Assessing Compaction Quality Using Light Weight Deflectometer: Laboratory versus Field Conditions

Abstract

This study was conducted to investigate the behavior of highway subgrade and subbase materials (2A, 2RC, and A-6) under varying compaction and moisture conditions using the light weight deflectometer (LWD). LWD tests were conducted across three testing scales: small-scale laboratory testing using Proctor molds, large-scale testing in a test pit, and full-scale field testing. Results demonstrate that LWD test response, for deflection and modulus, is highly sensitive to moisture content, increasing to an optimum point and then declining. Deflection measurements increased from laboratory to pit to field conditions, reflecting the influence of support stiffness and boundary conditions. Two correction factors ( $k_{1}$ ≈ 1.65 for laboratory-to-pit and $k_{2}$ ≈ 2.7 for laboratory-to-field) were developed and proposed to convert the laboratory LWD test results to the LWD deflection response in field applications. The study highlights the importance of scale effects, moisture variability, and soil behavior under impact loading, providing practical insights for using LWD as a compaction quality control tool in pavement construction. Because of the limited number of data points in this study, these preliminary correction factors $k_{1}$ and $k_{2}$ are proposed as exploratory values to bridge the gap between laboratory and field LWD measurements. These findings support improved interpretation of LWD data and its implementation in performance-based quality assurance practices.

Keywords

light weight deflectometer (LWD)subgrade compaction quality deflection field testing

Introduction

Compaction quality of highway materials, particularly in subgrade and subbase courses, is paramount for ensuring the long-term performance and structural integrity of pavement systems. While traditional assessment methods have primarily relied on density-based criteria, there is a growing interest in utilizing performance-based parameters, such as stiffness, for quality control (QC) and assurance ( 1 ). The light weight deflectometer (LWD) is a portable device gaining traction for its ability to rapidly evaluate the surface modulus or stiffness of unbound and partially bound materials ( 2 ). This research focuses on comparing the LWD response across three testing scales: small-scale laboratory testing using Proctor molds, larger-scale laboratory testing in a test pit, and actual field testing. By examining LWD deflection and modulus measurements under these conditions, this study aims to shed light on the applicability and challenges of using LWD as a compaction quality assessment tool.

Background

The LWD is a simplified version of the falling weight deflectometer, offering a portable device to determine the modulus of subgrade and subbase materials ( 3 ). The results obtained from the LWD provide a better understanding of the underlayer properties and the connection between long-term pavement performance and pavement design. The components of the LWD device include a falling weight, a load plate, a deflection sensor, a load cell, and a data processing and storage system. Various manufacturers (e.g., Zorn, Keros, Dynatest, Prima, Loadman, ELE, TFT, Olson, Humboldt, and CSM) produce LWD devices ( 4 ). Schwartz et al. reported that, while LWD measurements in large test pits correlated well with static plate load tests, the correspondence between Proctor mold LWD results and larger-scale tests was substantially weaker because of the mold’s confinement effects ( 4 ). Parameters such as the diameter of the loading plate, plate rigidity, plate contact stress, loading rate, buffer type, location, and deformation transducer influence the LWD’s measurement of compacted geomaterials. The LWD is designed to simulate a truck with 10 ton of axle weight moving at a speed of 80 km/h, applying a pressure of 0.1 MN/m² to the soil with a loading time of 18 ms ( 5 ).

To conduct an LWD test, the device is set up on a test site or in a laboratory. The weight is dropped from a standard height (guided down a rod) until it hits a buffer made of rubber pads or steel springs. The impact force is transferred to the plate and then to the ground, causing deflection. Typically, a total of six drops are applied: the first three are seating drops, and the last three are data-collecting drops. A velocity sensor or an accelerometer records the speed or acceleration of the plate’s downward movement ( 1 ). The modulus for field application is then calculated using the Boussinesq equation, which considers Poisson’s ratio, plate radius, stress distribution factor, and soil stiffness (calculated by the LWD device as the ratio of peak applied load to measured peak deflection). Because of the advantages, non-destructive tests based on modulus/stiffness, such as LWD, have gained attention as replacements for moisture and density tests ( 6 ). LWD is considered more functional for evaluating the quality of any compacted geomaterials because of its portability and rapid measurement of soil shear strength and stiffness in the field ( 7 ).

The LWD is increasingly employed as a tool for compaction QC in pavement construction. Compared with traditional density or moisture-based tests, LWD enables more practical, real-time monitoring of compaction quality during construction. Studies have demonstrated its effectiveness across various materials and construction layers. For example, Volovski et al. established statistical acceptance limits for LWD measurements in subgrade and subbase layers, highlighting its role in construction quality assurance ( 8 ). In LWD testing, the zone of influence extends to 1 to 1.5 times the plate diameter below the test level. Therefore, using a smaller plate size for thin unbound layers can help prevent the sublayer effects from influencing the surface layer moduli. Studies have shown strong correlations between LWD test results and compaction quality. For example, Kongkitkul et al. found a relatively good correlation between surface stiffness evaluated by LWD and the degree of compaction obtained from the sand cone test (R² = 0.73) ( 9 ). Similarly, Hariprasad et al. observed a robust correlation (R² = 0.98) between LWD modulus and the relative compaction of the base layer of a low-volume road, suggesting that increasing the degree of compaction leads to an increase in modulus ( 10 ). Furthermore, various empirical correlations have been developed between LWD and other devices such as the California bearing ratio (CBR) and unconfined compressive strength, often showing high coefficients of determination (greater than 80%). Schwartz et al. conducted comprehensive research to evaluate the compaction of subbase and base materials using three LWDs (Olson’s LWD-1, Dynatest 3031 LWD, and Zorn ZFG 3000 LWD) in laboratory test pits ( 4 ). Their study found a strong correlation between the predicted modulus from static load tests and the modulus obtained from LWDs, with coefficients of determination ranging from 78% to 97.7%. While a relatively strong correlation was observed between results from multiple LWDs, the correlation between LWD results from test pits and LWD results from Proctor molds was not strong. However, a strong correlation was observed between the triaxial compression test and LWD results on Proctor-mold-compacted soils, with coefficients of determination ranging from 73% to 89%. Makwana and Kumar investigated the correlation between LWD as a response variable and dynamic cone penetrometer (DCP) and CBR as explanatory variables for subgrade soils, reporting an R² of 0.86 between LWD and CBR, and an R² of 0.81 between LWD and DCP ( 11 ).

Several factors, such as material type, moisture content, and compaction energy, can influence LWD test outputs ( 12 ). Moisture content is one of the most critical parameters affecting soil behavior. While Hossain and Apeagyei reported that the effect of moisture on LWD modulus measurements did not follow a consistent trend and suggested further investigation, other research has demonstrated that LWD modulus is highly sensitive to moisture content, with drier conditions generally yielding higher moduli, even when dry density might be lower ( 13 ). This sensitivity highlights the complexity of interpreting LWD data, as direct correlation with density alone may be misleading without considering the moisture effect ( 14 ).

LWD tests can be conducted on soil compacted in a Proctor mold. These molds are used in laboratory standard tests, such as AASHTO T 99 and T 180, to establish optimum moisture content (OMC) and moisture-density relationships ( 15 , 16 ). Research at Purdue University indicated that the lubrication of the mold’s wall significantly affected deflection measurements when LWD tests were performed on compacted soil samples within the Proctor mold, leading to the recommendation of using test pits for experiments ( 17 ). Additionally, for open-graded materials with larger aggregate sizes, conventional 152 mm (6 in.) diameter Proctor molds may not provide enough room for laboratory compaction, potentially leading to increased surface deflection because of soil destabilization. To address this, the Purdue researchers also fabricated a custom mold with a 305 mm (12 in.) diameter and 292 mm (11.5 in.) height to better accommodate coarse aggregates ( 17 ). Furthermore, the scale of testing, from small Proctor molds to large test pits and field sites, introduces variations in underlying support conditions, which are known to significantly affect LWD deflection measurements ( 4 ). Understanding these scale-dependent variations in LWD response is crucial for developing practical LWD-based compaction quality assessment criteria. The repeatability of LWD measurements as a compaction QC tool requires improvement, and the same procedures could also be applied for quality acceptance applications. To achieve this, a well-trained operator should conduct the test consistently and cautiously, and incorporating a thin sand layer on the testing spot can provide a level surface for more uniform impact pressure distribution.

Scope

In this research, the LWD response of several highway materials across three distinct testing scales was investigated.

Small-scale laboratory testing (Proctor mold): LWD tests were conducted on compacted specimens prepared in 15.2 cm (6 in.) diameter Proctor molds. This phase included examining the LWD response of materials (PennDOT 2A, PennDOT 2RC, and AASHTO A-6) at different moisture contents and compaction levels. The effect of mold size and rest period on LWD response was also explored.

Large-scale laboratory testing (test pit): LWD tests were performed on compacted 2A and 2RC materials within a controlled test pit. Materials were placed in three 15.2 cm (6 in.) thick layers at three different moisture contents. LWD measurements were taken on each layer, with the top layer being tested after varying numbers of roller passes to simulate different compaction efforts.

Full-scale field testing: LWD tests were conducted on 2A and AASHTO A-6 materials at a field construction site (SR 3014). Measurements were taken at various compaction levels, corresponding to different roller passes.

This work was specifically focused on comparing LWD deflection and modulus results across these three testing scales to identify and quantify the differences in material response, particularly emphasizing the challenges in establishing direct correlations between laboratory and field LWD measurements.

Objectives

The primary objective of this research was to evaluate and compare the LWD test response when testing highway materials under laboratory (Proctor mold), large-scale laboratory (test pit), and full-scale field conditions.

More specific objectives include:

To measure and analyze the LWD modulus and deflection of 2A, 2RC, and A-6 materials compacted in a controlled test pit environment with different compaction efforts (roller passes).

To obtain LWD deflection and modulus data from in situ compaction (pit and field) of 2A, 2RC, and A-6 materials, considering varying roller passes and in situ moisture conditions.

To critically compare LWD results obtained from Proctor mold tests, test pit experiments, and field observations for 2A, 2RC, and A-6 materials, highlighting the discrepancies and influencing factors such as underlying support stiffness and moisture content variability.

To identify the challenges in establishing a direct relationship and calibrating LWD measurements between small-scale laboratory tests and field conditions.

Methodology

Materials

This research utilized three primary aggregate materials—2A, 2RC, and A-6—all sourced from Pennsylvania Department of Transportation (PennDOT) approved quarries.

2A is a widely specified aggregate for subbase and base course applications, as well as road fill and pipe bedding. It is characterized by a 2 in. top size and a variable fine content (material passing the #200 sieve) ranging from 0% to 10%. This variation allows 2A to function as either an open- or well-graded aggregate (Figure 1).

Figure 1.

Stockpiles utilized in this study: 2RC and 2A (left) and A-6 (right).

2RC also has a 2 in. top size but typically contains a significantly higher silt and clay content than 2A, and is free from organic matter. The presence of fine soil particles increases its plasticity. Its availability makes it a cost-effective alternative to 2A for applications such as pipe bedding or road fill (Figure 1).

A-6 soils are plastic clays with a plasticity index (PI) greater than 10 and may include a trace of organic matter, according to AASHTO classification (Figure 1). Figure 1 shows the original stockpiles of materials, where A-6 contained some large particles. However, for laboratory tests, the materials were systematically sieved and prepared according to AASHTO T 99/T 180 standards (e.g., removing or replacing oversize particles, depending on the method used). These soils typically have more than 75% passing the #200 sieve and may contain up to 64% sand and gravel. A-6 soils often experience significant volume changes between wet and dry conditions, showing low stability when wet.

The gradation characteristics of all materials used in this study, including 2A, 2RC, and A-6, are presented in Figure 2.

Figure 2.

Gradation chart for 2A, 2RC, and A-6 materials.

Laboratory Proctor Mold

Laboratory specimens were prepared in a 15.2 cm (6 in.) diameter Proctor mold. For each of the materials, three replicates were prepared to assess test variability at OMC and standard compaction levels. Note that, for laboratory testing, the materials were pre-sieved (e.g., removing particles larger than 19.0 mm for AASHTO T 99) to comply with standard test method requirements. Density and moisture content were determined following AASHTO T 99 and AASHTO T 265 specifications, respectively ( 18 ). All compaction procedures were conducted following AASHTO T 99 to ensure methodological consistency across material types. The resulting OMC and maximum dry density were not only used as laboratory benchmarks but directly guided the moisture selection for the test pit construction. By targeting the Proctor OMC in the test pit, the research team ensured a controlled comparison between laboratory and intermediate scales. For the field scale, these Proctor benchmarks were used to characterize the material state post-construction.

Test Pit

A dedicated test pit was constructed to simulate compacted aggregate layers for 2A and 2RC. For each aggregate type, experiments were conducted at three different moisture contents. At each moisture content, the material was placed and compacted in three 15.2 cm (6 in.) thick layers. LWD tests were performed on each layer. The final (top) layer was subjected to LWD testing at various compaction levels. In-place density and moisture content were determined using a nuclear moisture-density gauge ( 19 ). The nuclear density gauge was subjected to reference and stability checks following the manufacturer’s instructions and PennDOT standard procedures to ensure proper calibration and reliable measurements. The nuclear gauge was used to allow rapid, in situ measurements at the time of LWD testing, enabling timely sampling while minimizing disturbance to the compacted material.

Field Testing

Field testing was conducted at a single site involving 2A and A-6 materials. LWD tests were performed under varying roller passes. Field density measurements were obtained using nuclear density gauges. A summary of these methods and the measured parameters is provided in Table 1.

Table 1.

Summary of Material Testing Methods and Corresponding Standards

Experiment	Standard	Response
Moisture content	AASHTO T 265 ( 18 )	Moisture content
Moisture-density relations	AASHTO T 99, Methods B and D ( 15 , 16 )	Optimum moisture content and maximum dry density
Light weight deflectometer	ASTM E2583 ( 20 )	Deflection and modulus
In-place density and water content using a nuclear gauge	ASTM D6938 ( 19 )	Moisture content and density of the aggregates at the test pit
In-place density	ASTM D1556 ( 21 )	Verification of density measures at the test pit

Equipment

The primary equipment included a Proctor mold (Figure 3), vibratory roller, nuclear density gauge, and Dynatest LWD (Figure 4). The LWD was the primary device for evaluating surface modulus (stiffness). This portable device utilizes a 10 kg (22 lb) standard weight to apply an impact force up to 14.7 kN (3,300 lb) through 150 mm loading plates. Data were collected via Bluetooth to a tablet. LWD testing typically involved six weight drops, with the average of the last three used for material response.

Figure 3.

Soil compacted in the proctor mold: (a) before the light weight deflectometer (LWD) test and (b) after the LWD test.

Figure 4.

The light weight deflectometer testing at: (a) the Proctor mold and (b) the pit.

Experiment Procedure

LWD Testing of Proctor Mold Specimens

For laboratory experiments, specimens were prepared in a Proctor mold. Material density and moisture content were determined. When conducting LWD tests on Proctor mold specimens, the soil was compacted in three layers within the mold. To ensure proper LWD plate contact, any surface irregularities were smoothed out using fine sand (passing a #100 sieve). The LWD device was then carefully placed at the center of the specimen, as shown in Figure 4a. Each test involved six impact drops; the first three drops were for stabilization and were not used in calculations. The material’s modulus was reported as the average of the last three drops.

Testing at the Pit

Larger-scale testing was conducted in a dedicated test pit to simulate field conditions. The pit was divided into two parallel sections, each 102 cm (40 in.) wide, 178 cm (70 in.) long, and 46 cm (18 in.) high, designed to accommodate material placed to a thickness of 45.7 cm (18 in.) in three 15.2 cm (6 in.) thick layers, as depicted in Figure 5. Before placing the new materials, the existing soil was excavated to a depth of 45.7 to 50.8 cm (18 to 20 in.). The pit walls provided partial confinement, mimicking field boundary conditions.

Figure 5.

Dimensions of the soil after placement and compaction in the test pit (left: 3D configuration; right: layer thickness).

Side-by-side aggregate configurations were constructed in the test pit, as shown in Figure 6.

Figure 6.

Side-by-side aggregates prepared in the test pit showing light weight deflectometer test locations.

Water was added to the aggregate as needed and uniformly blended using a rotary mixer (Figure 7a) to achieve the target moisture levels. A plate compactor was used to compact the first two layers, while a vibratory roller compactor was used for the third (top) layer (Figure 7b). Compaction levels in the test pit were controlled by recording the number of roller passes before each LWD measurement.

Figure 7.

Photos showing: (a) uniform mixing and (b) vibratory compaction in the pit.

LWD tests were performed on the top layers immediately after 2, 4, 8, and 12 roller passes. In-place density and moisture content were determined using a nuclear density gauge (ASTM D6938). The compacted soil modulus was calculated using the Boussinesq equation:

E = \frac{2 k_{s} (1 - ϑ^{2})}{{Ar}_{0}}

(1)

where

k_s = $| \frac{F_{peak}}{ω_{peak}} |$ (calculated by the LWD device: $F_{peak}$ is the peak force applied by the LWD, $ω_{peak}$ is the corresponding peak deflection),

A = the stress distribution factor,

$ϑ$ = Poisson’s ratio, and

$r_{0}$ = the plate radius.

Data Collection from Field Projects

Field testing was conducted at a construction site to evaluate the in situ performance of 2A and A-6 material. At the site, LWD tests were performed on materials that had undergone varying degrees of compaction by a steel wheel roller compactor at three longitudinally spaced locations (3 ft apart) as shown in Figure 8. For the materials (approximately 8 to 10 in. thick), LWD tests were performed on previously compacted subbase and after additional passes of the roller compactor. The opportunity was not there to test the materials right after placement and before compaction; therefore, the first LWD test was conducted on material which had already been fully compacted. Nuclear density readings were recorded at LWD test locations using the surface method.

Figure 8.

Determination of three spots 3 ft apart longitudinally and light weight deflectometer testing. (left: determination of three spots spaced 3 ft longitudinally; middle: measurement setup; right: LWD testing)

Optimum Moisture Content (OMC) Determination

The OMC for each material was determined in the laboratory. The OMC represents the moisture content at which the highest dry density is achieved under a specified compaction energy. Because of specimen height limitations, all Proctor tests in this research were conducted using three compaction layers. The OMC for each material was derived from moisture content–dry density curves, established from four Proctor-compacted specimens at varying moisture contents. Table 2 and Figure 9 present a summary of the OMC results. Table 2 presents the measured individual moisture–density points used to establish the curve, from which OMC is determined via curve fitting; standard deviations are not reported, as the focus here is on curve construction rather than test variability.

Table 2.

Data From Study on Optimum Moisture Content (OMC) Based on AASHTO T 99

	Moisture content
Material	Target (%)	Measured (%)	Dry density (kg/m³)	OMC (%)
2RC	4.0	4.0	1,936.5	9.5
	8.0	7.6	2,101.2
	10.0	9.9	2,123.9
	12.0	11.1	2,069.9
2A	4.0	4.6	1,921.8	8.0
	6.0	7.0	2,065.7
	8.0	8.8	2,107.7
	10.0	11.4	2,013.1
A-6	6.0	5.5	1,498.5	15.0
	8.0	8.0	1,540.7
	15.0	15.1	1,733.1
	20.0	21.1	1,617.4

Figure 9.

Variation of dry density with moisture content for different soils.

The moisture content for the test pit was decided based on the laboratory Proctor OMC results. This target was selected to replicate the optimal laboratory conditions at a larger scale, with minor adjustments made during construction to reflect actual site realities. Conversely, the field moisture content was independent of the research control, reflecting the state achieved by the contractor’s standard operations. Consistency was maintained by targeting stabilized compaction states at all scales, with laboratory repeatability verified using two materials and three test locations used in both pit and field testing to capture variability.

Results and Discussion

Results from Different Test Scales

Table 3 presents the results obtained using the Proctor mold in the laboratory. The materials exhibited a typical trend: the LWD modulus increased with moisture content up to an optimum point, then declined at higher moisture levels. The 2A material consistently demonstrated higher stiffness than 2RC across all moisture levels. This higher stiffness of 2A is primarily attributed to its well-graded particle size distribution and generally lower fine content (0%–10% passing the #200 sieve) compared with 2RC, which typically contains significantly higher silt and clay. The combination of well-graded aggregates and low fines promotes denser packing, stronger interparticle contacts, and a more stable granular structure, resulting in greater resistance to deformation under load. A-6 is a cohesive soil, tested at a wide range of moisture content. Its response varied significantly under different moisture conditions, showing considerable differences in stiffness.

Table 3.

Laboratory Test Results Using the Proctor Mold

Material	Moisture content (%)	Average ELWD (MPa)	Deflection (micron)	Dry density (kg/m³)
2RC	4.0	50.7	402	1,936.5
	7.6	54.7	380	2,101.1
	9.9	36.3	624	2,123.9
	11.1	15.7	1478	2,069.9
2A	4.6	65.7	346	1,921.8
	7.0	74.2	324	2,065.7
	8.8	84.3	287	2,107.7
	11.4	68.1	346	2,013.1
A-6	5.5	43.0	648	1,498.5
	8.0	53.1	521	1,540.7
	15.1	82.3	337	1,733.1
	21.1	68.6	399	1,617.4

Note: ELWD = light weight deflectometer modulus.

Table 4 presents the results from testing in the test pit (only the highest LWD modulus achieved at each moisture content is reported) representing the optimal compaction condition. Results showed that LWD modulus generally increased with compaction effort, reaching a peak value that depended on moisture content. However, over-compaction at higher moisture levels led to a reduction in stiffness. These values were subsequently compared with the maximum modulus values obtained under different compaction passes in the field.

Table 4.

Pit Test Results under Varying Compaction and Moisture Conditions

Material	Moisture content (%)	Average ELWD (MPa)	Deflection (micron)	Dry density (kg/m³)
2RC	5.6	72.1 (12 passes)	744	2,122.6
	7.5	97.6 (12 passes)	533	1,960.3
	8.2	106.7 (8 passes)	513	1,953.3
2A	5.9	95.2 (12 passes)	551	1,917.0
	6.8	100.3 (8 passes)	504	1,999.3
	8.8	87.7 (4 passes)	591	2,118.7

Note: ELWD = light weight deflectometer modulus.

Tables 5 and 6 summarize the field test results for the 2A and A-6 materials. Additionally, corresponding deflection and dry density values were compared across the different test scales. These results were used to establish the stiffness-moisture-compaction relationships necessary for interpreting full-scale field results.

Table 5.

Field Light Weight Deflectometer Test Results of 2A Material with Different Compaction Passes

Number of passes	Moisture content (%)	Average ELWD (MPa)	Deflection (micron)	Dry density (kg/m³)
0	3.1	46.0	776.3	2,145.9
2		39.9	730.0	2,025.2
4		58.8	782.1	1,992.6
6		51.1	889.8	1,999.0

Note: ELWD = light weight deflectometer modulus.

Table 6.

Field Light Weight Deflectometer Test Results of A-6 Material with Different Compaction Passes

Number of roller passes	Moisture content (%)	Average ELWD (MPa)	Deflection (micron)	Dry density (kg/m³)
0	23.2	14.6	2,504	1,335.9
4		29.0	1,253	1,524.9
6		27.3	1,336	1,584.2

Note: ELWD = light weight deflectometer modulus.

Comparison Across Laboratory, Pit, and Field Tests

Since the field conditions are completely different from laboratory test conditions, it is important to determine how the test results compare. To compare the differences in LWD modulus and deflections between laboratory tests, pit tests, and field tests, results were selected under maximum compaction conditions with similar moisture content for the same material, as shown in Figures 10 and 11. In the field tests, the moisture content of the 2A material was 3.1%, and thus it can only be compared with the laboratory and pit test results obtained at the lowest moisture contents. It is important to note that the field placement and compaction of the 2A material were performed independently of this research project; therefore, the construction team did not target the laboratory-measured OMC (8.8%). Furthermore, the moisture content reported in Table 5 reflects the in situ condition measured after a period of time after the material had already been placed and compacted.

Figure 10.

Light weight deflectometer modulus (ELWD) comparison between laboratory tests, pit tests, and field tests.

Figure 11.

Light weight deflectometer deflections comparison between laboratory tests, pit tests, and field tests.

When comparing modulus values obtained from the Proctor mold and the test pit, one must exercise caution because of multiple influencing factors. Different equations are used to compute the modulus in each setup and, more importantly, the calculation is highly sensitive to the assumed Poisson’s ratio ( $ϑ$ ). However, the $ϑ$ value is not a constant—it is influenced by factors such as soil type, moisture content, degree of saturation, and density. This is particularly relevant in the field, where conditions are less controlled and material variability is greater.

As shown in the literature (e.g., Thota et al.), $ϑ$ values can range from as low as 0.1 for dry soils to nearly 0.5 for saturated ones, depending on soil classification ( 22 ). This variability significantly affects modulus calculations. For instance, for 2RC at 7.5% moisture content, deflection increases from 380 microns in the mold to 533 microns in the pit—a 40% increase—yet the modulus increases from 54.7 MPa to 97.6 MPa, even though the pit’s dry density is lower. A similar trend appears for 2A at 8.8% moisture, where modulus values from the mold and pit (84.3 MPa and 87.7 MPa) are very close, despite differing deflections.

This suggests that using a fixed Poisson’s ratio—especially in pit and field tests conducted at different moisture levels—may introduce errors. As such, direct comparison of modulus values across test conditions can be misleading if Poisson’s ratio variability is not considered. These observations underscore the limitations of relying solely on modulus values, which are affected by test method, moisture conditions, and assumptions on influencing factors such as Poisson’s ratio. Considering these uncertainties, deflection values provide a more consistent and reliable basis for comparing material behavior across laboratory, pit, and field testing environments.

Another key difference between laboratory, pit, and field LWD testing may lie in the distribution and dissipation of the applied force. Although the LWD drop hammer delivers a nearly identical total impact force in each test (because of consistent drop height and hammer mass), the measured peak force by the sensor can vary significantly depending on the contact area and the elastic and viscous properties of the tested material ( 23 ).

For the 2A field data, despite relatively consistent dry density and deflection values, the corresponding LWD modulus exhibited considerable variation. This is because the sensor captures the peak force, which is then used in a simplified work-based calculation that neglects the energy dissipated through plastic deformation. When the plastic phase is extended, the differences in stiffness and damping behavior across soil types become more pronounced, leading to increased errors in the calculated modulus ( 24 ).

Assuming equal contact area, a material such as A-6, with higher plasticity and cohesion, has a greater capacity to dissipate force through internal deformation. The assumption that all rebound energy is purely elastic is unrealistic, particularly for cohesive soils. This helps explain why the A-6 material exhibits greater divergence between field and laboratory moduli than the 2A material: in the field, energy is more extensively dissipated through plastic deformation and local yielding, which is not accounted for in the simplified LWD modulus calculation.

Moreover, deflection measurements inherently incorporate both elastic and plastic deformation components, offering a more comprehensive and stable representation of soil response. While the balance between elastic and plastic behavior varies with material type and loading conditions, deflection trends tend to reflect the integrated behavior of soils more reliably than modulus estimates.

This highlights a core limitation in current LWD modulus calculations: they rely on simplified models using peak force and peak elastic rebound, and fail to capture the nonlinear force–displacement relationship throughout the loading cycle. The force input is a momentary peak, while the displacement is a net dynamic response, leading to mismatch in temporal correspondence. Given that LWD imposes high-frequency impact loading (on the order of tens of milliseconds), and soil is a nonlinear, dissipative medium, a substantial portion of deformation occurs after the peak force has already dropped, because of delayed plastic flow or localized compaction. This dynamic mismatch introduces further inaccuracy in the modulus estimation.

To further enhance the practical applicability of these comparisons and quantify the relationship across testing scales, two correction factors were developed based on deflection comparisons between laboratory, pit, and field LWD results, as shown in Table 7. These factors, $k_{1}$ and $k_{2}$ , can help translate laboratory deflection values into practical field guidance for both performance expectations and QC:

k_{1} \times {Value}_{lab} \approx {Value}_{pit} \leq {Value}_{field} \approx k_{2} \times {Value}_{lab}

(2)

Table 7.

Deflections and Empirical Correction Factors

Material	Lab → pit (k1)	Averaged	Material	Lab → field (k2)	Averaged
2RC	533/380 ≈ 1.40	1.65 (SD: 0.28; COV: 17.2%)	2A	782/346 ≈ 2.26	2.70 (SD: 0.62; COV: 23.0%)
2A (Case 1)	551/346 ≈ 1.59		2A	782/346 ≈ 2.26
2A (Case 2)	504/324 ≈ 1.56		A-6	1,253/399 ≈ 3.14
2A (Case 3)	591/287 ≈ 2.06		A-6	1,253/399 ≈ 3.14

Note: COV = coefficient of variation; SD = standard deviation.

Correction Factor $k_{1}$ (Lab → Pit)

For 2RC and 2A materials, deflections observed in pit tests were consistently higher than those from laboratory tests, with ratios ranging from 1.40 to 2.06. The average correction factor was calculated as $k_{1}$ = 1.65. This value reflects the reduced confinement in pit conditions compared with the rigid boundary of Proctor molds and can be used to infer typical in situ performance from laboratory results.

Correction Factor $k_{2}$ (Lab → Field)

When comparing laboratory results with full-scale field measurements, deflections increased substantially. For example, the 2A material showed a field-to-laboratory deflection ratio of 2.26, while the A-6 material displayed a ratio of 3.14. These results support the use of $k_{2}$ = 2.7 as a general factor for defining maximum allowable deflections under realistic field compaction conditions. While some in situ moisture levels deviated from the laboratory OMC, the correction factors were derived from the maximum achievable response at each scale. This ensures that the factors primarily represent the physical change in support conditions and confinement rather than differences in compaction energy.

To evaluate the strength of predictability of the proposed factors, the coefficient of variation (COV) was calculated. The $k_{1}$ factor (laboratory-to-pit) exhibited a COV of 17.2%, while the $k_{2}$ factor (laboratory-to-field) showed a COV of 23.0%. These values indicate a reasonable level of consistency for empirical geotechnical correlations, especially considering the inherent variability in field boundary conditions and manual compaction efforts. The relatively narrow range of $k_{1}$ (1.40 to 2.06) suggests that the transition from rigid laboratory molds to larger test pits follows a predictable trend of reduced confinement.

The results from laboratory, pit, and field LWD tests highlight the important effects of moisture content and compaction effort on soil stiffness and deformation behavior. It is acknowledged that the comparison across scales relies on the Proctor-defined optimal compaction state as the consistent baseline. While matching the degree of saturation across all environments is practically unattainable because of different boundary conditions, basing the test pit moisture directly on the Proctor optimization ensures that the derived scale correction factors accurately capture the influence of confinement and compaction mechanism rather than arbitrary moisture fluctuations.

The differences observed across testing scales show that laboratory results cannot be directly applied to field conditions without adjustment. The identified OMC for maximum stiffness confirms the necessity of controlling moisture during compaction. The proposed correction factors provide a practical way to translate laboratory measurements to better predict field performance. These correction factors provide an empirical, single-value estimate of the scale effect and a practical means of translating laboratory-measured deflection targets into easy-to-use and field-applicable thresholds by accounting for the geometric and boundary-condition differences between Proctor mold testing and in situ LWD measurements.

The results from this research directly support the observations made by Schwartz et al., who found that the correlation between LWD measurements in Proctor molds and test pits was weak ( 4 ). By comparing the deflection increases from laboratory to pit (average $k_{1}$ = 1.65) and laboratory to field (average $k_{2}$ = 2.70), this study provides a numerical bridge for this previously identified gap. While Schwartz et al. highlighted the sensitivity of the setup, the data generated in this research quantifies this scale effect, showing that the increased confinement in the 15.2 cm mold artificially stiffens the material response compared with larger-scale conditions ( 4 ). The observed discrepancy between laboratory and field LWD measurements aligns with research by Tanyu et al., who evaluated the stiffness of unbound base aggregates across mold, bench-scale, and field environments ( 25 ). Similar to the findings from this research where deflection increased significantly from laboratory molds to field conditions, Tanyu et al. noted that LWD modulus estimates are highly sensitive to the material volume influenced by the loading plate and the lack of rigid lateral confinement in situ ( 25 ). This consistency across studies reinforces the conclusion that laboratory-derived benchmarks require empirical adjustment factors to be applicable for field QC.

When applied, $k_{1}$ and $k_{2}$ enable engineers and inspectors to establish realistic, field-specific maximum allowable deflections that reflect the expected scale effect, thereby supporting clearer interpretation of LWD results and more consistent QC decisions during construction. Although incorporating material properties such as PI or percent passing the No. 8 sieve could potentially improve the deflection correlation, evaluating these effects was outside the scope of this study. A comprehensive assessment would require a large set of fine aggregates with diverse PIs and gradations to isolate the influence of these parameters, which would constitute a separate research effort. These findings improve the understanding of soil response under impact loading and support more accurate QC in geotechnical and pavement engineering.

Conclusion

This study was conducted to investigate the behavior of subgrade and subbase materials (2RC, 2A, and A-6) under varying moisture conditions and compaction efforts using the LWD. Three conditions were considered: laboratory testing using Proctor mold, testing in a pit, and testing in situ. The findings provide critical insights into the interpretation and application of LWD data for quality compaction/assurance of subgrade and subbase materials.

The LWD modulus exhibits moisture-dependent behavior, increasing with moisture up to an optimum point before declining at higher moisture contents. The 2A material consistently showed higher stiffness than 2RC across all test conditions.

Test scale strongly affects measured values: LWD deflections obtained from pit tests are consistently higher than those from laboratory Proctor molds. Deflections increase progressively from laboratory to pit to field conditions. This scale effect reflects differences in boundary constraints, soil confinement, and energy dissipation mechanisms.

The simplified LWD modulus calculation, based on peak force and rebound, does not fully capture soil nonlinear and plastic behavior, especially for cohesive soils such as A-6. Deflection measurements integrate both elastic and plastic responses, thus providing a more stable indication of soil performance.

Direct deflection correlation across scales is limited. To bridge the gap between laboratory and field measurements, two empirical correction factors (k₁ ≈ 1.65 for laboratory-to-pit, k₂ ≈ 2.7 for laboratory-to-field) were proposed based on deflection ratios, offering practical guidance for interpreting LWD results in field applications.

The proposed correction factors offer practical guidance for applying laboratory-derived deflection benchmarks to field conditions by enabling the development of site-specific maximum allowable deflection thresholds. This approach helps practitioners better interpret LWD measurements, improve compaction QC evaluations, and adapt acceptance criteria to the inherent differences between laboratory and field testing environments. The authors emphasize that, given the limited dataset, these correction factors are site-specific and should be treated as an initial framework for future validation across more diverse materials and conditions. Given the empirical nature of the correction factors proposed in this study, further validation using a substantially larger dataset—incorporating a broader range of materials, gradations, moisture conditions, and field construction scenarios—is recommended. Such efforts would support the development of a more robust or even tiered system of correction factors based on a comprehensive field/conditions matrix, ultimately enabling broader applicability and improved reliability in field LWD interpretation.

Footnotes

Acknowledgements

The authors gratefully acknowledge the financial support by Pennsylvania Department of Transportation (PennDOT) for sponsoring this research. Special thanks are extended to Beverly Miller of PennDOT for her valuable support and guidance throughout the project. We are also grateful to Mr. Scott Milander of Penn State for coordinating laboratory activities.

Author Contributions

The authors confirm contribution to the paper as follows: study conception and design: M. Solaimanian, D. Neff; data collection: X. Guo, M. Solaimanian, D. Masouleh, D. Neff; analysis and interpretation of results: X. Guo, M. Solaimanian; draft manuscript preparation: X. Guo, M. Solaimanian. All authors reviewed the results and approved the final version of the manuscript.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Pennsylvania Department of Transportation (PennDOT) under Contract #512101, Work Order #PSU 02.

ORCID iDs

Xiaogang Guo

Mansour Solaimanian

References

Kessler

Use of DCP (Dynamic Cone Penetrometer) and LWD (Light Weight Deflectometer) for QC/QA on Subgrade and Aggregate Base. GeoHunan International Conference, 2009, pp. 62–67.

Riad

Zhang

Liu

Wang

Y. D.

Compaction Quality Assurance Specifications of Unbound Materials. Journal of Transportation Engineering, Part B: Pavements, Vol. 149, No. 1, 2022, p. 03122003. https://doi.org/10.1061/JPEODX.0000403.

Kuttah

Using Repeated Light-Weight Deflectometer Test Data to Predict Flexible Pavement Responses Based on the Mechanistic–Empirical Design Method. Construction Materials, Vol. 4, No. 1, 2024, pp. 216–237. https://doi.org/10.3390/constrmater4010012.

Schwartz

C. W.

Afsharikia

Khosravifar

Standardizing Lightweight Deflectometer Modulus Measurements for Compaction Quality Assurance. Maryland State Highway Administration, State Highway Administration, Baltimore, MD, 2017.

Jardine

R. J.

Potts

D. M.

Higgins

K. G.

Brandl

Adam

Load Plate Test with the Light Falling Weight Device. In Advances in Geotechnical Engineering: The Skempton Conference, Vol. 29, London, UK, 2004, pp. 29–31.

Mazari

Tirado

Abdallah

Nazarian

Mechanistic Estimation of Lightweight Deflectometer Target Field Modulus for Construction Quality Control. Geotechnical Testing Journal, Vol. 39, No. 6, 2016, pp. 965–976. https://doi.org/10.1520/GTJ20150266.

Fleming

P. R.

Frost

M. W.

Lambert

J. P.

Review of Lightweight Deflectometer for Routine in Situ Assessment of Pavement Material Stiffness. Transportation Research Record: Journal of Transportation Research Board, 2007. 2004(1): 80–87. https://doi.org/10.3141/2004-09.

Volovski

Arman

Labi

Developing Statistical Limits for Using the Light Weight Deflectometer (LWD) in Construction Quality Assurance. Purdue University, JTRP Technical Reports, FHWA/IN/JTRP-2014/10. West Lafayette, IN, 2014. https://doi.org/10.5703/1288284315504.

Kongkitkul

Saisawang

Thitithavoranan

Kaewluan

Posribink

Correlations Between the Surface Stiffness Evaluated by Light-Weight Deflectometer and Degree of Compaction. In Tunneling and Underground Construction, 2014, pp. 65–75.

10.

Hariprasad

Umashankar

Garala

T. K.

Lightweight Deflectometer for Compaction Quality Control. In Geotechnical Characterisation and Geoenvironmental Engineering: IGC 2016, Vol. 1, Singapore, 2018, pp. 35–42.

11.

Makwana

A. E. P.

Kumar

Correlative Study of LWD, DCP and CBR for Sub-Grade. International Journal of Engineering Trends and Technology, Vol. 67, No. 9, 2019, pp. 89–98. https://doi.org/10.14445/22315381/IJETT-V67I9P215.

12.

Stamp

D. H.

Mooney

M. A.

Influence of Lightweight Deflectometer Characteristics on Deflection Measurement. Geotechnical Testing Journal, Vol. 36, No. 2, 2013, pp. 216–226. https://doi.org/10.1520/GTJ20120034.

13.

Hossain

M. S.

Apeagyei

A. K.

Evaluation of the Lightweight Deflectometer for In-Situ Determination of Pavement Layer Moduli. Virginia Transportation Research Council, FHWA/VTRC 10-R6. Charlottesville, VA, 2010.

14.

Tirado

Gamez-Rios

K. Y.

Fathi

Mazari

Nazarian

Simulation of Lightweight Deflectometer Measurements Considering Nonlinear Behavior of Geomaterials. Transportation Research Record, 2017. 2641(1): 58–65. https://doi.org/10.3141/2641-08.

15.

AASHTO. Standard Method of Test for Resistance of Compacted Asphalt Mixtures to Moisture-Induced Damage. AASHTO T 283-14. American Association of State Highway and Transportation Officials, Washington, D.C., 2014.

16.

AASHTO. Standard Method of Test for Moisture–Density Relations of Soils Using a 2.5-Kg (5.5-Lb) Rammer and a 305-Mm (12-in) Drop (T 99). American Association of State Highway and Transportation Officials, Washington, D.C., 2017.

17.

Zhao

Yao

Jiang

Maximum Allowable Deflection by Light Weight Deflectometer and Its Calibration and Verification. Publication FHWA/IN/JTRP-2018/21. Purdue University, Joint Transportation Research Program, FHWA/IN/JTRP-2018/21, West Lafayette, IN, 2018.

18.

AASHTO. Standard Method of Test for Laboratory Determination of Moisture Content of Soils (T 265). American Association of State Highway and Transportation Officials, Washington, D.C., 2015.

19.

ASTM International. Standard Test Method for In-Place Density and Water Content of Soil and Soil-Aggregate by Nuclear Methods (D6938). ASTM International, West Conshohocken, PA, 2023.

20.

ASTM International. Standard Test Method for Measuring Deflections Using a Portable Impulse Plate Load Test Device (Light Weight Deflectometer, LWD) (E2583). ASTM International, West Conshohocken, PA, 2020.

21.

ASTM International. Standard Test Method for Density and Unit Weight of Soil in Place by the Sand-Cone Method (D1556). ASTM International, West Conshohocken, PA, 2015.

22.

Thota

S. K.

Coa

T. D.

Vahedifard

Poisson’s Ratio Characteristic Curve of Unsaturated Soils. Journal of Geotechnical and Geoenvironmental Engineering, Vol. 147, No. 1, 2021, p. 04020137. https://doi.org/10.1061/(ASCE)GT.1943-5606.0002424.

23.

Nazzal

M. D.

Abu-Farsakh

M. Y.

Alshibli

Mohammad

Evaluating the Light Falling Weight Deflectometer Device for in Situ Measurement of Elastic Modulus of Pavement Layers. Transportation Research Record, 2007. 2016(1): 13–22. https://doi.org/10.3141/2016-02.

24.

Nguyen

T. H.

Lee

Ahn

A Simulation of Lightweight Deflectometer Operation to Investigate Influences of Its Design on Soil Modulus of Elasticity. In International Conference on Geotechnics for Sustainable Infrastructure Development, Springer Nature Singapore, Singapore, 2024, pp. 2479–2487.

25.

Tanyu

B. F.

Ullah

Akmaz

Optimizing Reclaimed Asphalt Pavement (RAP) Content in Unbound Base Aggregate. Final Report VTRC 22-R9. Virginia Transportation Research Council, Charlottesville, V.A., 2021.

Assessing Compaction Quality Using Light Weight Deflectometer: Laboratory versus Field Conditions

Abstract

Keywords

Introduction

Background

Scope

Objectives

Methodology

Materials

Laboratory Proctor Mold

Test Pit

Field Testing

Equipment

Experiment Procedure

LWD Testing of Proctor Mold Specimens

Testing at the Pit

Data Collection from Field Projects

Optimum Moisture Content (OMC) Determination

Results and Discussion

Results from Different Test Scales

Comparison Across Laboratory, Pit, and Field Tests

Correction Factor k 1 (Lab → Pit)

Correction Factor k 2 (Lab → Field)

Conclusion

Footnotes

Acknowledgements

Author Contributions

Declaration of Conflicting Interests

Funding

ORCID iDs

References

Correction Factor $k_{1}$ (Lab → Pit)

Correction Factor $k_{2}$ (Lab → Field)