The impact of ggbs addition on sustainable form of concrete

literature review and Research background

Concrete as one of the most common materials building construction including cement, water, aggregates. However, the prediction of concrete properties cannot be investigated using detailed analysis methods. Based on the results obtained from experiments and experimental formulas, neural networks have important properties as one of the components of artificial intelligence that make them important in science and engineering issues. Neural network as a tool for regression (highly adapted to changes) and especially in cognition of its prototype and functions definition have found a special place [1]. These functions are nonlinear and capable of finding very complex relationships between input and output variables without any prior relationship to them, given the high cost of testing and wasted time consumed by previous methods, using neural networks and meta-heuristic algorithms to optimize rheological properties of self-compacting of the mechanical and concrete to present a SCC concrete mixing scheme would be useful. In this research, with the available information and Consider the concrete components (cement content, water to cement ratio, stone content and additives) as system variables (input parameters) and the results of each mixing scheme (compressive and tensile strength and slip by experiments) as desired output, using meta-heuristic algorithms and networks neural. This algorithms contain different types of information which would be analysed and finally optimal functions are obtained to predict the concrete properties using input data.

Ground blast furnace slag (GGBS or GGBFS) by removing molten iron slag  from a blast furnace in water or steam for glass and grain products. The product should be crushed in a fine powder. Konstantin Soblov’s study of a new method for the removal of high performance concrete mixtures [2]. The results of this research can be generalized from high performance concrete based on silica fume. The robustness and rheological properties of silica cloud grain plastic cement system are presented. Test results have shown that a super plasticizer ratio of silica (1:10) provides extremely dense packaging and high fluidity of the system. High performance concrete models derived from experimental data. These models provide W / C calculation equations for the required compressive strength (max. 130 MPa) as well as the amount of cement paste needed for stagnation (in the range of 40 to 200 mm). For modeling purposes, concrete stagnation is presented as a function of the aggregate ratio, volume, and fluidity of the cement paste. This method is probably a novel method that offers high performance for concrete mixing [2].

M.C.Nataraja et al study on a fuzzy-neuro model for ordinary concrete mix design [3].  The process used to fit the material in the proper ratio is called concrete mix design. Although based on sound technical principles and exploration, this whole path is not in the area of precision mathematics. This is due to inaccuracy, ambiguity, approximations and tolerance. In this research, the development of a new technique for approximate standard concrete mixtures is shown. The distinct fuzzy inference modules are framed in five layers for achieving ambiguity and approximation at different design stages as presented in IS. In this model, a trained three-layer neural network is integrated to generate experimental data on the pressure ratio w / cv / s with respect to 28-day compressive strength in three days from three popular cement brands. . The results are in good agreement with the values of cement, heavy aggregate, periodic aggregate and water. The present method has been used for different types of standard concrete mixtures with results obtained by the conventional method. Details of the system model and comparative diagrams are presented [3].

1. Alterman et al evaluatie concrete materials by application of automatic reasoning [4]. There were two targets for this research. One: to enable more or less automatic confirmation of the known linking– either quantitative or qualitative – between technological data and selected properties of concrete components. Even more important is the second purpose – presentation of expected feasibly of automatic identification of new such relationships, not yet identified by civil engineers. The association are to be obtained by technics of Artificial Intelligence, (AI), and according to actual consequences from experiments on concrete materials. The reason of applying the AI means is that in Civil Engineering the real data are usually non perfect, complicate, fuzzy, often with missing details, which means that their analysis in a conventional way, by constructing empirical models, is roughly possible or at least cannot be done quickly. The main idea of the proposed approach was to combine application of different AI methods in a one system, aimed at estimation, prediction, design and/or optimization of composite materials. The paradigm of the approach is that the unknown rules concerning the properties of concrete are hidden in experimental results and can be obtained from the analysis of examples. Different AI techniques like artificial neural networks, machine learning and certain techniques associated with statistics were applied. The data for the analysis originated from direct observations and from reports and publications on concrete technology. Among others it has been shown that by combining different AI technics it is feasible to improve the quality of the data, (e.g. when encountering outliers and missing values or in clustering difficulties), so that the whole data processing system will be giving better estimation, (when applying ANNs), or the newly discovered rules and regulations will be more effective, (e.g. with descriptions more complete and – at the same time – possibly more consistent, in case of ML algorithms) [4].

Seyed Jamaleddin Seyed Hakim et al. In an article investigated the efficiency of artificial neural networks for estimating the compressive strength of concrete [5]. They demonstrated a method for predicting 28-day compressive strength of high strength concrete (HSC) using MFNNs in this paper. The ANN model was trained and tested using available data. A total of 368 different data from HSC hybrid designs were collected from the technical. The data used to predict the compressive strength with ANN encompass eight input factors which include cement, water, coarse aggregate, fine aggregate, silica fume, superplasticizer, fly ash and granulated grated blast furnace slag. For the training state, different combinations of layers, number of neurons, learning rate, momentum and activation functions were considered. The training was terminated when the root mean square error (RMSE) became or was less than 0.001 and the results were tested with test data set. A total of 30 architectures were studied and the 8-10-6-1 architecture was the best possible architecture. The results indicate  that the relative percentage error (RPE) for the training set was 7.02% and the testing set was 12.64%. The ANNs models give high prediction accuracy, and the research results demonstrate that using ANNs to predict concrete strength is practical and constructive [5].

Another research has done by Arun Kumar Dwivedi in India(2013), entitled “GGBS as Partial Replacement of OPC in Cement Concrete – An Experimental Study”. In this paper an experimental study of compressive and flexural strength of concrete prepared with Ordinary Portland Cement, partially replaced by ground granulated blast furnace slag in different proportions varying from 0% to 40%. It is observed from the investigation that the strength of concrete is inversely proportional to the % of replacement of cement with ground granulated blast furnace slag. It is concluded that the 20% replacement of cement is possible without compromising the strength with 90 days curing(13).

Carbonation resistance of GGBS concrete is other research has done by Chao-Qun Lye and et.al(2016). This study presents an analysis of a 30 000 strong data matrix derived from 227 studies originating from 35 countries since 1968. Similar to the fly ash effect, the carbonation of concrete increases with the incorporation of ground granulated blast-furnace slag (GGBS), but the rate increases as GGBS content is increased. This effect is greater for concrete designed on an equal water/cement (w/c) basis to the corresponding Portland cement (PC) concrete than on an equal strength basis. The Eurocode 2 specification for XC3 carbonation exposure in terms of the characteristic cube strength of concrete (or its w/c ratio) may need to be increased (or decreased) with the addition of GGBS. Other influencing factors, including GGBS fineness, total cement content and curing, were also investigated. In some cases, the carbonation of in-service GGBS concrete has been estimated to exceed the specified cover before 50 years of service life. Measures to minimise the carbonation of GGBS concrete are proposed. Fully carbonated reinforced GGBS concrete is assessed to show a higher corrosion rate. In relation to PC concrete, the carbonation of GGBS concrete is essentially similar when exposed to 3–5% carbon dioxide accelerated or indoor natural exposure, and the conversion factor of 1 week accelerated carbonation equal to 0·6 year is established(14).

Other research which has written by Rami A.Hawileh and et.al entitle:” Performance of reinforced concrete beams cast with different percentages of GGBS replacement to cement”. In this study The goal was to study the impact of using Ground Granulated Blast Furnace Slag (GGBS) as a partial replacement to cement in reinforced concrete (RC) beams. A total of eight beams were cast with different percentages of GGBS replacement of 0%, 50%, 70%, and 90%, respectively. The performance of the tested specimens was evaluated and compared to that of a control beam without GGBS (0%). In addition, the concrete compressive and tensile strength of the different concrete mixes were evaluated and compared. Overall, test results indicated that the compressive and tensile strength of the different mixtures were quite similar. In addition, the performance of RC beams with GGBS replacement up to 70% is similar to that without GGBS. However, the stiffness and strength for the beam specimens with 90% GGBS were lower than that without GGBS by 16% and 6%, respectively. It was also concluded that the use of high percentage of GGBS up to 70% as a replacement to cement is practical and will not comprise the performance of RC beams. Furthermore, such replacement will contribute to the reduction in CO2 emission (carbon footprint) and therefore encourage the use of such sustainable and green concrete(15).

Vinay Chandwani et al worked on modeling slump of ready mix concrete using genetically evolved artificial neural networks [6]. Artificial neural networks (ANNs) have been the preferred choice for modeling the complex and nonlinear material behavior where conventional mathematical approaches do not yield the desired accuracy and predictability. Despite their popularity as a universal function approximator and wide range of applications, no specific rules for deciding the architecture of neural networks catering to a specific modeling task have been formulated. The research paper presents a methodology for automated design of neural network architecture, replacing the conventional trial and error technique of finding the optimal neural network. The genetic algorithms (GA) stochastic search has been harnessed for evolving the optimum number of hidden layer neurons, transfer function, learning rate, and momentum coefficient for back propagation ANN. The methodology has been applied for modeling slump of ready mix concrete based on its design mix constituents, namely, cement, fly ash, sand, coarse aggregates, admixture, and water-binder ratio. Six different statistical performance measures have been used for evaluating the performance of the trained neural networks. The study showed that, in comparison to conventional trial and error technique of deciding the neural network architecture and training parameters, the neural network architecture evolved through GA was of reduced complexity and provided better prediction performance [6].

The study of Sonebi, according to the required properties and parameters, a pattern of concrete mixing is set. In order to reduce production costs, it works on the effects of bonding on the properties of self-compressing concrete with moderate strength. The results show improved rheological properties, crossing ability as well as reduced resolution of GGBS and GGBS effects. This model illustrates a factorial design method to optimize mixtures and adjust efficient control quality [7].

Research (Manawadu, Wijesinghe and Abeyruwan, 2015) has been carried out on the identification of plastic viscosity and yield stress (Bingham parameters) of fresh concrete using coaxial type barometers. The effect of powder to water ratio and GGBS content on rheological properties was investigated to determine the optimal amount of GGBS composition. The results showed that plastic viscosity and yield stress increased with increasing GGBS especially when GGBS was replaced with 18% in the mixture. The study found that the optimal amount was 18% of the GGBS replacement when combined with the original SCC components, with the ratio between the GGBS and the total volume of the mixture estimated at about 12%. According to the results of the study, there is no relationship between the rheology of freshly mixed concrete and the hardening properties of concrete [8].

Experimental studies that by Bharali, was carried out to evaluate the effects of GGBS and GGBS, as a partial replacement (30%GGBS, 40%GGBS, 30%GGBS, 40%GGBS, 15%GGBS+15%GGBS and 20%GGBS+20%GGBS) in the M30 grade self-compacted concrete. The filling ability, passing ability and flow ability was considered by using slump flow test, V-Funnel test and L-Box test. Compressive strength test, flexural test and split tensile test was use at 3 days, 7 days and 28 days of the concrete age to identify the hardened properties of the concrete [9].

In an investigation, Anil and Roodari used limestone powder and GGBS to evaluate the effects on both fresh and hardened properties of self-compacted concrete. Cement was replaced with 10%, 20%, 30%, 40% and 50% GGBS, respectively. Results showed that optimum values ​​for both hardened and fresh properties were obtained when using 30% GGBS substitution. At this stage, the compressive strength is 428 kg / cm 2. Increasing GGBS by more than 30% results in decreased concrete strength. The maximum compressive strength of 460 kg / cm 2 was obtained at the time of mixing of 15% GGBS and 30% limestone. 50% GGBS replacement due to greater compressive strength than control (100% cement). Therefore, according to this study, 50% of GGBS substitution can be considered as a suitable ratio for moderate resistance application[9].

A study by (Ugwu et al., 2018) to evaluate the effects of partial replacement of cement (5%, 10%, 15% and 20%) by GGBS, carbide waste and mine dust both fresh and hardened properties. The W / C ratio and the amount of superplasticizer were constant in all mixtures. Various tests such as V-funnel, slenderness, annular current, J-ring and L-box were used to evaluate the passing ability, filling ability and separation strength of fresh concrete. To identify the compressive strength of the hard concrete, the compressive strength test, at 28 days of age of concrete, was used. The results showed that the new SCC features with GGBS, carbide waste and mine dust meet the guidelines of self-compacting Euro. Compressive strength of concrete increased from 318 kg / cm2 to 351 kg / cm2, 376 kg / cm2 and 435 kg / cm2 when exchanging 20% GGBS cement, 15% GGBS + 15% mine dust, respectively. . Also, replacement of 20% carbide waste resulted in a significant reduction in compressive strength up to 186 kg / cm 2. However, a combination of 15% GGBS and 15% mine dust was obtained at 28 days due to optimum compressive strength [11].

1      Purpose of Research

Our purpose of research is to evaluate impact of GGBS on the rheology, workability and compressive strength of self-compacted concrete. There are many reasons as to why we add GGBS to concrete:

Durability: GGBS cement is routinely specified in concrete to provide protection against both sulphate attack and chloride attack. GGBS has now effectively replaced sulfate-resisting Portland cement (SRPC) on the market for sulfate resistance because of its superior performance and greatly reduced cost compared to SRPC. Most projects in Dublin’s Docklands, consist of  Spencer Dock, are using GGBS in subsurface concrete for sulfate resistance.

we also use to support against chloride attack, GGBS  at a replacement level of 50% in concrete. Instances of chloride attack occur in reinforced concrete in marine environments and in road bridges where the concrete is exposed to splashing from road de-icing salts. In most NRA projects in Ireland GGBS is now specified in structural concrete for bridge piers and abutments for protection against chloride attack. The use of GGBS in such instances will increase the life of the structure by up to 50% had only Portland cement been used, and precludes the need for more expensive stainless steel reinforcing. GGBS is also routinely used to limit the temperature rise in large concrete pours. The more gradual hydration of GGBS cement generates both lower peak and less total overall heat than Portland cement. This reduces thermal gradients in the concrete, which prevents the occurrence of micro cracking which can weaken the concrete and reduce its durability, and was used for this purpose in the construction of the Jack Lynch Tunnel in Cork.

Appearance: By contrast to the stony grey of concrete made with Portland cement, the near-white color of GGBS cement permits architects to achieve a lighter color for exposed fair-faced concrete finishes, at no extra cost. achieving a lighter color finish, GGBS is usually specified at between 50% to 70% replacement levels, although levels as high as 85% can be used. GGBS cement also produces a smoother, more defect free surface, due to the fineness of the GGBS particles. Dirt does not adhere to GGBS concrete as easily as concrete made with Portland cement, reducing maintenance costs. GGBS cement prevents the occurrence of efflorescence, the staining of concrete surfaces by calcium carbonate deposits. Due to its much lower lime content and lower permeability, GGBS is effective in preventing efflorescence when used at replacement levels of 50% to 60%.

Strength: Concrete containing GGBS cement has a higher ultimate strength than concrete made with Portland cement. It has a higher proportion of the strength-enhancing calcium silicate hydrates (CSH) than concrete made with Portland cement only, and a reduced content of free lime, which does not contribute to concrete strength. Concrete made with GGBS continues to gain strength over time, and has been shown to double its 28-day strength over periods of 10 to 12 years.

Sustainability: Since GGBS is a by-product of steel manufacturing process, its use in concrete is recognized by LEED etc. as improving the sustainability of the project and will therefore add points towards LEED certification. In this respect, GGBS can also be used for superstructure in addition to the cases where the concrete is in contact with chlorides and sulfates. This is provided that the slower setting time for casting of the superstructure is justified.

Our main aim of this study is to achieve optimal mixing design of self-compacting concrete and improve its properties having GGBS in its ingredients by using neural network system and meta-heuristic algorithms as one of the artificial intelligence tools. Then predicting the mechanical and rheological properties of concrete (compressive strength, tensile strength, slip resistance) using systems that decline expense and is time-saving based on  the amount of information available. Given the abundance of laboratory data available in recent years it this study can save cost, time and manpower.

2      Research Methods/Methodology

The method  that we will use in this research is to collect the relevant information such as the mixing parameters and rheological properties of the self-compacting concrete and the results of the mechanical properties are obtain by the author in numerous experiments or, if necessary,  experiments would be performed to complete the input data and then use this information to select neural networks and meta-heuristic algorithms to start training on the aforementioned networks and modelling the data in the network to begin try and error many times. In order to reach the network by most optimized responses, and then performing a series of data on neural network validation. Finally, the results of the neural network are compared with the actual results of the experiments and the error rate will be determined.

1) True slump                      2) true slump                    3) collapsed slump

Figure 1. Some of this experiment samples that represents type of slump according to (The Constructor-Civil Engineering Home 2017).

• The method of data gathering that we choice in this study is experimental and also we use databases includes the results of concrete experiments with different mixing designs, as well as we can collect the meta-heuristic algorithms through library studies. It would be possible with below measures:
• performing Experiments  on SCC Concrete
• Fresh concrete test to obtain rheological properties including slip flow test, hopper test v, j ring test, and so on.
• Doing hardened concrete tests to determine the mechanical properties of concrete, including compressive strength tests of various ages and tensile strength tests.
• Searching for artificial intelligence algorithms using library studies.

Figure 2. Represents slump test tools used in this study[12].

Figure 3. Samples in programmed environmental chamber[12].

Table 1. Mix Design Proportion: term R means cement replaced by GGBS. While, A   refers to the addition of GGBS to the mix concrete without changing the cement portion.

3      Research Resources, Limitations and Constraints

1- In order to make samples, internally and externally valid standards will be used.

2. Some data will be used by authoritative experiments and experiments.

3. In this study, we use meta-heuristic algorithms that include useful algorithms such as genetic algorithm, ant colony optimization algorithm, bee colony algorithm, colonial competition algorithm and other algorithms. It is obvious that the neural network consists of perceptron neural network, recurrent neural network, cross-propagation neural network and other types. In this research, we try to study these networks and algorithms and if they are compatible with the topic, we use in the end. Their results will be compared and finally we can obtain the optimal functions

4. Input parameters in present research would be cement content, water to cement ratio, aggregates and additives in concrete mixing design and output parameters are compressive strength, slump and tensile strength which are the same mechanical and rheological properties of the self-compacted concrete due to the density and use to be selected based on available information.

4      Project Programme/Gantt Chart

For start implementing algorithms in MATLAB software and, considering the inputs and outputs of the experiments, we should gather information, Then the neural network should be trained. Then we will optimize the amount of hidden layers after that the number of neurons in each neurons that are layering and actually adapting the network to our model and this process comes with trial and error.

Fig. 4. Method for RCC mixture proportioning

Table 2. Gantt chart of project

List of References (new page)

1. Sobolev KG, Soboleva SV. High strength concrete mix design and properties optimization, concrete technology in developing countries. In: Proceedings of the 4th International Conference, Famagusta, TRNC; 1996. p. 189–202.
2. Konstantin Sobolev, The development of a new method for the proportioning of high-performance concrete mixtures, Cement & Concrete Composites 26 (2004) 901–907.
3. M.C.Nataraja, M.A.Jayaram, C.N.Ravikumar, A Fuzzy-Neuro Model for Normal Concrete Mix Design, Engineering Letters, 13:2, Advance online publication: 4 August 2006.
4. D. Alterman, J. Kasperkiewicz, Evaluating concrete materials by application of automatic reasoning, bulletin of the polish academy of sciences, technical sciences, Vol. 54, No. 4, 2006.
5. Seyed Jamalaldin Seyed Hakim, Jamaloddin Noorzaei, M. S. Jaafar, Mohammed Jameel, Mohammad Mohammadhassani, Application of artificial neural networks to predict compressive strength of high strength concrete, International Journal of the Physical Sciences Vol. 6(5), pp. 975-981, 4 March, 2011.
6. Vinay Chandwani, Vinay Agrawal, and Ravindra Nagar, Modeling Slump of Ready Mix Concrete Using Genetically Evolved Artificial Neural Networks, Hindawi Publishing Corporation, Advances in Artificial Neural Systems, Volume 2014, Article ID 629137, 9 pages.

7. Sonebi, M. (2014). Medium strength self-compacting concrete containing fly ash: Modelling using factorial experimental plans. Cement and Concrete Research, 34(7), pp.1199-1208.
8. Manawadu, A., Wijesinghe, W. and Abeyruwan, H. ,Rheological Behaviour of Cement Paste with Fly Ash in the Formulation of Self-Compacting Concrete (SCC). International Conference on Structural Engineering and Construction Management 2015, 2(13), pp.23-27.
9. Bharali, B, EXPERIMENTAL STUDY ON SELF COMPACTING CONCRETE (SSC) USING GGBS AND FLY ASH. International Journal of Core Engineering & Management (IJCEM), 2(6), pp.1-11.
10. Anil, K. and Chowdary, L., Study on Strength Properties of S elf Compacting Concrete using GGBS and Lime stone powder as Mineral Admixtures. International Journal of Innovative Research in Science, Engineering and Technology, Journal of Innovative Research in Science, Engineering and Technology, Vol. 6, Issue 3, March 2017 .
11. Ugwu, O., Nwoji2, C., Onyia, M., Gber, A., Tarzomon, T. and Happiness, O., Development of Self Compacting Concrete Using Industrial Waste As Mineral And Chemical Additives. IOSR Journal of Mechanical and Civil Engineering, 2018.
12. Morteza Khorami, Effect of PFA on Compressive Strength of Concrete in Cold Weather, 3023EXQ Research Dissertation, Coventry university, April 2018.
13. Arun Kumar Dwivedi and et.al: “GGBS as Partial Replacement of OPC in Cement Concrete – An Experimental Study” Volume: 2 |Issue: 11 | November 2013 • ISSN No 2277 – 8179.

14.  Chao-Qun Lye Ravindra K. Dhir Gurmel S. Ghataora: Carbonation resistance of GGBS concrete, https://doi.org/10.1680/jmacr.15.00449, August 16, 2016

15. Rami,A.Hawileh. Jamal,Abdalla. Fakherdine,Fardmanesh. Poya,Shahsana .AbdolrezaKhalili: “Performance of reinforced concrete beams cast with different percentages of GGBS replacement to cement”. Volume 17, Issue 3, May 2017, Pages 511-519. https://doi.org/10.1016/j.acme.2016.11.006 .