You are on page 1of 8

Materials Today: Proceedings 57 (2022) 859–866

Contents lists available at ScienceDirect

Materials Today: Proceedings


journal homepage: www.elsevier.com/locate/matpr

Prediction of compressive strength of self-compacting concrete using


four machine learning technics
Yousef El Asri a,⇑, Mouhcine Ben Aicha b, Mounir Zaher a, Adil Hafidi Alaoui a
a
Abdelmalek Essaâdi University, FST, Mechanical and Civil Engineering Laboratory, Tangier, Morocco
b
Structure and Materials Laboratory, National School of Architecture, Rabat, Morocco

a r t i c l e i n f o a b s t r a c t

Article history: In self-compacting concrete (SCC), admixtures and other modifiers are commonly used to decrease
Available online 2 March 2022 bleeding and segregation. Higher air content may be necessary to improve the mixture’s ability to with-
stand freezing and thawing. Based on the suggestions of the AFGC and EFNARC, several SCC mixes were
Keywords: created in this study. There was an air-entrainment admixture (AEA) applied in various percentages.
Self-compacting concrete Experimental studies have been performed to examine the mechanical characteristics of SCC’s, such as
Compressive strength sonic velocity (at 1 and 7 days) and compressive strength (at 1, 7, and 28 days) using different air content.
Multiple linear regression
This mechanical characterization is used taking into account void ratio and water absorption.
Support vector regression
Decision tree regression
In this sense, the objective of this work is to predict the compressive strength at 28 days from the
Random forest regression intrinsic parameters such as void ratio, water absorption, and the mechanical responses, at a young
age, such as sonic velocity and compressive strength (at 1 and 7 days).
To do this, we used the most famous machine learning algorithms to know: Multiple Linear Regression
(MLR), Random Forest Regression (RFR), Decision Tree Regression (DTR) and Support Vector Regression
(SVR). Finally, the best-proposed model is given on the basis of statistical comparison between the differ-
ent used algorithms.
Copyright Ó 2022 Elsevier Ltd. All rights reserved.
Selection and peer-review under responsibility of the scientific committee of the Third International Con-
ference on Aspects of Materials Science and Engineering.

1. Introduction as slump flow, V-Funnel, and L-Box, as well as rheological behavior


variables such as plastic viscosity and yield stress, are used to char-
In construction sites, concrete placing is made simpler with acterize the fluidity of the mixture. To put it another way, while
self-compacting concrete (SCC). Under the effect of gravity, SCC slump flow tests may be used to assess the yield stress of the mix-
will flow and condense easily even in the absence of vibration. ture, V-Funnel testing can be used to determine the plastic viscos-
SCC is considered a non-Newtonian fluid of the Bingham type ity [7–13].
[1–4]. Its ability to fill formwork and shapes, even narrow, deep, Recent decades have seen the rise of machine learning (ML)
and strongly reinforced, ensures stability against segregation and techniques as an appealing modeling method applicable to a wide
bleeding through to the absence of vibration [5,6]. range of scientific subjects, including the field of material science.
Despite its great fluidity compared to ordinary concrete, this These data points can be utilized to develop an accurate surrogate
type of concrete has good compressive strength. It is important model for pre-selected model parameters, which reduces the need
to know that depending on the desired use, there are SCC of ordi- for expensive and time-consuming experiments. Currently, several
nary resistance and SCC of ultra-high performance resistance. studies have used ML algorithm tools to predict compressive
Water, cementitious materials, minerals additives (such as strength, modulus of elasticity, and other properties of self-
limestone filler, and silica fume), granular skeleton (sand and compacting concrete [14–19].
gravel), and chemical additions (such as superplasticizers) are all In this sense, to better understand the correlation between
used to strengthen and plasticize the concrete. Empirical tests such mechanical strength at 28 days, and mechanical and intrinsic
parameters such as compressive strength at a young age (1 and
⇑ Corresponding author. 7 days), ultrasonic velocity, water absorption, and void ratio, this
E-mail address: elasriyousef@gmail.com (Y.E. Asri).
study will analyze, by ML techniques, the mathematical and

https://doi.org/10.1016/j.matpr.2022.02.487
2214-7853/Copyright Ó 2022 Elsevier Ltd. All rights reserved.
Selection and peer-review under responsibility of the scientific committee of the Third International Conference on Aspects of Materials Science and Engineering.
Yousef El Asri, Mouhcine Ben Aicha, M. Zaher et al. Materials Today: Proceedings 57 (2022) 859–866

statistical models developed to efficiently predict the long-term of samples. During modeling, each tree randomly selects fea-
mechanical resistance according to the parameters mentioned tures to divide the attributes of the internal nodes, thus forming
above. a forest. The result of the final output is the result of the com-
Consequently, the originality of this work lies in the use of the plete decision tree. For the regression algorithm, the final result
four most popular ML methods, namely: support vector regression of the prediction is the average value of the output results of all
(SVR), decision trees regression (DTR), random forest regression trees [24,25].
(RFR), and multiple linear regression (MLR).  Decision Tree Regression (DTR) model: In a ML system, a DT
The ultimate purpose of the proposed research is to find the algorithm separates the data into sub-groups in a machine
best mathematical model, with the most representative statistical learning (ML) system. When it comes to classification and
coefficients, of the four Machine Learning methodologies stated regression problems, Decision Trees can be utilized to execute
previously in order to predict the compressive strength of an SCC. split tests in their core nodes and forecast the target example
A dataset of 100 samples with similar physical and chemical in their leaf nodes [26].
characteristics will be used to find the best ML models for forecast-  Support Vector Machine (SVM) model: This model is a super-
ing SCC’s compressive strength at 28 days [20]. vised learning technique used for regression and classification
analysis. The data is represented as a point map, and the solu-
2. Machine learning algorithms tion is the hyperplane with the greatest potential deviation that
divides this data into two classes. In this interval, each point is
The study of computer algorithms that improve automatically represented by a support vector [27]. For more precision, the
based on previous experiences and data is commonly known as support vector regression (SVR) was used in this study.
machine learning [21]. Supervised Learning, Unsupervised Learn-
ing, and Reinforcement Learning are the three types of machine 3. Research methodology
learning described in the literature [22]. Because we already have
input and output data, we’ll be using supervised machine learning 3.1. Experimental data
in this paper. The following is a summary of the mathematical and
numerical features of the four ML algorithms utilized in this study: The data set used in this study is composed of 100 self-
compacting concrete mixtures [10,28,29]. The values of these data
 Multiple Linear Regression (MLR) model: It is based on the lin- are shown in Appendix A, with 6 input data such as water absorp-
ear regression, to find the correlation between two variables tion, void ratio, sonic velocity at 1 day, sonic velocity at 7 days,
(one input x and one output y), under this form: y = ax + b compressive strength at 1 day, and compressive strength at 7 days.
where ‘‘b‘‘ is the intercept and ‘‘a” is the slope of the line. Sim- The output data is compressive strength at 28 days.
ilarly, if there are more than one input and output variable, we
name it Multiple Linear Regression [23]. 3.2. Data pre-processing
 Random Forest Regression (RFR) model: This prediction
methodology addresses prediction problems based on the aver- In this step, the Min-Max scaler normalization techniques have
age of a large number of distinct trees;RFR uses the Bootstrap been applied only for the input data. However, the normalization
resampling technic to build a decision tree model for each set of the output data is not necessary since we have only one output

Fig. 1. Illustration of the normalization input parameters.

860
Yousef El Asri, Mouhcine Ben Aicha, M. Zaher et al. Materials Today: Proceedings 57 (2022) 859–866

Fig. 2. Illustration of 5-fold cross validation.

data for prediction (compressive strength at 28 days). This normal- All these steps are summarized in the following diagram
ization is very useful to facilitate the performance of modeling (Fig. 3).
since we have measures that have different units, the goal is to For evaluating and developing ML models, the results of this
homogenize all the input data between 0 and 1. study are obtained using Matplotlib, Numpy, Pandas, and Scikit-
The input data normalized is presented in Fig. 1. The calculation learn, which are libraries of open-source Python.
of this normalization is done using the Min-Max scaler of prepro-
cessing from the Scikit-learn library of Python.
4. Results and discussion

3.3. Splitting data


Admixtures and other modifiers are commonly used to decrease
the bleeding and segregation of self-compacting concrete. Higher
The dataset is split into equal parts for the K- fold cross-
air quantities may be necessary in this case, as well as to improve
validation [30], this prediction K = 5, where 20 percent of datasets
the mixture’s resilience to freezing and thawing.
are used for the testing set and 80 percent of datasets are used for
In this study, 100 mixtures were designed based on the recom-
the training set, repeated five times as shown in Fig. 2. Finally, the
mendations of EFNARC and AFGC [31,32]. An air entraining mix-
average of all performance metrics is calculated after 5-cross
ture (AEA) was added in several percentages, namely in 0.05 to
validation.
1% of the weight of cement, with a step of 0.05%.
Experimental studies have been performed to examine the
3.4. Proposed model mechanical and intrinsic characteristics of SCC, such as compres-
sive strength (at 1, 7, and 28 days), sonic velocity (at 1 and 7 days),
To predict the compressive strength at 28 days, four different water absorption, and void ratio of the SCC using different air
ML algorithms are used: MLR, RFR, DTR and SVR. content.
After training of each model, for every fold, we calculated the In this part of this work, for the prediction of the compressive
metrics of the testing set (correlation coefficient (R2), mean abso- strength at 28 days of SCC, the algorithms proposed previously
lute error (MAE), mean square error (MSE). were applied on the experimental data (Appendix A). In this
However, in order to determine the most accurate model for regard, Table 1 summarizes the number of datasets, their mean
predicting the compressive strength at 28 days of SCC, a compar- values as well as the standard deviation, minimum and maximum
ison between the average metrics of each model was made. values used in this study.

Fig. 3. Proposed organization chart to predict compressive strength at 28 days.

861
Yousef El Asri, Mouhcine Ben Aicha, M. Zaher et al. Materials Today: Proceedings 57 (2022) 859–866

Table 1
Data statistics: count, mean, std, min and max.

Input Output
Water Void Sonic velocity at Sonic velocity at Compressive strength at Compressive strength at Compressive strength at
absorption (%) ratio (%) 1 day (m/s) 7 days (m/s) 1 day (MPa) 7 days (MPa) 28 days(MPa)
Count 100 100 100 100 100 100 100
Mean 7.09 13.50 3762.82 4089.11 24.23 36.40 42.31
Std 3.64 6.63 373.62 420.02 10.77 13.07 12.40
Min 1.74 3.60 3286 3562 12.20 20.30 26.20
Max 12.72 24.35 4444 4790 44.00 57.00 65.60

Table 2
Results of test metrics of the variousML algorithmsto predictthe compressive strength at 28 days.

Models K-fold R2 MAE MSE Models K-fold R2 MAE MSE


MLR 1 0.99583 0.62352 0.63196 DTR 1 0.99868 0.35400 0.20017
2 0.99817 0.40383 0.26941 2 0.99812 0.44300 0.27753
3 0.99664 0.58090 0.51480 3 0.99833 0.39750 0.25562
4 0.98931 0.67035 0.66842 4 0.98714 0.68750 0.80413
5 0.99770 0.51660 0.39155 5 0.99869 0.33600 0.22325
Mean 0.99553 0.55904 0.49523 Mean 0.99619 0.44360 0.35214
RFR 1 0.99852 0.32422 0.22449 SVR 1 0,99124 0,99351 1,32748
2 0.99903 0.24079 0.14251 2 0,99629 0,57437 0,54643
3 0.99810 0.35061 0.29177 3 0,99350 0,91354 0,99557
4 0.99528 0.45079 0.29495 4 0,97860 0,88746 1.33783
5 0.99942 0.21796 0.09850 5 0,99519 0,72336 0,82019
Mean 0.99807 0.31687 0.21045 Mean 0,99096 0,81845 1,00550

Fig. 4. Mean of MAE, MSE and R2 for each model.

4.1. Prediction results performing model is RFR (with a large value of mean R2 and the
lowest errors), the second is DTR followed by MLR and lastly we
According to the ML algorithms discussed previously (MLR, RFR, find the SVR model.
DTR and SVR), the compressive strength of SCC (at 28 days) can be Therefore, the model which will be chosen for the prediction of
predicted using the results shown in Table 2. compressive strength at 28 days of SCC is Random Forest Regres-
The test metrics reported in Table 2 were obtained after divid- sion algorithm RFR (Fig. 5). This Figure analyzes the statistical coef-
ing our data into k fold cross-validation, and modeling each fold ficients mentioned previously. Thus, the comparison between the
using the 4 models for training and testing. true experimental and the predicted values of the adopted algo-
Table 2 shows that the mean R2 value and that of each fold and rithm model is presented in Fig. 6. The two graphs show that the
model exceed 0.99, which means that the four models predict com- predicted results and the experimental results are almost identical
pressive strength with high performance. with a higher correlation coefficient R2 = 0.9997 and minimal error
To choose the best-performing model, compared to the other with MAE = 0.134 MPa and MSE = 0.041, which means that the
three models, the histogram of Fig. 4 clearly shows that the best model is very efficient.

862
Yousef El Asri, Mouhcine Ben Aicha, M. Zaher et al. Materials Today: Proceedings 57 (2022) 859–866

Fig. 5. MAE, MSE and R2 of experimental and predicted compressive strength.

Fig. 6. Visualization of experimental and predicted values of compressive strength.

5. Conclusion strength at 28 days) as a function of intrinsic parameters (water


absorption and void ratio) and mechanical proprieties at a young
In this study, four different machine learning algorithms were age (compressive strength and sonic velocity at 1 and 7 days).
used to determine their ability to predict the compressive strength Given the observed results, the RFR model is the most efficient
of self-compacting concrete at 28 days. These algorithms included machine learning model for predicting the compressive strength of
Multiple Linear Regression (MLR), Random Forest Regression (RFR), SCC. In addition, the other three models also give good results with
Decision Tree Regression (DTR), and Support Vector Regression great performances, classified respectively in this study DTR, MLR,
(SVR). Following that, using statistical coefficients such as regres- and the SVR model.
sion coefficient R2, Mean Absolute Error (MAE), and Mean Square Machine learning approaches can be used to properly anticipate
Error (MSE) between predicted and experimental output, the work the mechanical behavior of SCC. Unlike typical prediction methods,
is focused on determining the best model among those used for the the model discovered does not require a precise equation. Addi-
prediction of the mechanical response at long-term (compressive tionally, this model is capable of retraining with fresh data, which

863
Yousef El Asri, Mouhcine Ben Aicha, M. Zaher et al. Materials Today: Proceedings 57 (2022) 859–866

distinguishes it from other models in terms of its capacity to learn Software, Visualization, Investigation. Hafidi Alaoui Adil: Visual-
from new data. ization, Supervision, Validation.

CRediT authorship contribution statement Declaration of Competing Interest

El Asri Yousef: Conceptualization, Methodology, Software, The authors declare that they have no known competing finan-
Writing – original draft. Ben Aicha Mouhcine: Data curation, Writ- cial interests or personal relationships that could have appeared
ing – review & editing, Investigation, Validation. Zaher Mounir: to influence the work reported in this paper.

Appendix A. Experimental data, results input and output parameters

Samples Input Output


Water Void Sonic velocity Sonic velocity Compressive Compressive Compressive
absorption ratio at 1 day (m/s) at 7 days (m/s) strength at 1 day strength at 7 days strength at 28 days
(%) (%) (MPa) (MPa) (MPa)
1 1.74 3.6 4432 4776 42 55 63.4
2 1.76 3.7 4438 4781 43 56 64.2
3 1.78 3.9 4440 4784 44 57 65.6
4 1.77 3.7 4442 4787 44 57 65
5 1.75 3.6 4444 4790 43 56 64.6
6 2.11 4.3 4230 4532 40 53 60
7 2.13 4.4 4232 4535 40 53 61
8 2.15 4.6 4234 4538 41 54 62
9 2.12 4.35 4236 4533 40 53 60.5
10 2.14 4.5 4238 4536 40 53.5 61.5
11 3.24 6.12 4199 4585 37 52 55.15
12 3.25 6.15 4202 4588 37 52 56
13 3.27 6.18 4204 4591 38 53 56.2
14 3.29 6.21 4206 4594 38 53 56.3
15 3.31 6.24 4208 4597 37 54 56.4
16 3.56 7.06 4185 4563 35 51 54
17 3.58 7.1 4187 4565 35 51 54.2
18 3.6 7.13 4189 4567 36 51.5 54.3
19 3.62 7.16 4191 4569 36 51.8 54.4
20 3.64 7.19 4193 4571 35 52 54.5
21 3.86 7.88 4167 4558 34 50 53
22 3.87 7.9 4170 4560 34 50 53
23 3.88 7.91 4173 4563 34.5 50.5 53.5
24 3.89 7.92 4176 4566 35 51 54
25 3.9 7.93 4179 4569 35.2 51.4 55
26 3.9 8.12 4134 4548 33 49 52
27 3.92 8.17 4136 4550 33.2 49.4 52.4
28 3.94 8.2 4138 4553 33.4 49.5 52.6
29 3.96 8.23 4140 4556 33.6 49.6 52.9
30 3.98 8.26 4142 4559 33.8 49.8 53
31 4.07 8.54 4103 4539 32 48 51
32 4.1 8.57 4105 4542 32 48.2 51.2
33 4.12 8.6 4107 4544 32.5 48.5 51.5
34 4.14 8.63 4109 4546 33 48.8 51.8
35 4.16 8.66 4111 4548 33.5 49.1 52.1
36 4.34 9.1 3943 4366 30 44 50
37 4.36 9.12 3946 4369 30.3 44.5 50.3
38 4.38 9.15 3949 4372 30.5 45 50.5
39 4.4 9.18 3952 4375 30.7 45.5 50.7
40 4.42 9.2 3955 4378 30.9 46 50.9
41 4.52 9.61 3855 4156 28 40 48
42 4.55 9.63 3857 4158 28.4 41 48.4
43 4.57 9.66 3859 4160 28.8 42 49
44 4.59 9.69 3861 4162 29.2 43 49.6
45 4.61 9.72 3863 4164 29.6 44 50.2

864
Yousef El Asri, Mouhcine Ben Aicha, M. Zaher et al. Materials Today: Proceedings 57 (2022) 859–866

Experimental data, results input and output parameters (continued)

Samples Input Output


Water Void Sonic velocity Sonic velocity Compressive Compressive Compressive
absorption ratio at 1 day (m/s) at 7 days (m/s) strength at 1 day strength at 7 days strength at 28 days
(%) (%) (MPa) (MPa) (MPa)
46 6.33 11.2 3688 4110 25 35 42
47 6.35 11.24 3693 4113 25.5 35.5 43
48 6.38 11.28 3695 4116 26 36 44
49 6.41 11.32 3697 4119 26.5 36.5 45
50 6.44 11.36 3699 4122 27 37 46
51 7.02 12.43 3583 3832 20.3 31 38
52 7.12 12.46 3585 3835 20.33 31.5 38.6
53 7.14 12.48 3588 3838 20.36 32 39
54 7.16 12.5 3591 3841 20.39 32.5 39.4
55 7.18 12.52 3594 3844 20.42 33 39.8
56 7.88 14.6 3544 3802 16.6 27 34
57 7.9 14.64 3546 3808 16.8 27.5 35
58 7.93 14.68 3549 3813 16.9 28 35.5
59 7.96 14.72 3552 3818 17 28.5 36
60 7.99 14.76 3555 3823 17.1 29 36.5
61 9.12 17.1 3524 3786 14.1 24.6 31
62 9.16 17.15 3528 3788 14.14 24.64 31.5
63 9.2 17.19 3530 3790 14.18 24.68 32
64 9.24 17.23 3532 3792 14.22 24.72 32.5
65 9.28 17.27 3534 3794 14.26 24.76 33
66 9.88 18.6 3488 3742 13.6 24.2 30.5
67 9.92 18.64 3490 3744 13.65 24.4 30.6
68 9.95 18.68 3493 3746 13.7 24.6 30.7
69 9.98 18.72 3496 3748 13.75 24.8 30.8
70 10.01 18.76 3499 3750 13.8 25 30.9
71 10.08 19.2 3422 3704 13.2 23.4 30.1
72 10.12 19.23 3425 3708 13.5 23.5 30.2
73 10.15 19.25 3428 3711 13.8 23.8 30.3
74 10.18 19.27 3431 3714 14.1 24.1 30.4
75 10.21 19.29 3434 3717 14.4 24.4 30.5
76 11 19.8 3388 3688 13 23.1 29.4
77 11.06 19.83 3390 3690 13.1 23.1 29.5
78 11.1 19.86 3392 3693 13.1 23.2 29.8
79 11.14 19.89 3394 3696 13.12 23.3 30.1
80 11.18 19.92 3396 3699 13.14 23.4 30.4
81 11.5 21.4 3364 3624 12.88 22.4 29
82 11.55 21.44 3366 3626 12.9 22.5 29
83 11.6 21.47 3368 3628 12.92 22.7 29.2
84 11.65 21.5 3370 3630 12.94 22.9 29.4
85 11.7 21.53 3372 3632 12.96 23.1 29.6
86 11.8 22.2 3323 3605 12.65 21.6 28.4
87 11.83 22.25 3325 3607 12.68 21.7 28.5
88 11.85 22.28 3327 3609 12.7 21.75 28.55
89 11.87 22.31 3329 3611 12.72 21.8 28.6
90 11.89 22.34 3331 3613 12.74 21.85 28.65
91 12.2 23.4 3302 3592 12.53 21.4 27.8
92 12.24 23.6 3304 3595 12.55 21.43 27.83
93 12.26 23.7 3306 3597 12.58 21.45 27.85
94 12.3 23.9 3310 3601 12.64 21.49 27.89
95 12.47 24.05 3294 3588 12.4 20.7 26.7
96 12.5 24.08 3296 3590 12.42 20.8 26.8
97 12.53 24.15 3298 3593 12.45 20.85 26.9
98 12.66 24.28 3286 3562 12.2 20.3 26.2
99 12.68 24.32 3288 3564 12.24 20.4 26.25
100 12.72 24.35 3290 3566 12.26 20.5 26.3

865
Yousef El Asri, Mouhcine Ben Aicha, M. Zaher et al. Materials Today: Proceedings 57 (2022) 859–866

References [17] T. Erdik, Discussion on ‘‘Prediction of compressive strength of concrete


containing fly ash using artificial neural networks and fuzzy logic” [Comput.
Mater. Sci. 41 (2008) 305], Comput. Mater. Sci. 44 (3) (2009) 1022, https://doi.
[1] J. Murata, Flow and deformation of fresh concrete, Matériaux Constr. 17 (2)
org/10.1016/j.commatsci.2008.07.031.
(1984) 117–129, https://doi.org/10.1007/BF02473663.
[18] H. Mashhadban, S.S. Kutanaei, M.A. Sayarinejad, Prediction and modeling of
[2] N. Roussel, A. Lemaître, R.J. Flatt, P. Coussot, Steady state flow of cement
mechanical properties in fiber reinforced self-compacting concrete using
suspensions: A micromechanical state of the art, Cem. Concr. Res. 40 (1) (2010)
particle swarm optimization algorithm and artificial neural network, Constr.
77–84, https://doi.org/10.1016/j.cemconres.2009.08.026.
Build. Mater. 119 (2016) 277–287, https://doi.org/10.1016/
[3] C. Hu, F. de Larrard, The rheology of fresh high-performance concrete, Cem.
j.conbuildmat.2016.05.034.
Concr. Res. 26 (2) (1996) 283–294, https://doi.org/10.1016/0008-8846(95)
[19] O. Belalia Douma, B. Boukhatem, M. Ghrici, A. Tagnit-Hamou, Prediction of
00213-8.
properties of self-compacting concrete containing fly ash using artificial neural
[4] I. Flatt, R.J., Schober, Superplasticizers and the rheology of concrete, in: N.
network, Neural Comput. Appl. 28 (S1) (2017) 707–718, https://doi.org/
Roussel (Ed.), Underst. Rheol. Concr., Woodhead Publishing Series in Civil and
10.1007/s00521-016-2368-7.
Structural Engineering, 2012, pp. 144–208.
[20] M. Benaicha, O. Jalbaud, A. Hafidi Alaoui, Y. Burtschell, Porosity effects on
[5] K.E. Alyamaç, R. Ince, A preliminary concrete mix design for SCC with marble
rheological and mechanical behavior of self-compacting concrete, J. Build. Eng.
powders, Constr. Build. Mater. 23 (3) (2009) 1201–1210, https://doi.org/
48 (2022) 103964, https://doi.org/10.1016/j.jobe.2021.103964.
10.1016/j.conbuildmat.2008.08.012.
[21] T.M. Mitchell, Machine Learning, McGraw-Hill, New York, 1997.
[6] P. Dinakar, K.P. Sethy, U.C. Sahoo, Design of self-compacting concrete with
[22] Y. Li, Deep Reinforcement Learning, (2018).
ground granulated blast furnace slag, Mater. Des. 43 (2013) 161–169, https://
[23] F. Khademi, M. Akbari, S.M. Jamal, M. Nikoo, Multiple linear regression,
doi.org/10.1016/j.matdes.2012.06.049.
artificial neural network, and fuzzy logic prediction of 28 days compressive
[7] L.D. Schwartzentruber, R. Le Roy, J. Cordin, Rheological behaviour of fresh
strength of concrete, Front. Struct. Civ. Eng. 11 (1) (2017) 90–99, https://doi.
cement pastes formulated from a Self Compacting Concrete (SCC), Cem. Concr.
org/10.1007/s11709-016-0363-9.
Res. 36 (7) (2006) 1203–1213, https://doi.org/10.1016/j.
[24] H.-V. Mai, T.-A. Nguyen, H.-B. Ly, V.Q. Tran, J. Zhang, Prediction compressive
cemconres.2004.10.036.
strength of concrete containing GGBFS using random forest model, Adv. Civ.
[8] M. Bartos, Sonebi, Tamimi, Workability and Rheology of Fresh Concrete:
Eng. 2021 (2021) 1–12, https://doi.org/10.1155/2021/6671448.
Compendium of Tests, 2002.
[25] L. Wang, T. Zeng, Y. Tao, X. Wu, H. Chen, R. Weerasinghe, C. Fang, Research on
[9] M. Benaicha, A. Belcaid, A.H. Alaoui, O. Jalbaud, Y. Burtschell, Rheological
prediction of concrete frost resistance based on random forest, E3S Web Conf.
characterization of self-compacting concrete: New recommendation, Struct.
237 (2021) 03033, https://doi.org/10.1051/e3sconf/202123703033.
Concr. 20 (5) (2019) 1695–1701, https://doi.org/10.1002/suco.201900154.
[26] A. Ahmad, F. Farooq, P. Niewiadomski, K. Ostrowski, A. Akbar, F. Aslam, R.
[10] M. Benaicha, X. Roguiez, O. Jalbaud, Y. Burtschell, A.H. Alaoui, Influence of
Alyousef, Prediction of compressive strength of fly ash based concrete using
silica fume and viscosity modifying agent on the mechanical and rheological
individual and ensemble algorithm, Materials (Basel). 14 (2021) 794, https://
behavior of self compacting concrete, Constr. Build. Mater. 84 (2015) 103–110,
doi.org/10.3390/ma14040794.
https://doi.org/10.1016/j.conbuildmat.2015.03.061.
[27] A.M. Abd, S.M. Abd, Modelling the strength of lightweight foamed concrete
[11] M. Benaicha, X. Roguiez, O. Jalbaud, Y. Burtschell, A.H. Alaoui, New approach to
using support vector machine (SVM), Case Stud. Constr. Mater. 6 (2017) 8–15,
determine the plastic viscosity of self-compacting concrete, Front. Struct. Civ.
https://doi.org/10.1016/j.cscm.2016.11.002.
Eng. 10 (2) (2016) 198–208, https://doi.org/10.1007/s11709-015-0327-5.
[28] M. Benaicha, A. Belcaid, A.H. Alaoui, O. Jalbaud, Y. Burtschell, Effects of
[12] M. Benaicha, O. Jalbaud, A. Hafidi Alaoui, Y. Burtschell, Marsh cone coupled to a
limestone filler and silica fume on rheology and strength of self-compacting
plexiglas horizontal channel: Rheological characterization of cement grout,
concrete, Struct. Concr. 20 (5) (2019) 1702–1709, https://doi.org/
Flow Meas. Instrum. 45 (2015) 126–134, https://doi.org/10.1016/j.
10.1002/suco.201900150.
flowmeasinst.2015.06.004.
[29] M. Benaicha, Y. Burtschell, A.H. Alaoui, K. Elharrouni, Theoretical calculation of
[13] M. Ben Aicha, The superplasticizer effect on the rheological and mechanical
self-compacting concrete plastic viscosity, Struct. Concr. 18 (5) (2017) 710–
properties of self-compacting concrete, in: New Mater. Civ. Eng., Elsevier,
719, https://doi.org/10.1002/suco.201600064.
2020, pp. 315–331, https://doi.org/10.1016/B978-0-12-818961-0.00008-9.
[30] T. Fushiki, Estimation of prediction error by using K-fold cross-validation, Stat.
[14] W.P.S. Dias, S.P. Pooliyadda, Neural networks for predicting properties of
Comput. 21 (2) (2011) 137–146, https://doi.org/10.1007/s11222-009-9153-8.
concretes with admixtures, Constr. Build. Mater. 15 (7) (2001) 371–379,
[31] François CUSSIGH, Recommandations pour l’emploi des bétons autoplaçants,
https://doi.org/10.1016/S0950-0618(01)00006-X.
2008.
[15] S.-C. Lee, Prediction of concrete strength using artificial neural networks, Eng.
[32] EFNARC, Specification and Guidelines for Self-Compacting Concrete, 2002.
Struct. 25 (7) (2003) 849–857, https://doi.org/10.1016/S0141-0296(03)00004-X.
[16] G. Trtnik, F. Kavčič, G. Turk, Prediction of concrete strength using ultrasonic
pulse velocity and artificial neural networks, Ultrasonics 49 (1) (2009) 53–60,
https://doi.org/10.1016/j.ultras.2008.05.001.

866

You might also like