Compressive strength analysis of fly ash-based geopolymer concrete using machine learning approaches
Cement is the primary component of concrete, an extensively used building material. When cement is manufactured or utilized, an excessive amount of gas is emitted into the atmosphere, which is detrimental to the environment. Using geopolymer concrete (GPC) is crucial in minimizing this flaw. Fly ash...
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2022-12-01
|
| Series: | Results in Materials |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2590048X22000954 |
| _version_ | 1817975706602176512 |
|---|---|
| author | Dina A. Emarah |
| author_facet | Dina A. Emarah |
| author_sort | Dina A. Emarah |
| collection | DOAJ |
| description | Cement is the primary component of concrete, an extensively used building material. When cement is manufactured or utilized, an excessive amount of gas is emitted into the atmosphere, which is detrimental to the environment. Using geopolymer concrete (GPC) is crucial in minimizing this flaw. Fly ash is used as the binder in geopolymer concrete and is commonly consumed to prepare composites of GPC. Compressive strength (C–S) is the most significant mechanical property for all types of concrete composites, including geopolymer concrete. Due to the importance of the C–S of concrete at 28 days in the design of structures, it is necessary to build a reliable model to predict the compressive strength of geopolymer concrete. A valid model for predicting the C–S of fly ash-based geopolymer concrete (FA-GPC) has been developed in this research for economic considerations. Therefore, the machine learning approach was used to predict the C–S of FA-GPC through artificial neural networks (ANNs), deep neural networks (DNNs), and deep residual networks (ResNet). In this regard, several academic research studies collected a comprehensive dataset of 860 samples and analyzed them to develop the three models. In the modeling process, twelve effective input parameters on the C–S of the FA-GPC, including the content of FA, fine aggregate (Fa), coarse aggregate (Ca), sodium silicate (SS), sodium hydroxide (SH), sodium silicate/sodium hydroxide ratio (SS/SH), silicon dioxide/alumina doped of FA (Si/Al), alkaline liquid/FA (AL/FA), sodium hydroxide concentration (M), curing temperature (CT), period of curing in ovens (CP), and sample ages (Sa), and the C–S of the FA-GPC as an output variable. Various statistical assessment criteria, such as the coefficient of determination, the mean-absolute percentage deviation, and root-mean-square deviation, were used to evaluate the efficiency of the developed models. The cross-validation technique (k-fold) confirmed the model's performance. The results indicated that the ResNet model predicted the C–S of FA-GPC mixtures better than the other models. Also, the sensitivity analysis of the three proposed models shows that the curing temperature, the ratio of alkaline liquid to the binder, and the amount of sodium silicate are the most important parameters for estimating the C–S of the FA-GPC. |
| first_indexed | 2024-04-13T21:53:35Z |
| format | Article |
| id | doaj.art-3832f9d9e80e49b193b0b3c688862e04 |
| institution | Directory Open Access Journal |
| issn | 2590-048X |
| language | English |
| last_indexed | 2024-04-13T21:53:35Z |
| publishDate | 2022-12-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Results in Materials |
| spelling | doaj.art-3832f9d9e80e49b193b0b3c688862e042022-12-22T02:28:21ZengElsevierResults in Materials2590-048X2022-12-0116100347Compressive strength analysis of fly ash-based geopolymer concrete using machine learning approachesDina A. Emarah0Construction Research Institute (CRI), National Water Research Center (NWRC), Delta-Barrage, 13621, EgyptCement is the primary component of concrete, an extensively used building material. When cement is manufactured or utilized, an excessive amount of gas is emitted into the atmosphere, which is detrimental to the environment. Using geopolymer concrete (GPC) is crucial in minimizing this flaw. Fly ash is used as the binder in geopolymer concrete and is commonly consumed to prepare composites of GPC. Compressive strength (C–S) is the most significant mechanical property for all types of concrete composites, including geopolymer concrete. Due to the importance of the C–S of concrete at 28 days in the design of structures, it is necessary to build a reliable model to predict the compressive strength of geopolymer concrete. A valid model for predicting the C–S of fly ash-based geopolymer concrete (FA-GPC) has been developed in this research for economic considerations. Therefore, the machine learning approach was used to predict the C–S of FA-GPC through artificial neural networks (ANNs), deep neural networks (DNNs), and deep residual networks (ResNet). In this regard, several academic research studies collected a comprehensive dataset of 860 samples and analyzed them to develop the three models. In the modeling process, twelve effective input parameters on the C–S of the FA-GPC, including the content of FA, fine aggregate (Fa), coarse aggregate (Ca), sodium silicate (SS), sodium hydroxide (SH), sodium silicate/sodium hydroxide ratio (SS/SH), silicon dioxide/alumina doped of FA (Si/Al), alkaline liquid/FA (AL/FA), sodium hydroxide concentration (M), curing temperature (CT), period of curing in ovens (CP), and sample ages (Sa), and the C–S of the FA-GPC as an output variable. Various statistical assessment criteria, such as the coefficient of determination, the mean-absolute percentage deviation, and root-mean-square deviation, were used to evaluate the efficiency of the developed models. The cross-validation technique (k-fold) confirmed the model's performance. The results indicated that the ResNet model predicted the C–S of FA-GPC mixtures better than the other models. Also, the sensitivity analysis of the three proposed models shows that the curing temperature, the ratio of alkaline liquid to the binder, and the amount of sodium silicate are the most important parameters for estimating the C–S of the FA-GPC.http://www.sciencedirect.com/science/article/pii/S2590048X22000954Geopolymer concreteCompressive strengthFly ashMachine learningSensitivity analysisANN |
| spellingShingle | Dina A. Emarah Compressive strength analysis of fly ash-based geopolymer concrete using machine learning approaches Results in Materials Geopolymer concrete Compressive strength Fly ash Machine learning Sensitivity analysis ANN |
| title | Compressive strength analysis of fly ash-based geopolymer concrete using machine learning approaches |
| title_full | Compressive strength analysis of fly ash-based geopolymer concrete using machine learning approaches |
| title_fullStr | Compressive strength analysis of fly ash-based geopolymer concrete using machine learning approaches |
| title_full_unstemmed | Compressive strength analysis of fly ash-based geopolymer concrete using machine learning approaches |
| title_short | Compressive strength analysis of fly ash-based geopolymer concrete using machine learning approaches |
| title_sort | compressive strength analysis of fly ash based geopolymer concrete using machine learning approaches |
| topic | Geopolymer concrete Compressive strength Fly ash Machine learning Sensitivity analysis ANN |
| url | http://www.sciencedirect.com/science/article/pii/S2590048X22000954 |
| work_keys_str_mv | AT dinaaemarah compressivestrengthanalysisofflyashbasedgeopolymerconcreteusingmachinelearningapproaches |