Accesso libero

Intelligent Models for Prediction of Compressive Strength of Geopolymer Pervious Concrete Hybridized with Agro-Industrial and Construction-Demolition Wastes

Shriram Marathe

Shriram Marathe

Department of Materials Engineering and Construction Processes, Faculty of Civil Engineering, Wrocław University of Science and TechnologyWrocław, Poland
Cerca questo autore su
Sciendo|Google Scholar
Marathe, Shriram
 e 
Anisha P Rodrigues

Anisha P Rodrigues

Department of Computer Science and Engineering, Nitte (Deemed to be University), NMAM Institute of Technology (NMAMIT)INDIA
Cerca questo autore su
Sciendo|Google Scholar
Rodrigues, Anisha P
  
26 set 2024
Intelligent Models for Prediction of Compressive Strength of Geopolymer Pervious Concrete Hybridized with Agro-Industrial and Construction-Demolition Wastes's Cover Image
INFORMAZIONI SU QUESTO ARTICOLO
Articolo precedente
Articolo Successivo
Cita
Scarica la copertina
Categoria dell'articolo: Original Study
Pubblicato online: 26 set 2024
Pagine: 349 - 376
Ricevuto: 07 mag 2024
Accettato: 15 lug 2024
DOI: https://doi.org/10.2478/sgem-2024-0020
Parole chiave
© 2024 Shriram Marathe et al., published by Sciendo
This work is licensed under the Creative Commons Attribution 4.0 International License.

Figure 1:

Flowchart showing the experimentation and development of soft computing models.
Flowchart showing the experimentation and development of soft computing models.

Figure 2:

Particle size distribution of binder materials and aggregates.
Particle size distribution of binder materials and aggregates.

Figure 3:

Preparation, air-curing, and testing sequence of geopolymer pervious concrete specimens.
Preparation, air-curing, and testing sequence of geopolymer pervious concrete specimens.

Figure 4:

Model architecture flow diagram of the soft computing adopted in the current investigation.
Model architecture flow diagram of the soft computing adopted in the current investigation.

Figure 5:

Average compressive strength and hydraulic conductivity of trial pervious GPC mixes.
Average compressive strength and hydraulic conductivity of trial pervious GPC mixes.

Figure 6: (a)

Corelation matrix showing the affiliation of individual parameters with the other parameters.
Corelation matrix showing the affiliation of individual parameters with the other parameters.

Figure 6: (b)

Pearson's correlation coefficients between the parameters.
Pearson's correlation coefficients between the parameters.

Figure 7:

Actual vs predicted compressive strength results from ML models.
Actual vs predicted compressive strength results from ML models.

Figure 8:

Results of RMSE and R2 values of developed ML models.
Results of RMSE and R2 values of developed ML models.

Figure 9: (a)

Results showing the errors in predicted vs actual values of compressive strength from the testing dataset.
Results showing the errors in predicted vs actual values of compressive strength from the testing dataset.

Figure 9: (b)

Results showing the feature score of the ML models for compressive strength.
Results showing the feature score of the ML models for compressive strength.

Figure 10:

Results of predicted values and actual values from the ensemble Voting Regressor ML model.
Results of predicted values and actual values from the ensemble Voting Regressor ML model.

Unit = kg per cubic meter Unit = MPa
Mix ID GGBS AWA AAS NCA RCA FA CS
M-0-0 290 0 143.58 1881.3 0 199.7 32.2
M-0-0 290 0 143.58 1881.3 0 199.7 31.0
M-0-0 290 0 143.58 1881.3 0 199.7 31.6
M-0-0 290 0 143.58 1881.3 0 199.7 32.9
M-0-0 290 0 143.58 1881.3 0 199.7 31.2
M-0-0 290 0 143.58 1881.3 0 199.7 30.4
M-0-0 290 0 143.58 1881.3 0 199.7 34.2
M-0-0 290 0 143.58 1881.3 0 199.7 31.1
M-0-0 290 0 143.58 1881.3 0 199.7 29.9
M-0-0 290 0 143.58 1881.3 0 199.7 34.0
M-0-0 290 0 143.58 1881.3 0 199.7 32.8
M-0-0 290 0 143.58 1881.3 0 199.7 31.0
M-0-25 290 0 143.58 1411.03 444.01 199.7 32.7
M-0-25 290 0 143.58 1411.03 444.01 199.7 27.5
M-0-25 290 0 143.58 1411.03 444.01 199.7 30.3
M-0-25 290 0 143.58 1411.03 444.01 199.7 31.9
M-0-25 290 0 143.58 1411.03 444.01 199.7 32.8
M-0-25 290 0 143.58 1411.03 444.01 199.7 31.2
M-0-25 290 0 143.58 1411.03 444.01 199.7 27.5
M-0-25 290 0 143.58 1411.03 444.01 199.7 27.4
M-0-25 290 0 143.58 1411.03 444.01 199.7 30.0
M-0-25 290 0 143.58 1411.03 444.01 199.7 30.5
M-0-25 290 0 143.58 1411.03 444.01 199.7 28.1
M-0-25 290 0 143.58 1411.03 444.01 199.7 30.6
M-0-50 290 0 143.58 940.68 888.03 199.7 22.6
M-0-50 290 0 143.58 940.68 888.03 199.7 25.8
M-0-50 290 0 143.58 940.68 888.03 199.7 26.2
M-0-50 290 0 143.58 940.68 888.03 199.7 28.0
M-0-50 290 0 143.58 940.68 888.03 199.7 24.4
M-0-50 290 0 143.58 940.68 888.03 199.7 28.7
M-0-50 290 0 143.58 940.68 888.03 199.7 24.9
M-0-50 290 0 143.58 940.68 888.03 199.7 25.0
M-0-50 290 0 143.58 940.68 888.03 199.7 26.8
M-0-50 290 0 143.58 940.68 888.03 199.7 27.9
M-0-50 290 0 143.58 940.68 888.03 199.7 24.3
M-0-50 290 0 143.58 940.68 888.03 199.7 23.2
M-0-75 290 0 143.58 470.34 1332.05 199.7 23.8
M-0-75 290 0 143.58 470.34 1332.05 199.7 23.2
M-0-75 290 0 143.58 470.34 1332.05 199.7 21.8
M-0-75 290 0 143.58 470.34 1332.05 199.7 22.8
M-0-75 290 0 143.58 470.34 1332.05 199.7 22.1
M-0-75 290 0 143.58 470.34 1332.05 199.7 22.5
M-0-75 290 0 143.58 470.34 1332.05 199.7 20.5
M-0-75 290 0 143.58 470.34 1332.05 199.7 23.2
M-0-75 290 0 143.58 470.34 1332.05 199.7 21.1
M-0-75 290 0 143.58 470.34 1332.05 199.7 20.8
M-0-75 290 0 143.58 470.34 1332.05 199.7 23.0
M-0-75 290 0 143.58 470.34 1332.05 199.7 21.4
M-0-100 290 0 143.58 0 1776.07 199.7 19.5
M-0-100 290 0 143.58 0 1776.07 199.7 18.6
M-0-100 290 0 143.58 0 1776.07 199.7 18.9
M-0-100 290 0 143.58 0 1776.07 199.7 18.3
M-0-100 290 0 143.58 0 1776.07 199.7 17.5
M-0-100 290 0 143.58 0 1776.07 199.7 16.2
M-0-100 290 0 143.58 0 1776.07 199.7 16.6
M-0-100 290 0 143.58 0 1776.07 199.7 15.0
M-0-100 290 0 143.58 0 1776.07 199.7 19.1
M-0-100 290 0 143.58 0 1776.07 199.7 20.0
M-0-100 290 0 143.58 0 1776.07 199.7 18.4
M-0-100 290 0 143.58 0 1776.07 199.7 18.0
M-5-0 275.5 14.5 143.58 1878.13 0 199.3 32.0
M-5-0 275.5 14.5 143.58 1878.13 0 199.3 36.2
M-5-0 275.5 14.5 143.58 1878.13 0 199.3 31.1
M-5-0 275.5 14.5 143.58 1878.13 0 199.3 36.0
M-5-0 275.5 14.5 143.58 1878.13 0 199.3 35.8
M-5-0 275.5 14.5 143.58 1878.13 0 199.3 36.1
M-5-0 275.5 14.5 143.58 1878.13 0 199.3 33.2
M-5-0 275.5 14.5 143.58 1878.13 0 199.3 39.0
M-5-0 275.5 14.5 143.58 1878.13 0 199.3 34.0
M-5-0 275.5 14.5 143.58 1878.13 0 199.3 35.8
M-5-0 275.5 14.5 143.58 1878.13 0 199.3 35.3
M-5-0 275.5 14.5 143.58 1878.13 0 199.3 36.1
M-10-0 261 29 143.58 1874.89 0 198.9 36.2
M-10-0 261 29 143.58 1874.89 0 198.9 37.5
M-10-0 261 29 143.58 1874.89 0 198.9 35.4
M-10-0 261 29 143.58 1874.89 0 198.9 38.0
M-10-0 261 29 143.58 1874.89 0 198.9 37.4
M-10-0 261 29 143.58 1874.89 0 198.9 34.0
M-10-0 261 29 143.58 1874.89 0 198.9 32.8
M-10-0 261 29 143.58 1874.89 0 198.9 39.6
M-10-0 261 29 143.58 1874.89 0 198.9 38.9
M-10-0 261 29 143.58 1874.89 0 198.9 37.0
M-10-0 261 29 143.58 1874.89 0 198.9 38.4
M-10-0 261 29 143.58 1874.89 0 198.9 39.8
M-15-0 246.5 43.5 143.58 1871.65 0 198.6 30.8
M-15-0 246.5 43.5 143.58 1871.65 0 198.6 33.0
M-15-0 246.5 43.5 143.58 1871.65 0 198.6 29.5
M-15-0 246.5 43.5 143.58 1871.65 0 198.6 31.0
M-15-0 246.5 43.5 143.58 1871.65 0 198.6 31.0
M-15-0 246.5 43.5 143.58 1871.65 0 198.6 30.9
M-15-0 246.5 43.5 143.58 1871.65 0 198.6 26.4
M-15-0 246.5 43.5 143.58 1871.65 0 198.6 31.8
M-15-0 246.5 43.5 143.58 1871.65 0 198.6 30.9
M-15-0 246.5 43.5 143.58 1871.65 0 198.6 29.8
M-15-0 246.5 43.5 143.58 1871.65 0 198.6 29.8
M-15-0 246.5 43.5 143.58 1871.65 0 198.6 27.3
M-20-0 232 58 143.58 1868.42 0 198.3 26.1
M-20-0 232 58 143.58 1868.42 0 198.3 28.9
M-20-0 232 58 143.58 1868.42 0 198.3 28.6
M-20-0 232 58 143.58 1868.42 0 198.3 24.4
M-20-0 232 58 143.58 1868.42 0 198.3 27.9
M-20-0 232 58 143.58 1868.42 0 198.3 28.9
M-20-0 232 58 143.58 1868.42 0 198.3 26.4
M-20-0 232 58 143.58 1868.42 0 198.3 30.7
M-20-0 232 58 143.58 1868.42 0 198.3 24.8
M-20-0 232 58 143.58 1868.42 0 198.3 22.8
M-20-0 232 58 143.58 1868.42 0 198.3 27.6
M-20-0 232 58 143.58 1868.42 0 198.3 25.0
M-5-50 275.5 14.5 143.58 939.065 886.505 199.3 26.5
M-5-50 275.5 14.5 143.58 939.065 886.505 199.3 24.4
M-5-50 275.5 14.5 143.58 939.065 886.505 199.3 26.4
M-5-50 275.5 14.5 143.58 939.065 886.505 199.3 26.9
M-5-50 275.5 14.5 143.58 939.065 886.505 199.3 27.2
M-5-50 275.5 14.5 143.58 939.065 886.505 199.3 28.9
M-5-50 275.5 14.5 143.58 939.065 886.505 199.3 28.6
M-5-50 275.5 14.5 143.58 939.065 886.505 199.3 26.8
M-5-50 275.5 14.5 143.58 939.065 886.505 199.3 26.1
M-5-50 275.5 14.5 143.58 939.065 886.505 199.3 27.3
M-5-50 275.5 14.5 143.58 939.065 886.505 199.3 26.9
M-5-50 275.5 14.5 143.58 939.065 886.505 199.3 27.7
M-10-50 261 29 143.58 937.45 884.98 198.9 29.3
M-10-50 261 29 143.58 937.45 884.98 198.9 30.4
M-10-50 261 29 143.58 937.45 884.98 198.9 29.1
M-10-50 261 29 143.58 937.45 884.98 198.9 31.4
M-10-50 261 29 143.58 937.45 884.98 198.9 30.8
M-10-50 261 29 143.58 937.45 884.98 198.9 29.5
M-10-50 261 29 143.58 937.45 884.98 198.9 31.4
M-10-50 261 29 143.58 937.45 884.98 198.9 31.2
M-10-50 261 29 143.58 937.45 884.98 198.9 28.3
M-10-50 261 29 143.58 937.45 884.98 198.9 29.6
M-10-50 261 29 143.58 937.45 884.98 198.9 30.2
M-10-50 261 29 143.58 937.45 884.98 198.9 30.2
M-15-50 246.5 43.5 143.58 935.83 883.45 198.6 24.9
M-15-50 246.5 43.5 143.58 935.83 883.45 198.6 24.7
M-15-50 246.5 43.5 143.58 935.83 883.45 198.6 24.9
M-15-50 246.5 43.5 143.58 935.83 883.45 198.6 24.8
M-15-50 246.5 43.5 143.58 935.83 883.45 198.6 25.9
M-15-50 246.5 43.5 143.58 935.83 883.45 198.6 25.3
M-15-50 246.5 43.5 143.58 935.83 883.45 198.6 25.3
M-15-50 246.5 43.5 143.58 935.83 883.45 198.6 24.9
M-15-50 246.5 43.5 143.58 935.83 883.45 198.6 25.3
M-15-50 246.5 43.5 143.58 935.83 883.45 198.6 25.0
M-15-50 246.5 43.5 143.58 935.83 883.45 198.6 26.2
M-15-50 246.5 43.5 143.58 935.83 883.45 198.6 25.0
M-20-50 232 58 143.58 934.21 881.92 198.3 20.7
M-20-50 232 58 143.58 934.21 881.92 198.3 19.9
M-20-50 232 58 143.58 934.21 881.92 198.3 19.7
M-20-50 232 58 143.58 934.21 881.92 198.3 20.5
M-20-50 232 58 143.58 934.21 881.92 198.3 19.9
M-20-50 232 58 143.58 934.21 881.92 198.3 19.7
M-20-50 232 58 143.58 934.21 881.92 198.3 20.4
M-20-50 232 58 143.58 934.21 881.92 198.3 19.8
M-20-50 232 58 143.58 934.21 881.92 198.3 20.0
M-20-50 232 58 143.58 934.21 881.92 198.3 20.4

Expressive statistics of the dependent and independent variables_

Variable Unit Count Mean std. dev Minimum 25% 50% 75% Maximum
GGBS kg 156 268.16 21.56 232.00 246.5 275.5 290.0 290.0
AWA kg 156 21.85 21.56 0.0 0.00 14.5 43.5 58.0
AAS kg 156 143.59 - 143.58 143.6 143.6 143.6 143.58
NCA kg 156 1230.18 601.67 0.00 935.8 940.7 1871.7 1881.3
RCA kg 156 610.13 569.48 0.00 0.00 881.9 886.5 1776.1
FA (WFS) kg 156 199.14 0.5321 198.30 198.6 199.3 199.7 199.7
CS MPa 156 27.73 5.544 14.96 24.37 27.79 31.2 39.81

Results on Machine Learning Models Applied on Input Data with the Performance Metrics

Statistical Parameters of ML Models Multiple Linear Regression XGBoost Tuned AdaBoost Tuned Gradient Boost Regressor Voting Regressor
RMSE 1.64 1.63 1.59 1.64 1.52
MAE 1.28 1.30 1.26 1.30 1.21
MSE 2.70 2.70 2.51 2.70 2.32
R2 Value 0.83 0.91 0.86 0.88 0.90
CVmean −0.14 −0.74 −0.91 −0.79 −0.11

Mix Proportion Details for 1 m3 Geopolymer Pervious Concrete Preparations in kg_

Mix ID GGBS AWA NaOH LSS Water AS NCA RCA FA
M-0-0 290 0 6.583 44.207 92.791 143.58 1881.3 0 199.7
M-0-25 290 0 6.583 44.207 92.791 143.58 1411.03 444.01 199.7
M-0-50 290 0 6.583 44.207 92.791 143.58 940.68 888.03 199.7
M-0-75 290 0 6.583 44.207 92.791 143.58 470.34 1332.05 199.7
M-0-100 290 0 6.583 44.207 92.791 143.58 0 1776.07 199.7
M-5-0 275.5 14.5 6.583 44.207 92.791 143.58 1878.13 0 199.3
M-10-0 261 29 6.583 44.207 92.791 143.58 1874.89 0 198.9
M-15-0 246.5 43.5 6.583 44.207 92.791 143.58 1871.65 0 198.6
M-20-0 232 58 6.583 44.207 92.791 143.58 1868.42 0 198.3
M-5-50 275.5 14.5 6.583 44.207 92.791 143.58 939.065 886.505 199.3
M-10-50 261 29 6.583 44.207 92.791 143.58 937.45 884.98 198.9
M-15-50 246.5 43.5 6.583 44.207 92.791 143.58 935.83 883.45 198.6
M-20-50 232 58 6.583 44.207 92.791 143.58 934.21 881.92 198.3

Input Data after Feature Standardization_

GGB S AWA AAS NCA RCA FA
1.02 −1.02 - 0.29 −0.28 1.06
0.32 −0.32 - −0.51 0.51 0.28
−1.76 1.76 - 1.06 −1.07 −1.66
−0.37 0.37 - 1.07 −1.07 −0.49
−0.372 0.37 - −0.51 0.51 −0.49

Thematic Categorization of Selected Soft Computing Models Used in AAC/GPC Research_

Ref. Model Key Findings Attributes Future Scopes
[7] ANN Effectively predicted the strength variation due to molar concentration changes in activator solutions with R2 values over 0.96 Predicting strength with the use of 70% results for training and 30% sample results for testing Further refine ANN models to enhance predictive accuracy
[8] GEP Developed numerical models to predict GGBS-based GPC strength, demonstrating high accuracy and validation with R2 values ranging from 0.97 to 0.99 Compressive strength prediction of GGBS-based GPC with the use of 351 samples Expand GEP models to include more variables influencing GPC properties
[9] GEP Predict the compressive strength of bacteria-incorporated GPC, showing minimal error against experimental data Modeling compressive strength of bacteria-incorporated GPC Explore GEP's application in other GPC types with different admixtures
[10] RFR and GEP RFR and GEP were applied to develop empirical models predicting fly-ash GPC strength, where RFR showed better performance through statistical error checks Strength prediction of GPC using advanced soft computing methods developed through 298 datasets Compare these models against other ML techniques for broader applicability
[11] AI tools AI techniques like GP, RVM, and GPR showed high accuracies in predicting GPC strength with R2 values in the range of 0.93–0.99 AI-assisted mix-design tool for GPC Test these AI models in real-world mix-design scenarios for validation
[12] GEP GEP provided an empirical equation for GPC strength prediction using FA, showing good model accuracy and generalization capability Estimating GPC compressive strength using GEP developed through 298 datasets Enhance the GEP model by incorporating more diverse datasets
[13] ANN, RSM, and GEP Comparative analysis of ANN, RSM, and GEP showed RSM and ANN outperformed GEP in accuracy for predicting the strength of engineered GP composite (EGC) Predictive modeling of EGC compressive strength. The RSM showed 96% accuracy, whereas the ANN had 93% Improve GEP models or explore hybrid approaches for better prediction in EGC
[14] ML Ensembled ML techniques, particularly AdaBoost and random forest, outperformed individual methods in predicting GPC strength, and the R2 values of 0.90 for ensemble methods were obtained. Applying ML for strength prediction of GP composites; AdaBoost and random forest showed superior predictions Further explore the potential of ensembling techniques in predictive accuracy improvement
[15] ANN, M5P-Tree, LR, and MLR ANN model excelled in predicting the compressive strength of GGBS/FA-based GPC, showcasing its potential over other models Compressive strength prediction for GPCcompositesdeveloped through 220 datasets Enhance model reliability with broader datasets and explore real-time prediction capabilities
[16] ANN ANN models showed promise in predicting strength characteristics of AAC masonry blocks, with significant accuracy in training and validation phases Strength prediction for alkali-activated masonry blocks developed through 108 datasets Validate ANN models in diverse AAC formulations and structural applications
[17] GEP GEP demonstrated high accuracy in predicting the compressive strength of FRGC, supporting its use in optimizing concrete mixes; R2 values in the range of 0.97–0.99 indicating GEP's robust performance and reliability Predictive modeling for fiber-reinforced geopolymer concrete (FRGC)developed through 393 datasets Apply GEP in broader FRGC applications and investigate other fiber types and contents
[5] ANN, MPR, and SA-LR Utilized ANN and advanced regression techniques for predicting the performance of high-strength GPC, focusing on sustainable and cost-effective solutions Optimization of high-performance GPC mixes, with the use of 81 sample data Extend analysis to include long-term performance and durability predictions
[18] NSGA-II and BPNN Introduced a multi-objective optimization approach using NSGA-II and BPNN for geopolymer mix design, balancing mechanical, environmental, and economic factors; R2 and other statistical tests were used for validation Mix design optimization for fly ash-based GPC mixes, with the use of 896 sample data Expand optimization frameworks to incorporate additional environmental and durability criteria
[19] LR, ANN, and AdaBoost AdaBoost model showcased superior prediction accuracy with the highest R2 value for the compressive strength of FlA-based GPC compared to conventional machine learning models Enhancing predictive accuracy for FlA-based GPC strength Investigate AdaBoost's application in predicting other relevant concrete properties
[20] SVR and GWO The study applied SVR combined with GWO to predict the compressive strength of GGBFS-based geopolymer concrete, showing high accuracy and potential for optimization; R2 value for SVR-GWO was 0.95 Prediction of compressive strength for GGBFS-based GPC developed through 268 datasets Explore the integration of GWO with other predictive models for enhanced optimization and prediction
[21] LSTM Employed LSTM to forecast the compressive strength of FAGC, introducing a novel approach with optimized LSTM parameters for better prediction accuracy Compressive strength prediction in FAGC using LSTM developed using 162 datasets Further refine LSTM models and explore their application in real-time monitoring and control of GPC properties
[22] XGB and SVM The study compared XGB and SVM for predicting the slumpand strength of AAC, finding XGB to perform significantly better with higher R2 values (respective R2 values of 0.94 and 0.97 for slump and strength), providing a robust tool for AAC mix design Slump and compressive strength prediction in AAC with a total of 193 datasets Investigate the applicability of XGB in broader contexts of AAC production and other performance parameters
Lingua:
Inglese
Frequenza di pubblicazione:
4 volte all'anno
Argomenti della rivista:
Geoscienze, Geoscienze, altro, Scienze materiali, Compositi, Materiali porovati, Fisica, Meccanica e dinamica dei fluidi
Feed RSS della rivista