Chapter 17 — Ground Improvement Methods

Ground improvement is a fundamental tool in the railroad geotechnical consultant's toolkit, employed when in-situ subgrade conditions cannot support track, structure, or facility loads without unacceptable settlement, slope instability, or bearing capacity failure. Railroad projects present a concentrated set of ground improvement challenges: soft coastal and alluvial clays crossed by CSXT and CN mainlines, expansive Prairie clays along BNSF and CPKC corridors, loose liquefiable sands requiring treatment for UPRR and Amtrak facilities, and organic deposits that underlie historic rail alignments across multiple clients.

This chapter presents methods for: (1) densification by compaction — mechanical, dynamic, and vibratory; (2) consolidation acceleration using prefabricated vertical drains (PVDs) with staged embankment loading; (3) stone column reinforcement with the Priebe method for composite stiffness; (4) deep soil mixing (DSM) for soft clay improvement and liquefaction mitigation; (5) grouting — permeation, compaction, and jet grouting; (6) lime and cement stabilization; and (7) geosynthetic reinforcement for embankment and subgrade applications. All equations are in English units.

Railroad ground improvement projects carry constraints not present in highway or building construction:

• Active operations: Most ground improvement beneath or adjacent to active track must be designed for minimal vibration transmission to avoid track geometry disturbance. Dynamic compaction and heavy drop weights are generally prohibited within 100 ft of active track without client-approved vibration monitoring protocols.
• Access limitations: Narrow railroad rights-of-way — often 50–100 ft between fence lines — limit crane swing radii, restricting some methods (stone columns, auger-cast piles) to compact equipment configurations.
• Track geometry tolerance: Post-improvement differential settlement criteria are driven by FRA track class tolerances (Chapters 12 and 18), not building codes. Methods must be designed to achieve residual differential settlement within those limits.
• Rapid execution: Emergency and programmatic track improvement projects often have compressed schedules. Methods selected must be compatible with railroad-standard maintenance windows (e.g., 4-hour possession on mainline, weekend gang windows on secondary).

Reference graphics · click to zoom

Improving the formation layer under an existing track

Many old lines were built without a formation layer. When speed or axle load is raised, the simplest fix is to thicken the track bed during maintenance, but clearance limits or stress targets often rule that out, and the formation layer itself must be improved or installed. The UIC methods compiled by Profillidis differ mainly in possession time, excavation depth and output:

Table 17-5. Methods for improving the formation layer under traffic (summary)

Method	Excavation below top of rail	Speed on improved track (mph)	Track removed?	Output (ft/h)
Earthmoving machinery, short sections (< 330 ft)	4 ft	25–50	Yes	500
Earthmoving machinery, longer sections (> 330 ft)	4 ft	43–62	Yes	500
Track-laying machinery	4 ft	43–62	No (yes with geotextile)	66–130
Ballast screening machine	33 in.	25–37	No	295
PUSCAL II formation-improvement machine	4.6 ft	31–50	Yes	59
PM 200 formation-improvement machine	3.6 ft	43	No	115–150
Thermal insulation sheets laid by ballast screener SPX	33 in.	25	No	330

Excavating deeper than 4 ft makes the stability of the adjacent track critical. Full table, with possession hours and man-hours, in Figure 17-11.

Formation rehabilitation machines insert a new sand–gravel formation layer at the top of the subgrade without removing the track. They are the substructure equivalent of ballast cleaning, for sites where subgrade bearing capacity, not ballast condition, is the problem.

Figures · click to zoom

After Profillidis (2014), §9.7.2, §16.8, Table 9.4; after UIC Code 722R. Full reference under Supplemental Sources.

Subgrade stabilization categories and costs

Subgrade stabilization follows the same logic as slope work. Rehabilitate drainage first, then reinforce the substructure below the ballast and the track structure above it, and improve the ground itself where the problem runs deep.

Table 17-6. Remedial categories and indicative costs (course figures, 2016 dollars)

Category	Examples	Indicative cost (2016 $)
Improve structure above track	Eliminate joints, heavier rail, swamp ties	Swamp ties ~$670,000/mile [2012: ~$600,000]
Improve substructure below ballast	Undercutting / ballast replacement, geogrid, HMA underlayment	Undercutting ~$280,000/mile [2012: ~$250,000]
Improve subgrade strength	Shear key, cross trench, cement or grout injection, remove and replace	Site-specific
Underpinning	Micropiles, screw piles	Site-specific
(Reference) new construction		~$6,700,000/mile [2012: ~$6,000,000]

Cost basis: The course figures were in 2012 dollars. They are escalated here to 2016 dollars with the ENR Construction Cost Index (20-city annual average): 10,338 (2016) ÷ 9,308 (2012) = 1.111, an increase of about 11%. Values are rounded; 2012 values are shown in brackets. To move to another year, multiply the 2012 value by CCI_year ÷ 9,308. Railroad work (track time, work trains, flagging, remote access) can vary by far more than the index change, so treat these as order-of-magnitude figures.

Figures · click to zoom

After Geotechnology for Railroaders Program, Module 6.2; §01. Full reference under Supplemental Sources.

17.2.1 Proctor Compaction Fundamentals

The relationship between dry unit weight and moisture content is defined by the Proctor test. The maximum dry unit weight γ_d,max and optimum moisture content w_opt are the design targets for compaction control (Chapter 15).

Zero Air Voids (ZAV) line:

γ_ZAV = G_s ⋅ γ_w1 + w ⋅ G_s    (pcf)

Where G_s = specific gravity (≈ 2.70 for most soils); γ_w = 62.4 pcf; w = water content (decimal).

Relative compaction (RC):

RC = γ_d,fieldγ_d,max × 100%

Standard Proctor (ASTM D698): 12,400 ft·lb/ft³ compactive effort; appropriate for light to moderate railroad facility construction. Modified Proctor (ASTM D1557): 56,250 ft·lb/ft³ compactive effort; required for intermodal yards, bridge approach fills, and all UPRR, BNSF, CPKC, and CN embankments.

Reference graphics

Reference graphics · click to zoom

17.2.2 Vibratory Compaction

Vibratory compaction (smooth drum, padfoot) densifies granular fills and coarse-grained subgrades. Compaction effectiveness is described by the influence depth Z_eff:

Z_eff = k √WA_f    (ft)

Where W = drum weight (lb); A_f = drum contact area (ft²); k = empirical constant (0.08–0.12 for granular materials).

Lift thickness limits by soil type:

• Granular (GW, GP, SW): 12–18 in. uncompacted
• Silty/clayey soils (SM, SC, ML, CL): 8–12 in. uncompacted
• Heavy clay (CH): 6–8 in. uncompacted (padfoot drum required)

Reference graphics

17.2.3 Dynamic Compaction

Dynamic compaction densifies loose fills and granular soils by repeated high-energy drops of a heavy pounder (10,000–30,000 lb) from heights of 30–90 ft.

Influence depth (Menard & Broise 1975):

D = n √W_h ⋅ H    (ft)

Where:

• W_h = weight of pounder (tons)
• H = drop height (ft)
• n = empirical efficiency factor (0.35–0.65; 0.50 typical for saturated granular soil)

Energy per unit area:

E = n_drops ⋅ W_h ⋅ HA_grid    (ton·ft/ft²)

For railroad applications, dynamic compaction is restricted to areas beyond client setback from active track (Table 17-1). Vibration monitoring (Chapter 15) with a peak particle velocity (PPV) threshold of 0.5 in./s at the nearest track is required by UPRR and BNSF.

Reference graphics

Reference graphics · click to zoom

Technology sheets for this method · see 17.16 Technology Library

17.16.42Traditional Compaction17.16.12Deep Dynamic Compaction17.16.37Rapid Impact Compaction17.16.28High-Energy Impact Rollers17.16.29Intelligent Compaction17.16.45Vibrocompaction17.16.4Blasting Densification

Case study: 17.17.1 Measuring roller force with pile-driving monitors (Look 2023), in §17.17 Case Studies.

Wick drains in rail work

Wick drains give trapped pore water a short vertical path out, accelerating consolidation of soft saturated ground that would otherwise take years to drain. Several mandrels can work a large site at once.

Figures · click to zoom

After Geotechnology for Railroaders Program, Module 6.2; §02. Full reference under Supplemental Sources.

Technology sheets for this method · see 17.16 Technology Library

17.16.36Prefabricated Vertical Drains and Fill Preloading17.16.43Vacuum Preloading with and without Prefabricated Vertical Drains17.16.15Electro-Osmosis

Case study: 17.17.2 Wick-drain rigs on dredged sand over 104 psf clay: four ways to size the platform (Look 2023), in §17.17 Case Studies.

17.7.1 Lime Stabilization

Hydrated lime (Ca(OH)₂) added to fine-grained soil produces immediate PI reduction through cation exchange and flocculation, and long-term pozzolanic strength gain.

PI reduction rule of thumb:

PI_final ≈ PI_initial - 5 ⋅ L_p

Where L_p = lime percentage (by dry soil weight, %). Target: PI_final < 12 for use as subgrade; PI_final < 20 for use as embankment fill. Minimum lime content: 3%; maximum practical: 8%.

Strength gain timeline:

Curing Time	q_u (psi) — Typical CL/CH Clay	q_u (psi) — ML/MH Silt
1 day	15–30	10–20
7 days	30–80	20–50
28 days	50–150	35–90
1 year	100–300	70–200

Sulfate precaution: For sulfate content SO₄ > 2,000 ppm (ASTM D4327), lime-treated clay may experience delayed ettringite formation causing swell of 3–10%. In such cases, use cement or fly-ash binder instead.

Reference graphics

17.7.2 Cement Stabilization

Portland cement (ASTM C150, Type I/II) or fly ash (ASTM C618, Class C or F) stabilizes both granular and fine-grained soils through cementation. Design target: q_u ≥ 200–300 psi at 7 days (ASTM D1633) for pavement base applications (Chapter 18).

Mix design proportioning:

q_u = f[c_cement, soil type, w_c, t_cure]

Determined from laboratory trial mixes at cement contents of 4%, 6%, 8%, 10% by dry weight. Minimum cement content for railroad fill: 4% (AREMA MRE Chapter 1 Specification).

Reference graphics

Technology sheets for this method · see 17.16 Technology Library

17.16.7Chemical Stabilization of Subgrades and Bases17.16.2Bio-Treatment for Subgrade Stabilization17.16.31Mechanical Stabilization of Subgrades and Bases17.16.16Excavation and Replacement17.16.3Beneficial Reuse of Waste Materials17.16.34Onsite Use of Recycled Pavement Materials

Case study: 17.17.3 Lime-treated local soil in a reinforced slope: choosing by total internalised cost (Look 2023), in §17.17 Case Studies.

Project management · Bull (2015)

Life-Cycle Selection Among No Treatment, In-Place Stabilization and Remove-and-Replace

In Bull’s volume, Praticò, Saride and Puppala (Chapter 3) build a life-cycle cost model for choosing among three ways of treating a weak subgrade:

• NS: leave the soil as it is.
• S: stabilize it in place.
• SUB: excavate and replace it.

Each alternative has an expected life that comes from its mechanical properties. Its present value sums three kinds of cost:

• Agency costs: construction, rehabilitation and maintenance.
• User costs: work-zone delay and vehicle operating costs.
• Externality costs: emissions from hauling, quarrying and landfill.

Over an infinite horizon, where the first rehabilitation falls at year E and repeats every O years, the agency present value is:

PV = C + C_RE × (1+i)^−E / [1 − (1+i)^−O]

The “gain” of one alternative over another is simply the difference in their present values.

The authors tested the model on two low-volume roads at a 4% discount rate. In the Italian case, a lime-treated CL subgrade was given a 20-year life against 15 years untreated. Stabilization beat substitution when stabilization was cheap. User costs were only about 1/50 of agency costs at those traffic levels. User gains turned negative once the treated alternative’s work zone lasted about 1.5 times as long as the untreated case. Stabilization paid at higher traffic. At very low traffic, remove-and-replace cost more than doing nothing. In the Texas case, lime–cement treatment turned a CH soil with 6.3% free swell into a non-swelling CL. Its unconfined strength rose from about 13 to 250 psi, and its resilient modulus from about 6 to 35 ksi.

For railroad work: the same three choices come up at most chronic subgrade sites:

• NS: keep surfacing.
• S: stabilize in place, by lime or cement treatment or by injection.
• SUB: undercut and replace with subballast.

The perpetual-cycle factor shows how much each added year of life is worth. At 4%, (1.04)⁻¹⁵/[1 − (1.04)⁻¹⁵] = 1.25, while the same factor at 20 years is 0.84. So a treatment that stretches the maintenance cycle from 15 to 20 years cuts the present value of all future rehabilitation by about a third. Replace the highway user costs with train-delay and slow-order costs. Count the haul trips, quarried aggregate and disposed spoil or fouled ballast as externalities. Then test how the ranking changes with treatment cost and traffic (MGT), as the authors do.

Source: Bull, J. W. (ed.) (2015). Life Cycle Costing for the Analysis, Management and Maintenance of Civil Engineering Infrastructure. Dunbeath, Scotland: Whittles Publishing, pp. 66–77. Summarized and paraphrased; figures reproduced for internal reference. Full reference under Supplemental Sources.

17.8.1 Geotextile Filtration and Separation

Geotextiles placed at subgrade/subballast and subgrade/subbase interfaces prevent fines migration (filtration) and maintain aggregate performance (separation). Key design criteria:

Apparent Opening Size (AOS) criterion:

AOS ≤ D_85,soil    (retention criterion)

Permeability criterion:

k_geotextile ≥ k_soil    (permittivity criterion)

AREMA MRE Chapter 1 specifies minimum permittivity of 0.02 s⁻¹ for track subgrade geotextiles. NSR, CSXT, and BNSF all reference this criterion in their earthwork specifications.

Reference graphics

Reference graphics · click to zoom

Geotextiles and geomembranes under track: what the field record shows

Use a geotextile for what it can do.

• It is poor at reinforcement. It needs large stretch to work, creeps, slips against ballast, and its openings enlarge as it stretches. Use geogrid for reinforcement.
• Its main role is separation, with filtration of seepage and in-plane drainage as secondary roles.
• It does not shed water from above.

Requirements.

• Only nonwoven fabrics are acceptable.
• AOS < D₈₅ of the protected soil. For a gap-graded soil, base this on the fines fraction. For new ballast over an undercut surface, use the gradation of the fouling material and ignore the ballast-size particles.
• High tensile, puncture, elongation, abrasion, chemical and UV resistance.
• Proven resistance to clogging, and a thick mat where transmissivity is needed.

Nonwovens fine enough for silt, and especially clay, may not exist.

Placement.

• At least 12 in. below the tie bottom, out of reach of tamping tines and undercutter chains.
• On a smooth surface sloped to drain, with no equipment running on the fabric and minimal UV exposure.
• Never in double layers, which create a slip plane.

Durability record.

• On Norfolk Southern, fabric permeability fell 92% and transmissivity 85% within 15 months. Tensile strength fell 54% within 48 months.
• At an AAR site near Caldwell, Texas, a clogged low-permittivity fabric trapped water, causing excess pore pressure and subgrade failure. It was believed to be the cause of a derailment.

“Undercut and add fabric” repairs often fail because the mud was not coming from the subgrade, or because the subgrade fines were too fine for the fabric. The book’s process is: diagnose the cause in the field and lab; compare geotextile and non-geotextile options; do an economic analysis; run a field trial; monitor and document. Only a properly graded, durable granular layer gives permanent separation and filtration (Chapter 2, §2.5.4).

Geomembranes. Uses:

• Blocking surface infiltration. Place the membrane between sand layers at least 2–4 in. thick, never against ballast, and drain the water off its top.
• Keeping expansive soil from drying.
• Cutting capillary rise in frost-susceptible embankments.
• Forming a cut-off below the water table in cuts.

Hazards: water trapped beneath the membrane creates a slip surface, and holes or leaky seams concentrate flow and erode soil, which is worse than having no membrane. Most leaks are at seams, so require seam testing. A bitumen emulsion sprayed into a sand or subballast layer, which can be done behind an undercutter, is a cheaper alternative with the same trapped-water risk.

After Selig & Waters (1994), ch. 12. Full reference under Supplemental Sources.

17.8.2 Geogrid Reinforcement

Biaxial geogrids (ASTM D6637) placed at subgrade/subbase interface improve load distribution and reduce differential settlement. The Giroud-Han (2004) method for granular layer thickness over weak subgrade:

h = P / (A_p ⋅ p_s) - 1K_R ⋅ (1 + ε ⋅ h/√A_p)    (iterative; ft)

Where:

• P = wheel load per wheel (lb)
• A_p = contact area per wheel (in.²); for intermodal reach stacker: A_p ≈ 100 in.²
• p_s = allowable subgrade stress = 3.2 × s_u for non-reinforced; 3.2 × s_u × TBR for reinforced
• TBR = traffic benefit ratio (1.5–3.0 for biaxial geogrid)
• K_R = load distribution factor (depends on h, A_p, and layer modular ratio)

For preliminary design, the simplified California chart method (USACE TM 5-818-8) is also acceptable when subgrade CBR and geogrid type are known.

Reference graphics

17.8.3 Geosynthetic-Reinforced Load Transfer Platform (GRLT)

Over piled foundations or column-supported embankments (stone columns, rigid inclusions), a Geosynthetic-Reinforced Load Transfer Platform (GRLT) distributes track loads to column heads. GRLT design follows BS 8006:2010 or FHWA-HRT-17-097 (Giroud-Han approach):

Membrane tension in geosynthetic over column caps:

T_geosynthetic = W_T ⋅ s²8 ⋅ δ_allow    (lb/ft)    (catenary approximation)

Where W_T = net soil weight transferred to geosynthetic (psf); s = clear span between column caps (ft); δ_allow = allowable deflection (in.); typically δ ≤ s/30 for railroad applications.

Reference graphics

Geotextiles in railway subgrades

• Types: woven (two perpendicular fiber sets, strongly anisotropic) and non-woven (random fibers, isotropic); polypropylene or polyester; 0.016–0.12 in. thick; 2.1–10 oz/yd².
• Functions: separation of granular layers, reinforcement of weak soil, filtration and drainage, and some frost protection. Most railway installations are simply separators between subballast and subgrade.
• Placement: under the subballast, never directly under ballast. A geotextile does not spread load, so it must not be used as a reason to reduce ballast or subballast thickness. Railways that did so saw puncture of the fabric and loss of the crossfall.
• Benefit: a finite-element study found about a 10% reduction in stress at the top of subgrade (Figure 17-12). Maintenance savings were reported to repay the geotextile cost quickly.
• Acceptance: verify tensile strength, elongation at failure, puncture and compressive strength, permeability and filtration against the soil before use. Filter criteria are in Chapter 2, §2.8.

Figures · click to zoom

After Profillidis (2014), §9.15, Fig. 9.20. Full reference under Supplemental Sources.

Geosynthetics, HMA underlayment and EPS: case histories

Geogrid can be placed before laying track or during out-of-face work with an undercutting chain. A hot-mix asphalt layer between 12 in. of ballast and 6 in. of compacted sub ballast spreads load and sheds water to the sides.

• Rosetown Sub Mile 63.65: subgrade fines contaminated the ballast. Finger tests confirmed plastic fines, and bearing calculations sized the fix: excavate, place full-width woven geotextile, and backfill with clean granular fill.
• Drummondville Sub Mile 42.7: decomposed corduroy and organic soil pumped mud. The fix was a shear-key trench along the shoulder backfilled with clean ballast (ideally with nonwoven geotextile), shoulders regraded to drain, and the trench left to drain ballast pockets toward the clay below.
• Bala 47: woven geotextile placed beneath a lifted track panel at night within a possession window.
• EPS lightweight fill (about $670/m³, or about $19/ft³, in 2016 dollars; about $600/m³ in 2012) reduces load on soft ground, but it can float in a flood unless ballasted and drained.

Figures · click to zoom

After Geotechnology for Railroaders Program, Module 6.2; §03–05. Full reference under Supplemental Sources.

Technology sheets for this method · see 17.16 Technology Library

17.16.21Geosynthetic Reinforced Embankments17.16.20Geosynthetic Reinforced Construction Platforms17.16.22Geosynthetic Reinforcement in Pavement Systems17.16.23Geosynthetic Separation in Pavement Systems17.16.25Geosynthetics in Pavement Drainage17.16.19Geocell Confinement in Pavement Systems17.16.17Fiber Reinforcement in Pavement Systems17.16.18GeoWEB17.16.24Geosynthetics - GeoTerra17.16.35Partial Encapsulation

Case study: 17.17.4 Working platform at the edge of a creek, designed for slope stability (Look 2023), in §17.17 Case Studies.

Worked Example 17-1: PVD Drain Spacing Design — CSXT Georgia Coastal Embankment

Project: CSXT mainline embankment widening, coastal Georgia. Soft CH clay, H = 30 ft. c_h = 0.012 ft²/day; c_v = 0.006 ft²/day. Target: U_total ≥ 90% in 120 days to allow next fill stage.

Step 1 — Vertical consolidation (Terzaghi):

U_v ≈ 6.4% at T_v = 0.0032 (negligible; vertical drainage almost zero in 120 days for 30-ft layer).

Step 2 — Required radial consolidation (from combined equation):

Step 3 — Required T_r:

Step 4 — Trial drain spacing (s = 5.0 ft, triangular grid):

d_e = 1.05 × 5.0 = 5.25 ft; d_w = 0.196 ft; ignore smear and resistance:

Step 5 — Tighten spacing to s = 2.0 ft, triangular grid:

d_e = 1.05 × 2.0 = 2.10 ft; μ = ln(2.10/0.196) − 0.75 = ln(10.71) − 0.75 = 2.371 − 0.75 = 1.621

Step 6 — s = 1.5 ft, triangular grid:

d_e = 1.05 × 1.5 = 1.575 ft; μ = ln(1.575/0.196) − 0.75 = ln(8.035) − 0.75 = 2.084 − 0.75 = 1.334

Step 7 — s = 1.25 ft triangular:

d_e = 1.05 × 1.25 = 1.3125 ft; μ = ln(1.3125/0.196) − 0.75 = ln(6.696) − 0.75 = 1.902 − 0.75 = 1.152

Recommendation: Install PVDs at 1.25-ft triangular grid spacing to achieve U_total ≥ 90% in approximately 102 days (within 120-day construction window). CSXT requires PVDs installed by displacement mandrel (static push, no vibration) to avoid disturbing soft clay structure. Piezometers at 3 depths to monitor dissipation; continue monitoring until U_r ≥ 90% confirmed before next fill stage.

Worked Example 17-2: Stone Column Design — BNSF Intermodal Yard, Gulf Coast

Project: BNSF Port Arthur container laydown area; CH clay subgrade, s_u = 450 psf; c_v = 0.005 ft²/day; H_clay = 18 ft. Applied stress from 4-high container stack: Δσ = 1,800 psf. Target: reduce post-improvement settlement to ≤ 6 in. (BNSF EI laydown area).

Step 1 — Untreated settlement estimate:

σ'_0 = 1,100 psf (mid-layer); C_c = 0.45; e_0 = 1.20:

Step 2 — Required improvement factor n₀:

Required: S_improved ≤ 6 in.; n₀ ≥ 18.6/6 = 3.1

Step 3 — Select stone column configuration:

Try D_c = 30 in. (2.5 ft), s_col = 5.0 ft square grid:

Using Priebe (1995) with φ'_col = 40°, K_a = 0.217; simplified improvement factor estimate:

For preliminary design, use empirical chart (Priebe 1995): a_s = 0.196, n₀ ≈ 2.5 (from chart, φ'_col = 40°).

Step 4 — Revised spacing:

Try D_c = 36 in. (3.0 ft), s_col = 5.0 ft: a_s = π(3.0)²/4/(25.0) = 7.069/25.0 = 0.283 → n₀ ≈ 3.2 (Priebe chart, a_s = 0.283)

Recommendation: 36-in. diameter stone columns at 5-ft square grid, installed to H_clay = 18 ft. Crushed stone fill: 3/4-in. to 3-in. angular aggregate, φ'_col ≥ 40°. Load test (ASTM D1143) on one production column per 100 columns, minimum 5 tests. Settlement monitoring plate at each quadrant of stacking area to confirm performance. BNSF EI requires post-improvement CBR ≥ 8% at subgrade surface; verify by plate bearing test (ASTM D1196).

Worked Example 17-3: Lime Stabilization Design — CPKC Prairie Clay, Moose Jaw, SK

Project: CPKC MOW access road subgrade; CH clay (LL = 66, PL = 20, PI = 46); sulfate content = 850 ppm (acceptable for lime). γ_d,max = 93 pcf (Modified Proctor); w_opt = 22%. Target: PI_final < 12 for use as stable subgrade.

Step 1 — Lime content selection:

PI reduction rule: PI_final ≈ PI_initial − 5 × L_p

Target PI_final = 12: 12 = 46 − 5 × L_p → L_p = (46 − 12)/5 = 6.8% → Use 7% hydrated lime (rounding up)

Step 2 — Proctor mix design check:

Laboratory Proctor on lime-treated mix at L_p = 7%: γ_d,max,lime ≈ 88 pcf (typically 5–7 pcf reduction from lime addition); w_opt,lime ≈ 24% (moisture optimum increases with lime).

Step 3 — Expected strength gain (28 days):

For CH clay with 7% lime at 28 days: q_u,28 ≈ 90–130 psi (from laboratory trial mixes). Target q_u ≥ 50 psi per CPKC TS for stabilized subgrade.

Step 4 — Field specification:

• Lime content: 7% by dry soil weight; hydrated lime (ASTM C977, Type S)
• Mixing depth: 12 in. (top of subgrade)
• Mellowing period: 24–48 hours before final compaction
• Compaction: 98% Modified Proctor RC (CPKC TS); moisture −2% to +1% of w_opt
• Curing: Protect from freezing for 3 days minimum (geotextile insulating blanket in autumn)
• QC testing: pH ≥ 12.4 post-mixing (confirms adequate lime reaction); PI at 1 day and 7 days; q_u at 7 days and 28 days (ASTM D2166)

Sulfate verification: Retest sulfate content from each borrow area lot ≥ 500 yd³. If SO₄ > 2,000 ppm encountered, immediately switch to Type II/V cement at 5–6% binder content.

Figure 17-1: Ground improvement method selection chart — applied stress vs. layer thickness for soft clay, showing PVD (staged), stone columns, DSM, and grouting zones.

Figure 17-2: PVD installation schematic — mandrel, drain card detail (a × b dimensions), smear zone d_s, equivalent drain diameter d_w, and influence zone d_e annotation.

Figure 17-3: U_r vs. time curves (Barron radial consolidation) for c_h = 0.012 ft²/day at drain spacings of 1.0, 1.25, 1.5, 2.0, and 3.0 ft (triangular grid). CSXT Georgia design point highlighted.

Figure 17-4: Priebe (1995) improvement factor chart — n₀ vs. a_s for φ'_col = 35°, 37.5°, 40°, 42°, 45°. BNSF Port Arthur case plotted at a_s = 0.283, n₀ = 3.2.

Figure 17-5: Stone column installation schematic — vibro-replacement wet process and dry process comparison, column grid plan view (square vs. triangular), and load transfer mechanism diagram.

Figure 17-6: DSM column cross-section — auger/mixing head, binder injection ports, column diameter range by system, and soil-cement micro-structure schematic.

Figure 17-7: Compaction grouting sequence — primary, secondary, and tertiary injection point grid; grout bulb growth schematic; vibration monitoring instrument location.

Figure 17-8: Lime stabilization PI reduction vs. lime content for typical CL and CH clays — laboratory data scatter with CPKC target PI < 12 line shown.

Figure 17-9: Geosynthetic-Reinforced Load Transfer Platform (GRLT) cross-section over stone column grid — geogrid layers, column caps, granular LTP, and track load distribution.

Figure 17-10: Staged embankment construction schedule — Stage 1 fill placement, PVD consolidation wait period, strength gain, Stage 2 fill, and settlement monitoring timeline with instrumentation triggers.

Reference graphics

1. Select the improvement method based on the governing constraint, not generic practice. For railroad applications, the governing constraint is almost always time (construction schedule), not cost. PVDs with surcharge are cost-effective but slow; DSM and stone columns are faster but more expensive. Always map the client's construction schedule against the consolidation time required for each method before recommending.
2. Conduct a pre-improvement investigation parallel to improvement design. Ground improvement design requires additional data beyond standard investigation: c_h (from horizontal permeability tests or piezocone dissipation testing), sensitivity (from VST — Chapter 7), OCR profile (from consolidation test or CPTu — Chapters 7, 8, 12). Schedule these tests during the same mobilization as standard borings to avoid a second mobilization.
3. Verify smear zone impact on PVD performance before finalizing spacing. In sensitive and structured clays (Amtrak NEC, CN Quebec), mandrel installation smears the horizontal permeability zone by a factor k_h/k_s of 3–8. Uncorrected Hansbo-Barron analysis that ignores smear will overestimate consolidation rate by 30–60%, leading to inadequate wait periods. Use the full μ formulation or correct from piezocone dissipation results.
4. Install pre-construction and post-improvement plate bearing tests for stone column and DSM projects. Improvement factor n₀ is a design estimate. UPRR, BNSF, and CN all require at least one post-treatment plate bearing or pressuremeter test per 5,000 ft² of treated area to verify that the design improvement factor has been achieved in the field. Document pre-improvement and post-improvement data in the same GDR section for direct comparison.
5. Do not specify lime treatment without sulfate testing. Ettringite swell from high-sulfate clays has caused major track heave failures on multiple Class I railroads. The roughly $28–55 cost (2016 dollars) of an ASTM D4327 sulfate test is trivial compared to remediation of failed lime stabilization. Test every distinct soil horizon and every new borrow area before specifying lime.
6. Coordinate vibration exclusion zones with railroad operations before mobilization. Dynamic compaction and vibratory pile driving within 150–200 ft of active track require written authorization from the client's chief engineer, a pre-activity vibration baseline, continuous PPV monitoring during work, and a defined shutdown threshold (typically 0.5 in./s at rail). Failure to coordinate this before mobilization results in work stoppages and day-rate equipment charges. Obtain written approval from UPRR, BNSF, or CN engineering at least 30 days before activity.
7. Provide a performance monitoring plan as a deliverable — not an afterthought. Ground improvement performance is verified by instrumentation: settlement plates, piezometers, inclinometers, and periodic plate bearing or vane shear tests. The monitoring plan, trigger levels, and reporting frequency should be finalized during design and included in the GDR. UPRR and BNSF expect electronic data transmittal to their infrastructure management systems weekly during active consolidation monitoring.

Project management · Kelly et al. (2004)

Paired Comparison and Weighted Scoring for Option Selection

Kelly, Male and Graham recommend weighting and scoring when several technically feasible options compete and none is clearly best. The first step is to agree on the criteria. Correlated criteria should be avoided because they count the same attribute twice; their example is that ease of cleaning and cost of cleaning should not both be used. Weights then come from a paired comparison. Each criterion is compared with every other, the more important of each pair is recorded, and each criterion’s number of “wins” becomes its weight. Each option is scored 1–5 against each criterion. Score is multiplied by weight and the products are summed. Sensitivity checks, made by changing a few weights and scores, show whether the ranking is robust.

Their worked example compares internal partitions for a building that is reorganized every two to six years. Eight criteria (demountability, noise attenuation, finish, support of fittings, concealing services, capital cost, maintenance cost and reliability of supply) received weights of 4, 6, 3, 0, 5, 3, 4 and 3. Stud-and-board partitions totaled 109 and plastered blockwork 104, while the proprietary systems scored 76–91, mainly because noise attenuation carried the highest weight. The authors say the result should prompt a second look rather than be accepted mechanically. Any preferred option must still be technically feasible, economically viable, functionally suitable and acceptable to the client.

For railroad work: The method suits ground improvement and subgrade remediation choices, such as undercut-and-replace versus geogrid-reinforced subballast versus lime or cement stabilization versus deep soil mixing. Typical criteria are track outage required, ability to work from the track or the right-of-way, settlement performance under the design traffic in MGT, construction risk in the expected soils, capital cost, maintenance demand and supply-chain or contractor availability. Derive the weights with the railroad’s engineering and operations staff in the room, because operations typically values outage time very differently from engineering. Keep capital cost as one weighted criterion, or compare it separately on a whole-life basis, and show the sensitivity run in the report so the choice can be audited.

Source: Kelly, J., Male, S. and Graham, D. (2004). Value Management of Construction Projects. Oxford: Blackwell Science, pp. 75–77, 308–310. Summarized and paraphrased; figures reproduced for internal reference. Full reference under Supplemental Sources.

Question 11-1: A CSXT coastal site has c_h = 0.015 ft²/day and soft clay layer H = 40 ft. PVDs are installed at s = 1.5 ft triangular grid (d_e = 1.575 ft; d_w = 0.196 ft). Compute U_r after 90 days (ignore smear and well resistance; ignore vertical consolidation for this problem).

Answer 11-1:

U_r = 96% far exceeds the 90% target — the 1.5-ft spacing is conservative. A 1.75-ft spacing would also achieve 90% within 90 days, reducing PVD quantities approximately 36%.

Question 11-2: A BNSF site has CH clay with s_u = 400 psf and proposes stone columns D_c = 30 in. at 6.0-ft square grid. Calculate a_s and estimate n₀ using the Priebe simplified approximation (assume φ'_col = 40°). If untreated S_c = 15 in., what is the post-improvement settlement?

Answer 11-2:

From Priebe (1995) chart or formula with φ'_col = 40° and a_s = 0.136: n₀ ≈ 2.1

This exceeds BNSF's 6-in. target. Reduce spacing to 5.5 ft square: a_s = π(2.5)²/4/(30.25) = 0.162; n₀ ≈ 2.4; S_improved = 15/2.4 = 6.25 in. — still marginally over. Try 5.0-ft spacing: a_s = 0.196; n₀ ≈ 2.7; S_improved = 15/2.7 = 5.6 in. ≤ 6 in. ✓ Recommend 30-in. columns at 5.0-ft square grid.

Question 11-3: A CPKC site has CH clay with PI = 52 and sulfate content = 1,200 ppm. What lime content is needed to achieve PI_final = 12? Is sulfate a concern?

Answer 11-3:

• Sulfate = 1,200 ppm < 2,000 ppm threshold → Lime treatment is acceptable. However, CPKC TS requires re-testing each lot ≥ 500 yd³ during construction. If any sample exceeds 2,000 ppm, switch to 5–6% Type II Portland cement (freeze-thaw durable) per CPKC TS sulfate procedure.
• Specified lime: 8.0% Ca(OH)₂ by dry soil weight, ASTM C977 Type S.

Question 11-4: A UPRR embankment compaction specification requires 95% Modified Proctor on CL fill. Laboratory Proctor gives γ_d,max = 105 pcf at w_opt = 16%. A nuclear density gauge reading shows γ_d,field = 99.5 pcf at w = 18%. Is the fill acceptable?

Answer 11-4:

RC = 94.8% < 95% minimum → Not acceptable. The fill is also 2% above optimum moisture (+2% over w_opt = 16%). The wet condition reduces shear strength and increases settlement potential. Required action: scarify and dry back to w_opt ± 2% (14–18%); re-compact; re-test. For UPRR intermodal yard applications, moisture must also fall within −2% to +2% of w_opt; the current reading at w = 18% (exactly at +2%) is at the limit — additional passes with padfoot drum recommended before accepting.

Question 11-5: A DSM ground improvement program at an Amtrak NEC site targets q_u,field = 75 psi. What laboratory design strength should be specified, and what is the expected composite modulus E_composite if a_DSM = 0.25 and E_soil = 500 psi for the soft clay?

Answer 11-5:

• Laboratory design strength using flab = 0.25: q_u,lab = q_u,field / f_lab = 75 / 0.25 = 300 psi (laboratory target)
• E_DSM = 300 × q_u,field = 300 × 75 = 22,500 psi (using E_DSM = 300 × q_u,field, midpoint of 150–500 range)
• Composite modulus:
• Settlement reduction factor: E_composite/E_soil = 6,000/500 = 12× (substantial improvement)
• Verify: post-improvement settlement S_improved = S_untreated/12. If untreated S = 3 in. (typical NEC soft clay), S_improved ≈ 0.25 in. — well within Amtrak's 0.50-in. NEC Class 7 track tolerance.

End of Chapter 17 — Ground Improvement Methods

Sections 17.2 through 17.8 describe how each family of ground improvement works and how it is designed for railway loading. This library carries the supporting technology sheets: 45 techniques, numbered 17.16.1–17.16.45 in alphabetical order and grouped below by category. Each sheet opens with a how-it-works schematic and key figures, followed by its fact sheet, design guidance, QC/QA, specifications, cost information and case histories. Use the Previous / Next buttons at the bottom of each sheet to step through the library in order.

The sheets most relevant to each method are also linked at the end of Sections 17.2–17.8. Soil nailing, MSE walls and deep foundation sheets support Chapter 16 — Retaining Structures and Chapter 11 — Bearing Capacity.

Open the selection guide on its own page (for printing) →

Full table: choosing a technology for a railway problem

Vertical Columns & Deep Mixing (7)

17.16.1Aggregate Columns17.16.8Column-Supported Embankments17.16.9Combined Soil Stabilization with Vertical Columns17.16.13Deep Mixing Methods17.16.26Geotextile Encased Columns17.16.38Sand Compaction Piles17.16.44Vibro-Concrete Columns

Subgrade & Base Stabilization (3)

17.16.2Bio-Treatment for Subgrade Stabilization17.16.7Chemical Stabilization of Subgrades and Bases17.16.31Mechanical Stabilization of Subgrades and Bases

Materials Reuse & Recycling (3)

17.16.3Beneficial Reuse of Waste Materials17.16.16Excavation and Replacement17.16.34Onsite Use of Recycled Pavement Materials

Densification & Compaction (7)

17.16.4Blasting Densification17.16.12Deep Dynamic Compaction17.16.28High-Energy Impact Rollers17.16.29Intelligent Compaction17.16.37Rapid Impact Compaction17.16.42Traditional Compaction17.16.45Vibrocompaction

Grouting & Injection (4)

17.16.5Bulk-Infill Grouting17.16.6Chemical Grouting & Injection Systems17.16.10Compaction Grouting17.16.30Jet Grouting

Piles & Deep Foundations (3)

17.16.11Continuous Flight Auger Piles17.16.27Helical Pile17.16.33Micropiles

Soil Nailing & Retaining Structures (5)

17.16.14Drilled-Grouted and Hollow Bar Soil Nailing17.16.32Mechanically Stabilized Earth Wall System17.16.39Screw-in Soil Nailing17.16.40Shoot-in Soil Nailing17.16.41Shored Mechanically Stabilized Earth Wall System

Preloading, Drainage & Dewatering (3)

17.16.15Electro-Osmosis17.16.36Prefabricated Vertical Drains and Fill Preloading17.16.43Vacuum Preloading with and without Prefabricated Vertical Drains

Geosynthetics & Pavement Systems (10)

17.16.17Fiber Reinforcement in Pavement Systems17.16.18GeoWEB17.16.19Geocell Confinement in Pavement Systems17.16.20Geosynthetic Reinforced Construction Platforms17.16.21Geosynthetic Reinforced Embankments17.16.22Geosynthetic Reinforcement in Pavement Systems17.16.23Geosynthetic Separation in Pavement Systems17.16.24Geosynthetics - GeoTerra17.16.25Geosynthetics in Pavement Drainage17.16.35Partial Encapsulation

Technology sheets compiled from FHWA, SHRP2 GeoTechTools and manufacturer literature; each sheet lists its own sources.

Field case studies by Burt G. Look, from Earthworks: Theory to Practice – Design and Construction (CRC Press/Balkema, 2023), chosen for this chapter’s topics. Each is summarized from the book with its key figures and a closing note on railroad application added by the authors of this manual. Credit for the case material belongs to Look; see item S12 under Supplemental Sources.

Case Study 17.17.1 — Measuring roller force with pile-driving monitors

Case study · Look (2023) §7.10

Related section: 17.2 Compaction Methods

Three trial embankments were built from extremely to slightly weathered rock taken from site cuttings: sandstone, basalt, and interbedded siltstone/sandstone. All of it broke down to meet “soil” specifications during placement and rolling (Look and Lacey, 2017). A 20.6-ton vibrating padfoot and a 19-ton smooth drum roller compacted each fill at field and targeted moisture.

Pile-driving monitors (PDM), which normally record set and energy for each hammer blow, tracked reflectors on the rollers. They measured each pass’s set, vibration frequency, amplitude, peak velocity and acceleration, and so the peak compaction force, which is not the same as static weight. Findings:

• Dynamic force was well above static weight and changed with material and with each pass as the ground stiffened: about 26–40 tons equivalent for the padfoot and up to 76 tons for the smooth drum on the interbedded fill.
• Maximum force and best material performance came after five to six passes; beyond that the values cycled and some rebound occurred.
• Over-compaction broke the material down, which showed both in the force record and later in borehole shear test strengths.

Table 17-LK1. Padfoot roller, manufacturer’s values against PDM measurements (after Look 2023, Table 7.13)

Parameter	Manufacturer	Measured with PDM, three fills
Vibration frequency	29 Hz high amp., 31 Hz low	27.2–30.9 Hz (range)
Roller speed	0–7 mph	median 2.8 mph
Amplitude	0.071 in. high, 0.047 in. low	medians 0.055–0.087 in.
Force	71 kips high amp., 52 kips low	medians 53–83 kips

For railroad work: a pass-count specification without a check of the roller’s actual output can over-work weathered-rock fill. A trial embankment with measured rebound is a sound way to set the pass count for a large rail embankment.

Source: Look, B. G. (2023). Earthworks: Theory to Practice – Design and Construction. CRC Press/Balkema. Case study §7.10, pp. 232–236. Summarized and paraphrased; figures reproduced for internal reference. Full reference under Supplemental Sources.

Case Study 17.17.2 — Wick-drain rigs on dredged sand over 104 psf clay: four ways to size the platform

Case study · Look (2023) §12.7

Related section: 17.3 Prefabricated Vertical Drains (PVDs) with Staged Loading

At the Port of Brisbane reclamation, wick-drain rigs had to work for nine months on 31 acres of dredged sand, about 8.2 ft thick, over Holocene clay with a design undrained strength of 104 psf in the top 20 ft (Look and Honeyfield, 2016). Mud waves during filling left the sand thickness uncertain. BR470 is outside its range here, both for clay under 420 psf and for platform thickness to load width (D/W) above 1.5; it was used for the preliminary design for want of anything better.

Loads. The heaviest rig did not exert the highest pressure: the 50-ton CAT on narrower tracks produced 3,500 psf against 2,030 psf for the 89-ton machine. Check every rig.

Parameters. 27 LFWD tests, 8 DCPs and 4 borehole shear tests gave the sand properties. Two proof-roll passes raised the LFWD modulus by about 20% and reduced its scatter (10th percentile 4,400 psi rolled against 3,260 psi unrolled). Groundwater was found at 6.2–10 ft, not the 20 in. assumed. Clay directly under the sand measured 42–380 psf.

Four answers.

• BR470 deterministic: 7.5 ft required.
• BR470 with a PERT distribution and Latin Hypercube simulation: baseline 7 ft; 7.5 ft has under 5% probability of being exceeded. Subgrade strength dominates the tornado chart.
• Finite element displacement, calibrated to proof-roll ruts, with a 3 in limit: 6.9 ft.
• Finite element stress at the sand/clay interface against a lower-bound bearing factor N_c = 3.14 for combined shear, as mud waves imply: 7 ft.

Proof rolling as the check. A loaded proof truck will bog rather than topple, while an eccentrically loaded rig can topple, so the truck is the safer test. Proof rolling only means something with a defined vehicle and an acceptance deflection. A rock-filled truck loaded more evenly than a water truck (whose load sloshes) and was preferred. Mean ruts were 3.7 in. after one pass and 5.2 in. after two, and 4.7 in was adopted as the acceptance criterion.

Table 17-LK2. Wick-drain rigs and track pressures (after Look 2023, Table 12.20)

Rig	Weight (tons)	Track width (ft)	Critical track pressure
Komatsu excavator	65	3	2,700 psf
CAT excavator	50	3	3,500 psf
Komatsu excavator	89	4.3	2,030 psf

For railroad work: ground improvement under new embankments on soft soil, such as wick drains for a yard expansion or a new grade across a marsh, puts the rigs on exactly this kind of platform. Size it by more than one method and calibrate the proof roll before the rigs move on.

Source: Look, B. G. (2023). Earthworks: Theory to Practice – Design and Construction. CRC Press/Balkema. Case study §12.7, pp. 432–446. Summarized and paraphrased; figures reproduced for internal reference. Full reference under Supplemental Sources.

Case Study 17.17.3 — Lime-treated local soil in a reinforced slope: choosing by total internalised cost

Case study · Look (2023) §12.8

Related section: 17.7 Lime and Cement Stabilization

Prolonged rain in 2010/11 triggered a 656 ft landslide under a Sunshine Coast hinterland road, in colluvium up to 20 ft thick below a steep escarpment (Look 2012; Chew et al. 2012; Boyer et al. 2013). Five remediation options were compared on total internalised cost: construction cost plus embodied energy, greenhouse-gas emissions from fuel, transport and materials, and the cost of damage to local roads from heavy trucks hauling material. Environmental costs ranged from 2% to 23% of construction cost.

Options 3 and 4 had the lowest totals. Option 3 was chosen because the road authority wanted the fewest heavy trucks on local roads, and lime treatment let the high-plasticity “unsuitable” colluvium be screened of boulders and reused. With 4% hydrated lime cured seven days, shear box cohesion rose from about 420 psf to 1,250–2,200 psf at a similar friction angle; the design still used a softened cohesion of 104 psf.

The work was done in five stages with soil-nailed temporary slopes, drainage slots and regular inclinometer and survey monitoring. When inclinometer 2 showed excessive movement, inspection found a tree at the crown destabilised by the excavation. Once it was felled, the movement stopped.

Table 17-LK3. Remediation options, 2011 AU$ (after Look 2023, Table 12.23, from Chew et al. 2012)

Option	Construction	Environmental	Total internalised
1. Cast-in-place pile wall with ground anchors*	3,300,000	67,000	3,367,000
2. Rockfill buttress	3,800,000	628,000	4,428,000
3. Reinforced fill slope, lime-treated local soil	2,400,000	278,000	2,678,000
4. Reinforced fill slope, imported granular fill	2,000,000	452,000	2,452,000
5. Earth fill embankment with rock toe	3,200,000	444,000	3,644,000
* Greatest technical uncertainty.

For railroad work: slide repairs on remote lines face the same haul-cost and community questions. Reusing treated local soil cuts truck traffic on rural roads, and staged excavation with inclinometers is the standard of care on an active slide.

Source: Look, B. G. (2023). Earthworks: Theory to Practice – Design and Construction. CRC Press/Balkema. Case study §12.8, pp. 446–448. Summarized and paraphrased; figures reproduced for internal reference. Full reference under Supplemental Sources.

Case Study 17.17.4 — Working platform at the edge of a creek, designed for slope stability

Case study · Look (2023) §12.6

Related section: 17.8 Geosynthetic Reinforcement

A bridge pier at the edge of a creek, with the next pier 8 ft into it, needed a platform for a BG28 piling rig for a day and an outrigger crane for beam lifting. The temporary-works designer rightly declined to design it with the standard method, since slope stability over soft clay governed. Sheet piling would have cost almost five times the roughly AU$100,000 (2015) platform. Moving the pier a few meters would have removed the problem, which shows why constructability belongs in design.

A deep rock-fill replacement moved the critical slip surface deeper, with high-strength geogrids within the rock fill. Stability was checked separately for standing (uniform 3,800 psf, FS 1.5, above the 1.3 usually accepted for temporary works), penetrating (uniform and trapezoidal, FS 1.3) and extracting. The extracting case was the most critical (FS 1.2) even though its 5,500 psf design pressure is lower than penetration’s 6,400 psf, because of where the load sits. Large plate load tests were run but cannot test a deep slip surface, so an excavator was walked along the edge first. The work went without incident, though rig depressions had to be repaired before the outrigger came on.

Table 17-LK4. Design pressures for the BG28 piling rig (after Look 2023, Table 12.19)

Operation	Design pressure	Factor of safety adopted
Handling / transporting / standing	3,780 psf	1.5
Penetration	6,390 psf	1.3
Extracting	5,470 psf	1.2 (critical, trapezoidal)

For railroad work: bridge replacement platforms beside streams and next to embankment slopes need a global stability check for each rig operation, not only a bearing check. The extract phase and the outrigger reactions are often the governing cases.

Source: Look, B. G. (2023). Earthworks: Theory to Practice – Design and Construction. CRC Press/Balkema. Case study §12.6, pp. 430–432. Summarized and paraphrased; figures reproduced for internal reference. Full reference under Supplemental Sources.

Authors: Daniel P. Colwell and Mario Ruel. This chapter was written and compiled by Daniel P. Colwell and Mario Ruel for Railway Ground Hazard Risk Management.

Technical content is drawn from the published sources below, and credit for that material belongs to their authors and publishers. Tables, equations and figures taken from a specific source are also cited where they appear.

Primary sources

• USACE EM 1110-1-1804, Geotechnical Investigations (2001)
• FHWA NHI-06-019, Ground Improvement Methods (2006)
• Barron (1948); Priebe (1995)
• Kirsch, K. and Bell, A. (2012). Ground Improvement
• ASTM D698, D1557
• AREMA MRE Chapter 1

Supplemental sources (LEVI Library)

• Blacklock, J. R. and Lawson, C. H. (1977). Handbook for Railroad Track Stabilization Using Lime Slurry Pressure Injection. FRA/ORD-77/30, Federal Railroad Administration
• Pengelly, A. D. Injection for Railroad Subgrades. Hayward Baker, Inc. (scanned; text recovered by OCR)
• Hayward Baker, Inc. Railroad Subgrade Stabilization (brochure)
• National Lime Association (2004). Lime-Treated Soil Construction Manual: Lime Stabilization & Lime Modification, Bulletin 326
• SHRP2 R02 / GeoTechTools (2012). Column-Supported Embankments — fact sheet, design guidance and photographs
• SHRP2 R02 / GeoTechTools (2012). Rapid Impact Compaction and Jet Grouting fact sheets
• Schaefer, V. R. et al. (2017). Ground Modification Methods Reference Manual, Vol. I. Geotechnical Engineering Circular No. 13, FHWA-NHI-16-027
• Look, B. G. (2023). Earthworks: Theory to Practice – Design and Construction. CRC Press/Balkema. Case studies §7.10, §12.6, §12.7, §12.8

File locations for these documents are listed under Supplemental Sources — LEVI Library.

Project management sources

• Sears, S. K., Sears, G. A., Clough, R. H., Rounds, J. L., and Segner, R. O., Jr. (2015). Construction Project Management: A Practical Guide to Field Construction Management, 6th ed. Hoboken, NJ: John Wiley & Sons
• Bull, J. W. (ed.) (2015). Life Cycle Costing for the Analysis, Management and Maintenance of Civil Engineering Infrastructure. Dunbeath, Scotland: Whittles Publishing
• Miller, D. P. (2009). Building a Project Work Breakdown Structure: Visualizing Objectives, Deliverables, Activities, and Schedules. ESI International Project Management Series. Boca Raton, FL: Auerbach Publications (CRC Press, Taylor & Francis Group)
• Kelly, J., Male, S. and Graham, D. (2004). Value Management of Construction Projects. Oxford: Blackwell Science
• ASCE Construction Institute Committee on Specifications (L. Eckersley, Chair) (2013). Preparation of Construction Specifications for Civil Projects. Reston, VA: American Society of Civil Engineers
• Sullivan, W. G., Wicks, E. M. and Koelling, C. P. (2015). Engineering Economy, 16th ed. Upper Saddle River, NJ: Pearson
• Levy, S. M. (2007). Project Management in Construction, 5th ed. New York: McGraw-Hill