1214 kPa (176 psi) (assuming sh after driving is two times svo).

3.4.6 Connector Housing

The connector housing gathers all the wires from up to 10 MDMP sensors and connects them to a main cable that extends through the drill rods to the surface. The connector housing must also be watertight to prevent water from entering the MDMP. The cable needs to be at least 45 m (150 ft) long.

3.4.7 Slip Joint

The slip joint is used to measure local displacements. The slip joint needs to be extended immediately after driving to allow measurements of displacement during subsequent compression load tests. The slip joint needs to be able to measure a total displacement of up to 5 cm (2 in), allowing for four load tests to be performed with 1.25 cm (0.5 in) of displacement for each test. The displacement of 1.25 cm (0.5 in) is assumed to be enough to accommodate the soil quake (approximately 0.1 in along the shaft) and provide adequate information of the post-peak and residual soil resistance. Wires for the bottom load cell below the slip joint must be able to pass through the slip joint without affecting the slip joint or bottom load cell.

3.4.8 Loading Frame

Following installation, several tension and compression static load tests need to be performed over time. The static load frame needs to be attached to the drill rods, which are connected to the MDMP. The frame must be easy to assemble and position over the drill rod string and must be capable of conducting both extension and compression load tests of up to 445 kN (50 tons). The load application system must be capable of performing the tests by displacement or load control techniques. The loading system also requires sufficient throw to be connected to the drill rods and to conduct several 1.25-cm (0.5-in) load tests in succession.

3.4.9 Summary of the Instrumentation Range Requirements

Table 6 presents a summary of the required instrumentation as outlined in section 3.4.

Table 6. Summary of the MDMP Required Instrumentation Ranges.

Component

Location

Type

Installation Condition

Range

Accelerometers

Top of Rods

Piezoresistive

N/A

0-2000 g's

Top Load Cell

Piezoelectric

N/A

0-2000 g's

Middle Load Cell

Piezoresistive

N/A

0-2000 g's

Bottom Load Cell

Piezoelectric

N/A

0-2000 g's

Load Cells

Top, Middle,

Electric

Soft/Medium Soil

89 kN

and Bottom

Strain

Hard/Stiff Soil

445 kN

Load Cells

Gauges

Dynamic

2.5 times static cap.

Pore Pressure Transducer

Transducer Housing

Electric Strain Gauges

All

1070 kPa

Total Pressure Cell

Transducer Housing

Electric Strain Gauges

All

1214 kPa

Slip Joint

Slip Joint

DC-LVDT

All

5 cm

Loading Frame

N/A

N/A

All

445 kN

3.5 Design

3.5.1 General

The MDMP is composed of several components that are screwed together (Figure 17). All of the major components are made of stainless steel to inhibit oxidation and other possible chemical reactions. Rubber O-rings are used to seal all components in order to create a watertight environment in the interior of the pile. The outside diameter of all components remains constant (76.2 mm) throughout the entire MDMP length, except at the lower slip joint, where it is 57.15 mm (2.25 in). The overall length of the closed-ended MDMP with tip extension is 2.87 m (9.42 ft). The major components of the MDMP (referring to Figure 17) include: N-rod adapter, connector housing, upper extension, load cells, couplings, transducer housing, slip joint, lower extension, and interchangeable tip segment. Table 7 lists the different components, their description, length (when applicable to the total length), material, and related detail. The following sections provide details of the different components presented in Figure 17 and listed in Table 7. Appendix C presents the shop drawings and details of the individual components.

3.5.2 N-Rod Adapter

The N-Rod adapter is 21.59 cm (8.5 in) long and attaches the MDMP (76.2 mm diameter) to N drill rods (60.3 mm diameter) that are used to advance the model pile to the desired test depth. Type 316 stainless steel is used to machine the component. The end that is connected to the drill rod is a female modified Box thread (three threads per inch). The opposite end is a female 2.50-5 Stud ACME thread that is attached to the connector housing.

3.5.3 Connector Housing and Mount

The connector housing is 10.24 cm (4.03 in) long and is machined from Type 316 stainless steel. The end that connects to the N-rod adapter is a male 2.50-5 Stud ACME thread. The opposite end is a female 2.500-12-2 thread. Six slots that are evenly spaced about the circumference allow water to enter and/or drain from the drill rods.

The connector mount is made of 17-4PH @H1050 stainless steel. The connector mount has two functions: (1) providing a watertight seal to protect the instrumentation within the MDMP from water intrusion and (2) providing a waterproof cable connection enabling the instrumentation wiring within the MDMP to be connected to the data acquisition cable. A set screw is used to attach the connector mount to the upper extension to ensure that the mount does not rotate and the sensor wires do not shear.

3.5.4 Upper Extension

The upper extension is 31.27 cm (12.31 in) long and ensures that the instrumentation is in a zone of radial dissipation and is an adequate distance below the drill rods so that the sensors are not affected by the change in cross-sectional area from the rods to the pile. The lead wires from the various MDMP sensors are gathered together and combined in the interior of the upper extension. The component has a male and female 2.500-12-2 thread on either end and is machined from Type 316 stainless steel.

Table 7. MDMP Component List.

Component

Part

Length

(cm)

Material

See Figure

N-Rod Adapter

N-Rod Adapter

21.59

Type 316 Stainless Steel

Appendix C

Connector Housing

Connector Housing

10.24

Type 316 Stainless Steel

Appendix C

Connector Mount

N/A

Type 17-4PH @H1050 Stainless Steel

Appendix C

Upper Extension

Upper Extension

31.27

Type 316 Stainless Steel

Appendix C

Load Cell

Load Cell

14.60

assembled

Type 17-4PH @H1050 Stainless Steel

Appendix C

Load Cell Cover

N/A

Type 316 Stainless Steel

Appendix C

Accelerometer Mount

N/A

Type 6160 Aluminum

Appendix C

Coupling

Coupling

25.74

Type 316 Stainless Steel

Appendix C

Transducer Housing

Transducer Housing

5.99

Type 17-4PH @H1050 Stainless Steel

Appendix C

Retainer

N/A

Brass

Appendix C

Slip Joint

Upper Slip Joint

20.40

Type 17-4PH @H1050 Stainless Steel

Appendix C

Lower Slip Joint

10.21

Type 17-4PH @H1050 Stainless Steel

Appendix C

LVDT Mount

N/A

Type 17-4PH @H1050 Stainless Steel

Appendix C

LVDT Pad

N/A

Type 17-4PH @H1050 Stainless Steel

Appendix C

Lower Extension

Lower Extension Adapter, Male

82.04

assembled

Cold-Rolled Type 1018 Round

Appendix C

Lower Extension Adapter, Female

N/A

Cold-Rolled Type 1018 Round

Appendix C

Lower Extension Tubing

N/A

Mechanical Tubing

Appendix C

Open-Ended Cutting Shoe

varies

varies

varies

Appendix C

60° Tip

60° Tip

10.08

Type 316 Stainless Steel

Appendix C

N/A - not applicable

3.5.5 Load Cells

Three load cells are installed in the MDMP. Figure 22 is a photograph of a load cell with a sleeve. Each load cell utilizes four foil strain gauges arranged in a full Wheatstone bridge formation. The total nominal bridge resistance is 350 ohms. This formation of strain gauges cancels the effects of bending and measures only axial loads. The strain gauges are attached to the center section of the load cell, which has a uniform cross-section of either 4.75 cm2 (0.7363 in2) or 13.06 cm2 (2.0249 in2) for the 89- and 445-kN (10- and 50-ton) load cells, respectively. The dynamic to static overload ratio (dynamic loading/static loading) for these load cells is 2.5, which is sufficient to accommodate the maximum anticipated driving stresses. The load cell is made from Type 17-4PH @H1050 stainless steel. There are two O-rings that maintain a watertight seal between the adjacent modular parts. Four holes allow air to flow to both sides of the uniform cross-section, thereby avoiding temperature and pressure variations that might affect the sensitivity of the load cells. Both ends are male ends with 2.500-12-2 threads. At one end, the inside circumference is threaded with a 1.820-20-2 thread to attach the LVDT mount.

Figure 22. View Alternative Text

Figure 22. Photograph of the MDMP Load Cell With Sleeve.

Each load cell is protected by a load cell cover. The load cell cover slides over the load cell and provides a watertight seal using three additional O-rings. When screwing the adjacent components to the load cell, they exert compression on the flange of the load cell cover, holding it in place. The load cell cover is 13.03 cm (5.130 in) long, with an outside diameter of 7.62 cm (3 in). There is a small gap at one end of the load cell cover to allow the load cell to strain. The load cell cover is made from Type 316 stainless steel. The assembled load cell with cover has a combined length of 14.60 cm (5.75 in).

Each load cell is fitted with an accelerometer, which is mounted in the interior of the load cell. The maximum accelerations (in terms of gravity "g") that the accelerometers will be subjected to range between approximately 500 g's to 2,000 g's. To maintain compatibility with the Pile-Driving Analyzer (to be used for monitoring the accelerometers), two types of accelerometers (piezoelectric and piezoresistive) are employed. A piezoresistive accelerometer was selected for the middle load cell due to space constraints dictated by the DC-LVDT. The upper and lower MDMP load cell accelerometers are of the piezoelectric type.

3.5.6 Couplings

There are two couplings in the MDMP. These couplings are used to center the transducer housing between the two load cells and provide a known surface area for measuring the skin resistance along the friction sleeve. Both ends of each coupling are female 2.500-12-2 threads. Type 316 stainless steel is used to form the couplings, which are 25.74 cm (10.134 in) long and have a cross-sectional area of 22.83 cm2 (3.54 in2).

3.5.7 Transducer Housing

The total stress cell and pore pressure transducer are located in the transducer housing. Figure 23 is a photograph of the transducer housing with pore pressure transducer and total radial stress cell. The transducer housing has an outside diameter of 7.62 cm (3 in) and is machined from Type 17-4PH @H1050 stainless steel. Both ends are male 2.500-12-2 threads. Two O-rings at either end form a watertight seal with adjacent components. Two holes 2.54 cm (1 in) in diameter and 0.953 cm (0.375 in) deep are aligned 180º apart. The holes are fitted with porous aluminum oxide stones. Behind each stone is a duct that allows the free flow of pore water from the porous stones to the pore pressure transducer located in the center of the transducer housing. A 2.54-cm (1-in) through-hole, located 90º from the porous stones, houses the total stress transducer.

Figure 23. View Alternative Text

Figure 23. Photograph of the Transducer Housing With the Pore Pressure Transducer and the Total Radial Stress Cell.

The porous stones for the pore pressure transducer are pressed into place. The outer surface is shaped to the same curvature as the model pile so that no discontinuities will exist around the stone. The pore water flows through the pores in the stone to the transducer so that only fluid pressures are recorded. Since below-freezing temperatures were anticipated during field testing, a mixture of glycerin and water was examined for use in the transducer housing water ducts and stones. Following laboratory tests, a solution consisting of 30% glycerin and 70% water was used to saturate the stones. This solution would not freeze up to a temperature of approximately -7°C (20°F) and did not appear to separate when the temperature increased back to room temperature. An added benefit to the use of a glycerin solution is that due to the increased viscosity of the solution, the porous stones will remain saturated even if exposed to air for a limited time. A Kistler Model 4140A20 pressure transducer is used to measure the pore fluid pressure. The nominal transducer output is 29.99 mV/bar using an excitation current of 4 mA and its range is 0 to 2000 kPa (0 to 290 psi). A Kistler Model 4670V signal conditioner was installed to supply the current excitation and amplify the output. This signal conditioner was incorporated into the connection box (to be described in Chapter 4).

The total stress transducer principle of measurement is similar to the load cells. Four foil strain gauges are mounted to a "dog bone"-shaped piece of aluminum (Type 2024-T4), the ends of which have a circular cross-section. The foil strain gauges are arranged in a 350-ohm full Wheatstone bridge formation to measure only axial load. Two end plugs fit over the circular ends of the aluminum "dog bone" and are fitted with O-rings to ensure watertightness. The outer surfaces of the end plugs have the same curvature as the model pile so that they are flush with the pile wall. The nominal transducer output is 0.35 mV/bar using an excitation voltage of 10 V.

3.5.8 Slip Joint

The slip joint consists of upper and lower components and is 30.61 cm (12.05 in) long when fully compressed. Figure 24 is a photograph of the components of the slip joint. The upper slip joint (20.40 cm long) is made of Type 17-4PH @1050 stainless steel with a female 2.500-12-2 threaded end. Four holes 90° apart with counterbores are used to insert guides that slide in the slots on the lower slip joint. The lower slip joint (10.21 cm long) is also made of Type 17-4PH @1050 stainless steel. The lower slip joint has four slots that the guides from the upper slip joint slide in. The lower end of the lower slip joint is threaded with 2.500-12-2 female thread as well.

The DC-LVDT that measures the local displacement of the slip joint is held in place with the LVDT mount. The LVDT mount screws into the middle load cell and a set screw securely holds the LVDT in place, providing a reference point for the top portion of the model pile. The LVDT pad is attached to the lower slip joint utilizing two #8-32 screws. The LVDT pad provides the reference point for the lower slip joint.

The DC-LVDT is manufactured by Macro Sensors and is model no. GHSD 7500-1000. The transducer measures up to 4.9 cm (2 in) of movement. The excitation voltage required is ±15 V DC and the output range is ±10 V DC. Specifications indicate the transducer has a shock survival of 1000 g's over 11 ms.

Figure 24. View Alternative Text

Figure 24. Photograph of he Slip Joint With DCDT.

3.5.9 Lower Extension

The lower extension connects the lower slip joint to the lower load cell for "long" piles. The lower slip joint screws directly into the bottom load cell for "shorter" piles. Male and female adapters are used to attach the lower extension to the lower slip joint and lower load cell. The length of the lower extension and adapters is 82.04 cm (32.3 in). This length is based on the following two criteria: (1) the additional length required to reduce end effects near the slip joint and (2) the provision of a larger frictional area for the "anchored" portion of the pile below the slip joint. The lower extension consists of male and female 2.500-12-2 threaded end pieces that are welded to a 6.35-mm- (¼-in) thick mechanical tubing.

3.5.10 Tip Segment

An interchangeable tip segment 10.2 cm (4.0 in) long for a conical tip screws into the lower load cell. Various tip attachments can be fabricated so that different driving modes (such as open-ended and closed-ended penetration) could be investigated. Figure 18 presents three different possible extensions for the MDMP based on the following tip segment configurations:

  • "Short" closed-ended pile, where the tip segment (angled or flat tip) screws directly into the lower load cell.

  • Open-ended pile, where a threaded open-tip segment cutting shoe screws directly into the slip joint.

  • "Long" closed-ended pile, where an 82.0-cm (32.3-in) extension connects the lower slip joint to the lower load cell. The tip segment (angled or flat tip) screws directly into the lower load cell.

3.6 Calibration

3.6.1 Overview

The components of the MDMP were tested and calibrated before and after the testing program. To ensure that system errors did not affect the calibration process, the MDMP was completely assembled utilizing the required cables, connection box (section 4.3), and data acquisition system (section 4.2). The three load cells, pore pressure cell, total pressure cell, and slip joint DC-LVDT were all calibrated in the Geotechnical Laboratory at UMass-Lowell. The calibration process included testing the load cells, the pore pressure transducer, total pressure cell, and slip joint DC-LVDT under static loading conditions. The accelerometers were tested to ensure that they were wired correctly and the drill rods were examined to ensure that the discontinuities at the joints would not inhibit the measurement of dynamic response during driving. A 222.4-kN (25-ton) load cell and two DC-LVDTs were tested in the laboratory to be used in the static testing program. Additional instrumentation, including a strain gauge and accelerometer attached to the drill rod, was also checked in the laboratory.

3.6.2 Load Cell Calibration

The three MDMP load cells that were used in the Newbury, MA test program were calibrated five times. Refer to Appendix D for calibration plots describing the relationship between output voltage and load. The initial factory calibrations were performed by Technology & Calibration, Inc. (TechCal) before the assembly of the MDMP. Each load cell was calibrated under compression loading in 11.12-kN (2,500-lb) increments to a maximum compressive load of 111.2 kN (25,000 lb). During the calibration process by TechCal, the outer sleeve was not in position over the load cell. At the conclusion of the Newbury testing program, the load cell calibrations were rechecked at the UMass-Lowell Geotechnical Laboratory with the pile disassembled and the outer sleeves removed.

Before and after the Newbury testing program, the load cells were also calibrated at the UMass-Lowell Geotechnical Laboratory with the model pile fully assembled (the bottom load cell was not recalibrated after the testing program). The procedure for calibrating the load cells when the model pile was assembled consisted of placing the MDMP in a reaction frame and applying a compressive load with a hydraulic jack. Figures 25a and b are a schematic and a photograph of the system used for the load cell calibration. The reaction frame was constructed of two vertical W sections, 152.4 mm in depth (W6x15), and one horizontally oriented W section with the web aligned vertically, 152.4 mm in depth (W6x15), with steel members with reinforced welded joints. Appendix E provides details about the frame analysis and construction. A 222.4-kN (50,000-lb) Lebow load cell was placed in line with the model pile to record the applied compressive load. A ball connection was placed between the jack and the bottom of the model pile to eliminate moments during calibration (see Figure 25). The jacking system was comprised of an hydraulic ram and a hand-operated pump. A compressive load of 53.4 to 57.8 kN (12,000 to 13,000 lb) was applied gradually at a near constant rate and was then allowed to return instantaneously to zero. This loading procedure was repeated four times and the data were used to develop the calibration factors.

The calibrations for each of the three MDMP load cells at the different times are listed in Tables 8 through 10. The obtained calibration factors suggest that the calibration results vary depending on the timing of the calibration process with respect to testing. This can be explained in a number of ways. The factory calibration was done before the load cells were exercised and the outer sleeve was not in place. In a perfect design, the O-rings used to develop a watertight seal around the load cell instrumentation would not transfer any load. Practically, however, the O-rings are compressed to form the watertight seal and some load transfer does occur. Another reason for different factory calibrations and assembled calibrations is the voltage drop that occurs through the 60-m (200-ft) length of cable and connections.

The recalibration of the load cells at the end of the testing program was not completely linear, as the curve appeared to be bi-linear with a change in slope at 20.91 kN (4,700 lb) (Appendix D). This bi-linear behavior may be due to dried soil in the gap that allows for the load cell and O-ring expansion. The difference between the two calibrations with the outer sleeve removed (Appendix D, Factory Calibration and Recalibration) may be due to a physical change in the load cell, possibly caused by the dynamic forces or residual forces that the pile was subjected to during installation and removal.

Table 8. Top Load Cell Calibration Results.

Test

Calibration Equation: (V sub out divided by V sub in) divided by lb

Coefficient of Determination

Remarks

Factory Calibration

5.0720x10-8

0.99752

Without outer sleeve

Calibration Before Test

5.1127x10-8

0.99969

Model Pile Assembled

Recalibration After Test Trial #1

6.1317x10-8

0.98933

Model Pile Assembled 0-53.4 kN

Recalibration After Test Trial #1

4.9085x10-8

0.9941

Model Pile Assembled 0-20.9 kN

Recalibration After Test Trial #2

6.4326x10-8

0.98800

Model Pile Assembled 0-53.4 kN

Recalibration After Test Trial #2

5.4595x10-8

0.97649

Model Pile Assembled 0-20.9 kN

Recalibration After Test

5.332x10-8

0.9988

Without outer sleeve

 



Figure 25a. View Alternative Text

Figure 25a. Schematic of the Calibration Frame for the MDMP.

Figure 25b. View Alternative Text

Figure 25b. Photograph of the Calibration Frame for the MDMP.

Table 9. Middle Load Cell Calibration Results.

Test

Calibration Equation: (V sub out divided by V sub in) divided by lb

Coefficient of Determination

Remarks

Factory Calibration

6.3654x10-8

0.99999

Without outer sleeve

Calibration Before Test

6.0946x10-8

0.99998

Model Pile Assembled

Recalibration After Test Trial #1

7.9535x10-8

0.98848

Model Pile Assembled 0-53.4 kN

Recalibration After Test Trial #1

6.2198x10-8

0.99563

Model Pile Assembled 0-20.9 kN

Recalibration After Test Trial #2

7.5772x10-8

0.98476

Model Pile Assembled 0-53.4 kN

Recalibration After Test Trial #2

5.9933x10-8

0.98349

Model Pile Assembled 0-20.9 kN

Recalibration After Test

6.7737x10-8

0.99998

Without outer sleeve



Table 10. Bottom Load Cell Calibration Results.

Test

Calibration Equation: (V sub out divided by V sub in) divided by lb

Coefficient of Determination

Remarks

Factory Calibration

6.2911x10-8

0.99996

Without outer sleeve

Calibration Before Test

5.9639x10-8

0.99998

Model Pile Assembled

Since the three MDMP load cells will be monitored during driving using the Pile-Driving Analyzer (PDA), a calibration factor for the PDA was determined. The PDA calibration factor is a function of the static calibration. A multiplication factor of 0.288 is used to transform the static calibration for use as a PDA calibration factor. This multiplication factor, also called a Pile Dynamics Inc. (PDI) factor, is based on excitation voltage and the internal circuitry of the PDA (based on correspondence with Pile Dynamics, Inc.). The modulus of elasticity used for stainless steel is 2.05 x 105 MPa (29.7 x 106 psi) as provided by the manufacturer of the load cells. Based on the above, the PDA calibration factor is related to the static calibration factor by the following relationship and is summarized in Table 11:

There are two equations listed. In the first equation the General Factor equals a quotient with the strain multiplied by the volts-in in the numerator and millivolts out in the denominator. This is then shown to be equal to the product of two quotients. The first with pounds multiplied by microvolts-in in the numerator and millivolts-out in the denominator and the second with 1 in the numerator and the cross sectional area of the pile multiplied by the modulus of elasticity in the denominator.


Table 11. Dynamic Calibration Results of the MDMP Load Cells.

Load Cell

Serial Number

Channel on PDA

Static Calibration Static Calibration column: Equation: (volts-out divided by the product of lb and volts-in)

General Factor General Factor column: Equation: (product of microstrain and volts-in divided by millivolts-out)

PDA Calibration Factor
PDA Calibration column: Equation: (microstrain divided by volts)

Top

D10T-02

F1

5.1127x10-8

894.4

257.59

Middle

D10T-01

F3

6.0946x10-8

750.3

216.08

Bottom

D10T-03

F2

5.9639x10-8

766.8

220.83

3.6.3 Pore Pressure Transducer Calibration

The pore pressure cell is composed of a pressure transducer connected to the porous stones via ducts. Figure 26 is a schematic of the pressure instrument calibration layout. A custom-made cylindrical calibration chamber was placed over the pore pressure cell. The cables, DAS, and all other devices that were used in the field testing program were also used in the calibration process. The chamber utilizes O-rings to form a seal that maintains a vacuum during the de-airing process and withstands pressurization during the calibrating process. The porous stones and internal channels were filled with glycerin and de-aired water mixture. A vacuum was applied to the chamber to ensure de-airing and complete saturation. A proportional integration differentiation (PID) circuit was then used to apply pressure to the fluid in the calibration chamber. An additional accurately calibrated pressure transducer was placed in line between the PID and the pressure chamber to measure the actual pressure applied to the fluid in the chamber. This additional pressure transducer was used to record the reference pressure on which the calibration factors were based.

In order to simulate field conditions, three different pressure application procedures were used. The pressure was increased and decreased at a constant rate (for four or five cycles) during all the procedures. These procedures were:

  1. Opening the system to the atmosphere and holding for a short duration after each ramp up and down sequence.

  2. Opening the system to the atmosphere before and after the entire ramp up and down sequence without any pauses between ramps.

  3. Holding the pressure steady after each ramp up and opening the system to the atmosphere after each ramp down.

Procedures 1 and 2 both simulate driving conditions because the pressure application is rapid and the transducer can measure the pressure only if the response time is quick. Procedure 3 represents the period after driving when the pressure changes are slower and the response time of the system is not as important a factor. The different pressure application procedures did not affect the performance of the pore pressure transducer.


Figure 26. View Alternative Text

Figure 26. Pressure Instrumentation Calibration Setup.

The pore pressure cell was calibrated three times before the testing program at Newbury using each pressure application procedure. After the completion of the testing program, the calibration of the pore pressure cell was checked twice using the third pressure application procedure. Table 12 summarizes the pore pressure transducer calibration results of the various calibration procedures. An average of all five calibration factors were used in the test result data reduction (7.0652 PSI over Volt).

3.6.4 Total Pressure Cell Calibration

The total pressure cell was calibrated along with the pore pressure cell using the same pressure application procedures previously outlined in section 3.6.3. The pressure application procedure was an important factor when calibrating the total pressure cell as the response time of the cell was affected by the O-rings. As a result, procedures 1 and 2 were not adequate as not enough response time was provided after each loading ramp. Table 13 summarizes the results of the various calibrations of the total pressure cell. The calibration factor for test #3 (64527.16PSI over Volt) was used in the data reduction. The total pressure cell was damaged during the testing program and the calibration could not be checked after testing.

Table 12. Pore Pressure Transducer Calibration Results.

Test

Procedure

Calibration PSI over Volt

Zero Pressure Voltage (volts)

Coefficient of Determination

Remarks

Test #1

#1

7.0209

1.5120

0.99998

Before Testing Program

Test #2

#2

7.0687

1.4786

0.99999

Before Testing Program

Test #3

#3

7.0838

1.5218

0.99999

Before Testing Program

Test #4

#3

7.0849

1.5623

0.99997

After Testing Program

Test #5

#3

7.0678

1.5510

0.99998

After Testing Program

Table 13. Total Pressure Cell Calibration Results.

Test

Procedure

Calibration PSI over Volt

Zero Pressure Voltage (volts)

Coefficient of Determination

Remarks

Test #1

#1

67114.9912

0.0004323

0.9997

Before Testing Program

Test #2

#2

65261.119

0.0004187

0.9972

Before Testing Program

Test #3

#3

64527.1586

0.0005621

0.9999

Before Testing Program

3.6.5 Displacement Transducer Calibration

The assembled MDMP was suspended from the reaction frame using the same setup used to calibrate the load cells. Two displacement transducers were aligned 180° apart to measure the movement of the hydraulic ram. With the slip joint in the fully extended position, the ram was advanced approximately 5 cm to close the slip joint. The displacements monitored by the two displacement transducers (that measured the ram movement) were averaged together to determine the movement of the slip joint. The calibration constant determined for the DC-LVDT displacement transducer obtained via this procedure is 0.097735 in/Vout.

3.6.6 Accelerometer Response During Dynamic Loading

The MDMP was not calibrated under dynamic loading. The actual calibration of the accelerometers was performed by Pile Dynamics, Inc., of Cleveland, Ohio. The response of the MDMP instrumentation under dynamic loading was examined using a custom-designed support system that was fabricated for this purpose by George Saliby, a graduate research assistant at UMass-Lowell. Figures 27a and b are a schematic and photograph of the configuration used for the support of the MDMP and associated loading equipment. Four supports were constructed out of steel wiring and circular clamps. These supports were fixed to a ceiling beam.

The MDMP was oriented horizontally and placed within two of the supports in order to simulate a completely free pile (i.e., without frictional or end-bearing resistance). The responses of the three accelerometers and load cells within the MDMP were tested in this system. The two remaining supports were clamped to a steel ram that was used as a hammer. Different rams, which varied in weight (based on lengths between 15.2 and 61.0 cm (6 to 24 in)), were machined from a 76.2-mm- (3-in-) diameter solid steel cylinder. Alternatively, a 5-lb (22.2-N) sledgehammer was used to impact the top of the MDMP. A 12.7-mm- (1/2-in-) thick piece of plywood and/or a 3.175-mm- (1/8-in-) thick piece of plastic were used for pile cushions. In some cases, drill rods were connected to the top of the MDMP. An additional strain gauge and accelerometer from the Pile-Driving Analyzer (PDA) system were attached to these drill rod segments to measure force and acceleration applied to the rods. The output from the three MDMP accelerometers and strain gauges, and the additional drill rod strain gauge and accelerometer, was recorded and routed to the PDA via a connection box.

Figure 27a. View Alternative Text

Figure 27a. Photograph of the Dynamic Instrumentation Testing Setup.



Figure 27b. View Alternative Text 

Figure 27b. Schematic of the Dynamic Instrumentation Testing Setup.

Previous    Table of Contents    Next

 

ResearchFHWA
FHWA
United States Department of Transportation - Federal Highway Administration