Rapid Airfield Damage Recovery Next Generation Backfill Technologies Comparison Experiment (TPD2102011)

Contents

Abstract………………………………….ii

Figures and Tables……………………..v

Preface……………………………………viii

1      Introduction ………………………1

1.1  Background ………………………..1

1.1.1 Airfield Damage Repair……………1

1.1.2 RADR repairtasks…………………..2

1.1.1 Need for new backfill methods………3

1.2  Objective………. 6

1.3 Sope………..6

2      Crater Repair Test Material…………..7

2.1  Constructedsubgrade soil…………. 7

2.2  Backfi11 soils ……………………… 7

2.3  Polymers………………… 7

2.4  Port iand Iimestone cement (PLC)…………………… 8

2.5  Geocells…………………………………… 8

2.6  Expanding polyurethane foam………………….. 9

2.7  Rapid-setting flowable fill……………………. 10

2.8  Rapid-setting concrete………………………. 10

3    Experimental Program…………………………….. 12

3.1 Experiment layout………………………………….. 12

3.2 Crater preparation………………………………. 15

3.2.1       Construction of uniform subgrade…………… 15

3.2.2        Instrumentation (EPCs) …………………..19

3.3           Backfill construction ………………………. 20

3.3.1       Chemically stabilized soil backfill (Repairs 1 and 5)………20

3.3.2        Rapid-setting flowable fill backfill (Repairs 2 and 6 )………….25

3.3.3        Geoce/1-reinforced backfill (Repairs 3 and7)…………….28

3.3.4        Foam backfill(Repairs 4 and 8) ……………………30

3.4 Repair backfill assessment using light weight deflectometers………………. 34

3.5 Rapid-setting concrete cap placement ………………….. 37

3.6 Simulated aircraft traffic operations………………………….. 43

3.7 Data collection………………………………. 43

4     Results……………….46

4.1    Repair data summary……………………….. 46

4.2    Backfill placement times…………………………..47

4.3    Post-traffic results…………………………….. 49

 4.3.2         Elevationchanges………………………….56

4.3.3        Impact stiffness modulus …………….56

4.3.4       Earth press ure cell data…………………….57

5    Comparison of Backfill Technologies ………………………………… 60

6      Conclusions and Recommendations……………………… 62

 6 . 2    Recommendations………………………… 63

References………………………… 64

AppendixA:.ConstructedSubgrade SOii Classification Data……………………………….. 67

Appendix B: Backfnr Soils’ Classification Data………………………. 72

Appendix C: Pavement Distress Data…………………………77

Acronyms…………………………..81

Report Documentation Page

Figures and Tables

Figures

Figure 1.1. RADR backfill method of flowablefill capped with rapid-setting concrete……..2

Figure 2.1. Benefits of foam expansion ………………..10

Figure 3.1. Simulated 8.5-ft square craters………………………..12

Figure 3.2. Repair layout for the test sequences………………………….. 13

Figure 3.5. Subgrade construction…………………………. 17

Figure 3.8. Placement of soil in 21-ft by 10-ft area……………………….21

Figure 3.13. Compacting soil…………………………….. 25

Figure 3.15. Rapid-setting flowable fi11 dry placement method……………. 27

Figure 3.16. Adding water to rapid-setting flowable fill during dry placement method…………… 27

Figure 3.17. Finished rapid-setting flowable fill surface…………………. 28

Figure 3.18. Cutting geocelI section to fit crater size…………….. 28

Figure 3.19. Geocell-reinforced backfill placement…………………….29

Figure 3.20. CRADR 1 -L prototype foam dispensing system………………..30

Figure 3. 21. Polyethyleneliner applied to Repair 4 ……………………. 31

Figure 3.22. Foam placement…………………….32

Figure 3.23. Placing additionaI foam in Repair 4 to achieve the desired backfill thickness……….32

Figure 3.24. Final foam backfill surface of Repair 4……………33

Figure 3.25. Expanded foam backfill surface of Repair 8………………….34

Figure 3.26. Final foam backfill surface of Repair 8 (after cutting)……………….34

Figure 3.29. Volu metric mixer …………………………………………….. 3. 8

Figure 3 .30 . Rapid-setting concrete cap screeding with straight-edgebar…………………. 39

Figure 3.31. Finished rapid•setting concrete cap on Repair 1. ……………………………….. 3. 9

Figure 3.32 . Finished rapid-setting concrete cap on Repair 2………………………………… 40

Figure 3.33 . Finished rapid-sett ing concrete cap on Repair 3………………………………… 40

Figure 3.34 . Finished rapid-sett ing concrete cap on Repair 4………………………………… 41

Figure 3 .35 . Finished rapid-sett ing concrete cap on Repair 5………………………………… 41

Figure 3.36. Finished rapid-sett ing concrete cap on Repair 6………………………………… 42

Figure 3.37. Finished rapid-sett ing concrete cap on Repair 7…………………………………. 42

Figure 3.38. Finished rapid-sett ing concrete cap on Repair 8………………………………… 42

Figure 3.39. F-15E load cart and typical F-15 normally distributed traffic pattern………… 43

Figure 3.40. Measuring a spall on repair edge……………………………………………………. 44

Figure 3.41. Survey data collection………………………………………………………………… 45

Figure 3 .4 2. FWD testing on the surface of a repair…………………………………………….. 45

Figure 4.1. Backfill ti mes ……………………………………………………………………………… 4. 9

Figure 4.2. Traffic data summary…………………………………………………………………… 50

Figure 4.3. Post-traff ic photographs of Repair 1  after 3.600 passes ……………………….. 5.1

Figure 4.4. Post-traff ic photographs of Repair 2 after 3.600 passes……………………….. 51

Figure 4.5 . Post-traff ic photographs of Repair 3 after 1,520 passes……………………….. 52

Figure 4.6. Post-traff ic photographs of Repair 4 after 830 passes………………………….. 52

Figure 4.7. Post-traffic photographs of Repa ir 5 after 1,520  passes ……………………….. 5. 3

Figure 4.8. Post-traffic photographs of Repair 6 after 3.600 passes……………………….. 54

Figure 4.9. Post-traff ic photographs of Repair 7 after 1,792 passes………………………….. 54

Figure 4.10. Post-t raffic photographs of Repa ir 8 after 1,520 passes……………………… 55

Figure 4.11. Shrinkage cracks on Repair 4…………………………………………….55

Figure 4.12. FWD resuIts……………………………………………… 57

Figure 4.1 3. Peak pressure values for Repairs 1 t hrough 4 during each pass interval applied………………………………………… 58

Figure 4.14. Peak pressure values for Repairs 5 through 8 during each pass interval applied……………………………………………………….. 59

Tables

Table 2.1. GW2 0V8 mechanical properties……………………………………………… 9

Table 3.1. Weather conditions during the test period each day………………………………. 14

Table 3.2. Crate r natural foundation and compacted subgrade DCP tests results……….. 18

Table 3.3. CL-ML constructed subgrade average nuclear moisture and density………….. 18

Table 3.4 . Backfill LWD deflection and modulus results……………………………………….. 36

Table 4 .1. Repair data summary…………………………………………………….. 46

Table 4.2. Maximum changes in elevation……………………………………………………….. 56

Table 4.3. Average peak pressures measured on each repair……………………………….. 5.8

Table 5.1. Technology trade-off analysis………………………………………………. 61

Preface

This study was conducted for the U.S. Air Force Civil Engineer Center (AFCEC) under the Rapid Airfield Damage Recovery Research and Development Program, Project  Number  P246334. The technical  monitor was Dr. Bobby Diltz, AFCEC. The Program Manager was Mr. Jeb S. Tingle.

This work was performed by the Airfields  and  Pavements  Branch  (GMA) of the Enginee1ing Systems and Materials Division (GM), U.S. Army Engineer Research and Development Center, Geotechnical and Structures Laboratory (ERDC-GSL). At the time of publication, Ms. Anna Jordan was Chief, GMA; Mr. Justin S. Strickler was Chief, GM; and Mr. R. Nicholas Boone , CEERD-GZT, was the Technical Director for Force Projection and Maneuver Support . Mr . Charles W. E1tle II was Deputy Director of ERDC­ GSL, and Mr. Brutley P. Durst was the Director .

COL Teresa A Schlosser was the Commander of ERDC, and Dr . David W. Pittman was the Director.

1            Introduction

1.1        Background

1.1.1      Airfield Damage Repair

The airfield damage repair (ADR) process consists of engineer personnel activities in response to an attack on an airbase to provide adequate launch and  recovery surfaces for the  mission aircraft . Although  ADR criteria are based on repairing airfields in friendly territory, recent military operations require repairing airbases occupied previously by hostile forces damaged during forcible entry or purposely sabotaged by departing forces.

Legacy ADR methods for repairing bomb damage consist of using debris to fill the crater, followed  by placement  of  crushed  stone backfill with a foreign object debris (FOD) cover. ADR encompasses  other  areas  besides the repair of bomb damage including damage assessment, identification of candidate minimum operating strips, and the safe disposal of unexploded ordnance (Department of the Army, Navy, and Air Force 2003).

In  2002 , the U.S. Army Engineer Research and Development Center began the most prominent research program addressing ADR. The program included several load  cart tests and flight tests that  revealed none of the legacy repair technologies were capable of supporting C-17 aircraft on a runway where high, horizontal breaking forces caused the systems to fail. For C-17 operations only, the repairs can be performed without the protective mat surfacing or FOD cover. However, protective surfacing or FOD covers are required for fighter aircraft operations.

Thus, the legacy repair technologies cannot be used on airfields subjected to both fighter and heavy cargo aircraft , due to the inability of the legacy repairs to sustain C-17 operations and the inability of fighter aircraft to operate on unsurfaced repairs.

During the following decade, ERDC efforts were dedicated to address the operational limitations oflegacy repair methods. Research was conducted to develop capability to return damaged runways to full operational sortie production in less than 8 hr. Evaluations included new materials , tactics, techn iques, and procedures (TI Ps), and prototype equipment for rapid automated airfield damage assessment and crater repair (Tingle et al.2009) _ Multiple operational utility assessments (OUAs) were conducted to demonstrate the updated ADR methods and to collect feedback from the end users for further refinement as part of the Critical Runway AssessmenT and Repair (CRATR) Joint Capabilities Technology Demonstration (JCfD)_ Following the CRATR JCTD, multiple other troop and full-scale tests have been conducted to address other gaps identified within the scope of the Rapid Airfield Damage Recovery (RADR) modernization  program. The  RADR program focused on  refming techniques and reducing the logistical footprint of the ADR process.

1.1.2      RADR repair tasks

The crater repair process developed during the CRATR JCTD and further refined during the RADR modernization program is well-documented in many references (Bell et al. 2013; Priddy et al. 20 13 ; Edwards et al. 2013; Carruth et al. 2015)_ First , large debris is removed with compact track loaders and large front-end loaders. Next, the stanchion method is used (USAF 1992) to mark the extent of upheaval before a square repair area on the pavement is marked. The marked area is then saw-cut, followed by breaking and removal of the existing material with wheeled excavators.

Once the repair is excavated, it is backfilled, typically with rapid-setting flowable fill (Figure 1.1) or compacted crushed stone. The main purpose of the backfill is to fill the void and provide enough support to the repair cap to withstand the loads from aircraft tr affic. Rapid-setting flowable fill has proven to be the best backfill altern ative for RADR.

Figure 1.1. RADR backfill method Of flowable fill capped wtth rapid-setting concrete.

Lastly, the repair is capped with rapid-setting concrete by using the U.S. Air Force Simplified Volumetric Mixer, or with recycled asphalt cookies using a Bagela recycler and appropriate large pneumatic and steel­ wheel compaction equipment.

1.1.1         Need for new backfiill methods

While the current RADR crater repair method provides successful results, it is heavily dependent on specific material and equipment resources. Other repair alternatives are required in the event that large quantities of rapid-setting flowable fill and rapid-setting cement super sacks cannot be transported to the repair site. Backfill alternatives are not included in the TIPs.

Three repair alternatives for backfill that have proven their suitability for use in RADR are: 1) expandable polyurethane foam, 2) mechanical soil reinforcement, and 3) chemical soil stabilization. However, each backfill technology requires additional refinement to ultimately meet all RADR requirements. Also, there are no clear indications as to how the performance of each of these methods compares to each other.

This project consisted of evaluating the three backfill technologies in full-scale controlled test environments and comparing the performance of each technology to the standard rapid-setting flowable-fill backfill method described in the RADR ITPs.

1.1.1.1  Expandable polyurethane foam

The foam backfill technology consists of mechanically pouring liquid polyurethane foam into a crater void, then waiting for the foam to expand and set. The foam expands six times its original liquid volume; therefore, enough liquid volume is poured so that when the foam expands, it leaves room for a minimum 10-in.-thick rapid setting concrete cap. The polyurethane foam provides a uniform , low-quality backfill to support the rapid-setting concrete cap. Its primary function is to fill the crater void while providing uniform support to the rapid-setting concrete slab.

Previous research  efforts  and  end-user  demonstration  exercises  showed two  main  issues associated  with  this technology, i.e.,  (1)  material sensitivity to moisture and (2) clogging of the reacted foan1 material within the mechanical  distribution  system  (Priddy et  al. 2013). This technology was recently refined by redesigning the foam distribution equipment, improving foam material formulations , and upgrading the foam backfill placement procedures (Figure 1.2) (Johnson et al. 20 1 7 ) The refined technology was successfully  demonstrated  at  the  Silver  flag  Exercise Site at Tyndall Air Force Base in20 16 , where ten airmen evaluated the technology by petforming nine crater  repairs and  provided valuable feedback for further technology refinement (Mejias-Santiago et al. draft*) _

Figure 1.2. Foam backfill demonstration.

1.1.1.2  Geocell mechanical  stabilization

Fom1erly known in the Department of Defense (DOD) community as “sand grid,” geocell is a geosynthetic cellular confinement system used to reinforce geotechnical materials (Figure 1.3). Within the scope of ADR, geocells have been used to reinforce backfill materials for expedient crater repairs (Read and Dukes 1988, Webster 1986). One of the primary benefits of using geocells is that lower quality backfill materials can be used rather than the traditional well-graded crushed limestone material. This application is particularly beneficial at deployed locations where the use of indigenous materials is desired since large stockpiles of quality materials are not routinely available. At permanent installations, the use of geocells to  reinforce lower  quality geotechnical  materials  minimizes  the requirement for prepositioned stockpiles of high-quality aggregates. The original TfPs forthe useof geocells (sand grid) were developed in the 1980s, and the original geocell material has significantly changed. New commercial geocell products have emerged since the expiration of the original government patent, and many of the new products were evaluated against desired petformance standards. In addition, the petformance of different  soil  types  reinforced with geocell for  RADR missions had  not been adequately quantified and documented. Investigations petformed in 2016-2017 were aimed at addressing these gaps and evaluating new products with different backfill materials for selecting optimum geocell materials and geometries. Successful performance was observed for several products (Garcia et al. draft*)

*Mejias-Santi ago, M.,J_ S. Tingle, J_L    Johnson. L  A. Gurt owski and C. s_ Griggs_ Draft_ Refinementof foambackfilltechnology tor expedient airf ield damage recove,y Phase //1: Technology demonstration and end user evaluation. ERDC/ GSL TR. Vicksburg, MS: U.S. Army Engineer Research and Development Center

Figure 1.3. Geocell mechanical stabilization method.

1.1.1.3  Chemically stabilized soil

The chemically  stabilized  soil backfill  method (Figure 1.4) utilizes the locally available soil mixed with an additive, typically cement, lime, fly ash, or polymers. The soil stabilization additive(s) are used to modify the soil properties and improve strength and stiffness. Other additives include sodium silicate and polyurethane-based products. According to UFC 3- 250-11, a minimum 7-day unconfined compressive strength (UCS) of 500 psi is required for a rigid pavement base course. Normally, with chemically stabilized soils, additives are placed in bulk with a truck with spreader bar or from bags. ‘The additives are typically mixed in the soil using a rotary mixer and then compacted. However, the equipment  used for traditional soil stabilization  operations (i.e. highway applications) is not available or feasible for RADR scenarios. In 2017, laboratory tests were conducted to evaluate viable chemical stabilization alternatives for use in crater repairs (Johnson et al. 2018). With limited equipment available, an alternative method for mixing and compacting was fielded and proven as a reduced-logistics technology.

Garc,ia  L,  J_ Rowland and J_ s_ Tingle_ Draft_ Rapid Airf ie ld Damage Recovery Program, evaluation of geoee/1-reinforced backfill_ ERDC/GSL TR. Vicksburg MS: U.S. Anny Engineer Research and Development Cente r_

Figure 1.4. Chemically stabilized soill backfill method.

1.2        Objective

The main objective of the work presented in this report was to demonstrate three different crater repair backfill technologies under similar operational conditions and compare their performance. The technologies evaluated included (1) expandable polyurethane foam, (2) mechanical soil reinforcement, and (3) chemical soil stabilization.

1.3        Scope

The three backfill technologies were evaluated during a full-scale field experiment conducted at ERDC in Vicksburg, MS, in July 20 17 _ The exercise consisted of repairing eight simulated craters in two days. Four craters were repaired consecutively each day, and only the backfill and capping steps of the repair procedure were conducted during this experiment for simplicity. All repairs had the same geometric characteristics, i.e., volume and layer thicknesses, as well as the same capping material to capture the relative performance of each technology and support a fair comparison between the technologies.

Simulated F-15E aircraft traffic was applied to each of the completed repairs to verify their ability to meet the operational RADR mission requirements . Detailed performance data were gathered in terms of timing, manpower requirements, and logistical requirements to support a trade-off analysis of these technologies.

2           Crater Repair Test Ma terila  ls

2.1        Constructed subgrade soil

A silty clay (CL-ML) was used as the constructed subgrade material for all of the repairs. This material was purchased from a local quarry and classified according to ASTM D2487 (2017a). Classification data are shown in Appendix A.

2.2        Backfilll soils

Two different backfill soils were used for the mechanical soil reinforcement and chemical soil stabilization meth ods. The mechanical soil reinforcement method used a poorly graded sand (SP) or a silty sand (SM) blend as the backfill materials for individual repairs. For the chemical soil stabilization methods, the SM blend was used as the backfill material. The SP was obtained from a local vendor. The SM blend was produced at ERDC by mixing 33% of ML with 67% of SP (by dry weight of total soil blend). Classification data for the SM and SP are shown in Appendix B.

2.3        Polymers

Polymer emulsions are commonly used for soil stabilization  and  dust control. These matelials coat and physically bind soil particles together to provide stiffness, toughness, and  water resistance. Previously, an in-depth study conducted at ERDC (Johnson et al. 2018) tested four products that included Soil-Sement®, GRT 9000®, Soiltac®, and Top Seal White®.

According to Johnson et al. (2018) , these products produced the highest 7-day UCS of the different polymers tested in the lab, ranging from 401 to 583 psi at 6% moisture content and approximately 3% to 4% polymer by dry weight of soil. Only polymers that were able to produce high st rengths, i.e., UCS greater than 400  psi, at  a concentration  of  3% to  4% were selected for inclusion in the crater repair tests to minimize the amount required in the field. Soil stabilization  with  polymers is a challenge in a small crater environment because the traditional equipment used to mix polymers with soil in place, such as a soil stabilizer reclaimer , is  too  large for use in craters and not available for RADR.

Soil-Sement@, manufactured by Midwest Industlial Supply, is a polymer emulsion  normally used for dust control and soil stabilization. Soil­ Sement@ was chosen for use in this expeliment because of its performance in prior field testing (Edwards draft*) _ Soil-Sement@produced the best results in terms of California bearing ratio (CBR) values as measured by the in-situ CBR method and with a Dynamic Cone Penetrometer (DCP).

2.4        Portlandlimestonecement(PLC)

Portland limestone cement (PLC) is a combination of calcium silicate­ based clinker, limestone, and gypsum and is designated as Type 1L according to ASTM C595 (2019)_ Portland cement can be used to stabilize a vaiiety of so il types; however, high-plasticity soils may not be as beneficial (UFC 3-250-11 20 0 4) _ PLC differs from regular portland cement in that it contains more than 5% limestone and is made by grinding portland cement clinkers with limestone   up to 15%. The   advantage of using PLC is the   reduced carbon footprint during the manufacturing process and increased hydration for cementitious efficiency (http://www.theconcreteproducer.com/how-to/ concrete-production/the­ advantages-of-portland-limestone-cement o).

2.5        Geocells

The geocells used for this investigation were manufactured by Presto – Geosystems. The product, called Geoweb@, is made from high-density polyethylene (HDPE) with cell dimensions of approximately 8.8 in. x 10.2 in. and a cell height of 8 in. The specific product designation is GW20V8. Mechanical properties for the GW20VS geocell used in this experiment are shown in Table 2.1.

Edward, s  L  Dra Fieldevaluationofnext.generationbackfillmaterialsandmethodsfor Airfield Damage Repair_ ER DC/ GSL TR Vicksbur,g MS: U.S. Army Engineer Research and Development center.

Table 2.1. GW20V8 mechanical properties.

2.6              Expanding polyurethane foam

Rigid polyurethane foam consists of two components, i.e., (a) polymeric isocyanate compound and (b) a blend of polyol compounds, catalysts, proprietary additives, and water. The two components are mixed in a 1: 1 ratio (by volume) to form a closed-cell expanding polyurethane foam that rises and sets in approximately 15 min. The foam product used for RADR is Foam-iT! 10 Slow@, which is a water-blown, rigid polyurethane designed to expand approximately six times its initial volume to produce a final density of approximately 10 pcf. At this density, the rigid closed cell polyurethane foam produces a minimum unconfined compressive st rength (UCS) of 200 psi, which is adequate to provide uniform load support for the rapid-setting concrete cap. This material is formulated with a relatively slow reaction rate for easier handling during large-pour operation.

The main benefit of using expanding polyurethane foam for crater backfill is the significant reduction in material requirements, which reduces the logistical footprint of the repair process. As shown in Figure2 .1 , material reductions are such that 4,500 lbofliquid foam would  replace  30,000 lb of backfill soil This is especially important during base recovery operations in remote locations where quality backfill materials are not available and air-deployable technologies are required.

Figure 2.1. Benefits of foam expansion.

As noted, previous research conducted at the ERDC showed that polyurethane foam material is capable of providing sufficient compressive strength (200 psi) to support aircraft traffic under a concrete cap. The material has also been certified under live-flight testing (Priddy et al. 2013, Gurtowski et al. 2016).

2.7              Rapid-setting flowable fill

Rapid-setting flowable fill is a highly fluid mixture of rapid-setting cement, fly ash, fine aggregate, and water. Common applications of commercial flowable fill are for backfilling utility cuts and for use in pipe bedding.  However, traditional  commercial  flowable fi]] products are not rapid setting. The rapid-setting material used in this experiment generally provides an UCS of 250 psi after 30 min of cure time and 750 psi after 3 hr of cure time. Post-trafficking investigations typically reveal a CBR of approximately 100% after 24 hr of cure time. It is packaged in 3,000-lb (1-yd3) super sacks and can be used as backfill material beneath rapid-setting concrete or asphalt.

2.8              Rapid-settingconcrete

The surface material used in this experiment was CTS Rapid Set Concrete Mix®. It is a pre-blended, rapid-setting concrete material that has been identified by the ERDC as appropriate for crater capping due to its high early strength and load carrying capacity after only 2 hr of curing. The main cementitious component is Rapid Set Cement@, a proprietary material.

Since the material is preblended and does not require the addition of local aggregates, it can be mixed in a variety of concrete mixing equipment when water is added. The aggregate in the mix is 3/8-in. maximum-size pea gravel. Due to the fast set time of the mate1ial (10 to 20 min depending on temperature), bulk citric acid can be added to the mix water at elevated temperatures to increase the working time of the material and to prevent it from flash setting in the mixer. CTS Rapid Set Concrete Mix® achieves a UCS in excess of 3,000 psi after 2 hr and over 5,000 psi after 28 days. The material is packaged in 2,900-lb (0.93 yd3) super sacks.

3            Experimental Program

3.1        Experiment layout

The experiment was conducted during the week of July 24-28, 2017, at the Outdoor Pavement Test Facility, on the East Campus of the Wateiways Experiment Station, Vicksburg, MS. A total of eight (8) simulated craters were cut to 8.5-ft square and excavated to a depth of 34 in. (Figure 3.1).

The simulated craters were arranged in two lanes of four (4) craters. Each lane represented a test sequence consisting of one repair per backfill method as shown in Figure 3.2. Figure 3-3 shows the repair profiles. An 11-in.-thick layer of compacted CL-ML was placed on top of the natural foundation at the bottom of the craters to provide a consistent subgrade throughout the craters. Some repairs were replicates, and others had specific variations between two sets of testi ng. For example, the chemic.al

soil-stabilization backfill in Test Sequence 1 consisted of cement-stabilized silty sand (SM), while the repair in Test Sequence 2 was polymer­ stabilized SM. The geocell backfill in Test Sequence 1 used sand (SP) while the geocell repair in Test Sequence 2 used SM. The foam and flowable fill backfills were intended to be replicates for both test sequences. However, some variations occurred during testing as discussed later in this report.

Figure 3.1. Simulated 8.5-ft square craters.

Figure 3.2. Repair layout for the test sequences.

Figure 3.3. Repair profiles.

The target layer thicknesses for this experiment  were 16 in. for the backfill and 7 in. for the rapid-setting concrete cap. These thicknesses deviated from those established in the RADR TfPs (AFCEC 2018). The purpose of using 7 in. of rapid-setting concrete was to increase the impact of the backfill performance for each of the backfill technologies under a thinner cap. A thinner cap was expected to fail quicker and demonstrate the relative performance of the different backfill materials. ERDC already had data for some of the technologies under 6, 8, and 10 in. of ra pid-setti ng concrete.

The 16-in. backfill thickness was selected based on the required geocell heights in UFC 3-270-07, which states that two layers of 8-in. geocells were required underneath a thinner (7 in.) rapid-setting concrete cap. Therefore, a 16-in.-thick backfill was placed for all technologies to provide consistency between repairs. It important to note that this was done to allow meaningful comparisons between the backfills, and it was not intended to be used as recommendations for all backfill layer thicknesses.

The repairs were completed in three days with a team of seven (7) tech ni­ cians, and the repair process consisted of only backfilling and capping the craters. Repairs 1 through 4 were completed on July 26. The backfills in Repairs 6 and 8 were completed on July2 7 , and the backfills in Repairs 5 and  7 were completed on  July 28. The rapid-setting  concrete caps in Repairs 5 through 8 were completed on July 28. The weather conditions during the experiment were generally hot and humid with cloudy skies.

Table 3.1 provides a summary of weather conditions and repair tasks completed each day of the experiment.

Taole 3.1. weatl’ler conditions during the test period each day.

During each test sequence, the teams were divided into two teams of three (3) to work simultaneously on two backfill technologies. Once all four backfills were completed, the full team placed the rapid-setting concrete cap in all craters consecut ively. Simulated F-15E aircraft traffic (112 passes) was applied to the four finished repairs 2 hr after the last repair cap was completed. Additional traffic was applied in the following days, as weather and scheduling allowed.

The following sections provide more details on the crater preparation, backfill construction, capping, and evaluation procedures.

3.2        Crater preparation

3.2.1      Construction of uniform subgrade

Each simulated crater was initially prepared by cutting the PCC surface with a 60 -in . wheel saw attachment (Caterp illar SW6o) on the Caterpillar compact track loader (CTL 279C), which was  representative of cutting around upheaval for a crater repair. Once the repair was cut to size (approximately 8.5-ft square), the  PCC was broken into smaller pieces with a moil-tipped hydraulic breaker (Bobcat HB1180) on a mini excavator (Bobcat E45 T4) _ Large debris and material  underneath the PCC surface were then removed with the bucket attachment on the mini excavator. Each crater was excavated to a minimum depth of 34 in. to ensure that a uniform natural foundation (subgrade) was reached at the bottom of the crater  repair. Dynamic cone penetrometer  (DCP)  (ASTM  D6951 2018) data were collected on the subgrade and correlated to CBR strength values (Table 32. ) . Photographs of this process are shown in Figure 3-4.

A silty clay (CL-ML) material was added in each simulated crater to provide a consistent and stable subgrade throughout the experiment. Each compacted CL-ML lay er was at least 11-in. thick and was finished to a depth of 23 in. below the surface. The material was placed in approximat ely 3-to 4-in.-thick compacted lifts (5-to 6-in.-thickloose) to produce a uniform, constructed subgrade. Compaction of each lift was carried out by applying two coverages with a rammer, commonly called a jumping jack (Multiquip Mikasa MTX70 HD Rammer) compactor. A coverage consisted of start ing compaction in one corner of the repair and working towards the center of the repair in a circular motion, and then returning in a circular motion towards the comer where compaction began. Photograph s of subgrade construction are shown in Figure 3-5.

Figure 3.4. Cutting breaking and excavation process.

Figure 3.5. Subgrade construction.

Nuclear moisture and density measurements (ASTM D6938 2017b) and DCP tests (ASTM D6951 2018) were conducted to verify the stiffness of the natural foundation (natural soil) and constructed subgrade layers and to ensure they were consistent throughout all the craters. Data are shown in Tables 3.2 and 3-3- Natural foundation CBR values ranged from 5 to 11%. The compacted subgrade CBR values ranged from 19 to 27%.

Table 3.2. crater natura,I foundation and compacted subgrade DCP tests results.

Table 3.3. CL:-ML constructed subgrade average nuclear moisture and density.

The surface of the constructed subgrade was surveyed with a rod and level to later verify the thickness of the backfill. Data were collected as shown in Figure 3.6, which is the recommended method reported by Priddy et al. (20 16) . S urvey points were located along three lines Oabeled N, C, and  S  representing North, Center,  and  South, respectively) in the direction of traffic. Survey points were also collected for the cross section, labeled CS in Figure 3.6.

Figure 3.6. Location of survey points.

3.2.2      Instrumentation (EPCs)

One earth pressure cell (EPC) was installed on the surface of the constructed subgrade in each repair to measure the magnitude and changes in pressure below the backfill  material during traffic. The EPCs were installed at the center of each crater, as shown in Figure 3.7, and were connected to a data acquisition box that was placed nearby to the craters. During each test sequence, data were continuously recorded while simulated traffic was being applied to the repairs.

Figure 3. 7. EPC being installed at the center of a repair over the constructed subgrade.

3.3        Backfilll  construction

3.3.1      Chemically stabilized soil backfill! (Repairs 1and 5)

Chemically stabilizing soil with either polymer or cement was used to potentially increase the strength of low quality backfill with minimal equipment and materials. The ability to use locally available soil decreases the logistical footprint in tenns of material required. A polymer, Soil­ Sement@, and PLC were chosen to be the additives for the two chemically stabilized soil backfill crater repairs. A blended SM was used as the soil for the chemical-stabilized backfill, and both the polymer and PLC additives were placed using a similar methodology.

The mixing process required a team of three (3) people. To ensure the correct amount of soil was mixed, the soil was carefully laid out in a 21-ft by 10-ft blend area, at a height of 1 ft with the front-end loader (Figure 3.8). The SM was at a moisture content of approximately 8.2 %, which was slightly above the optimum moisture content of 7.9% (Appendix C).

According to ETL 12-7 (AFCEC 20 12) , it is recommended to use 1% more moisture content than the optimum moisture content. However, based on previous laboratory testing, 5% PLC at an 8% moisture content was recommended for field testing (Johnson et al 2018). To achieve the 5% PLC, 3-34 lb of PLC was added per cubic foot of loose soil A total of 700 lb of PLC was added to the soil for this testing. The PLC was packaged in 50-lb buckets, and half of the total PLC, seven buckets, were incorporated at a time, as shown in Figure 3-9 –

Figure 3.8. Placement of soil in 21-ft by 10-ft airea.

Figure 3.9. Addition of PLC.

To blend the PLC and so il, a tiller attachment (Caterpillar LT13B) on the CTL was utilized (Figure 3.10). Once the CTL with the tiller attachment mixed the  soil with half  of the  required cement , the  front-end loader bucket was used to flip the soil to ensure homogeneity. The remaining half of the  cement was distributed  evenly on  the  soil, and  the CTL with the tiller attachment resumed mixing the soil. After mixing, the front-end loader was once again utilized to flip the stockpile.

Figure 3.10. CTL tiller attael’lrnent (Caterpillar LT13B) used to blend soil with additive.

To blend the soil with the Soil-Sement@ polymer emulsion, a 1/6-hp submersible utility pump was placed into the polymer drum, and a hose was used to distribute the product (Figure 3.11). The polymer was added at a rate of 3% by weight of soil — approximately 0.225 gal of polymer  per cubic foot of loose soil. The SM soil moisture content was approximately 7.5%, whi ch was just under the optimum moisture content of 7.9%. It was important to make sure that the soil is at or below optimum because the polymer emulsion would add moisture to the mixture. To mix the polymer emulsion, the loose soil was laid out similar to the cement-mixing method in a 21-ft-long by 10-ft-wide area at a height of 1 ft.  For  this test, a total of 47 gal of Soil-Sement@ polymer emulsion was used. Initially, approximately 27 gal was sprayed using the submersible utility pump and hose (Figure 3.12). The CTL with tiller attachment mixed the soil, and the front-end loader was used to flip the soil. This process was repeated with the remaining 20 gal of polymer emulsion .

Figure 3.11. Submersible utility pump in 55-gal drum of polymer.

Figure 3.12. Spraying polymer emulsion on loose soil.

For both PLC and polymer mixed soil, the backfill placement process was executed by placing the  mixed soil in four 4-in.-thick compacted lifts to obtain a total compacted thickness of 16 in. The mixed soil was placed by a team of three (3) people using a front-end loader and shovels.  For each lift, the stabilized soil was placed into the crater above the desired level to account for compaction  and  was compacted  with two coverages of a jumping jack (Multiquip Mikasa MTX70 HD rammer) (Figure 3.13). A coverage is defined as starting compaction in one corner, working in a spiral pattern towards the center, and spiraling  back around  to the starting point. The final top surface was compacted with a plate compactor (Northern Industrial JPC-80) for two additional coverages to ensure a flat smface was achieved (Figure 3.14). The  plate compactor was not  a required step but, since this was a controlled test , more attention was focused on the final surface of the backfill to ensure thickness consistency between repairs.

The cement-stabilized soil was placed in the morning and was capped later that afternoon , giving it approximately 4 hr to cure. This cm·e time was a result of the test sequence and not considered a requirement. The polymer­ stabilized soil was placed on one day and that repair was not capped until the following day. In this case, the curing time was approximately 24 hr, but also a result oflogistical planning and not a requirement.

Figure 3.13. Compacting soil.

3.3.2      Rapid-setting flowable fill backfilll {Repairs 2 and 6)

Approximately 3.5 super sacks of dry material were used to complete 16 in. of rapid-setting flowable fill backfill in the 8.5-ft  x  8.5-ft repairs. The material was placed in 6-in. lifts. Each lift consisted of placing one super sack of dry material in the excavated  repair , and  then applying approximately  40  to 50  gal of water to the  surface. The water was allowed to percolate through the dry material by gravity before placing the next lift_ Figure 3-15 shows flowable fill being placed inside Repair 2 using the dry placement method. Figure 3.16 shows the process of adding water to the rapid-setting flowable fill. The finished rapid-setting flowable fill surface is shown in Figure 3.17. Equipment required for rapid-setting flowable fill placement included one telehandler for transporting the super sacks of rapid-setting flowable fill from the staging area to the excavated hole, a water truck with a water flow meter, and assorted hand tools such as rakes and shovels to distribute the dry material in the repair.

Figure 3.14. Final surface preparation with plate compactor.

 Figure3.15. Rapid-setting flowable filll dry placement method.

Figure 3.16. Adding water to rapid-setting flowable filll during dry placement method.

Figure 3.17. Finished rapid-setting flowable fill surface.

3.3.3      Geocell-reinforced backfill (Repaiirs3and7)

Once the constructed subgrade was compacted for Repairs 3 and 7, the horizontal dimensions of the crater were measured so that the geocell sections and geotxtile could be cut to size (Figure 3.18). Tools for cutting the geocell sections and geotextile to the appropriate size included utility knives, a measuring tape, rebar, and hammers.

Figure 3.18. Cutting geocell section to fit crater size.

Once geocell sections were cut, the installation of the geocell-reinforced backfill for each repair was as follows (Figure 3.19).

1. The bottom of the excavation was lined with the nonwoven, 6-oz, needle-p unched, polypropylene geotextile.
2. The first layer of geocells was expanded parallel to the centerli ne. Rebar was placed at all corners of the repair to hold the expanded geocell section in place. Backfill material was used to fill some of the geocell pockets along the edges to help hold the geocell section in place.

3- The geocell section was backfilled using a CTL with a bucket attachment to drop the material vertically into the repair to avoid displacing the geocell section. The geocell section was overfilled by approximately 2 in., and the material was spread evenly with shovels and rakes.

4. The SP backfill in Repair 3 was compacted using the Northern Industrial JPC-80 plate compactor until the sand appeared to be nearly level with the geocell section. The SM backfill in Repair 7 was compacted using the Multiquip Mikasa MTX70 HD Rammer, or jumping jack. Any excess material was struck level with the top of the geocell using shovels and rakes.

Figure 3.19. Geocell-reinforced backfill placement.

5 – The surface of the first layer of geocells was then lined with the non­ woven, 6-oz, needle-punched polypropylene geotextile.

6. The second layer of geocells was expanded perpendicular to the center­ line and held in place, as described in Step2 .

7_   The second layer of geocells was backfilled and compacted, and excess material was removed, as described in Steps 3 and 4.

Data collection on each layer of geocell-reinforced soil was subjected to nuclear moisture and density and OCP test. The surface was also surveyed with a rod and level to determine the thickness of the geocell-reinforced base. The target thickness was 16 in. (two 8-in.-thick layers of geocells). Each repair was capped with Rapid-Set Concrete as described later in this chapter.

3.3.4      Foam backfill (Repairs 4 and 8)

The foam backfill was placed using a prototype automated foam distribution system (Figure 3.20) ,  to  dispense liquid foam material into the excavated repair and then allowing the foam to expand and set before the rapid-setting concrete cap was applied.

Figure 3.20. CRADR 1-L proto type foam dispensing system.

The foam backfill process was as follows.

1. 1.     The excavated repair dimensions were measured and entered into the automated foam distribution system through the touch screen control panel. The required measurements , in inches, were the total width, length, and depth of the repair, as well as the desired cap thickness.

1. 2.     The bottom of the excavation was lined with 10-mil polyethylene plastic sheeting, making sure that the liner was in contact with the excavation walls and comers. Rebar was installed to hold the liner in place (Figure 3.21).

3-   The  foam dispenser (Figure 32.2   ) was operated by following the procedures indicated in the user’s manual. All of the fluid lines and the mixing head were connected in advance to save time during the repair process.  Crater  dimensions  varied slightly between  Repair  4  and Repair 8, thus requiring different  foam  pour times for each crater (Figure 3.23 and Figure 3.24)_ The liquid polyurethane foam mix was dispensed  in  approximately  4 to  6  min. Once the  liquid foam placement was completed , the mixing head was disconnected from the fluid lines and placed in a bucket for draining.

Figure 3.21. PO!yethylene liner applied to Repair 4.

Figure 3.22. Foam placement

Figure 3.23. Placing additional foam in Repair 4 to achieve the desired backfill thickness.

Figure 3.24. Final foam backfill! surface of Repair 4.

1. . The foam was allowed to expand and cure for about 10 min. At that point, the excess liner was cut and removed. When the foam expanded in each excavated repair , an uneven and cracked domed surface was formed (Figure 32. 5). Therefore, the foam surface was inspected to identify areas that required leveling or cutting so that a smooth and level surface was provided for the required 7-in. cap thickness. These domed areas were cut with a reciprocating saw, and the excess material was discarded.

1. On Repair 4, the dispensed foam liquid was not enough to produce a 16-in.-thick layer. Therefore, an additional layer of foam was poured on top. Since the foam surface on the first layer was uneven, the liquid foam was deposited in the low areas creating more of an uneven overall surface. The second layer foam was soft to the touch and did not bond very well to the first layer. No cutting was required on Repair 4.
   1. On Repair 8, the foam was dispensed in a single pour. However, the expanded foam surface was very uneven and cracked. Cutting was required to meet the thickness requirement (Figure 32. 6).

1. The target backfill thickness was 16 in. However, since the expanded foam backfill surfaces were uneven, the backfill thicknesses varied throughout the repairs. The average foam layer thicknesses were 14 in. and 15 in. in Repair 4 and Repair 8, respectively.

Figure 3.25. Expanded foam backfllll surface of Repair 8.

Figure 3.26. Final foam backfill surface Of Repairs (after cutting).

3.4        Repair backfill assessment using light weight deflectometers

After each backfill was completed for each repair, two lightweight deflectometers (LWDs) were used to quantify the relative backfill stiffness by measuring the deflections resulting from a known force. This method consisted of dropping a mass to incur a dynamic load to a plate on the ground, and a buffer was used to simulate vehicle traffic by modifying the impact of the load. The maximum deflection along with the peak force was used to estimate the dynamic modulus, as shown in Equation 1 , where A is the stress distribution factor, v is the soil Poisson ‘s ratio, Wo is the peak vertical deflection, Fpk is t he peak applied load, and ro is the radius of the load plate (Stamp and Mooney2013).

formula

Although the two LWD devices measure the vertical deflection differen tly, Equation 1 was utilized to provide an estimated dynamic modulus for both. Data were collected on all of the crater repairs, with the exception of the foam backfill repai rs. The foam backfill returned an error on both devices, possibly because of the uneven surface.

The Dynatest LWD Model 3031(Figure 3.27) follows ASTM E2583-07 (ASTM 2015a) and ASTM E2835-11 (ASTM 20 15b )_ To operate the device, a 10 – kg load was dropped from a maximum height of 850 mm onto a 300-mm­ diam plate with a rubber buffer. The load cell measured the applied load and peak magnitude, while a geophone measured the deflection in the center of the  device. To run the Dynatest LWD, the load was dropped three times to seat the device, and the subsequent two readings were recorded. The LWDmod software was used to determine the modulus, E, and is reported in Table 3.4 along with the maximum deflection, s.

The Zorn ZFG 3000 GPS (Figure 3.28) follows ASTM E2835-11 and has a 10-k g load that is dropped onto a 20 0 – mm – di am , 40 – mm -thi ck plate to estimate the dynamic  deflection modulus using the measured deflection, s (mm). The maximum pulse force is 70. 7 kN, and the pulse duration is 17 ms. To run the Zorn LWD, three seating drops were completed to seat the device, and the readings from following three drops were averaged by the device to provide the deflection and dynamic deflection modulus for the test location. The ZFG 3000 software was used  to  determine  the  modulus, E, and is reported in Table 3.4 along with the maximum deflection, s.

Figure 3.27. Dynatest LWD Model 3031.

Table 3.4. Backfill LWD deflection and modulus results.

s Measured maximum deflection_

b Una vailable data_

c No data were recorded for the foam backfilled craters of Repairs 4 and 8.

Figure 3.28. Zorn ZFG 3000 GPS.

Results from both devices indicated that cementitious-type backfills, the rapid-setting flowable fill and cement-stabilized backfills, were the stiffest. The Zorn  LWD showed  that both of the  rapid-setting  flowable fill  repairs (2 and 6) were the st rongest, with similar modulus values of 402 and 403 MPa and small deflection values of0 0. 80 and 0.076 mm.The Dyna­ test LWD showed that the stabilized cement and flowable fill have small deflection values of 0.074 mm or less, and modulus values greater than 758 MPa. The   remaining backfill repairs with the  polymer-stabilizer and the geocells had much lower values than the rapid-setting flowable fill and cement-stabilized backfills. The polymer-stabilized backfill showed approximately double the modulus values than the geocell backfill, but the polymer backfill was much less stiff than the cementitious-type backfill_ The polymer backfill required more curing time, and for water to evaporate and increase strength in the polymer-stabilized backfill.

3.5        Rapid-settingconcretecap placement

The rapid-setting concrete cap was placed by a team of seven (7) people including a mixer operator, a forklift operator, two screed operators, and three hand-tool workers. The volumet ric mixer (Figure 3.29) was filled with approximately six super sacks of Rapid Set Concre te Mix® prior to the starting the pour. However, a forklift was required to consistently feed the mixer with Rapid Set Concrete Mix® material to keep enough material weight on the dry material belt to get a consistent mix. The mix was poured into each crater until reaching the surface level. Concrete rakes were used to distribute the  mix throughout  the crater. The screeding process was performed by hand using a straight-edge bar (Figure 3.30). Concrete hand-finishing was performed after screeding by using aluminum or magnesium trowels. Once all repairs in a test sequence were capped, the mixer was pressure-wash ed to  remove residual material. No intermediate wash was required between crater repairs. Figures 3.31 through 3.38 show the finished rapid-setting concrete surfaces for all eight repairs before traffic was applied.

Figure 3.29. Volumetric mixer.

After  trafficking, core samples were  collected from three different locations in each repair to determine the rapid-setting concrete unconfined compressive strength (UCS) and to verify the cap th icknesses. The core samples were collected 56 days after construction.

Figure 3.30. Rapid-setting concrete cap screeding with straight-edge bar.

Figure 3.31. Finished rapid..settingconcrete cap on Repair 1.

Figure 3.32. Finishedrapid-setting concrete cap on Repair 2.

Figure 3.33 . Finished rapid-setting concrete cap on Repair 3.

Figure 3.34. Finishedrapid-setting concrete cap on Repair 4.

Figure 3.35. Finishedrapid-setting concr;ete cap on Repair 5.

Figure 3.36. Finishedrapid-setting concrete cap on Repair 6.

Figure 3.37 . Finished rapid..settingconcrete ea,p on Repait 7.

Figure 3.38. Finished rapid-setting concrete cap on Repair 8.

3.6       Simulated aircraft traffic operations

Simulated aircraft traffic was applied to the repairs 2 hr after the rapid­ setting concrete caps were completed. A single-wheel F-15E load cart (Figure 3.39) equipped with a 36-in.-diam, 11-in.-wide, 18-ply tire inflated at 325 psi was used. The gross wheel load was approximately 35,235 lb. A normally distributed  pattern  of  simulated traffic was applied in a 3_75-ft-wide traffic lane as shown in  Figure 3-39- Traffic was applied by driving the load cart forward and then backward in the same wheel path, then moving laterally approximately one tire width (9 in.) on each forward pass. A total of 16 passes constituted one traffic pattern. After the first 112 passes were completed and  data were collected, the repairs were trafficked until failure or until 3,600 passes. Failure of the repairs was defined as a high-severity shattered slab or spalls that were 2 in. deep. If at any point during trafficking high FOD potential or tire hazards were identified in a given repair, it was considered failed, and traffic was discontinued on that repair.

Figure 3.39. F-15E load cart and typical F-15 normally distributed traffic pattern.

3.7        Data collection

Prior to trafficking, the surface of each repair was inspected for any pre­ traffic distresses. After 112 passes, the surface of each repair was inspected again. The repairs were then inspected at 200- to 500-pass intervals until they failed, or until 3,600 passes were completed. Cracks were measured and marked using paint sticks, and photographs were taken to document the crack/ failure progression. Spalling was measured for length, width , and depth (Figure 3.40 ) . Loose, spalled material was also measured, and the average size of particles was recorded.

Each repair was also surveyed using a  rod and  level prior to trafficking, at 112 pass es , and  at  failure or 3,60 0  passes. The survey data  were used to measure any permanent deformation that occurred and to determine elevation changes at the joint due to removal of spalled material during trafficking. Data were collected as shown in Figure 3.6, with survey points along three lines in the direction of traffic and along the cross section (Figure 3.41).

Figure 3.40. Measuring a spall on repair edge.

A Falling Weight Deflectometer (FWD) was employed in the center of each repair and on a control slab within the test site in order to monitor changes in stiffness of Repairs 1-7 (Figure 3.42). FWD data were not collected on the foam-backfilled repairs 4 and 8. The control slab was not trafficked. FWD data were collected prior to trafficking, after 112 pas ses, and then at 200- to 50 0 -pas s intervals until failure, or when 3,60 0 passe s were complet ed.

Figure 3.41. Survey data, collection.

Figure 3.42. FWD testing on the surface of a repair.

4           Results

4.1        Repairdata summary

Table 4.1 provides a summary of  the  finished  repair layer thicknesses and strength properties. All repairs met the required thicknesses, except for the foam backfill repairs (4 and 8). As was discussed previously, the foam surfaces were uneven with overall thicker center areas and thinner edges. This caused the average thicknesses to be a little lower than required, consequentially causing the rapid-sett ing concrete cap to be thicker around the repair edges. Most rapid-setting concrete cap thicknesses determined by measuring extracted core samples match the thicknesses calculated from the survey data, except for Repair 4_ For this repair, two of the three core samples were extracted near the edges where the rapid-setting concrete was thicker. The thickness from the survey data was calculated by averaging the measurements throughout three profiles and one cross section of the repair .

Table 4.1. Repair data summary.

s    Tested 56 days after cons tr uction.

When performed, DCP tests in the rapid-setting flowable fill backfills resulted in refusals, i.e., penetration was less than 0.08 in. after five (5) blows or the rod was deflected more than 3 in. from the vertical position due to the presence of  aggregat  es. Therefo re, the  rapid-sett ing  flowable fill was considered 100-CBR material. DCP test data showed that the cement-stab ilized SM was  the  second strongest backfill after the flowable fill. The data also indicate that SM soil stabilization using cement produced a better quality backfill than when polymer or geocells were used to stabilize the SM soil. DCP testing may not be indicative of the bearing capacity of geocell confined materials due to the nature of the reinforcement and the small diameter cone.

DCP tes ting was not conducted on the foam backfills (Repairs 4 and 8) during this experiment. However, DCP data from a subsequent field experiment using foam backfill in similar repair conditions showed DCP penetration rates in the foam backfill of 20 to 25   mm /b lows, which is typical of soils with CBR values ranging from 8 to 10% (Bell et al. 20 19 ) _ UCS testing results of core samples obtained from the rapid-setting concrete cap showed that all caps had appropriate strength. Howeve r, in Repair 1,  the  UCS after 56 days was  4,653  psi, which was a lot lower than expected. Typically, CTS Rapid Set Concrete Mix® produces UCS values of over 8,000 psiafter 28  days.

4.2        Backfilll placement times

The backfill placement times are shown in Figure 4.1. The TTPs established a threshold time of 11 min for placing 14 in. of flowable fill using the dry method in 8.5-ft crater repairs. However, the backfills placed during this experiment were 16 in. thick. It is important to note that in this experiment, the repairs were conducted by a crew of engineering technicians under expetimental conditions and not during a simulated operational scenario with a sequential repair process. Therefore, these times will be used in this report only for comparing between backfill technologies.

The standard backfill method for dry placement of rapid-setti ng flowable fill was completed in 20 min in Repair 2 and 22 min in  Repair 6. The methods that took the longest time to complete were the chemical soil­ stabilization backfills, which were completed in 75 min and 90 min in

Repairs 1 and s, respectively. These methods involved laying out soil, adding the chemicals (cement or polymer), mixing the chemicals with the soil using a tiller, flipping the soil to complete the mixing process, placing the soil in the crater in layers, and applying compaction. This process was very time-consumin g and labor-int ensive. The geocell backfill method required less time than the chemical-stabilized soil backfill (43 and 66 min) . However, the process of installing the geocells, dumping the soil in the geocell pockets, compacting, and then repeating these steps for the second layer was still more time-consuming and labor-intensive than the standard rapid-setting flowable fill backfill method.

The fastest method of backfill placement was the foam backfill. In both Repair 4 and Repair 8, it took less than 15 min to complete the foam backfill process of setting up the dispensing equipment, measuring the crater dimensions, installing the plastic liner, and dispensing the liquid foam into the crater. After the  foam expanded , the  excess liner was  cut , and the crew was ready to move forward to the next task. As discussed in Section 3.3.4 on Repair 8, the  foam smface was uneven and  required cutting. This task added 25 min to the process. However, cutting was performed by the project engineers  while trying different  cutting techniques. Therefore, this time was not considered realistic. Previous experiments during simulated operational scenarios (Bell et al. 2019, Mejias-Santiago et al. draft*) have shown that cutting could take from 5 to 10 min. if only the high spots on the foam backfill are  cut  down to ensure an even rapid-setting concrete cap thickness (10 in. minimum).

• • Mejias-Santiag_o  M., J. S. Tingle, J_ L. Johnson. L. A. Gurtowski and C. S. Griggs_ Draft _ Refinementof foam backfill technology for expedient airf ield damage recovery Phase //1: Technology demon stration an d end user evaluation. ERDC/ GSL TR. Vic ksburg, MS: U.S. Army Engineer Research and Development Center.

4.3    Post-traffic results

4.3.1    Pavement distress survey

A summary of the condition and distress data collected throughout traf­ ficking is provided in Appendix C. The results of the pavement condition for all repairs are reported in Figure 4.2. Therepairs withstood varying levels of traffic prior to failing. The repairs with cement-stabilized SM and rapid-setting flowable fill (Repairs 1, 2,  and 6) met the RADR objective pass level for sustainment level repairs. These did not fail, but trafficking was stopped at 3,600 passes. The remaining  repairs sustained at least 500 passes, therefore meeting the RADR threshold pass level for expedient repairs. The geocell-reinfo rced SM and  the  polymer-stabilized SM were the best performers in this category. Some repairs were trafficked after failure was identified to capture additional data.

Figure 4.2.Traffic data summary.

The  main mode of failure was high-severity  spalling along the transverse joints (perpendicular to the direction of traffic). Minor distresses, such as shrinkage cracking after curing , were noted prior to traffic for many of the repairs but did not appear to contribute to their failures. Figures 4.3 through 4.10 show post-traffic  photos for each repair. Figure 4.11 shows shrinkage cracking on Repair 4 prior to trafficking (marked with orange paint stick).

Figure 4.3. Post-traffic photographs of Repair 1 after 3,600 passes.

Figure 4.4. Post-traffic photographs of Repair 2 after 3,600 passes.

Figure 4.5. Post-traffic photographs of Repair 3 after 1,520 passes.

Figure 4.6. Post-traffic photographs of Repair 4 after 830· passes.

Figure 4.7. Post-traffiic photographs of Repair 5 after 1,520 passes.

Figure 4.8. Post-traffic photographs of Repair 6 after 3,600 passes.

Figure 4.9. Post-traffic photographs of Repair 7 after 1,792 passes.

Figure 4.10. Post-traffic photographs of Repair 8 after 1,520 passes.

Figure 4.11. Shrinkage cracks on Repair 4.

4.3.2      Elevation changes

Permanent deformation and changes at the joint due to removal of spalled material were monitored using rod and level measurements in the longitudinal and transverse directions. Maximum measured changes in elevations for each repair are  reported  in Table 4.2. The maximum  change in elevation was recorded for all repairs along the edges where spalls were located (either on the West or East edges). Note that the values in Table 4.2 are less than that measured during the inspection process with a ruler, which are reported in Appendix C. This is because of the  inability to place the rectangular base of the rod in the deepest portions of the spalled areas.

However, Repairs 4, 5, 7, and 8 have maximum changes in elevation that are greater than 1.25 in., which can result in gear damage for the F-15E aircraft. This generally matches the information in Figure 4.2, where these repairs failed by high-severity spalling before reaching 3,60 0 passes.

Repairs 1, 2, and 6 were the best performers and had the lowest values for maximum change in elevation. The survey data also show that there was no evidence of settlement or faulting within the repairs.

Table 4.2. Maximum changes in elevation.

* Survey data was only take n when traffic was discontinued on each repair and not necessarily when the repair was considered failed_

4·. 3 . 3        Impact stiffness modullus

Figure 4.12 shows the impact stiffness modulus (ISM) results obtained from the FWD test data. The plot shows that the repairs lost about 50 % of the stiffness at around 1,000 passes.

 Figure 4.12. FWD results.

After that point, the ISM values stabilized, and no significant change was observed. Repairs with  cement-stabilized soil  and  flowable  fi]l  backfills had similar trends and were the stiffest compared to the rest of the repairs. This was expected, considering the cementitious nature of the  flowable fill and cement-stabilized materials. The polymer emulsion-stabilized SM showed a slightly higher stiffness than the geocell-reinforced repairs.

However, it showed a much lower stiffness than the PLC-stabilized repair. For the geocell-reinforced repairs, Repair 7, which was backfilled with SM soil, was stiffer than Repair 3, which was backfilled with a uniform SP soil.

4.3.4        Earth pressure cell data

Table 4-3 shows a summary of the average peak pressures measured  below the backfill in each repair. Figures 4.13 and 4.14 present the peak pressures measured during each traffic interval applied to the repairs. Each traffic interval represents  a different  day when traffic was applied with the  load cart. Before traffic began each day, the data collection system was turned on and the initial pressure values were recorded. These pressures were then used to  baseline the  data to account  only for the  pressure effects on the EPCs caused by the applied traffic and, thus, to neglect the overburden pressures caused by the weight of the backfill and capping layers.

Table 4.3. Average peak pressures measured on each repair.

Figure 4.13. Peak pressure values for Repairs 1 through 4 during eaeh pass interval applied.

It is important to note that data from Repairs 5 through 8 during passes o through 200 are not available due to technical issues with the instrumentation, as shown in Figure 4.14.

In general, the data show significantly low pressures (less than 10 psi) below the backfill layer across all repairs. This indicates that most of the pressu re applied by the simulated F-15 traffic was well-dissipated by the rapid-setting concrete cap and the backfill. The pressures measured were generally consistent during traffic for most of the repairs.

Figure 4.14. Peak pressure values for Repairs 5 through 8 during each pass interval applied.

Some of the changes in observed pressure were attributed to instrumentation technical issues rather than to repair failure. Since the traffic was applied in an alternate pattern between the two test sequences, the data collection system was moved back and forward to capture the EPC readings during traffic at each test sequence. This resulted in inconsistent baseline pressure values due to inherent equipment error.

It was observed that Repairs 1 and 5 had the highest overall pressures measured. These were the cement- and polymer-stabilized soils. The high pressure could be attributed to the fact that these stabilization methods required longer curing times.

5            Comparison of BackfiH Technolog ies

The test observations indicated that the rapid-setting flowable fill and cement-stabilized soil backfills met the RADR objective of 3,500 passes under a 7-in.-thick rapid-setting concrete cap. However, for the geocell and foam backfills the 7-in.-thick cap was not adequate, as these repairs did not reach the 3,50 0 – pass es RADR objective. It is recommended to use the foam backfill technology with a minimum rapid-setting concrete cap thickness of 10 in. (Johnson et al. 2017; Priddy et al. 2016; Bell et al. 2019; Mejias­ Santiago et al. draft*)_ Further research is required to characterize the geocell­ reinforced backfill performance with a 10-in.-thick rapid-setting concrete cap.

In general, the experiment observations and results indicated that the three backfill technologies evaluated have potential for use in RADR Each technology has trade-offs, and their applicability will depend on the RADR operation scenario. For comparison purposes, a trade-off analysis was conducted to determine the rankings of all backfill technologies evaluated. The criteria used for the analysis included repair logistics (such as equipment, materials, manpower, and repair times, repair performance under simulated aircraft traffic, and material costs). The rapid-setting flowable fill backfill technology was included in this analysis for comparison purposes. Each technology received a ranking value from 1 to 4, where a ranking of 4 means that technology is the best of the four in that specific parameter.

Table 5.1 presents the results of the trade-off analysis. On average, all four technologies share similar rankings. The use of each individual technology depends on the RADR scenario and the availability of backfill materials.

For example, the best performing material under simulated aircraft traffic is the rapid-setting flowable fill. If there were access to take the material in to the RADR location, the recommended backfill technology would be rapid-setting flowable fill. This is the current backfill technology used for RADR, but its limitation is the material footprint and weight. A lighter solution in this case would be the foam backfill technology.  However, it needs to be understood that this technology requires specialized equipment and training. In the case where there is no access to take in

*Mejias-Santi ago, M..J_ s_ Ting,le  J_L    Johnson. L  A_ Gurt owsk i and c_ s_ Griggs_ Draft_ Refinement of foam backfill technology tor expedient airfield damage recovety Phase Ill:Technology demonstration and end user evaluation. ERDC/GSL TR Vicksburg, MS: u_s_ Army Engineer Research and Development Center.

materials and indigenous materials are to be used, cement-stabilized soil would be the recommended backfill technology. Cement is cheap, and it is generally available  in  most countries.  However, the  soil-stabilization process is labor-intensive and time-consuming. A quicker approach for stabilizing indigenous materials would be using geocells for mechanical stabilization. The geocells are lightweight and easy to transport and store. Furthermore, UFC 3- 270 -0 7 allows the  use ofFOD cover in combination with geocells for expedient repairs, requiring a much smaller logistical footprint than rapid-setting concrete. Using a FOD  cover requires less material and less specialized equipment than rapid-setting concrete, which utilizes a volumetri c mixer for placement.

Table5.1.Technologytrade-offanalysis.

a Material costs for 16 in. of backfill in one small crater (96 ft3): flowable fill: $3.200

cement $100 geocells: $600 foam: $3.100

6            Concllusion sand   Recommendations

Tiris expe1iment evaluated three different crater backfill technologies during a full-scale field expeliment to compare their performances and conduct a technology trade-off analysis. The technologies evaluated included (1) expandable polyurethane foam, (2) mechanical soil reinforcement, and (3) chemical soil stabilization. After the repairs were completed, simulated F-15E aircraft traffic was applied to the repairs to verify their ability to meet the operational RADR mission requirements. Detailed performance data were gathered in terms of timing, manpower requirements, and logistical requirements to support a trade-off analysis of these technologies.

The following conclusions and recommendations were determined from the experiment and trade-off analysis.

6.1        Conclusions

1. All of the backfill tested demonstrated an ability to support aircraft traffic. Each technology has tra de-offs, and their applicability will depend on the RADR scenario.
2. The foam backfill technology required the least manpower, time, and logistical footprint of all backfill methods evaluated in this experiment. However, repairs with foam backfill did not meet the 3,500-pass sustainment level requirement for RADR. Foam backfill technology provides a small logistical footprint, but it requires specialized equipment and training.

3- A layer of 16 in. of compacted silty sand stabilized with 5% cement is a backfill material comparable to rapid-setti ng flowable fill for 3,500 simulated  F-15Epasses  under  a 7-in.-thick rapid-setting concrete cap. The cement-stabilized soil and rapid-setting flowable fill backfill were placed on the same day during this testing and were allowed to cure for approximately 4 hr prior to trafficking the initial 112 passes after cap placement. While the time to mix and place the cement-stabilized backfill was longer, the cement-stabilized soil can be used  when flowable fill is  not  available. This method  requires less added material but an available source of soil

4_ This experiment provided new information on the performance of geocell-reinforced soil under a rapid-setting concrete cap. The repair with the geocell-reinforced silty-sa nd backfill performed well under simulated F-15E traffic but did not meet the sustainment-level performance requirements under 7 in. of rapid-setting concrete cap. Geocells provide a small logistical footprint but require significantly more manpower for installation than other backfill repair methods described in this report.

5. Data indicated that SM soil stabilization using cement produced a better quality backfill than when polymer or geocells were used to stabilize the same SM soil.

6.2        Recommendations

1. Only the rapid-setting flowable fill and cement-stabilized soil backfills are recommended for crater repairs with a 7-in. rapid-setting concrete cap thickness to meet the sustainment-level RADAR performance requirement (3,500 passes). The minimum rapid-setting concrete cap thickness recommended for geocell and foam backfills is 10 in.
2. For cement-stabilized silty sand, the moisture of the soil should be on the wet side with up to 10% moisture content . If not enough moisture is present , water can be added when compacting. The lack of adequate water content causes problems in achieving a good compaction and also may not provide enough moisture for hydration of PLC.

References

Air Force Civil Engineer Center (AFCEC). 2012. Repair of cement-stabilized soil surfaces.

Engineering Technical Letter, ETI..12-7. Panama City, FL: Tyndall AFB.

Air Force Civil Engineer Center (AFCEC). 2018. Interim tactic.s, techniques, and proce­ dures (1TP), interim process for Rapid Airfield Damage Recovery (RADR).

Revision 11.2. Panama City, FL: Tyndall AFB.

American Society for Testing and Materials International (ASTM). 2018. Standard test method for use of the dynamic cone penetrometer in shallow pavement applications. Designation: D6g51. West Conshohocken, PA: American Society for Testing and Materials.

ASTM International. 2015a. Standard test method for measuring deflections with a light weight deflectometer (LW’l))_ Designation: E2583-07. West Conshohocken, PA: American Society for Testing and Materials.

ASTM International. 2015b. Standard test method for measuring deflections using a portable impulse plate load test device. Designation: E2835-11. West Conshohocken, PA: American Society for Testing and Materials.

                . 2017a. Standard practice for classification of soils for engineering purposes (Unified Soil Classification System). Designation: D2487. West Conshohocken, PA: American Society for Testing and Materials.

                . 2017b. Standard test methods for in-place density and water content of soil and soil-aggregate by nuclear methods (shallow depth). Designation: D6g38. West Conshohocken, PA: American Society for Testing and Materials.

                 . 2019. Standard specification for blended hydraulic cements. Designation: C595. West Conshohocken, PA: American Society for Testing and Materials.

Bell, H.P., L. Edwards, W. D. Carruth , J. S. Tingle, andJ. R Griffin. 2013. Wet weather craterrepairtestingatSilver F1.agexercisesite,TyndallAirForceBase, Florida. ERDC/ GSL TR-13-42. Vicksburg, MS: U.S. Army Engineer Research and Development Center.

Bell, H.P., B. C. Cox, L. Edwards, L. I. Garcia, N. R  Hoffman, M. Mejias-Santiago, and J.

L. Johnson. 2019. Rapid airfield damage recovery technology integration experiment. ERDC/GSL TR-19-8. Vicksburg, MS: U.S. Army Engineer Research and Development Center.

Carruth, W. D., L. Edwards, H.P. Bell, J. S. Tingle, J. R Griffin, and C. A. Rutland. 2015. Large crater repair at Silver Flag exercise site, Tyndall Air Force Base, Florida. ERDC/ GSLTR-15-27. Vicksburg, MS: U.S. Army Engineer Research and Development Center.

Concrete Construction. 2014. The advantages of Portland-limestone cement. http://www.theconcreteproducer.com/how-to/concrete-production.

Department of the Army, the Navy, and the Air Force.200 3 . Ai ,fi.elddamage repair. Unified Facilities Criteria UFC 3-2 70-0 7. Washington, OC: Department of the Army, Navy, and Air Force.

          . 2004. Soil stabilization for pavements. Unified Facilities Criteria UFC 3-250-

11. Washington, DC: Departm ent of the Army, Navy, and Air Force.

Edwards, L , H. P. Bell, W. D. Carruth, J. R. Griffin, and J. S. Tingle. 2013. Cold weather crater repair testing at MalmstromAir Force Base, Montana. ERDC/ GSLTR- 13-32. Vicksburg, MS: U.S. Army Engineer Research and Development Center.

Gurtowski. L, M. Mejias-Santiago, C. Griggs, J _ J ohns       on , B. Ruiz, and D. Felt. 2016.

Refinement of foam backfill technology for expedient ai,fi.eld damage recovery Phase I: Laboratory evaluation of foam materials. ERDC/ GSLTR-16-16. Vicks­ burg, MS: U.S. Army Engineer Research and Development Center.

Johnson, J. L, M. Mejias-Santiago, L.A. Gurtowski, C. Griggs, and C. A Rutland. 2017. Refinement of foam backfill technology for expedient airfield damage recovery Phase II: Development of prototype foam dispensing equipment and improved tactics, techniques and procedures. ERDC/ GSLTR-17-14. Vicksburg, MS:

U.S. Army Engineer Research and Development Center.

Johnson, J. L , C. A. Weiss , L Edwards, J. S. Tingle, W. S. Hammond , and T. J. Talbot.

2018. Laboratory evaluation of next-generation backfill materials and methods for ai,fi.eld damage repair. ERDC/ GSLTR-18-33. Vicksbur g, MS: U.S. Army Engineer Research and Development Center.

Priddy, LP., H. P. Bell, L. Edwards, W. D. Carru th , an d J . F. Rowland. 2016. Evaluation of the structural pe1fonnance of C1S Rapid Set Concrete Mix®. ERDC/ GSL

TR-162-0 . Vicksburg, MS: U.S. Army Engineer Research and Development Center.

Priddy, L P., J _ S. T ingle, M. C. Edwards, J _ R. Griffin, and T. J . McCa:ffrey.2 01 3 . CRA1R technology   demonstration: Limited operational utility assessment2     ,   Tyndall Air Force Base, Fl.orida. ERDC/ GSLTR-13-39 . Vicksb urg , MS: U.S. Army Engineer Research and Development Center.

Read, D. L., and P. E. Dukes. 1988. Sand grid base course Phase I,for fiberglass mat crater repair. Report RRR-TR-88 -01. Tyndall Air Force Base, FL: The Air Force Engineering and Services Center .

Stamp, D., and M. Mooney. 2013. Influence of lightweight  deflectometer  characteristics on deflection measurement . Geotechnical Testing Journal 36(2).

Tingle, J. S., L. P. Priddy, M. C. Edwards, C. A Gartrell, and T. J . McCaffrey. 2009.

Critical runway assessment and repair (CRA1R) technology demonstration: Limited operational utility assessment 1 (LOUAl), Tyndall Air Force Base, Fl.orida. ERDC/ GSLTR-09-12. Vicksburg, MS: U.S. Army Engineer Research and Development Center.

United States Air Force (USAF). 1992. Crushed-stone cater repair and line-of-sight profile measurement for rapid runway repair. TO 35E2-5-1. Washington, DC: USAF.

Webster, S. L. 1986. Sand-grid demonstration roads constructedforJLOTS II tests at Fort Story, Virginia. Technical Report GL-86 -19. Vicksburg, MS: U.S. Army Engineer Waterways Experiment Station .

Appendilx A: Constructed Subgrade Soill Classification Data

Appendix B: Backfill Soils’ Classification Data

Appendix C: Pavement Distress Data