Influence of coal ashes on fired clay brick quality: Random forest regression and artificial neural networks modeling

https://doi.org/10.1016/j.jclepro.2023.137153Get rights and content

Highlights

  • 303 cases of brick clays and coal ashes mixtures from the literature are studied.

  • Random forest and Artificial Neural Network models are built to find the link between raw materials and final products.

  • The amount of Na2O in the raw clay and K2O in the ash are decisive factors influencing the quality of ash-clay bricks.

  • Additional research regarding the use of pond and bottom ashes in industrial products is needed.

  • The hand-molded products with 50% pond ash fired at 950 °Care proven non-resistant to frost.

Abstract

Finding a solution to the problem of the large buildup of coal ashes is a vital necessity. Although the use of coal ashes in fired clay bricks has been thoroughly investigated, there is insufficient information on their industrial utilization and researchers do not agree on whether or not this addition improves the quality of the final products. Therefore, a database has gathered 20 years of research containing key factors related to the quality of the bricks (i.e., chemical composition, firing temperature, soaking time, open porosity, water absorption and compressive strength). Then, random forest regression and artificial neural networks (ANN) modeling were used to separately predict the parameters concerning the quality of the final products. The overall conclusions were that the compressive strengths were the highest when using fly ashes and that class F ashes were highly suitable to be used in the brick industry as a replacement material for brick clay. In addition, the ANN models showed higher coefficients of determination and an overall better fit to the experimental data. By changing the chemical makeup of the initial materials and their proportions, the particle size of the ashes, the firing temperature and soaking time, as well as the size of a product, the created models can be used to estimate the quality of the brick containing coal ash. That is crucial because the inconsistent chemical composition of ash is generally the main obstacle to its utilization. The local sensitivity analysis revealed the highest influence of the content of the alkali oxides in the initial clay on the fired clay bricks due to their fluxing effect. In the case of ash-clay bricks, the decisive factors were the type of furnace used, the ashes' class, the Na2O content in raw clay, and the K2O introduced with the ash. The F class ashes containing about 2–3% of K2O and <5% of CaO gave the highest compressive strength in bricks fired at 1000–1100 °C.

Additional analyzes were made for 50% pond ash and 50% clay bricks to test the best-suited model and fill in the knowledge gap. The results obtained in this study are important for supporting the decision in the selection of materials and process parameter values that will increase the quality of the ash-clay-fired bricks.

Introduction

With the expansion of thermal power plants to supply the worldwide electricity demand, the volume of coal ash continues to rise (Vasić et al., 2021). The ashes differ significantly depending on the age of coal used in the production, the type of incinerator and the combustion process. The most common classification of fly, bottom and pond ashes is based on the content of calcium which divides the ashes into the classes F and C (ASTM C618–19:2022), which present the extent of their possible pozzolanic activity important in the concrete industry. Higher calcium content is seen when younger sub-bituminous coal deposits are used. Additional important differences are seen in the granulation of the ashes and the position of the collection (fly and bottom ashes). The ashes also strongly differ in the quantities of the chemical constituents, most importantly silica and alumina. Furthermore, a highly significant factor is the leftover quantity of the unburned particles, which changes the chemical composition (Sarkar et al., 2007; Bįlgįl et al., 2017; Vasić et al., 2021). The differing composition and particle size distribution of coal ashes limit their industrial application (Vasić et al., 2021).

Coal ashes contain heavy metals in significant quantities (especially mercury, lead, chromium, cadmium, and arsenic), and as such represent a huge threat to the environment due to the fine fraction that is easily distributed by wind and can penetrate the lungs of living organisms. Moreover, the high demand for water needed for transport to the ponds (15:1), the potential leaching of heavy metals to the groundwaters, problematic and expensive disposal, and the taking up of large areas of land, are all seen as significant problems (Sena da Fonseca et al., 2015; Vasić et al., 2021).

Fly ash is usually used in the construction and building industry, mainly as a partial replacement in cement and concrete formulations, but this implementation in final products is falling far short of its yearly output rate. The pozzolanic activity in these mixtures is of high importance, and that, being higher in fly ashes, leaves the bottom ashes problems unsolved. Although intensive research on the effective utilization of the ashes has been carried out during the last decades, it seems that there is still insufficient practical application, since it constitutes the second largest waste worldwide. There are about 35,000 thermal power plants in the world in 167 countries (Global Power Plant Database, 2021), of which about 8500 are coal-generated and account for 38% of global electric energy production (Abbas et al., 2020). Although the emerging trend in decreasing CO2 emissions led to alternative electricity production, there is still a huge quantity of leftover coal ashes. Worldwide, 730 and 800 million tons of bottom and fly ashes respectively are generated annually (Ahmed et al., 2016; Abbas et al., 2020), of which most are in China, India and the United States. According to the US EPA, only up to 40% of the coal ashes that are produced worldwide are reused as of 2012. However, in the nations with the highest production rates, primarily China (86%) (Abbass et al., 2022), more than 50% of the coal ashes are recycled.

The advantages of the usage of coal ashes in the brick industry are numerous (Vasić et al., 2021). Maybe the most important one is that the expected quantity that may be introduced is high given the similar chemical composition of both the materials and their same major constituents (SiO2, Al2O3, and Fe2O3) (Taki et al., 2020; Vasić et al., 2021). Furthermore, the usage of these kinds of ashes significantly reduces the environmental footprint in the industry (Ncube et al., 2021). There is no concrete evidence to support the assertion that ash consumption in the brick sector has increased recently (Mukhtar et al., 2022). The small amount of literature on this subject focuses primarily on hand-molded industrial-sized samples (Andreola et al., 2005; Sarkar et al., 2007; Sonawane and Dwivedi, 2013; Makaka, 2014; Pawar and Garud, 2014; Abbas et al., 2017). In addition, the batches used in factories are also shrouded in secrecy and industrial studies in the literature are scarce. As a result, there is a significant knowledge gap that should be filled during future research.

Statistical and mathematical methods are applicable in all areas of the research (Arsenović et al., 2015a), and they have been successfully intensively implemented in fired bricks investigations for more than a decade (Arsenović et al., 2013a). Based on the 20 years of accessible data, this study uses random forest regression (RFR) methods and artificial neural networks (ANN) to forecast the quality of the ash-clay bricks. The input parameters used in modeling and further analysis were those that were most frequently found in the literature (chemical composition of the raw clay and ashes, weight percentage and the particle size of the ashes, the dimensions of formed products, firing temperature, and soaking time). Moreover, the effects of the features of the original materials and the processing settings are contrasted. The most decisive factors determined are the chemical compositions of the employed raw clays and ashes, notably the content of alkali metals, and the furnace employed. The share of the ashes, particle size and firing temperature all seemed to be of a lower significance. It must be kept in mind that most of the analyzed cases included fly ashes which are of fine particle sizes (below 200 μm) (Dondi et al., 1997). To improve the predictions and findings, more information is required, particularly on the use of bottom and pond ashes and testing of the industrial probes. Also, further research on hollow bricks is required. Moreover, corrosion resistance and durability (freeze-thaw resistance) testing are lacking in this area and samples must be more intensively tested to investigate the available micropores in the matrix when coal ash is used as secondary raw material.

As a piece of additional information in this study, to fill in the knowledge gap on the behavior of the ash-added bricks in extreme conditions showing their durability after firing at 950 °C, the laboratory samples of 50% of raw clay and 50% pond ash were molded. The mixed material was first characterized to determine its chemical and mineralogical compositions, particle size distribution and functional groups detection. The results are also used as another check of the best-behaved mathematical model. After the fired samples underwent freeze-thaw cycles, their microstructure is compared to those not subjected to extreme conditions.

The main objective of this research was to construct a reliable mathematical model that, when given significant input factors including the chemical composition of raw clay and ash, particle size, sample size, soaking time and firing temperature, can accurately predict the quality of the final brick described by compressive strength, water absorption and open porosity. The ultimate goal is to support the manufacturers in their effort to recycle highly represented waste and gain high-quality bricks by decreasing a troublesome carbon footprint. Additional experiments are conducted on bricks containing pond ash, and the findings are used to evaluate and prove the predictive power of the models.

Section snippets

The database

The database includes 303 cases that have been published in the literature within the last 20 years (Vasić et al., 2021). The parameters gathered and employed in modeling are shown in Fig. 1. The foundation also contains the previous findings from the Institute for testing of materials in Belgrade, Serbia (36 of the samples) research (Arsenović et al., 2015b, 2015c). Based on the previously provided data, the chemical composition of the used clays and coal ashes was taken into consideration.

Random forest regression modeling

In this investigation, the collected dataset was randomly divided into two homogenous subsets: a training subset and a test subset, which referred to 60% and 40% of the total data, respectively. From the input sample dataset, new sub-samples were chosen. To fit these sub-samples, a 1000 trees architecture was chosen for the RFR structure. To reduce prediction error, the RFR model averaged the output of the built trees during the training cycle. The random test data percentage was set to 40%

Conclusions

Finding a solution to the ever-growing ponds of the electricity industry-generated waste ash remains imperative. This study provides some new aspects of the application of fly, bottom and pond ashes in the brick industry. To understand the quality of the fired bricks, a database compiled from over 20 years of research was created. It contained information on the chemical makeup of brick clays and ashes as well as other pertinent factors including peak firing temperature and soaking duration.

CRediT authorship contribution statement

Milica Vidak Vasić: Conceptualization, Methodology, Statistical Analysis, Mathematical modeling, Writing – original draft, Visualization, Writing – review & editing, Supervision. Heli Jantunen: Conceptualization, Writing – review & editing. Nevenka Mijatović: Conceptualization, Writing – review & editing. Mikko Nelo: Visualization, Writing – review & editing. Pedro Muñoz Velasco: Conceptualization, Methodology, Writing – review & editing, Supervision.

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.

Acknowledgments

This research was funded by the Ministry of Science, Technological Development and Innovation of the Republic of Serbia [Contract No. 451-03-47/2023-02/200012]. The authors also wish to thank the Chilean National Commission on Research and Development (CONICYT) [FONDECYT REGULAR grant number 1180414]. The authors are grateful to Smilja Marković from the Institute of Technical Science of the Serbian Academy of Sciences and Arts, Belgrade, Serbia for providing particle size and FT-IR analyses.

References (57)

  • S. Marković et al.

    Designing, fabrication and characterization of nanostructured functionally graded Hap/BCP ceramics

    Ceram. Int.

    (2015)
  • N. Mijatović et al.

    Fired pressed pellet as a sample preparation technique of choice for an energy dispersive X-ray fluorescence analysis of raw clays

    Talanta

    (2023)
  • W. Mozgawa et al.

    Investigation of the coal fly ashes using IR spectroscopy

    Spectrochim. Acta Mol. Biomol. Spectrosc.

    (2014)
  • A. Ncube et al.

    Moving towards resource efficiency and circular economy in the brick manufacturing sector in Zimbabwe

    J. Clean. Prod.

    (2021)
  • B. Sena da Fonseca et al.

    Technical feasibility of reusing coal combustion by-products from a thermoelectric power plant in the manufacture of fired clay bricks

    Appl. Clay Sci.

    (2015)
  • M. Sutcu et al.

    Recycling of bottom ash and fly ash wastes in eco-friendly clay brick production

    J. Clean. Prod.

    (2019)
  • K. Taki et al.

    Utilization of fly ash amended sewage sludge as brick for sustainable building material with special emphasis on dimensional effect

    J. Clean. Prod.

    (2020)
  • S. Abbas et al.

    Recycling untreated coal bottom ash with added value for mitigating alkali–silica reaction in concrete: a sustainable approach

    Sustainability

    (2020)
  • W. Abbass et al.

    Manufacturing of sustainable untreated coal ash masonry units for structural applications

    Materials

    (2022)
  • S. Ahmed et al.

    Geographical spread of fly ash generation and residual potential for its utilization in India

    Int. J. Innov. Res. Rev.

    (2016)
  • M.N. Amin et al.

    Prediction of strength and CBR characteristics of chemically stabilized coal gangue: ANN and random forest tree approach

    Materials

    (2022)
  • F. Andreola et al.

    Recycling industrial waste in brick manufacture. Part I

    Mater. Construcción

    (2005)
  • M. Arsenović et al.

    The main factors influencing canine demodicosis treatment outcome and determination of optimal therapy

    Parasitol. Res.

    (2015)
  • Standard Specification for Coal Fly Ash and Raw or Calcined Natural Pozzolan for Use in Concrete

    (2022)
  • Standard Specification for Building Brick (Solid Masonry Units Made from Clay or Shale)

    (2017)
  • Testing Method of Clay Bricks, Blocks and Slabs

    (2002)
  • A. Bįlgįl et al.

    Manufacture of fired bricks containing an industrial waste (bottom ash)

    OHU J. Eng. Sci

    (2017)
  • M.M. Chou et al.

    Chemical and engineering properties of fired bricks containing 50 weight percent of class F fly ash

    Energy Sources

    (2010)
  • Cited by (5)

    View full text