V.

Ariza et al

CEMENT DOSAGE AND COMPRESSIVE STRENGTH CORRELATION

IN RAMMED EARTH WALLS: A CASE STUDY
CORRELACIÓN ENTRE LA DOSIFICACIÓN DEL CEMENTO Y LA
RESISTENCIA A LA COMPRESIÓN EN MUROS DE TIERRA
APISONADA: UN ESTUDIO DE CASO
Victor Ariza-Flores1 Elizabeth Ortega-Hilario2
1 Universidad Tecnológica del Perú, Lima, Perú
1
Universidad de Huánuco, Huánuco, Perú
Recibido (Received): 27 / 03 / 2023 Publicado (Published): 20/12/2023

ABSTRACT

This study investigated the correlation and development of a linear regression model between the variables "added cement"
and "compressive strength of rammed earth walls" built with aggregates obtained from Colpa Alta, Huánuco, Peru through
bivariate analysis. This analysis was motivated by the growing difficulty of the local population to build confined masonry housing
due to the increase in construction material prices caused by the COVID-19 pandemic. In this scenario, rammed earth walls offer
a more affordable alternative to the traditional system, although it is necessary to improve their structural capacity. 60 aggregate
samples were collected in situ, following the Peruvian Technical Standard E-080. Subsequently, they were divided into four
groups of 15 samples each, where 5%, 10% and 15% of the aggregate was replaced with cement. Compressive strength tests were
carried out and the results were analyzed using statistical techniques. The findings revealed a significant increase in compressive
strength in samples containing cement compared to conventional rammed earth blocks. It was found that there is a strong
correlation between the "added cement" variable and the "compressive strength of rammed earth walls". The linear regression
model quantitatively explained the influence of cement on compressive strength.

Keywords: rammed earth walls, cement, compressive strength, correlation, linear regression

RESUMEN

Este estudio investigó la correlación y el desarrollo de un modelo de regresión lineal entre las variables "cemento agregado" y
"resistencia a la compresión de paredes de tierra apisonada" construidos con agregados obtenidos en Colpa Alta en Huánuco,
Perú mediante análisis bivariado. Este análisis fue motivado por la creciente dificultad de la población local para construir
viviendas confinadas de mampostería debido al aumento de los costos de los insumos de construcción provocados por la
pandemia de COVID-19. En este escenario, los muros de tierra apisonada ofrecen una alternativa más asequible al sistema
tradicional, aunque es necesario mejorar su capacidad estructural. Se recolectaron 60 muestras de agregados in situ, siguiendo
la Norma Técnica Peruana E-080. Posteriormente, se dividieron en cuatro grupos de 15 muestras cada uno, donde el 5%, 10% y 15%
del agregado fue reemplazado por cemento. Se efectuaron ensayos de resistencia a la compresión y los resultados se analizaron
mediante técnicas estadísticas. Los hallazgos revelaron un aumento significativo en la resistencia a la compresión en muestras
que contienen cemento en comparación con los bloques de tierra apisonada convencionales. Se obtuvo una alta correlación para
la variable "cemento incorporado" y la "resistencia a la compresión de las paredes de tierra apisonada". El modelo de regresión
lineal explicó cuantitativamente el efecto del cemento en la resistencia a la compresión.

Palabras clave: bloques de tapial, cemento, resistencia a la compresión, correlación, regresión lineal

1. INTRODUCTION low construction cost, because it involves the utilization

The construction of rammed earth houses or of materials specific to the area and does not require
compressed earth block (BTC) involves the use of clay skilled labor; In addition to having an excellent thermal
soil, which is compacted using wooden rammers. On and acoustic capacity.
this system, Roux and Espuna mention that "vestiges
found in the Asian, European and American continents Based on the most recent population survey
confirm the use of earth construction techniques for conducted by the National Institute of Statistics and
many years" [1, pp.23] And its expanded use is due to its Informatics, there are 104 930 private homes with
* Corresponding author.: E-mail: adobe or rammed earth, which represents 55.3% of the
E19200075@postgradoutp.edu.pe total private homes in the region [2, pp.31]. Although

DOI: https://doi.org./ 10.21754/tecnia.v33i2.1685 TECNIA Vol.33 N° 2 Julio-Diciembre 2023

 V. Ariza et al

the number of homes with noble material in 2017 stable compacted lump containing 10% cement is 39.02
increased by 53.3% compared to 2007 [2, pp.26], due to kg/cm2 at 7 days and 76.96 kg/cm2 at 14 and 21 days. In
the fact that families have migrated to the confined addition, increases of 52.66%, 154.83% and 252.20% were
masonry system that provides greater structural safety, evidenced in 7 days, 14 days and 21 days respectively
better finish and comfort, its implementation is compared to standard BTC. Because of this, it is
expensive due to the materials required and the concluded that the addition of cement to BTC increases
participation of professionals throughout the process, its strength.
worsening following the COVID-19 pandemic; that is
why there is still a significant number of users with 2.1 TAPIAL
preference to the BTC system. Due to this reality, there The Peruvian Technical Standard E.080 [6, pp. 5]
is a need to increase the compressive strength of defines rammed earth as a "construction technique that
conventional BTC, adding certain percentages of uses wet earth poured into firm molds (boards), to be
cement to the aggregate obtained from the area, to compacted by layers using wooden mallets or
reduce structural failures in the walls built with this rammers".
system for the benefit of the safety of the population.
2.2 RAMMED EARTH AND FORMWORK UNIT
2. BACKGROUND Peruvian Technical Standard E.080 [6, pp. 18]
A comprehensive review of the relevant scientific stipulates that rammed earth units must have specific
literature, related to the scope of our research, was dimensions: a minimum width of 0.40 m, a maximum
conducted with the aim of obtaining a solid theoretical height of 0.60 m, a maximum length of 1.50 m, and the
context and understanding the experimental wood used for formwork must have a minimum
background prior to conducting the field tests. One of thickness of 20 mm.
these references was that of Samaniego and Sarmiento
[3] who clarified to what extent the additives used to 2.3 AGGREGATE
modify the mechanical properties of concrete alter the In the present work it has been called as an
compressive strength and density of cement-stabilized aggregate to the earth used in the construction of BTC,
rammed walls. Research requires an experimental the Peruvian Technical Standard E.080 [6, pp.5] defines
quantitative approach. The values of the results the earth as "construction material composed of four
obtained from the test of the compressive strength of basic components: clay, silt, fine sand and coarse sand".
the rammed earth gave as a standard of 10.71 kg / cm2;
In addition, in the methodology of replacing the earth 2.4 PORTLAND CEMENT TYPE 1
with cement in 6%, 8% and 10% a resistance of 7.2 kg / The main components of this type of cement are
cm2, 10.73 kg / cm2 and 13.47 kg / cm2 was obtained tricalcium silicate, Ca 3 SiO5, beta dicalcium silicate, Ca 3
respectively. It is concluded that mixture No. 12, SiO 4, and lime (CaO, 60%) and alumina (Al2O3) and
composed of 10% cement and even air, has the best Portland Clinker. In addition, his theories of structure,
compressive strength of 29.48 kg/cm2; This obtained constitution and the process of formation are diverse
22% more resistance when contrasting with the [7]. According to Sánchez, this type of cement is used in
stabilized rammed earth with 10% cement and an different works in general since no peculiar properties
increase of 175% compared to the standard tapial. are requested to this type of cement [8, pp. 49].

On the other hand, Garcia [4] focused on testing the 3. METHODOLOGY

compressive strength of unburned masonry by
replacing 3%, 6%, 9% and 12% of the soil with grade I 3.1 AGGREGATE EXTRACTION
Portland cement and lime. The result obtained from the The optimal location for obtaining aggregates for
compressive strength of standard adobes is on average the manufacture of BTC in the region of Colpa Alta,
11.3 kg/cm2, and when compared to test samples Huánuco, was identified. It was considered important
containing 9% lime and 12% cement, the results are 63 that the soil used meets the evaluation criteria
kg/cm2 and 73.47 kg/cm2 respectively; Therefore, it is corresponding to the presence of clay, to ensure its
concluded that the addition of cement and lime suitability in the construction of rammed earth walls
presents higher performance compared to stipulated in NTP E0.80 [6].
conventional adobe.
According to [6, pp.19], the first test performed was
Another important research work for data the "Clay Ribbon" in which a 12mm diameter cylinder
processing was that of Chávez and Medina [5] whose was molded with a wet mud sample, and then flattened
objective was to produce blocks of compacted earth with fingers., forming a 4mm thick tape. If you hang it
mixed with cement to be used in the construction of as much as possible between 20cm and 25 cm long, the
houses in rural areas in the province of San Martín. The soil has a high clay content.
results show that the design compressive strength of a

DOI: https://doi.org./ 10.21754/tecnia.v33i2.1685 TECNIA Vol.33 N° 2 Julio-Diciembre 2023

 V. Ariza et al

The second test applied was the "Dry Resistance" The compressed samples were reserved for 28 days
test where with the minimum amount of water, four in an area away from moisture and fresh, so that they
pellets were formed: Then these dried for 48 hours, have a slow drying in order to prevent cracking. [6, p.
protecting them from moisture and water in general. 18]
Once dry they were pressed with the fingers, in our case
none broke or cracked, so the aggregate could be used 3.4 COMPRESSIVE STRENGTH TEST
as a building material. If this had not happened, the test Breaking stress tests were conducted by applying
would have to be performed again, if they still do not axial loads or compressive forces to previously
pass the test, the quarry is discarded. prepared and mixed cubes at a set speed, aiming to
measure the compressive strength of the soil block until
The third test was "Moisture content" where a fist- failure was induced. The sample resistance was
sized aggregate ball was formed, compressing it determined by dividing the peak force attained in the
strongly; It was then released to a firm surface of1.10m test, as specified by NTP-339.034, by the specimen's
high. The earth ball broke into more than 5 pieces, so cross-sectional area. [11].
the amount of moisture was indicated. [6, p. 20]
Keep in mind that in [6, p.15] explains that the
3.2 SOIL MECHANICS TESTS average of the four best samples of 6 cubes must be
The granulometric analysis test was conducted to greater than or equal to the last resistance observed.
assess the percentage of clay, silt, and gravel content in
the extracted aggregate, following the guidelines of 3.5 DATA ANALYSIS AND PROCESSING
NTP-400.012 [9] The plasticity index (PI) was also After collecting the data from applying the test [11]
determined, whose definition according to NTP 339.129 using laboratory records, the data were processed
[10, pp. 4] is the range of soil moisture content in which using Excel spreadsheets, calculating the compressive
soil behaves plasticly; to determine this value, the liquid strength of the samples. The statistical evaluation of
limit and the plastic limit were previously calculated in the data of the studied samples was carried out using
order to make an arithmetic subtraction of the values in the statistical program SPSS V.26, with which the
the same order mentioned. The Liquid Limit symbolizes measures of central tendency of each group of samples
the soil moisture content percentage at the transition were determined. First, we assessed that the data
from its liquid to plastic states, determined by the samples proceed from a Gaussian distribution using the
Casagrande test. Conversely, the Plastic Limit denotes Shapiro-Wilk test since no more than 50 samples were
the moisture content percentage at the boundary analyzed and the Student’s parametric t test to
between the plastic and semisolid states. To calculate compare the sample measurements.
LP, approximately 20 grams of the material prepared
for the LL calculation are kneaded. This material is then 4. ANALYSIS OF RESULTS
allowed to lose moisture until cylinders with a diameter In the TABLE I, the compressive strengths obtained
of 3.2 mm form. The process involves gradually reducing from the standard samples are presented, which is
the diameter until the cylinder starts to crack or formed only by the aggregate.
crumble, indicating the need to measure the material's
weight to assess moisture content. This procedure is TABLE I
Compressive strength of conventional rammed earth blocks
repeated with another soil sample, and the average
moisture content from both tests is calculated to Conventional Compress Area (cm2) Compressive
rammed ion force strength
determine the LP. earth blocks (Kg) (Kg/cm2)
or standard
(sample)
The IP value in ranges of IP>20, 20≥IP≥7, 7>IP>0 and
1 2046 102.01 20.06
IP=0 indicates the presence of very clay soils, clay soils, 2 2056 100.00 20.56
little clay soils and clay-free soils, respectively. 3 2038 104.04 19.59
4 2239 100.00 22.39
5 2137 98.01 21.80
3.3 HANDLING AND PROCESSING OF SAMPLES 6 2120 100.00 21.20
7 2048 98.01 20.90
After the soil mechanics tests, 4 groups of 15 8 2069 96.04 21.54
samples were formed each, making up a total of 60 soil 9 2139 100.00 21.39
samples. One of the groups was left unchanged; to the 10 2024 102.01 19.84
11 1970 102.01 19.31
other three, 5%, 10% and 15% of Port Cement were added 12 2146 100.00 21.46
to the type I, respectively. 13 2026 102.01 19.86
14 2126 100.00 21.26
15 1988 100.00 19.88
Then the compressed samples were elaborated, Note: Calculation of the compressive strength after division of the
according to NTP E.080 [6, p.15], in molds of 0.1 x 0.1 x compressive force obtained from the test by the cross-sectional area
of the sample.
0.15 m to which it was compacted by applying 10 blows
with a mallet of 5 Kg.

DOI: https://doi.org./ 10.21754/tecnia.v33i2.1685 TECNIA Vol.33 N° 2 Julio-Diciembre 2023

 V. Ariza et al

14 21.26 25,47 32,30 39,02

Considering TABLE I, Fig. 1 was elaborated, which 15 19.88 25,96 30,34 39,83
shows the behavior of the compressive strength of the Note: The compressive strengths of the standard samples and the
blocks. In addition, TABLE II shows the measures of compressive strengths of the samples with 5%, 10% and 15% Portland
central tendency of the data collected. cement type I are displayed.

Considering TABLE III, Fig. 2 showing the behavior of

the compressive strength of the blocks compared to
the averages of the resistances of the blocks with
addition of cement of 5%, 10% and 15%.

Fig. 1. Graph of the compressive strength of conventional rammed

earth blocks [12]

TABLE II
Central tendency measures of compressive strength data from
conventional rammed earth blocks
Sample Valid 15 Fig. 2. Frequency graph of conventional rammed earth blocks'
number compressive strength versus earth blocks with 5%, 10%, and 15%
Lost 0 cement by dry weight [12]
Stocking 20.7360
Fashion 19.31to
Note: Average and mode of 15 valid data. In the figure above you can see a marked
improvement in the compressive strength of the
The mean for the analyzed data of the compressive samples with cement content compared to the results
strength of conventional rammed earth blocks at 28 obtained from the standard samples. The first step in
days is 20.74 kg/cm2. determining whether cement is influenced by
aggregates extracted from Colpa Alta is to verify
From the altered samples, in which 5%, 10% and 15% whether the compressive strength values for the
of cement were added, the data of compressive datasets meet the normality hypothesis for data
strength were obtained, which on average were distribution using the Shapiro-Wilk test. (n<50
f'c=25.27 Kg/cm2, f'c=30.75 Kg/cm2 and f'c=39.43 samples).
Kg/cm2, respectively. To this end, the average of these TABLE IV
Compressive strength normality test
was determined, whose values are shown in TABLE III
together with the compression resistance of the Shapiro-Wilk
Statistical Gl Gis.
unaltered samples. Compressive 0.942 15 0.412
strength of
TABLE III conventional
rammed earth
Compressive strength of cement earth blocks from 5% to 15% with
blocks
respect to the dry weight of the mixture
Sample Compressive Compressive Compressive Compressive Compressive 0.940 15 0.387
strength of strength of strength of strength of strength of
standard samples blocks with blocks con blocks with 15% rammed earth
(Kg/cm2) 5% cement 10% cement cement blocks with 5% by
addition addition addition weight of cement
(Kg/cm2) (Kg/cm2) (Kg/cm2)

Compressive 0.982 15 0.982

1 20.06 25,18 31,53 39,32 strength of
2 20.56 24,58 29,33 37,36 rammed earth
3 19.59 24,88 30,44 37,45 blocks with 10% by
4 22.39 24,90 31,07 39,72 weight of cement
5 21.80 24,93 30,41 38,51
6 21.20 25,26 31,39 39,18 Compressive 0.967 15 0.805
7 20.90 24,01 31,01 38,24 strength of
8 21.54 24,63 29,60 42,09 rammed earth
9 21.39 26,43 30,94 39,69 blocks with 15% by
10 19.84 24,23 30,62 40,58 weight of cement
11 19.31 26,55 29,87 39,07
12 21.46 26,74 31,52 40,00
13 19.86 25,24 30,85 41,44

DOI: https://doi.org./ 10.21754/tecnia.v33i2.1685 TECNIA Vol.33 N° 2 Julio-Diciembre 2023

 V. Ariza et al

Note: Applied to conventional rammed earth blocks and average occurred, Welch's t-test could have been used, which
compressive strength for rammed earth blocks with 5%, 10% and 15%
does not assume equality of variances.
by weight of Portland cement type.

After applying the Shapiro-Wilk as indicated by the Considering that all samples complied with Gaussian
TABLE IV, a p-Value = 0.412 was obtained for the distribution. The parametric test Student's t of two
compressive strength of the conventional rammed independent samples was applied, pairing the standard
earth blocks, a p-Value = 0.3 87 for the compressive samples with each of the samples added with cement.
strength for the rammed earth blocks with 5%, a p-Value Here, the null hypothesis (H0) posits that the difference
= 0. 982 for compressive strength for rammed earth in the averages of the two populations from which the
blocks with 10% and a p-Value=0. 805 for compressive samples are drawn is not statistically significant. In
strength for rammed earth blocks with 15% by weight of essence, it asserts that the disparity between the means
cement, so the null hypothesis is accepted (H0: the of these populations amounts to zero; H0: μ1 - μ2 = 0
samples comply with normal distribution) given that and H1: μ1 - μ2 ≠ 0. The test is either a bilateral
p>0.05, for each of the four cases studied (pattern, 5%, hypothesis or a two-tail test. The outcomes of the
10% and 15% addition) Student's t-test are displayed in Table VI.

TABLE VI
After that, the test of equality of variances or Student's t-test
hocedasticity of variances is carried out by applying the
t Gl Gis.
parametric Levene test between the standard sample
and each of the samples added with cement as shown Standard vs 5% -14.076 28 3.1612E-14
added cement
in the TABLE V. Levene's test's foundational
assumption, or null hypothesis, is that all groups being
analyzed have equal variances, suggesting the absence
of notable variance discrepancies among them. The Pattern vs 10% added -31.767 28 1.6524E-23
alternative hypothesis (H1) posits that at a minimum, cement

one of the population variances deviates from the rest.

𝐻0 ∶ 𝜎12 = 𝜎22 Pattern vs 15% added -45,136 28 1.066E-27

𝐻1 ∶ 𝜎12 ≠ 𝜎22 cement

Should the Levene test produce a statistically

significant outcome (that is, the p-value falls below the
established significance threshold, typically 0.05), the Note: Paired differences between conventional rammed earth
null hypothesis would be dismissed, leading to the blocks and the compressive strength of rammed earth blocks of 5%,
conclusion that the population variances are not 10% and 15% by weight of cement.
equivalent.
When applying the t test, the significance level
TABLE V indicator P-Value close to zero was obtained, which is
Test of independent samples less than 0.05; so the null hypothesis is rejected in all
Levene test cases, concluding that cement has an influence on the
F Gis.
compressive strength of BTC made with aggregates
Pattern – 5% cement 0.979 0.331
from Colpa Alta, Huánuco.

Pattern – 10% 1.409 0.245

However, to quantitatively determine the degree of
cement influence or level of correlation between the added
cement and the compressive strength of the reinforced
earth walls, it is necessary to introduce multivariate
Pattern – 15% 0.498 0.486 analysis taking into account that samples have been
cement
made for different percentages of incorporated cement
Note: Applied to conventional rammed earth blocks and average (standard, 5% cement, 10% cement and 15% cement; that
compressive strength for rammed earth blocks with 5%, 10% and 15% is, 4 study groups). To introduce us to multivariate
by weight of portland cement type.
analysis, the homogeneity of variances test or Levene
test is applied for the four groups of studies. The null
As shown in TABLE V, for all cases the significance hypothesis (H0) and the alternative hypothesis are as
value p>0.05; that is, the null hypothesis in which the follows:
variances are equal is satisfied. With this, you can
proceed to apply the parametric test Student's t, which 𝐻0 ∶ 𝜎12 = 𝜎22 = 𝜎32 = 𝜎42
assumes that the variances are equal. Had this not

DOI: https://doi.org./ 10.21754/tecnia.v33i2.1685 TECNIA Vol.33 N° 2 Julio-Diciembre 2023

 V. Ariza et al

𝐻1 ∶ 𝑎𝑡 𝑙𝑒𝑎𝑠𝑡 𝑜𝑛𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑠 𝑖𝑠 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡

Note: The ANOVA test allows us to compare the measurements of
the four study groups.
The TABLE VII shows the results of the Levene test. The
significance level p is 0.298 > 0.05, so the null hypothesis It is important to note that the one-factor ANOVA
is accepted. does not identify which specific groups differ from each
other. To determine which groups have different
TABLE VII
Homogeneity of Variances Test means, the Tukey test and the Bonferroni test, after
performing the one-factor ANOVA, were performed to
Levene Gl1 Gl2 Gis.
statistics compare results as shown in Figure 3.
Resistance Based on 1.258 3 56 0.298
the
average

Note: Levene test for all four study groups.

The results of Table VIII are important since when

performing the bivariate analysis between the standard
sample and another sample added with cement, they
resulted with a value p<0.05, for which the null
hypothesis of equality of variances had been accepted.
Now when performing the multivariate analysis, the
equality of variances is reaffirmed.

After that, the next step is to apply the Analysis of

Variances known as one-factor ANOVA. This is based on
the principle of decomposition of the total variance into
two components: the variance attributable to the effect
of the factor under study (variance between groups)
and the variance attributable to random error (variance
within groups). The null hypothesis (H0) in the one-
factor ANOVA holds that all the means of the groups are Fig. 3. Tukey and Bonferroni test results among study groups
equal, that is, there are no significant differences
between them: From the analysis of Figure 3, a greater influence of
cement added with 15% by weight in the variable
𝐻0 ∶ µ1 = µ2 = µ3 = µ4 compressive strength of reinforced earth walls is
evident. However, as shown in TABLE IX, the deviation
In this context, μ1, μ2, μ3, and μ4 denote the mean in the results is greater in this study group, an effect
values for the four investigated groups (pattern, 5% that does not occur in other groups where even the
cement, 10% cement, and 15% cement addition). The deviation is usually less than the pattern.
alternative hypothesis (H1) postulates that a minimum
of one population mean differs from the remaining TABLE IX
Descriptive data of samples
ones:
Resistance
𝐻1 ∶ 𝑎𝑡 𝑙𝑒𝑎𝑠𝑡 𝑜𝑛𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑠 𝑖𝑠 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡 Stocking Standard deviation
Category Boss 20.74 0.93
Where μ1, μ2, μ 3, μ4 represent the means of the 5% Cement 25.27 0.83
populations of the 4 groups under study (pattern, 5%
cement, 10% cement and 15% added cement). TABLE VIII 10% Cement 30.75 0.79
shows the results of the test and the null hypothesis is
verified because p = 1.422E-18 < 0.05. This also reaffirms 15% Cement 39.43 1.31
the bivariate analysis between two samples.
TABLE VIII
One-factor ANOVA test
To determine the correlation, the variable "cement
Sample Sum of Gl Quadrati F Gis.
squares c mean addition" can be classified as an ordinal type (0%, 5%, 10%
and 15%). When performing this analysis, a correlation
Between 2912.070 3 970,69 999.165 1.422E-48 factor of 0.968 with Spearman's Rho factor and a
groups
Within 54.404 56 0.972 significance level of 1.1475E-36<0.05 are identified, so it
groups is considered that there is a "strong" correlation
Total 2966.474 59 30,44

DOI: https://doi.org./ 10.21754/tecnia.v33i2.1685 TECNIA Vol.33 N° 2 Julio-Diciembre 2023

 V. Ariza et al

between the variable "cement addition" and Where:

"compressive strength. Y: Compressive strength expected (kg/cm2)
However, it is also convenient to adapt the variable X: % cement added (expressed in decimal places)
"cement addition" as a quantitative variable (0, 0.05,
0.10 and 0.15) to build a linear regression model. The R2 coefficient, indicative of the goodness of fit
in linear regression models, has a value of R2 = 0.959.
This indicates that the regression model can explain
95.9% of the variability in compressive strength through
the variable of cement addition. The remaining 4.1% is
attributable to other variables not incorporated into the
model. An R2 value approaching 1 signifies that the
regression model exhibits a strong fit to the data and
elucidates a significant portion of the variability in the
independent variable. Figure 5 shows the linear
regression model including confidence intervals of 95%

Fig. 4. Scatter plot considering the variables "added cement" and

"compressive strength."

When making the determination by Pearson's

correlation factor, it is identified that the factor is
equivalent to 0.979 with a significance 8.6049E-42 so it
is accepted that there is a strong correlation between
the variables. With this, we proceed to build the linear
regression model. The model equation has the
structure:
𝑌 = 𝛽0 + 𝛽1 𝑥 (1) Fig. 5. Linear regression model with confidence intervals

Where: Figure 6 depicts the regression model alongside

𝛽0 ∶ 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑦 𝛽1 : 𝑙𝑖𝑛𝑒𝑎𝑟 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡
prediction intervals. Contrary to confidence intervals,
prediction intervals offer a more cautious approach to
TABLE X shows the results of the determined model,
estimating at the 95% level.
where β0 = 19.81 and β1 = 123.418. This means that a
minimum average strength of 19.81 kg/cm2 is expected
for samples without cement addition and 1.23148
kg/cm2 for each cement addition percentage unit.

TABLE X
Determination of the linear regression model

Non-standardized Standardized
coefficients coefficients

Dev.
Model β Error Beta t Gis. Fig. 6. Linear regression model with prediction intervals
1 (Constant) 19,810 0,315 62,982 ,000
CONCLUSIONS
Cement 123,148 3,362 ,979 36,624 ,000
- The 3 tests according to [6], carried out in the
Dependent variable: Compressive strength Colpa Alta - Huánuco quarry for the prior
verification of the adequate clay content,
The structure of the linear regression model complied as indicated by the same.
describing the average compressive strength with the - It was verified through the tests of granulometric
addition of Portland cement type I using High Colpa analysis, liquid limit and plastic limit, described
aggregates in Huánuco is: in [9], [10], that the samples extracted comply
with the provisions for the preparation of BTC
𝑌 = 19.81 + 123.148𝑥 (2) according to NTP-E.080 [6].

DOI: https://doi.org./ 10.21754/tecnia.v33i2.1685 TECNIA Vol.33 N° 2 Julio-Diciembre 2023

 V. Ariza et al

- The stabilization designed with 5%, 10% and 15% [12] E.F. Ortega Hilario, “Influencia del cemento en la resistencia a
la compresión en muros de tapial, hecho con agregados
are adequate to quantify the impact of this
extraídos de Colpa Alta – Huánuco – 2021”, B.S. thesis,
material on increasing compressive strength, Universidad de Huánuco, Huánuco, 2022.
showing that there is a strong correlation
between the added percentage of cement and
the compressive strength. The results of this
investigation are limited to this range of
application (5%, 10% and 15%)
- The rise in the average compressive strength of
the BTC correlated with the amount of cement
added by weight; that is, a higher cement
content in the aggregate used will allow a
greater compressive strength. The
mathematical model of linear regression
determined is 𝑌 = 19.81 + 123.148𝑥, where
“x" is the added % of cement while “Y" is the
compressive strength of rammed earth walls
expressed in kg/cm2.
- ~It was determined that cement significantly and
positively influences the compressive strength
of BTC; being that samples with addition of 15%
to weight influence greater incidence but are
those that have results with greater deviation
from samples with 5% and 10% by weight.

REFERENCES

[1] J. A. Espuna and R. S. Roux, Bloques de Tierra Comprimida

adicionados con fibras naturales, México: Plaza y Valdés
Editores, 2012.
[2] Instituto Nacional de Estadística e Informática (INEI),
"Características de las viviendas particulares censadas" in
Perú: Características de las viviendas particulares y los
hogares, Perú: INEI, 2017.
[3] C. E. Samaniego and N. Sarmiento, "Evaluación de tapial
estabilizado con cemento ante la adición de aditivos
comerciales", B.S. thesis, Universidad de Cuenca, Ecuador,
2019.
[4] A. W. Garcia, " Resistencia a la compresión de una unidad de
albañilería de adobe, sustituyendo el 3%, 6%, 9% y 12% de
tierra por cal y cemento portland tipo I. Distrito de Huaraz
- Huaraz 2017", B.S. thesis, Universidad San Pedro, Peru,
2018.
[5] S. Chavez and O. Medina, "Diseño de bloque de tierra
comprimida estabilizada con cemento, como material
sustentable, para su uso en la construcción de vivienda
rural en la provincia de San Martín, Región San Martín", B.S.
thesis, Universidad Nacional de San Martin, Peru, 2020.
[6] Norma E.080 Diseño y Construcción con Tierra Reforzada, R.M
Nº 121-2017-VIVIENDA, Ministerio de Vivienda, Construcción
y Saneamiento, Perú, 2017
[7] M. A. Guiraldo and J. I. Tobón, "Evolución mineralógica del
cemento portland durante el proceso de hidratación", Dyna,
vol. 73, no. 148, pp. 69–81, mar. 2006.
[8] D. Sanchez de Guzman, Tecnología del Concreto y del Mortero.
Colombia: Bhandar Editores, 2001.
[9] Dirección de Normalización - INACAL, "AGGREGATES.
Granulometric analysis of fine and coarse aggregate. Test
method", NTP 400.012, Perú, 2021.
[10] Dirección de Normalización - INACAL, "SOILS. Standard Test
Method for Liquid Limit, Plastic Limit, and Plasticity Index
of Soils", NTP 339.129, Perú, 2019.
[11] Dirección de Normalización – INACAL, “Concrete Standard
Test Method for Compressive Strength of Cylindrical
Concrete Specimens”, NTP 339.034, Peru, 2021.

DOI: https://doi.org./ 10.21754/tecnia.v33i2.1685 TECNIA Vol.33 N° 2 Julio-Diciembre 2023