Content uploaded by Patricia Stouter
Author content
All content in this area was uploaded by Patricia Stouter on Jul 04, 2021
Content may be subject to copyright.
60% Scale Contained Earth Earthbag Shear
Performance
Patricia Stouter
Build Simple Inc., www.BuildSimple.org
May 10, 2017, revised December 30, 2020
Introduc"on
Low-cost earthbag is spreading worldwide into high seismic risk regions (Geiger n.d.) whether
engineered or not. Earthbag walls can cost 60% less than concrete block walls in regions with low
labor costs (Geiger 2011, 5) because they require little or no Portland cement and little water. The
less than 5% manufactured components are light and portable to rural and remote building sites and
can be used with many types of subsoil. Better structural data is needed so engineers and designers
can optimize reinforcement and ensure safe buildings.
Earthbag uses fabric or mesh bags or tubes as forms for damp subsoil stacked in a running bond
pattern over barbed wire strands and tamped into walls at least 380 mm (15”) thick. Barbed wire
increases bed joint friction and adds tensile strength for shear transfer and out-of-plane strength
(Ross et al. 2013). A bond beam unites the wall top. Plaster or stucco protects bag fabric from UV
damage, and roof overhangs protect wall fill from rain in wet regions. Builders often hammer steel
rod into damp fill at half height and at the wall top (Gieger 2011, 41) but the 1.5 m (5’) lengths
overlap in the walls, unconnected to each other or to footings.
Weak earthquake motion 0.3 g (30% of gravity) can destroy brittle conventional earthen buildings.
Earthbag of barely cohesive fill built without rebars is less brittle but also vulnerable to earthquake
damage. Safe earthbag buildings can be low-cost if builders use the best details and build plans like
safe plans for unreinforced adobe with intersecting walls or buttresses evenly distributed for
stability (Morris and Walker 2000). Earthbag shows potential for earthquake resilience because its
fabric forms can contain wall material to prevent loss if damaged (Rao, Raghunath, and Jagadish
2004), its barbed wire and inserted rebar offer tensile reinforcement (Ross 2013), and the surface
plaster can delay damage to internal wall materials (Blondet et al. 2011).
More than 50 earthbag schools and homes in Nepal survived 2015 quakes with little or no damage
near villages where unreinforced masonry buildings were destroyed (Nordquist n.d.). However,
future earthquake motion could be 2- 3 times stronger than the estimated 0.7- 0.8 g forces
experienced by these buildings (Stouter 2016, 43). With innovative detailing will CE equal code
strengths for highly reinforced adobe (Morris and Walker) in 1.2 g risk areas?
Many studies of earthbag structural performance used sand as fill starting with Pelly at the
University of Bath who assisted designers considering a non-conventional project in a clay-poor
Namibian desert (2009). Vadgama and Croft continued research with the same fill. Proposed
earthbag research is often hindered by this precedent and engineering students’ lack of earth
building skills. Limited time and funding often prevent adequate selection and strength testing of
soil materials and adequate sample drying times.
1
With sand fill, earthbag units fail in compression when bag fabric splits (Vadgama 2010, 15, 37).
Tamped bag units failed under compression between 1.2- 1.8 MPa (174- ± 260 psi) (Pelly 2009, 12,
29). Fabric surfaces between courses have a coefficient of static friction without barbed wire of
0.43, and with barbed wire of 0.66 with 8.2 kPa (1.2 psi) adhesion (ibid., 21) caused by dowel
action of the barbed wire barbs. Barbs translating through loose fill cause small slits in fabric
(Vadgama, 40). In shear box tests, bag stacks, and 1 m² (11 ft²) wall portions, both barbed wire
between courses and plaster on mesh increased shear strength of sandbags. Although shear tests
failed from twisting under normal loads, the 1.1 m (43“) long sandbag wall panels resisted
horizontal forces from 8.5- 20 kN (1910- 4500 lb-f) under 4- 15 kN (900- 3370 lb-f) normal load
when stucco covered and 2- 2.6 kN (450- 580 lb-f) under 1.4- 2.2 kN (315- 495 lb-f) loads when
unplastered (Croft 2011). These values for 1.07 m (42”) high sandbag walls are equivalent to
plastered shear walls resisting 7.7 kN/m (528 plf) with N=4 kN (900 lb-f). When barbs prevented
slipping between bed joints, fabric stretched and rotated around fill (Vadgama, 19) causing low
vertical stability. Computer modelers have proposed structural design guidelines for domes
(Canadel, Blanco and Cavalero, 2015) based on strength values from sand fill.
Poorly dried soil fills have been tested. Daigle used ‘topsoil’ in six unit samples with results
indicating cohesive properties but only half dried since brick size samples require 4- 8 weeks in 40-
80% humidity to dry completely (Minke 2006). Daigle’s samples had an unconfined compressive
strength of 2.3- 3.0 MPa (340- 430 psi) and failed without splitting bags (Daigle 2008, 171-172). A
partly dried silty sand soil fill in an unplastered 8.4 m² (90 ft²) full-scale bag wall curved under out-
of-plane force with bed-joints shearing between courses until barbed wire pulled taut and became
‘critical components of force transfer’ (Ross). This author compared undried, partly dried and fully
dried unit samples for pullout forces of barbed wire between bags, finding 50% lower force required
for undried and 34% lower for partly dried samples. Barbs bent at 177 N (40 lbs) per barb before
pulling out of dry cohesive soil without damaging soil masses or fabric (Stouter, 14, 15).
Earthbag research using fully dried samples with strongly cohesive soil (denoted by this author as
contained earth or CE) included units, small samples and some 80% scale walls. This author’s cured
units reached 2.5 MPa (370 psi) compressive strength (see Appendix B), 40% higher than sand-
filled bags at Bath and similar to Daigle’s partly cured CE. Undergraduates at Dartmouth College
found modulus of elasticity values for bag stacks (Connoly, Jin and Malik, 2012) and shake table
results for four plastered 1:6 scale samples which showed that corners with piers survived 13%
higher motion with 33% less deformation (Malik 2013, 16, 22). This author tested three plastered
80% scale 0.8- 1.5 m² (9- 16 ft²) wall portions under in-plane horizontal force. Weak silt soil
without rebar reached 3.3 kPa1 (69 psf) and barbs chipped cones of crushed soil loose. One 1.2 m
(49”) high strong soil sample with wire mesh pins near wall ends and a 16 mm (5/8”) inserted rebar
reached 8.8 kPa1 (184 psf) without yielding, equivalent to 1.1 m (42”) high earth-plastered shear
walls resisting 5 kN/m (343 plf). A 1.2 m (4’) high strong soil hyperadobe sample (built in mesh
tubing) was much stronger, resisting up to 24 kPa1 (501 psf). Barbs did not slit fabric and translate
through dried CE fill in any of these tests. (Stouter).
1 Calculated as horizontal force/ wall face square area
2
Experimental Work
The importance of soil strength to traditional earthen wall construction is recognized by engineers
(Blondet et al. 2011, 5) but has never been researched for earthbag building with components fully
dried to equilibrium moisture levels (Minke) before strength testing. Current research explores how
compressive strength of dried soil fill affects different types of earthbag reinforcement using static
monotonic testing with simple equipment. Four wall portions were tested under horizontal in-plane
force and 25 elongated unit samples under three point flexile testing. The earth plastered wall
portions of medium or strong cohesive fill were 60% scale and each reinforced with two full-length
rebars either inserted from the top (Figure 1) or base-anchored and partially embedded in the end of
the panel (external) (Figure 2a). The 360 mm (14”) long unit samples each compared a rebar in a
different location relative to surface stucco of cement mortar. A single full-length rebar was located
at the sample surface (external), or 50 mm below the surface (interior), or in a formed 35 mm (1-
3/8”) deep channel (in earthen or mortar fill) separated by fabric from the rest of the sample.
Overlapped interior rebar used 2 separate rebar pieces 100 mm (4”) long placed 50 mm (2”) below
the surface. Both woven and mesh fabric alternates were used for most rebar locations.
Figure 1: End and side views of inserted sample layouts.
Materials
Soils used had compressive strengths confirmed by laboratory tests of three cured full-size unit
samples each following ASTM C-67 (see Appendix B). Medium soil was a very coarse-grained fill
dirt with a sandy loam texture (NRCS 2017) averaging compressive strength of 1.8 MPa (260 psi)
with almost equal proportions of coarse sand, fine aggregate, coarse aggregate and of silt and clay
3
combined. Although handfuls dropped from 1.5 m (5’) height split into many small piles even when
very damp so would not be recommended under standard builder training (Geiger 2011), this soil
firmed up well in bags. Strong soil was a sandy clay loam texture (NRCS) used to manufacture
adobe blocks that averaged 2.2 MPa (320 psi) compressive strength. Damp handfuls met standard
builder guidelines when dropped from 1.5 m (5’) height by splitting into only 2- 3 pieces.
Barbed wire was 4 point, 15.5 gauge high tensile strength. Rebar were 9.5 mm (3/8”) diameter
deformed (textured) steel.
Bags were standard solid-weave polypropylene fabric sandbags, machine sewn 280 mm (11”) wide.
Electricians’ polypropylene pull string of ±95 kg (210 lb) tensile strength was used for strapping. A
light plastic mesh sold as bird netting was used under plaster.
Wall bases and bond beams were recyclable wood.
Procedures
Soils for samples, walls and plasters were dampened 24 hours before using.
Sample walls ± 1 m² (11 ft²) and ±230 mm (9”) thick were built between wooden forms to maintain
panel sides plumb and ends vertical. Courses of one long and one medium length bag used uniform
soil moisture and with 20% compaction reached heights of 83- 99 mm (3.3- 3.9”). One short angled
rebar pin in the center of each wall and the longer vertical rebars about 4”/ 10 mm from each wall
end were all inserted within 24 hours of construction. The inserted vertical rebars fit tightly into
holes in a 50x200 mm (2x8) wood bond beam. The external samples had rebar hammered into
slightly smaller holes in the wood base and fill formed around the rebar (Figure 2a- sample on
right). One strand of barbed wire was used per course and vertical strapping tied to it every four
courses. The external sample wire was looped around one rebar and bent back around the other. The
top of the external rebar was screwed to the end of the bond beam.
Samples dried at an average temperature of 10° C (50° F) for 2.5 months at humidity levels near
45%. A curing proof stack of four 250x250 mm (10” x 10”) bags was built a week after walls were
finished and testing delayed until the proof stack was dry inside. Wall portions were plastered with a
rough (Figure 2b) and then a finish coat of earthen plaster which dried 10 days before testing.
Figure 2 (left to right): a) Samples on base with testing post. b) Rough plaster on weak fill external.
4
The four walls shared a single central wood post of three spaced 50x250 mm (2x10s) interlaced
with and bolted to the wood bases. Fiberglass straps rated for 22 kN (5,000 lb-f) were used as a
hold-down at the near-pressure side and rated 44 kN (10,000 lb-f) ran diagonally from the post to
the base of the wall. Motion targets of a grid of centimeters tacked on samples received a laser spot
light. A hydraulic jack rated for 44 kN (10,000 lb-f) with a 63 mm (2.5”) diameter gauge extended a
12.6 cm² (1.96 inch²) diameter piston. A piece of 25x150 mm (1x6) hardwood under an 8 mm
(0.25”) thick steel plate spread pressure on the samples. Blocks of wood were bolted together as
extenders and screwed to a small shelf to reduce piston twisting at high pressures.
Pressure was increased by hand, then read and recorded manually and deformation photographed,
taking about 15 seconds for each pressure step. Plaster damage was noted at the first appearance of
a crack extending more than 1/3 of the sample width. Most pressure levels were visited at least
twice. When the piston reached full extension or the wall twisted out-of-plane pressure was released
to reset equipment. Samples were tested until pressures declined or reached a persistent plateau.
Unit samples were tamped lightly in wood forms 90 x 100 x 360 mm (3.5 x 4 x 14”) and dried for
one week in shade before adding mesh and cement mortar to one surface. Samples cured in total for
1 month until test weights on a scale sensitive to 1 gram stabilized for 2 days in a row. Units were
placed in a testing frame with rebar at the lower edge to be under tensile force.
Results
Wall Panel Results
All wall samples were still stable after testing.
MI- Medium Fill Inserted Sample (Figure 3a, b): Pressure was reduced to zero twice during testing.
After maximum force, deformation continued at a pressure plateau.
ME- Medium Fill External Sample (Figure 3c, d): Pressure was reduced 5 times during testing. A
small pressure rise was followed by a sudden shock before pressure dropped and plateaued.
Figure 3 (left to right): a) MI before test. b) MI at maximum. c) ME before test. d) ME at maximum.
SI- Strong Fill Inserted Sample (Figure 4a, b): Pressure was reduced 11 times during testing as the
sample flexed out-of-plane under pressure as much as 10 degrees from vertical, but returned to
vertical without visibly increased damage at pressure release. Flexure under pressure was reduced
when horizontal rotation of the piston was reduced by attachment to a rigid support shelf.
SE- Strong Fill External Sample (Figure 4c, d): Pressure was reduced 8 times during testing because
of twisting out-of-plane similar to that seen on SI. Flexure was less extreme than on SI, possibly
5
due to the use of the piston shelf throughout. Near the end of the test process, at 175 mm (6.9”)
deformation, the near side vertical rebar broke out of the wooden base.
Figure 4 (left to right): a) SI before test. b) SI at maximum. c) SE before test. d) SE at maximum.
First plaster crack ocurred at the highest force for the SI sample (Table 1), and this sample also
reached the highest maximum force. The force at first plaster crack for external samples was
significantly lower than the inserted sample of the same fill strength. Maximum forces were similar
on medium fill samples but differed more for the strong fill samples.
Table 1: Sample Dimensions and Test Results
Sample Courses Average
Course Ht.
Mm (“)
Wall Face
Area m²
(s.f.)
First Plaster
Crack
kN (lb-f)
Deformation at
First Plaster
Crack mm (“)
Maximum
Force
kN (lb-f)
MI 9 83 (3.28) 0.84 (8.9) 5.99 (1346) 30 (1.18) 6.98 (1569)
ME 10 99 (3.90) 0.97 (10.4) 3.99 (897) 21 (0.83) 6.65 (1495)
SI 10 94 (3.72) 0.91 (9.8) 6.80 (1529) 32 (1.26) 10.47 (2354)
SE 10 96 (3.78) 0.97 (10.4) 5.23 (1176) 28 (1.10) 8.20 (1843)
Inserted rebar elements were too tightly bonded to fill to be rotated or lifted after testing. Inserted
pins and rebar did not bend, but the medium fill external rebar near the pressure bent slightly near
the center of the first course. The strong fill external rebar near the pressure broke away from the
base by splitting the wood element in the base that held the rebar.
Some bags on lower courses had more barb hole stretching than upper courses, but no barb holes
were slit. Most barb to fill attachment points lacked any signs of chipping although some air
pockets reduced barb embedment in fill on the undersides of courses around some barb clusters.
None of the longer barbs in each cluster were damaged but some of the short barbs were bent. More
of the strong fill samples’ barbs were bent than samples with medium fill.
Unit Sample Results
Small flexural samples all failed by earth cracking off the rebar and separation of the stucco. Tensile
strength of the rebar was not engaged in the external or mortar fill samples. The rebars disconnected
completely from solid woven fabric (Figure 5a) but showed weak connection to mesh fabric (Figure
6
5b). Earthen fill samples lost less soil from a diagonal crack at a shallower angle. Overlapped rebar
samples lost little fill but cracked between the rebar ends (Figure 5c) and showed flexural strengths
1/3 lower than the interior full-length rebar sample. Strengths increased more from deeper rebar
embedment in dried earth than from cement mortar, with highest strength for fully embedded
interior rebar (Figure 6).
Flexural strength in strong fill units was highest for a full-length rebar fully embedded in dried
earth. Rebar in shallow filled external channels was stronger when mesh allowed some connection
between the main soil mass and the channel, but was surprisingly stronger with earthen fill than
with cement mortar. Overlapped internal rebar was weaker than the mortar filled rebar channel, but
external rebar was weakest in all cases.
Figure 5 Flexural unit sample results (left to right): a) External rebar not connected through woven
fabric to sample. b) External rebar weakly connected through mesh. c) Overlapped rebar damage.
Figure 6 Unit flexural results: Samples with woven fabric at left, samples with mesh at right.
7
Evalua"on
Wall Panels
Medium and strong fill CE prevented damage to container fabric. Unlike earthbag built with non-
cohesive or weakly cohesive fill no fabrics split under pressure or were slit by barbed wire. No
obvious hole enlargement was seen in rebar impressions in dried CE soil masses despite racking to
angles of 14- 18° from vertical. Walls failed by incremental slipping between all bag-to-bag
courses but without gaps between head joints. The central shorter pin through the upper 4 courses
reduced motion in the top courses, causing the most motion to occur below the end of the pin in
most samples, ± 12 mm (0.5”) compared to 6- 8 mm (0.23- 0.31”) for other courses at maximum
force.
Shear stress was calculated using Equation 1 (Engineer’s Edge 2017) in all samples. Although stress
is usually based on the cross-sectional wall area (length x thickness), for best contrast between
slightly different sized small-scale samples in a material that probably resists stress by the number
of barbed wire barbs and area of vertical rebars piercing low-friction bed-joints, stress was
calculated for Table 2 using the wall face area of each sample.
τ = F / A τ = Shear stress as force per unit area (1)
F = Force
A = area (perpendicular to applied force)
Table 2: Shear strength Results
Sample Yield Force/ area
kPa (psf)
Deformation at
Yield mm (“)
Maximum Force/
area kPa (psf)
Deformation at
Maximum Force mm (“)
MI 6.34 (132.4) 20 (0.79) 8.3 (173.3) 80 (3.15)
ME 3.42 (71.4) 8 (0.31) 6.9 (144.1) 50 (1.97)
SI 7.84 (163.7) 40 (1.58) 11.5 (240.2) 153 (6.02)
SE 3.41 (71.2) 5 (0.20) 8.4 (175.4) 156 (6.14)
Stronger soil samples showed higher strengths and lower deformation for the inserted rebar samples
at yield and higher strength for both inserted and external rebar at maximum. All samples showed
strain hardening after yield (Figure 7) and initial angles were stiffer for external than for inserted
rebar samples. Strong fill samples showed greater pressure increases after yield than medium fill
and showed a clear strength decline after maximum strength.
Composite materials can have complex stress-strain response curves as different elements weaken
at different pressure levels. Bimodal behavior was least noticeable in the medium fill inserted
8
sample. External rebar samples yielded at low force levels before strain hardening, followed by
strain softening before maximum, altogether resulting in large deformations.
Figure 7: Stress-strain envelope curves of all samples from raw data.
Yield values were indicated by graph curves (Figure 8) were confirmed by elastic or plastic
behavior evident during pressure unloading cycles. The first noticeable plaster cracks did not occur
until near 4% strain for inserted and near 2% strain for external samples, although building codes
often limit walls or ceilings to less than 1% deformation (Underwood and Chiuni 1998) to avoid
sheetrock cracking. Serviceability limits for plastered CE walls may deserve to be reconsidered for
developing and/ or developed world settings since plastered walls are more usable during repair and
more easily repaired than sheetrock.
Comparing inserted samples shows that stronger fill increased strengths of samples with embedded
rebar (Table 3). Comparing medium fill samples showed significant reduction in strength from
external rebar at yield and maximum. Rebar in earthen material may perform similar to rebar in
concrete, requiring good bond strength for best component strength, which derives from a
combination of surface adhesion, confining effect caused by shrinkage of concrete, and frictional
resistance of the bar’s surface deformations (University of Memphis Civil Engineering 2017, 193)
to prevent buckling of steel. Pairs of external rebar as originally recommended by engineers (Geiger
2010) may not improve performance and are often considered too costly for developing world use.
The greatest contribution of a 23% strong soil is 39% improvement at maximum strength and 24%
at yield for wall portions with a single entire inserted rebar. Increases of this magnitude may not be
seen with the conventional practice of inserting two separate pieces of rebar, since unit tests indicate
overlapping separate rebars also reduce flexural strength.
9
Figure 8: Detail of initial angles of stress-strain envelope curves.
SE compared to other samples performed poorly. Apparently its lack of barbed wire on the second
course from the top caused lower strengths and more deformation. SE was no stronger at yield than
ME, and only 22% stronger at maximum. Although SI deformation was 1.9 times greater than MI,
SE deformation was 3 times greater than ME.
Table 3: Comparison of similar sample pairs
Strong fill 23% stronger than
medium fill
Yield strength Yield
deformation
Maximum
strength
Max.
deformation
SI/ MI +24% +100% +39% +91%
SE/ ME No change -37% +22% +200%
External and base-anchored
rebar/ inserted rebar
ME/ MI -47% -60% -17% - 37%
SE/ SI -57% -87% -27% Minimal
change
Unloading and reloading cycles of all samples displayed some time-dependent responses including
hysteresis as deformation retracted more at testing pauses and relaxation as pressures ‘slipped’ after
reaching pressure levels. Constant strain resulting in decreasing stress (Banks et al. 2011, 21) and
10
exaggerated deformation ocurred in samples with strong fill because higher toughness required
more pressure step repetitions.
Performance probably varied due to variation in barbed wire barb embedment. Viscoplastic
materials repeat the same stiffness angle at increasing levels of strain, but in almost every current
and past shear test, CE stiffness increased after 2- 5 low pressure cycles. Barbed wire barbs may
seat themselves more deeply as horizontal motions repeat. The average deformation per course at
maximum in the current shear tests was 9 mm (0.35”), or 4.5 mm (0.17”) per course surface. Each
barb on high tensile strength barbed wire can flex up to 5 mm (0.2”). Partial embedment of barbs
may increase deformation and allow barb failure at lower forces than when fully embedded from tip
to wire. Barbed wire with larger barbs or pins with several teeth cut from wire mesh added over
barbed wire at stress locations may increase wall initial stiffness.
Conclusions
Stronger fill increases sample strength significantly.
Poor behavior of the SE sample with missing barbed wire on one course indicates that every strand
of barbed wire is critical to transfer of in-plane stresses between courses.
Base-anchored rebar should increase initial wall stiffness, but best wall strength and least
deformation of CE earthbag walls can only be achieved when every detail is optimized. The goal of
earthbag reinforcement should be to reduce deformation with better detailing so performance is not
lowered by some ‘softer’ elements or looser connections.
Because many separate scale issues influence sample performance, it is not yet possible to
accurately estimate full size CE earthbag wall performance without reference to identical full-size
testing. Although the current samples used reinforcement with 50% as many barbs and 60% of the
cross-sectional area as standard 12 mm (0.5”) rebar, this strength reduction may not offset the
longer moment arm that causes 2.4 m (8’) high building walls to typically resist only 37% as much
force as 0.9 m (3’) high in-plane samples of the same aspect ratio, since lateral loads are
proportional to height as well as weight (Underwood and Chiuni).
Calculating the force per length of wall and adjusting force by the ratio of sample heights to 1.1 m
(42”) average height, the strength per length of earthbag research can be approximately compared
(Table 4).
Table 4: Approximate shear strength of current samples compared to previous, adjusted to
estimated 1.07 m (42”) height.
Croft stucco covered
sandbag ULS (Croft
2011)
Current medium
fill CE ULS
(avg)
Previous strong
CE before yield
(Stouter 2016)
Current strong
fill CE SI
sample at yield
Current strong
fill CE SI
sample at ULS
N = 4 kN (900 lb) 0 0 0 0
kN/m 7.8 5.5 6.8 7.1 10.4
(plf) (537) (378) (469) (488) (714)
11
Because increased normal force increases shear force required, the results for sand fill reported by
Croft are overstated in comparison to the other results. Reinforcement detailing also differed, but
with the current results this data indicates an approximate range of strengths for different fill and
reinforcement types of small earthbag panel samples with poorly connected vertical reinforcement.
Wall strength of CE built with medium to strong fill is unlikely to be based on the tensile strength of
the fabric container, indicated by the lack of fabric splitting or being cut during testing. Shear
results of the current CE samples also reached comparatively higher strengths (considering the lack
of normal force) than that reported for the sand-filled samples of similar sizes.
Recommenda"ons
To resist higher seismic forces contained earth needs strong cohesive soil with embedded
reinforcement chosen for risk levels. More than 50 mm (2”) of dried soil must completely surround
reinforcement rebar for good bond. Full embedment of a single entire vertical rebar per course is the
strongest detail and techniques must be found to stiffly connect them to footings. Provide spot
footings for and interconnect vertical rebar so that CE earthbag walls can fail structurally by steel
bending (Morris and Walker) rather than by less predictable failure of earthen fill. Since most
earthen walls in a building are needed to brace heavy adobe (Walker and Morris 1997, 7) similar to
CE walls, check the distribution and size of bracing walls.
Many earthbag buildings with all sorts of shapes, soil strengths, and qualities of reinforcement
detailing survived 0.6- 0.8 g horizontal forces in Nepal in 2015. CE walls may resist collapse well
even past ultimate strength if their barbed wire is continuous around corners, but structural design is
needed to specify buildings that will not be significantly deformed after earthquake stresses, and
will require greater stiffness than conventional earthbag.
Engineers should stop using structural data generated from sand-filled samples to design CE, and
require tests of soil fill strength. Designers should also consider all effects of reinforcement and
ductility. The unusual twisting of strong fill samples shows high toughness and the excessive
deformation indicates possible high ductility with strong soil and base-anchored rebar. Viscoplastic
behavior seen merits vibration testing. Morris and Walker’s adobe research also showed that
‘unreinforced walls provide considerably less bracing capacity without the vertical and horizontal
reinforcement’ (2000, 5). Engineers shouild not use standard design formulas based on conventional
brittle masonry, but consider the contributions of both horizontal and vertical reinforcement to
earthbag wall performance as a unique composite structural system.
Further research is needed and should focus on:
•Buildable techniques using only low-cost parts easily available in the developing world.
•Fully dried samples. Use multiple proof stacks to allow confirmation of equilibrium
moisture content before testing.
•Soil fill should be described by hand-texturing (NRCS 2017) whether particle size analysis
is available or not so that designers as well as engineers can understand research results.
•Samples with levels of soil fill strengths common to raw local soils, ranging from 1.1- 1.9
Mpa (160- 275 psi) with proven unconfined compressive strengths.
12
References
Banks, Harvey T., Shuhua Hu and Zackary R. Kenz. 2011. “A Brief Review of Elasticity and
Viscoelasticity for Solids.” Advances in Applied Mathematics and Mechanics, 3 (February), 1- 51.
https://www.cambridge.org/core/services/aop-cambridge-core/content/view/
1DFF46A2298C59CAEAC9BE72F419BFE8/S207007330000151Xa.pdf/
brief_review_of_elasticity_and_viscoelasticity_for_solids.pdf.
Blondet, Marcial, Daniel Torrealva, Julio Vargas, José Velasquez and Nicola Tarque. 2006. “Seismic
Reinforcement of Adobe Houses Using External Polymer Mesh.” Presented at First European
Conference on Earthquake Engineeering and Seismology, Geneva Switzerland, 3- 8 September
2006.
Canadel, Samuel, Ana Blanco and Sergio H. P. Cavalero. 2015. “A Comprehensive Design Method
for Earthbag and Superadobe Structures.” Materials and Design, 96, 270- 282.
Conolly, David, Steven Jin and Awais Malik. 2012. “Earthbag Walls Part 2.” Course Presentation,
Dartmouth College, Hanover, NH, USA.
Croft, Chris. 2011. “Structural Resistance of Earthbag Housing Subject to Horizontal Loading.”
Masters of Engineering, Bath University, 2011. https://www.semanticscholar.org/paper/Structural-
Resistance-of-Earthbag-Housing-Subject-%3A-Croft/
64dc03097a2fef1e41047ad787faecce85c0917c.
Daigle, Bryce Callaghan. 2008. “Earthbag Housing: Structural Behavior and Applicability in
Developing Countries.” MS Engineering Thesis, Queen's University, Kingston Ontario CA.
https://qspace.library.queensu.ca/bitstream/handle/1974/1421/Daigle_Bryce_C_200809_MscEng.pd
f?sequence=1&isAllowed=y.
Engineer’s Edge. 2017. “Mechanics of Materials.”
https://www.engineersedge.com/material_science/shear-stress.htm.
Geiger, Owen. 2010. “Reinforced Earthbag Specifications.” Earthbag Building Blog.
https://earthbagbuilding.wordpress.com/2010/12/04/reinforced-earthbag-specifications/.
Geiger, Owen. 2011. Earthbag Building Guide: Vertical Walls Step-by-Step. Thailand: Geiger
Research Institute of Sustainable Building. Retrieved from
https://store.payloadz.com/details/1049860-ebooks-architecture-earthbag-building-guide.html.
Geiger, Owen. n.d. “Expanding Earthbag Building in Nepal.” Retrieved March 2, 2017, from
Natural Building Blog: http://www.naturalbuildingblog.com/expanding-earthbag-building-in-nepal/.
Malik, Awais. 2013. “Structural Analysis of Earthbag Systems.” Honors Thesis, Dartmouth College,
Hanover, NH, USA.
Minke, Gernot. 2006. Building With Earth: Design and Technology of a Sustainable Architecture.
Basel, CH: Birkhauser.
Morris, Hugh W. and Richard Walker. 2000. “Aseismic Design and Construction of Earth Buildings
in New Zealand.” 12th World Conference on Earthquake Engineering, 2193. Auckland, New
Zealand.
13
Nordquist, S. n.d. “Earthbag Building Still Standing After Nepal Quake.” Retrieved March 2, 2017,
from Newshub NZ 3 News. http://www.newshub.co.nz/world/earth-bag-building-still-standing-
after-nepal-quake-2015050316#axzz3t6L5ULwP.
NRCS Natural Resources Conservation Service, Soils. 2017. “Guide to Texture by Feel.” USDA.
https://www.nrcs.usda.gov/wps/portal/nrcs/detail/soils/edu/?cid=nrcs142p2_054311.
Pelly, Ralph. 2009. “Plastic Limit Analysis of Earthbag Structures.” Masters of Engineering, Bath
University, 2009. https://www.istructe.org/journal/volumes/volume-89-(published-in-2011)/issue-
13/plastic-limit-analysis-of-earthbag-structures/.
Rao, K. S. Nanjundra, S. Raghunath and K. S. Jagadish. 2004. “Containment Reinforcement for
Earthquake Resistant Masonry Buildings.” 13th World Conference on Earthquake Engineering,
1968. Vancouver, B.C., Canada. http://www.iitk.ac.in/nicee/wcee/article/13_1968.pdf.
Ross, Brandon, Michael Willis, Peter Datin and Ryan Scott. 2013. “Wind Load Test of Earthbag
Wall,” Buildings, 3 (August): 532–544. doi: 10.3390/buildings3030532.
Standards New Zealand. 1998. NZS 4298: Materials and Workmanship for Earth Buildings.
Wellington, New Zealand: New Zealand Code Council.
Stouter, Patricia. 2016. “Structural Information for Earthbag and Contained Earth.” Albuquerque,
New Mexico: Build Simple Inc..
https://www.academia.edu/25067309/Structural_Information_for_Earthbag_and_Contained_Earth_
2016.
Underwood, Rod and Michele Chiuni. 1998. Structural Design: A Practical Guide for Architects.
New York: John Wiley and Sons.
University of Memphis Civil Engineering. n.d. “Chapter 10: Bond and Anchorage: Reading notes
CIVL 4135. http://www.ce.memphis.edu/4135/PDF/Notes/Chap_10_1_Bond_&_Anchorage.pdf.
Vadgama, Nikul. 2010. “A material and structural analysis of earthbag housing.” Masters of
Engineering, Bath University, 2011. http://earthbagbuilding.com/pdf/vadgama.pdf.
Walker, Richard and Hugh Morris. 1997. “Development of New Performance Based Standards for
Earth Building.” Department of Civil and Resource Engineering, University of Auckland.
https://archive.org/details/DevelopmentOfNewPerformanceBasedStandardsForEarthBuilding/
page/n1/mode/2up.
14
Appendix A
Test results
Table 10: MI Results Table 11: ME Results
Medium soil, inserted rebar Medium soil, external rebar
% deformation kPa % deformation kPa
0 0 0
1.81 3.96 0.85 3.42
1.97 5.55 2.23 4.11
3.11 7.13 3.51 4.11
4.46 7.13 3.30 4.11
5.08 7.53 4.10 4.79
5.44 7.93 5.16 6.16
5.59 5.95 5.21 6.16
5.80 6.23 5.32 6.84
8.29 8.31 5.85 5.38
11.91 8.31 8.83 5.38
15
Table 9: SI Results Table 9: SE Results
Strong soil, inserted rebar
Strong soil, external rebar
% deformation kPa % deformation kPa
0 0 0 0
1.94 1.91 0.51 3.41
2.16 1.91 0.82 4.10
3.45 7.46 0.92 4.44
4.21 7.65 2.15 4.78
4.31 7.84 2.15 4.78
6.47 7.84 3.17 6.83
6.47 8.60 3.68 6.83
7.12 8.60 4.81 5.37
8.20 9.56 8.79 5.37
9.06 9.56 12.58 7.16
10.79 10.51 14.62 7.16
12.84 9.56 15.24 8.05
14.45 9.56 15.95 8.41
16.50 11.47 17.59 7.16
20.49 8.60 18.30 6.26
23.30 8.60 18.00 6.89
25.02 11.47 19.22 6.89
16
Appendix B
17
