Elsevier

Measurement

Volume 152, February 2020, 107309
Measurement

An evolutionary approach to formulate the compressive strength of roller compacted concrete pavement

https://doi.org/10.1016/j.measurement.2019.107309Get rights and content

Highlights

  • An evolutionary approach is introduced to formulate the compressive strength of RCCP.

  • Three formulas are presented based on different inputs combination for the compressive strength.

  • The volumetric/weighted (dimensional) model has better performance than the other models.

  • The proposed models are evaluated by parametric study, sensitivity analysis, and external validation.

Abstract

The construction and maintenance of roads pavement was a critical problem in the last years. Therefore, the use of roller-compacted concrete pavement (RCCP) in road problems is widespread. The compressive strength (fc) is the key characteristic of the RCCP caused to significant impact on the cost of production. In this study, an evolutionary-based algorithm named gene expression programming (GEP) is implemented to propose novel predictive formulas for the fc of RCCP. The fc is formulated based on important factor used in mixture proportion in three different combinations of dimensional form (coarse aggregate, fine aggregate, cement, pulverized fly ash, water, and binder), non-dimensional form (water to cement ratio, water to binder ratio, coarse to fine aggregate ratio and pulverized fly ash to binder ratio) and percentage form of input variables. A comprehensive and reliable database incorporating 235 experimental cases collected from several studies. Furthermore, mean absolute error (MAE), root mean square error (RMSE), correlation coefficient (r), average absolute error (AAE), performance index (PI), and objective function (OBJ) as the internal standard statistical measures and external validation evaluated proposed GEP-based models. Uncertainty and parametric studies were carried out to verify the results. Moreover, sensitivity analysis to determine the importance of each predictor on fc of RCCP revealed that fine aggregate content and water to binder ratio is the most useful predictor in dimensional, non-dimensional and percentage forms, respectively. The proposed equation-based models are found to be simple, robustness and straightforward to utilize, and provide consequently new formulations for fc of RCCP.

Introduction

The high cost of producing roads asphalt pavement and the large volume of oil products polluting the environment have necessitated the use of alternative technologies. Roller-compacted concrete pavement (RCCP) is a rigid pavement that has been of interest to researchers due to the accelerated construction and reduced maintenance costs [1], [2]. The lower cement pastes and higher aggregates make RCCP has low consistency, and to be compacted, the roller and compactor are used. Compared with the tropical and temperate asphalt, the long-term durability of RCCP has received interest due to the resistance to rising temperatures, low water absorption, and proper compressive strength, and consequently, it experiences less deformation under loading [3]. Also, in cold areas, the concrete pavement is resistant to frost cycles in the face of possible damage. In addition, due to the impermeability of the constituting materials, as consistent pavement, there is no environmental problem in its application range [4]. The gray and neutral color of the RCCP also has a good temperature absorption coefficient and helps lower the ambient temperature. Due to the lower cost of production than cement and improved strength properties, the use of pozzolanic materials has also been considered in the production of the RCCP [4]. Therefore, in this study, the RCCP incorporating binder-based pulverized fly ash (PFA) powder is investigated. The PFA is well combined with the gel produced in concrete and increases the hydration of concrete [5] and, as a result, increases the density of the produced concrete and improves the mechanical and chemical properties of RCCP.

The mechanical properties of RCCP required important factors [6]. The compressive strength (fc) as a prominent mechanical property of RCCP is necessary to investigate experimental and computational intelligence analysis in construction materials. Developing models with an accurate estimation for this key characteristic caused to an eco-friendly benefit, saving cost and time of production and designing an optimal blend. In the last years, an evolutionary approach presented an enhanced potency for solution of civil engineering challenges [7]. Some of the data-driven methods named multiple linear regression (MLR) [8], [9], artificial neural network (ANN) [10], [11], [12], [13], support vector machine (SVM) [14], [15], adaptive neuro-fuzzy inference systems (ANFIS) [16], [17] that have been implemented to predict the properties of concretes. Chen [8] employed the MLR method to predict fc based on physical properties of concrete. He investigates five series of input variables and reports results of all scenarios in terms of error metrics. The capability of the ANN methodology for prediction of properties of SCC investigated by Asteris and Kolovos [11] and Getahun et al. [12]. They reported the ANN technique as an alternative to the experimental program to simulate characteristics of concretes. Wang et al. [13] Compared MLR, ANN and ANFIS techniques to estimate the expansion behavior of self-stressing concrete. The detail of the result illustrates the ANFIS model outperforms ANN and MLR in terms of acceptability and accuracy. Sonebi et al. [14] derived SVM method for modeling of self-compacting concrete using kernel functions named radial basis and polynomial. The modeling process demonstrates that the SVM based radial basis method robustness and accurate in their study. Regarding the entire aforementioned, machine learning methods generally considered as black-box methodology since they are not capable to formulate linear relation between predictor and target variables. Genetic Programming (GP) [18], [19], [20] is another heuristic regression method that automatically generates computational models on the basis of evolutionary genetic [21]. González-Taboada et al. [19] focused on a comprehensive database to forecast a high accuracy model of mechanical properties of structural recycled concrete using GP technique. GEP is the extended form of GP, which evolves intelligence process with different shapes and sizes that are encoded chromosomes in the form of simple linear with a fixed length [22]. Several other studies have proved the estimation capability and potential of the evolutionary data-driven models in the field of concrete technology and material design [23], [24], [25].

As stated above, the classical data-driven models provide reliable tools to predict the compressive strength of concrete. However, it is not a robust method to formulate this fundamental property. The main focus of this attempt is presented initially as a new interpretation and conceptualization GEP-based formulas in three forms of inputs combination to calculate the fc of RCCP for the first time. Performance criteria of the proposed formula-based GEP models are compared with the previous investigation. Moreover, the uncertainty analyses inspired by Monte-carlo simulation (MCS) method of all of the developed GEP models are performed. To verify the developed models, external validations based on some performances indicators and parametric studies are evaluated.

Section snippets

Overview of gene expression programming

Based on evolution biological theory of Darwin, evolutionary algorithms are a type of computational techniques used for problem-solving through natural selection and heuristic search amongst a population of solutions to find the “best-fit”. Each candidate solution which belongs to the population is referred to as an individual. Poor solutions are eliminated through the iteration of an evolutionary algorithm, which includes a competitive selection by evaluating the quality of solutions based on

Modeling of the compressive strength of RCCP

In this study, three different models based on volumetric/weighted content (dimensional) consisting six predictors, ratio of different constituents (non-dimensional) of the variables consisting five predictors and percentage form of variables consisting six predictors are implemented by the GEP-based method to clarify between the fc and each important predictors in the functional relationship as bellow:fcGEP-1=(CA,FA,C,PFA,W,B,AS)fcGEP-2=WB,WC,PFAB,CAFA,ASfcGEP-3=(%CA,%FA,%C,%PFA,%W,%B,AS)

Result and discussion

In the present study, the performances of the GEP-based models for training and testing phases were compared. To evaluate the accuracy of the models, statistical metrics in terms of r, MAE, RMSE, AAE, PI, and OBJ have been performed. Moreover, the validity of the derived GEP-based models performed using external validation and uncertainty analysis. The error measures of the developed models for the training and testing performances are indicated in Table 4. Considering Table 4, the GEP approach

Limitations and future work

Although, this study was the first attempt to propose artificial intelligence (AI)-based formula for estimation of the compressive strength of RCCP, several limitations should be addressed for future work. As can be seen from computational process, GEP-based models could predict the fc of RCCP in different inputs combination forms with high accuracy. One of the disadvantages of evolutionary-based model (e.g., GEP) is that its implementation is time-consuming because it optimized user manual

Conclusion

In this study, the evolutionary approach based on GEP was applied as a soft computing technique proposed new predictive formulations of the compressive strength of roller compacted concrete pavement in three forms of inputs combination. The models were implemented using 235 datasets of RCCP mixture proportions. The wide range database containing components of RCCP caused to ensure that the GEP models will be suitable and generalized. Implemented models based on weighted/volumetric content,

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References (54)

  • Ali Ashrafian et al.

    Compressive strength of Foamed Cellular Lightweight Concrete simulation: new development of hybrid artificial intelligence model

    Constr. Build. Mater.

    (2020)
  • H.M. Azamathulla et al.

    Linear genetic programming to scour below submerged pipeline

    Ocean Eng.

    (2011)
  • A. Mardani-Aghabaglou et al.

    Mechanical properties of high-volume fly ash roller compacted concrete designed by maximum density method

    Constr. Build. Mater.

    (2013)
  • C.D. Atiş et al.

    Strength properties of roller compacted concrete containing a non-standard high calcium fly ash

    Mater. Lett.

    (2004)
  • S.K. Rao et al.

    C, Experimental studies in Ultrasonic Pulse Velocity of roller compacted concrete pavement containing fly ash and M-sand

    Int. J. Pavement Res. Technol.

    (2016)
  • C. Cao et al.

    The analysis on strength and fly ash effect of roller-compacted concrete with high volume fly ash

    Cem. Concr. Res.

    (2000)
  • A.H. Gandomi et al.

    A, Assessment of artificial neural network and genetic programming as predictive tools

    Adv. Eng. Softw.

    (2015)
  • A. Tahmassebi et al.

    Building energy consumption forecast using multi-objective genetic programming

    Measurement

    (2018)
  • A.H. Gandomi et al.

    Formulation of shear strength of slender RC beams using gene expression programming, part II: With shear reinforcement

    Measurement

    (2017)
  • A.H. Gandomi et al.

    Genetic programming for experimental big data mining: a case study on concrete creep formulation

    Autom. Constr.

    (2016)
  • S.K. Babanajad et al.

    New prediction models for concrete ultimate strength under true-triaxial stress states: an evolutionary approach

    Adv. Eng. Softw.

    (2017)
  • ACI 325.10. State of art Report on Roller Compacted Concrete Pavements. American Concrete Institute report ACI...
  • N. Banthia et al.

    Permeability of roller compacted concrete

    J. Mater. Civ. Eng.

    (1992)
  • C. Atiş

    D, Strength properties of high-volume fly ash roller compacted and workable concrete, and influence of curing condition

    Cem. Concr. Res.

    (2005)
  • H. AzariJafari et al.

    Ternary blended cement: an eco-friendly alternative to improve resistivity of high-performance self-consolidating concrete against elevated temperature

    J. Clean. Prod.

    (2019)
  • ACI 211 3R-02, Guide for Selecting Proportions for No Slump Concrete,...
  • Mohammad Rezaie Balf et al.

    Evolutionary polynomial regression approach to predict longitudinal dispersion coefficient in rivers

    J. Water Supply: Res. Technol.-Aqua

    (2018)
  • Cited by (53)

    • Enhancing Compressive strength prediction of Roller Compacted concrete using Machine learning techniques

      2023, Measurement: Journal of the International Measurement Confederation
    View all citing articles on Scopus
    View full text