Analytical and genetic programming model of compressive strength of eco concretes by NDT according to curing temperature
Introduction
Obtaining a good estimation of the compressive strength using non-destructive testing, has both long-term and short-term benefits. In the short term (during construction), it can allow an adjustment of the work deadlines, and at the same time, improve structural safety. In the long term (service, upgrading, modification and/or extension) it is of great importance, since it can predict the strength of concrete without having to take cores and consequently deteriorate concrete structures.
In the case of eco-concrete with recycled fine aggregate, recycled coarse aggregate (for percentages higher than 20%) or biomass ashes, getting greater insight and control over the compressive strength is of particular interest, since its use for the manufacture of structural concrete is either prohibited or conditioned to the development of complementary studies in multiple regulations [1], [2], [3], [4].
The influence of curing temperature on the correlations between ultrasonic pulse velocity and compressive strength was observed in a previous study [5], concluding that the curing temperature is a critical factor when estimating the compressive strength through its correlation with ultrasonic pulse velocity (UPV), noting that for a same velocity, the lower the curing temperature is, the lower the corresponding strength. There are some previous studies [6], [7], [8] in which some influence of temperature on this correlation was observed. As can be seen in Table 1, in the usual estimation models combining non-destructive testing, curing temperature or historic internal temperatures are not usually considered as possible factors influencing the correlation between non-destructive tests and compressive strength.
Models which are more frequently used in this correlation are linear and exponential [9], [10], [11], although some authors consider other possibilities such as the potential or parabolic [9], [12], [13].
Huang et al. [14] develop a study comparing the results of 11 models, some of them collected in Table 1. They get the best fit with the most complex model they used, in which the input data are ultrasonic pulse velocity (UPV), age (t), the correlation water/cement (WC) and the rebound number (R). With this model a fit is obtained with an R2 of 0.89.
Martínez-Molina [15] performed a comparative study of 10 models. Eight of these models use UPV, R, WC and t. Other models proposed also use the resonant frequency and the electrical resistivity, but the results do not improve significantly with respect to the formulation of Huang (who does not use resonant frequency or electrical resistivity).
Some investigations study the possibility of finding a better prediction of the compressive strength by using ultrasonic pulse velocity or by combining several non-destructive tests, using artificial neural networks [16], [17], and they suggest that promising results and improving the accuracy of estimations can be obtained, however, an estimation equation including some important factors that can be applied easily in practice is not proposed. In a recent study, Ayaz et al. [18] using the M5 rule and M5P Tree proposes an equation for estimating the compressive strength using the ultrasonic pulse velocity modified, depending on the amount of cement, the amount of fly ash, amount of slag and the curing period. 10 different mixes are studied in which the amount of cement, fly ashes and slag are different. Data for each mix at 4 ages (3, 7, 28 and 120 days) is obtained. Therefore, the results of a total of 40 tests are used. Good estimation results are obtained, although it is necessary to note that in this investigation decisive factors on correlations such as the curing temperature [5], [8] or the density of the resulting concrete among others are not considered. Other techniques used for estimating the mechanical properties of concrete and eco-concrete are, for example, the genetic programming, model tree and artificial neural network. As a result of using these techniques, significant improvements in the performance of predictive models can be obtained [19], [20], [21], [22], [23], [24], [25], [26]. Therefore, one of the goals of this research is the analytical development and proposal of new models to estimate compressive strength, by combining non-destructive testing, considering the curing temperature, the history of internal temperatures and maturity of concrete. Other factors that can influence the correlations between non-destructive testing will be considered, especially in the case of eco-concrete, such as the density of the concrete (very important in the case of concrete with recycled aggregate) or the percentage of cement replacing biomass ashes. In addition, it aims to develop an equation for estimating the compressive strength on the safe side for conventional concrete and eco-concrete, for this purpose, analytical results are considered and the oriented genetic programming technique is used in which overestimation will be very sharply penalized.
Section snippets
Materials
A total of 11 concretes are studied: 3 self-compacting concrete (SCC) and 8 vibrated concrete (VC). For each group, one reference concrete was set without recycled materials. The main intention was that these concretes were as different as possible in order to have diverse concretes. So, the following parameters were varied: type of concrete, quantity of cement, correlation water/cement, aggregate, maximum aggregate size, percentage of substitution of recycled aggregate and quantity of ashes.
Results and analysis
To avoid repetition of the legend on each of the Fig. 5, Fig. 6, Fig. 7, Fig. 8, Fig. 9, Fig. 4 legend is applicable to all figures comparing the estimation of the compressive strength with its real value. In addition to all the graphs of Fig. 5, Fig. 6, Fig. 7, Fig. 8, Fig. 9, the error of ±10% and ±20% is indicated.
Conclusions
It concludes that the proposed models substantially improve estimates compared to the reference ones. To estimate the compressive strength by combining non-destructive testing, it is important to note the curing temperature, age, the modified cement water ratio (Eq. (3)) and density.
Estimations with temperature-time factor between the reference temperature (FTT/Tr) have a similar accuracy to those with the equivalent age at which the energy activation is needed. Therefore the use of MN6 and MN10
Acknowledgment
M.V-L., I.M-L and P. V-B. were partially supported by he projects of Program FEDER-INNTERCONECTA ITC-20133075 “Use of paper and tire industry wastes other than their use in building lightweight technological embankments and other construction materials” and ITC-20113055 “Development of value adding technologies for RCDs for innovative applications”, convened by the Center for Industrial Technological Development (CDTI, for its initials in Spanish), dependent on the Ministry of Economy and
References (40)
- et al.
Influence of temperature in the evolution of compressive strength and in its correlations with UPV in eco-concretes with recycled materials
Constr. Build. Mater.
(2016) - et al.
Compressive strength evaluation of structural lightweight concrete by non-destructive ultrasonic pulse velocity method
Ultrasonics
(2013) - et al.
Non-destructive testing techniques for the forensic engineering investigation of reinforced concrete buildings
Forensic Sci. Int.
(2007) - et al.
Prediction of concrete strength using ultrasonic pulse velocity and artificial neural networks
Ultrasonics
(2009) - et al.
Concrete properties evaluation by statistical fusion of NDT techniques
Constr. Build. Mater.
(2012) - et al.
Modeling of compressive strength and UPV of high-volume mineral-admixtured concrete using rule-based M5 rule and tree model M5P classifiers
Constr. Build. Mater.
(2015) - et al.
Optimal adjustment of EC-2 shear formulation for concrete elements without web reinforcement using Genetic Programming
Eng. Struct.
(2010) - et al.
Optimization of existing equations using a new Genetic Programming algorithm: ppplication to the shear strength of reinforced concrete beams
Adv. Eng. Software
(2012) - et al.
Prediction of the mechanical properties of structural recycled concrete using multivariable regression and genetic programming
Constr. Build. Mater.
(2016) - et al.
Evaluation of the splitting tensile strength in plain and steel fiber-reinforced concrete based on the compressive strength
Constr. Build. Mater.
(2015)
Predicting modulus elasticity of recycled aggregate concrete using M5′ model tree algorithm
Constr. Build. Mater.
Prediction of the compressive strength of normal and high-performance concretes using M5P model tree algorithm
Constr. Build. Mater.
Concretes and mortars with waste paper industry: biomass ash and dregs
J. Environ. Manage.
Genetic programming and floating boom performance
Ocean Eng.
Instrucción de Hormigón Estructural EHE-08
Concrete. Complementary British Standard to BS EN 206. Specification for Constituent Materials and Concrete
Guia para a utilização de agregados reciclados grossos em betões de ligantes hidráulicos, Portugal
Norme tecniche per le costruzioni
Ultrasonic assessment of concrete strength at early ages
Mag. Concr. Res.
Pulse velocity as a measure of concrete compressive strength
Spec. Publ.
Cited by (53)
Estimating compressive strength of concrete containing rice husk ash using interpretable machine learning-based models
2024, Case Studies in Construction MaterialsForecasting the strength characteristics of concrete incorporating waste foundry sand using advance machine algorithms including deep learning
2023, Case Studies in Construction MaterialsFresh and mechanical performances of recycled plastic aggregate geopolymer concrete modified with Nano-silica: Experimental and computational investigation
2023, Construction and Building MaterialsPrediction of compressive strength of rice husk ash concrete: A comparison of different metaheuristic algorithms for optimizing support vector regression
2023, Case Studies in Construction MaterialsGEP-based models for estimating the elastic shear buckling and ultimate loads of cold-formed steel channels with staggered slotted web perforations in shear
2022, StructuresCitation Excerpt :For both models, 80 % of the whole data repository was used in the predictive model training, while the remaining data was employed to test the models' generalization performance. Similar approaches were previously used in [21,24,31,32,36–39,41,43–45,47,48]. Fig. 4a and b present statistical comparisons of the data distributions for the training, testing, and entire sets in the form of box plots, where the top, middle, and bottom lines respectively indicate the maximum, mean, and minimum values of each input variable, and the standard deviations are denoted by red points.
Introduction of a novel evolutionary neural network for evaluating the compressive strength of concretes: A case of Rice Husk Ash concrete
2022, Journal of Building Engineering