Elsevier

Expert Systems with Applications

Volume 40, Issue 17, 1 December 2013, Pages 6856-6862
Expert Systems with Applications

Prediction of high performance concrete strength using Genetic Programming with geometric semantic genetic operators

https://doi.org/10.1016/j.eswa.2013.06.037Get rights and content

Highlights

Abstract

Concrete is a composite construction material made primarily with aggregate, cement, and water. In addition to the basic ingredients used in conventional concrete, high-performance concrete incorporates supplementary cementitious materials, such as fly ash and blast furnace slag, and chemical admixture, such as superplasticizer. Hence, high-performance concrete is a highly complex material and modeling its behavior represents a difficult task. In this paper, we propose an intelligent system based on Genetic Programming for the prediction of high-performance concrete strength. The system we propose is called Geometric Semantic Genetic Programming, and it is based on recently defined geometric semantic genetic operators for Genetic Programming. Experimental results show the suitability of the proposed system for the prediction of concrete strength. In particular, the new method provides significantly better results than the ones produced by standard Genetic Programming and other machine learning methods, both on training and on out-of-sample data.

Introduction

Concrete is a composite construction material made primarily with aggregate, cement, and water. There are many formulations of concrete, which provide varied properties, and concrete is the most-used man-made product in the world (Lomborg, 2001). Modern concrete mix designs can be complex. The choice of a concrete mix depends on the need of the project both in terms of strength and appearance and in relation to local legislation and building codes. The design begins by determining the requirements of the concrete. These requirements take into consideration the weather conditions that the concrete will be exposed to in service, and the required design strength. Many factors need to be taken into account, from the cost of the various additives and aggregates, to the trade offs between, the “slump” for easy mixing and placement and ultimate performance. A mix is then designed using cement, coarse and fine aggregates, water and chemical admixtures. The method of mixing will also be specified, as well as conditions that it may be used in. This allows a user of the concrete to be confident that the structure will perform properly. As reported in Yeh (1998), high-performance concrete (HPC) is a new terminology used in the concrete construction industry. In addition to the basic ingredients in conventional concrete the making of HPC needs to incorporate supplementary cementitious materials, such as fly ash and blast furnace slag, and chemical admixture, such as superplasticizer (Kumar, Singh, & Singh, 2012). High-performance concrete is such a highly complex material that modeling its behavior is a difficult task.

The Abrams’ water-to-cement ratio (w/c) law (Abrams, 1927, Nagaraj and Banu, 1996) has been described as the most useful and significant advancement in the history of concrete technology. According to Abrams’s law, the compressive strength of concrete varies inversely with the W/C ratio. Hence, an increase in the w/c decreases the concrete strength, whereas a decrease in the w/c ratio increases the strength. The implication, therefore, is that the strengths of various but comparable concrete are identical as long as their w/c ratios remain the same, regardless of the details of the compositions.

The Abrams rule implies that only the quality of the cement paste controls the strength of comparable cement. The paste quantity does not matter. Analysis of a variety of experimental data shows that this is not quite true (Popovics, 1990). For instance, if two comparable concrete mixtures have the same w/c ratio, the strength of the concrete with the higher cement content is lower (Popovics, 1990).

As reported in Yeh (1998), several studies independently have shown that concrete strength development is determined not only by the w/c ratio, but that it is also influenced by the content of other ingredients (Bhanja & Sengupta, 2005). Therefore, although experimental data have shown the practical acceptability of this rule within wide limits, a few deviations have been reported. The current empirical equations for estimating compressive strength are based on tests of concrete without supplementary cementitious materials. The validity of these relationships for concrete with supplementary cementitious materials (fly ash, blast furnace slag, etc.) should be investigated (Bhanja & Sengupta, 2005). The more we know about the concrete composition versus strength relationship, the better we can understand the nature of concrete and how to optimize the concrete mixture.

All these aspects highlight the need of reliable and accurate techniques that allow modeling the behavior of concrete materials.

In this paper, for the first time, we propose an intelligent system based on Genetic Programming for the prediction of the concrete strength. More in detail, in this work we use recently defined geometric semantic genetic operators for Genetic Programming. These operators allow to include the concept of semantics in the search process and have several advantages with respect to standard genetic operators used in Genetic Programming.

The paper is organized as follows: Section 2 introduces basic concepts about Genetic Programming with a particular focus on the standard operators used in the evolutionary process; Section 3 presents the geometric semantic operators used in this work and outlines some of their properties that have a direct effect on the search process; Section 4 presents a literature review on using computational intelligence methods in simulating the behavior of concrete materials; Section 5 describes the data used in this work, the experimental settings and proposes an accurate discussion and analysis of the results; Section 6 concludes the paper summarizing the results that have been achieved.

Section snippets

Genetic Programming

Models lie in the core of any technology in any industry, be it finance, manufacturing, services, mining, or information technology. The task of data-driven modeling lies in using a limited number of observations of system variables for inferring relationships among these variables. The design of reliable learning machines for data-driven modeling tasks is of strategic importance, as there are many systems that cannot be accurately modeled by classical mathematical or statistical techniques.

Geometric semantic operators

In the last few years, GP has been extensively used to address problems in different domains (Aslam et al., 2013, Gusel and Brezocnik, 2011, Guo et al., 2011, Kayadelen, 2011, Tsakonas et al., 2006) and it has produced a wide set of results that have been defined human-competitive (Koza, 2010). While these results have demonstrated the suitability of GP in tackling real-life problems, research has recently focused on developing new variants of GP in order to further improve its performances. In

Related works

Prediction of concrete strength is an important problem in the engineering field and several works have been proposed to address this problem. In this section a brief literature review about the use of computational intelligence techniques for facing this problem is proposed. In Saridemir, Topcu, Ozcan, and Severcan (2009), artificial neural networks and fuzzy logic models for prediction of long-term effects of ground granulated blast furnace slag on compressive strength of concrete under wet

Data set information

Following the same procedure described in Yeh (1998), experimental data from 17 different sources was used to check the reliability of the strength model. Data were assembled for concrete containing cement plus fly ash, blast furnace slag, and superplasticizer. A determination was made to ensure that these mixtures were a fairly representative group governing all of the major parameters that influence the strength of HPC and present the complete information required for such an evaluation. The

Conclusions

High-performance concrete is a highly complex material that makes modeling its behavior a difficult task. In this study an intelligent GP-based system to predict the compressive strength of HPC has been proposed. The system uses recently defined geometric semantic genetic operators that present several advantages with respect to standard GP operators. In particular, they have the extremely interesting property of inducing a unimodal fitness landscape for any problem consisting in matching input

Acknowledgments

This work was supported by national funds through FCT under contract PEst-OE/EEI/LA0021/2013 and by projects EnviGP (PTDC/EIA-CCO/103363/2008) and MassGP (PTDC/EEI-CTP/2975/2012), Portugal.

References (44)

  • S. Lai et al.

    Concrete strength prediction by means of neural network

    Construction and Building Materials

    (1997)
  • S.-C. Lee

    Prediction of concrete strength using artificial neural networks

    Engineering Structures

    (2003)
  • W. Dias et al.

    Neural networks for predicting properties of concretes with admixtures

    Construction and Building Materials

    (2001)
  • S. Akkurt et al.

    The use of gaanns in the modelling of compressive strength of cement mortar

    Cement and Concrete Research

    (2003)
  • M. Sebastiá et al.

    Neural network prediction of unconfined compressive strength of coal fly ash cement mixtures

    Cement and Concrete Research

    (2003)
  • L. Ren et al.

    An optimal neural network and concrete strength modeling

    Advances in Engineering Software

    (2002)
  • F. Demir

    A new way of prediction elastic modulus of normal and high strength concrete fuzzy logic

    Cement and Concrete Research

    (2005)
  • F.-L. Gao

    A new way of predicting cement strength fuzzy logic

    Cement and Concrete Research

    (1997)
  • B. Lomborg

    The skeptical environmentalist: Measuring the real state of the world

    (2001)
  • D.A. Abrams

    Water-cement ration as a basis of concrete quality

    ACI Materials Journal

    (1927)
  • S. Popovics

    Analysis of concrete strength versus water-cement ratio relationship

    ACI Materials Journal

    (1990)
  • Poli, R., Langdon, W. B., & Mcphee, N. F. (2008). A field guide to genetic programming....
  • Cited by (124)

    View all citing articles on Scopus
    View full text