Grzegorz Kacprzak
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Grzegorz Kacprzak has a total of 34 co-authors. Here are the top co-authors by number of shared publications:
- Mateusz Frydrych
Civil Engineering Faculty, Warsaw University of Technology, 00-637 Warszawa, Poland
6 shared publications - Semachew Molla Kassa
Faculty of Civil Engineering, Warsaw University of Technology, 00-637 Warsaw, Poland
6 shared publications - Betelhem Zewdu Wubineh
College of Engineering and Technology, Wachemo University, Hossaena P.O. Box. 667, Ethiopia
4 shared publications - Paweł Nowak
Civil Engineering Faculty, Division of Production Engineering and Construction Management, Warsaw University of Technology, 00-661 Warsaw, Poland
4 shared publications - Seweryn Bodus
Arcadis Sp. z o.o., Al. Jerozolimskie 142B, 02-305 Warsaw, Poland
4 shared publications - Claude Boutin
Laboratoire de tribologie et de dynamique des systèmes (LTDS), ENTPE, Vaulx-en-Velin, France
2 shared publications
Latest Publications
- Open Access
- 2 September 2026
- Published by Springer Nature in Scientific Reports
Explainable ensemble machine learning for reliable concrete compressive strength prediction: a validation and uncertainty assessment framework
- 2 September 2026
- Published by Springer Nature in Scientific Reports
Open AccessExplainable ensemble machine learning for reliable concrete compressive strength prediction: a validation and uncertainty assessment framework
- Open Access
- 30 June 2026
- Published by Polish Academy of Sciences Chancellery in Archives of Civil Engineering
- Vol. 72 (2)
Application of metakaolin and cement in the stabilisation of expansive soil: an experimental study
Expansive soils have the inherent property of swelling when they have adsorbed water and shrink when dried. This property of expansive soil causes damage to structures constructed on it. To minimise the impact of the expansive soil, it is necessary to improve existing soil before commencing infrastructural construction activity. Samples were collected from Debre Tabor City from three test pits and tested for identification and characterisation. The soil was improved with cement content of 3%, 5%, 8%, and 10% by the dry weight of the soil. The results show that the optimum cement content was obtained at 10% and that half of the optimum cement content was replaced by metakaolin with 5%, 8%, 11%, and 14%, respectively, 7%, 20%, 23%, 26%, and 29%. The optimum combined additive decreases the plasticity index from 64.76% to 27.5%, free swell index from 134% to 45%, maximum dry density from 1.33 g/cm3 to 1.36 g/cm3, California bearing ratio value from 1.28% to 5.15% at 5% cement and 17% metakaolin.
- 30 June 2026
- Published by Polish Academy of Sciences Chancellery in Archives of Civil Engineering
- Vol. 72 (2)
Open AccessApplication of metakaolin and cement in the stabilisation of expansive soil: an experimental study
Expansive soils have the inherent property of swelling when they have adsorbed water and shrink when dried. This property of expansive soil causes damage to structures constructed on it. To minimise the impact of the expansive soil, it is necessary to improve existing soil before commencing infrastructural construction activity. Samples were collected from Debre Tabor City from three test pits and tested for identification and characterisation. The soil was improved with cement content of 3%, 5%, 8%, and 10% by the dry weight of the soil. The results show that the optimum cement content was obtained at 10% and that half of the optimum cement content was replaced by metakaolin with 5%, 8%, 11%, and 14%, respectively, 7%, 20%, 23%, 26%, and 29%. The optimum combined additive decreases the plasticity index from 64.76% to 27.5%, free swell index from 134% to 45%, maximum dry density from 1.33 g/cm3 to 1.36 g/cm3, California bearing ratio value from 1.28% to 5.15% at 5% cement and 17% metakaolin.
- Open Access
- 8 February 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (4) , 1700
An Artificial Intelligence-Based Data-Driven Method for Predicting Soil Shear Strength
Accurate prediction of soil shear strength is critical for safe and cost-effective geotechnical design. This study investigates the application of four machine learning (ML) models—Random Forest (RF), Decision Tree (DT), Artificial Neural Network (ANN), and Support Vector Regression (SVR)—to predict the shear strength of soils from Bahir Dar city using laboratory-obtained geotechnical data. A total of 298 soil samples and 13 geotechnical parameters were collected from depths of 0.13–35 m, encompassing both disturbed and undisturbed conditions. The dataset was divided into training (80%) and testing (20%) sets, and models were trained with optimized hyperparameters. The RF model achieved the highest accuracy (R2 = 0.9992, RMSE = 0.0983), followed by DT (R2 = 0.9974, RMSE = 0.1812). ANN and SVR showed lower predictive accuracy, with SVR demonstrating the largest maximum errors. Predicted vs. actual plots, kernel density estimates, and absolute error per sample analysis confirmed that tree-based models provide the most reliable predictions, while ANN and SVR exhibited sporadic large deviations. SHAP analysis revealed that Cohesion, Clay content, and Plasticity Index are the most influential factors in predicting shear strength. The results demonstrate that ensemble tree-based ML models offer a robust and accurate tool for geotechnical prediction, capturing complex nonlinear relationships in soil behavior.
- 8 February 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (4) , 1700
Open AccessAn Artificial Intelligence-Based Data-Driven Method for Predicting Soil Shear Strength
Accurate prediction of soil shear strength is critical for safe and cost-effective geotechnical design. This study investigates the application of four machine learning (ML) models—Random Forest (RF), Decision Tree (DT), Artificial Neural Network (ANN), and Support Vector Regression (SVR)—to predict the shear strength of soils from Bahir Dar city using laboratory-obtained geotechnical data. A total of 298 soil samples and 13 geotechnical parameters were collected from depths of 0.13–35 m, encompassing both disturbed and undisturbed conditions. The dataset was divided into training (80%) and testing (20%) sets, and models were trained with optimized hyperparameters. The RF model achieved the highest accuracy (R2 = 0.9992, RMSE = 0.0983), followed by DT (R2 = 0.9974, RMSE = 0.1812). ANN and SVR showed lower predictive accuracy, with SVR demonstrating the largest maximum errors. Predicted vs. actual plots, kernel density estimates, and absolute error per sample analysis confirmed that tree-based models provide the most reliable predictions, while ANN and SVR exhibited sporadic large deviations. SHAP analysis revealed that Cohesion, Clay content, and Plasticity Index are the most influential factors in predicting shear strength. The results demonstrate that ensemble tree-based ML models offer a robust and accurate tool for geotechnical prediction, capturing complex nonlinear relationships in soil behavior.
- Open Access
- 16 November 2025
- Published by MDPI AG in Sustainability
- Vol. 17 (22) , 10249
Sustainable Stabilization of Expansive Soils Using Metakaolin and Cement: Evaluation Through Soil–Water Characteristic Curve Analysis
The study examines the use of environmentally friendly materials, such as metakaolin and cement, in various proportions to stabilize expansive plastic soils and assess their effects on the soil–water characteristic curve (SWCC). Metakaolin, a supplementary cementitious material with a lower carbon footprint than ordinary cement, enhances soil behavior through pozzolanic reactions. The incorporation of metakaolin and cement reduced the fitting parameter “a,” linked to the air-entry value (AEV), indicating that treated soils desaturate at lower suction values due to improved aggregate formation and pore structure. With increasing stabilizer content, the SWCC shifted toward lower suction values, reflecting improved hydraulic performance and reduced moisture sensitivity. The fitting parameter “n,” representing desaturation capacity and pore size distribution, increased with stabilizer content, suggesting a more uniform and durable soil structure. Overall, using metakaolin and cement enhances expansive soils’ structural and hydraulic behavior while conserving cement and reducing CO2 emissions. Machine learning models, Random Forest (RF), Decision Tree (DT), and Artificial Neural Network (ANN), were developed to predict SWCC. The RF model achieved the best accuracy (R2 = 0.9063, adjusted R2 = 0.8631), demonstrating the reliability of ML in evaluating green soil stabilization methods.
- 16 November 2025
- Published by MDPI AG in Sustainability
- Vol. 17 (22) , 10249
Open AccessSustainable Stabilization of Expansive Soils Using Metakaolin and Cement: Evaluation Through Soil–Water Characteristic Curve Analysis
The study examines the use of environmentally friendly materials, such as metakaolin and cement, in various proportions to stabilize expansive plastic soils and assess their effects on the soil–water characteristic curve (SWCC). Metakaolin, a supplementary cementitious material with a lower carbon footprint than ordinary cement, enhances soil behavior through pozzolanic reactions. The incorporation of metakaolin and cement reduced the fitting parameter “a,” linked to the air-entry value (AEV), indicating that treated soils desaturate at lower suction values due to improved aggregate formation and pore structure. With increasing stabilizer content, the SWCC shifted toward lower suction values, reflecting improved hydraulic performance and reduced moisture sensitivity. The fitting parameter “n,” representing desaturation capacity and pore size distribution, increased with stabilizer content, suggesting a more uniform and durable soil structure. Overall, using metakaolin and cement enhances expansive soils’ structural and hydraulic behavior while conserving cement and reducing CO2 emissions. Machine learning models, Random Forest (RF), Decision Tree (DT), and Artificial Neural Network (ANN), were developed to predict SWCC. The RF model achieved the best accuracy (R2 = 0.9063, adjusted R2 = 0.8631), demonstrating the reliability of ML in evaluating green soil stabilization methods.
- Open Access
- 30 September 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (19) , 10613
Domain-Driven Teacher–Student Machine Learning Framework for Predicting Slope Stability Under Dry Conditions
Slope stability prediction is a critical task in geotechnical engineering, but machine learning (ML) models require large datasets, which are often costly and time-consuming to obtain. This study proposes a domain-driven teacher–student framework to overcome data limitations for predicting the dry factor of safety (FS dry). The teacher model, XGBoost, was trained on the original dataset to capture nonlinear relationships among key site-specific features (unit weight, cohesion, friction angle) and assign pseudo-labels to synthetic samples generated via domain-driven simulations. Six student models, random forest (RF), decision tree (DT), shallow artificial neural network (SNN), linear regression (LR), support vector regression (SVR), and K-nearest neighbors (KNN), were trained on the augmented dataset to approximate the teacher’s predictions. Models were evaluated using a train–test split and five-fold cross-validation. RF achieved the highest predictive accuracy, with an R2 of up to 0.9663 and low error metrics (MAE = 0.0233, RMSE = 0.0531), outperforming other student models. Integrating domain knowledge and synthetic data improved prediction reliability despite limited experimental datasets. The framework provides a robust and interpretable tool for slope stability assessment, supporting infrastructure safety in regions with sparse geotechnical data. Future work will expand the dataset with additional field and laboratory tests to further improve model performance.
- 30 September 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (19) , 10613
Open AccessDomain-Driven Teacher–Student Machine Learning Framework for Predicting Slope Stability Under Dry Conditions
Slope stability prediction is a critical task in geotechnical engineering, but machine learning (ML) models require large datasets, which are often costly and time-consuming to obtain. This study proposes a domain-driven teacher–student framework to overcome data limitations for predicting the dry factor of safety (FS dry). The teacher model, XGBoost, was trained on the original dataset to capture nonlinear relationships among key site-specific features (unit weight, cohesion, friction angle) and assign pseudo-labels to synthetic samples generated via domain-driven simulations. Six student models, random forest (RF), decision tree (DT), shallow artificial neural network (SNN), linear regression (LR), support vector regression (SVR), and K-nearest neighbors (KNN), were trained on the augmented dataset to approximate the teacher’s predictions. Models were evaluated using a train–test split and five-fold cross-validation. RF achieved the highest predictive accuracy, with an R2 of up to 0.9663 and low error metrics (MAE = 0.0233, RMSE = 0.0531), outperforming other student models. Integrating domain knowledge and synthetic data improved prediction reliability despite limited experimental datasets. The framework provides a robust and interpretable tool for slope stability assessment, supporting infrastructure safety in regions with sparse geotechnical data. Future work will expand the dataset with additional field and laboratory tests to further improve model performance.
- Open Access
- 3 September 2025
- Published by Polish Academy of Sciences Chancellery in Archives of Civil Engineering
- p. 527
Analysis of soil-structure interaction using 2D/3D FEM calculations in geotechnical design
Geotechnical design is a complex, interdisciplinary field involving soil science, mechanics, and rheology. This article presents a comprehensive solution for geotechnical design using digital data from various programs, focusing on Ultimate Limit State (ULS) and Serviceability Limit State (SLS) analysis for a storage hall foundation in challenging geotechnical conditions in eastern Germany. The study utilizes the groundwater conditions and layer layout from the Leapfrog software to create a spatial computational model of the subsoil using Plaxis 3D FEM based on nearly 400 virtual boreholes. Data on hall geometry and load combinations were stored in digital IFC format and tabular data in Excel files respectively. The analysis involved both analytical and FEM 3D/2D methods. In the analytical method, all digital information was collected into one Excel file database for further calculations. Once the consistent digital project database was collected, the same data was implemented into the Plaxis 3D program for FEM numerical calculations. The hall’s foundation solution used footings supported by drilled piles. Soil layer parameters were calibrated based on static pile tests during design calculations. Due to part of the hall being located in a high embankment area (about 5 meters high), consolidation analysis of cohesive soils in the area was considered. The impact of consolidation was factored into negative skin friction in the ULS pile design process. Pile geometry was tailored to existing soil conditions and weak bearing soil thickness. The final foundation geometry and pile lengths were designed to meet SLS conditions set by the structural designer and to address uneven building settlement.
- 3 September 2025
- Published by Polish Academy of Sciences Chancellery in Archives of Civil Engineering
- p. 527
Open AccessAnalysis of soil-structure interaction using 2D/3D FEM calculations in geotechnical design
Geotechnical design is a complex, interdisciplinary field involving soil science, mechanics, and rheology. This article presents a comprehensive solution for geotechnical design using digital data from various programs, focusing on Ultimate Limit State (ULS) and Serviceability Limit State (SLS) analysis for a storage hall foundation in challenging geotechnical conditions in eastern Germany. The study utilizes the groundwater conditions and layer layout from the Leapfrog software to create a spatial computational model of the subsoil using Plaxis 3D FEM based on nearly 400 virtual boreholes. Data on hall geometry and load combinations were stored in digital IFC format and tabular data in Excel files respectively. The analysis involved both analytical and FEM 3D/2D methods. In the analytical method, all digital information was collected into one Excel file database for further calculations. Once the consistent digital project database was collected, the same data was implemented into the Plaxis 3D program for FEM numerical calculations. The hall’s foundation solution used footings supported by drilled piles. Soil layer parameters were calibrated based on static pile tests during design calculations. Due to part of the hall being located in a high embankment area (about 5 meters high), consolidation analysis of cohesive soils in the area was considered. The impact of consolidation was factored into negative skin friction in the ULS pile design process. Pile geometry was tailored to existing soil conditions and weak bearing soil thickness. The final foundation geometry and pile lengths were designed to meet SLS conditions set by the structural designer and to address uneven building settlement.
- Open Access
- 17 June 2025
- Published by Polish Academy of Sciences Chancellery in Archives of Civil Engineering
- p. 503-519
Model studies and quantitative evaluation of the reduction of gyttja consolidation settlement by different types of piled raft foundations
Combined Pile-Raft Foundations (CPRF) are widely employed to mitigate vertical settlement in building and engineering structures. For structures resting on cohesive and organic soils, understanding the time required for complete stabilization of settlement is crucial. The total settlement involves instantaneous settlement (elastic ground deformation), consolidation settlement (water squeezing from pore space), and secondary settlement (structural changes in the ground skeleton, known as secondary consolidation or soil creep). Urban expansion, notably in cities like Warsaw, compels developers to construct in previously overlooked areas, potentially containing organic carbonate sediments like gyttja and chalk. Buildings on such organic soils often settle due to gyttja consolidation during construction and operation. Analyzing long-term settlement, especially of CPRF on reconsolidated organic soils, becomes paramount. Model tests on a laboratory scale offer a cost-effective alternative to large-scale tests, providing quantitative insights into CPRF settlement reduction through piles. This study presents results from model tests conducted on natural organic soils, enabling the prediction of CPRF settlement solely based on gyttja’s geotechnical parameters.
- 17 June 2025
- Published by Polish Academy of Sciences Chancellery in Archives of Civil Engineering
- p. 503-519
Open AccessModel studies and quantitative evaluation of the reduction of gyttja consolidation settlement by different types of piled raft foundations
Combined Pile-Raft Foundations (CPRF) are widely employed to mitigate vertical settlement in building and engineering structures. For structures resting on cohesive and organic soils, understanding the time required for complete stabilization of settlement is crucial. The total settlement involves instantaneous settlement (elastic ground deformation), consolidation settlement (water squeezing from pore space), and secondary settlement (structural changes in the ground skeleton, known as secondary consolidation or soil creep). Urban expansion, notably in cities like Warsaw, compels developers to construct in previously overlooked areas, potentially containing organic carbonate sediments like gyttja and chalk. Buildings on such organic soils often settle due to gyttja consolidation during construction and operation. Analyzing long-term settlement, especially of CPRF on reconsolidated organic soils, becomes paramount. Model tests on a laboratory scale offer a cost-effective alternative to large-scale tests, providing quantitative insights into CPRF settlement reduction through piles. This study presents results from model tests conducted on natural organic soils, enabling the prediction of CPRF settlement solely based on gyttja’s geotechnical parameters.
- Open Access
Assessment of the Interaction of the Combined Piled Raft Foundation Elements Based on Long-Term Measurements
Understanding the complex phenomena of interactions between the elements of a combined piled raft foundation (CPRF) is essential for the proper design of such foundations. To evaluate the effects of mutual influence among the CPRF’s elements, a series of long-term measurements of selected physical quantities related to the performance of the foundation were conducted on a building with a frame structure, stiffening walls, and monolithic technology, consisting of seven aboveground stories and one underground story. The analysis distinguishes the real deformations resulting from temperature changes and from stress strains resulting from load changes. The two types of deformations were subjected to further interpretation of only changes in the stress and strain over time. Changes in stress values in the subsoil, as well as strain measurements in the vertical direction of concrete columns, were recorded to assess the load distribution between the CPRF’s components. The numerical analysis results obtained for a fragment of the monitored foundation were compared with actual measurement results to verify the numerical model of interaction between the structure and the soil. Field monitoring and FEA methods were used to compare the long-term deformation analysis, and they helped to minimize the monitoring time. This comparison also served to supplement and simultaneously expand the dataset of test results on a real-world scale.
Open AccessAssessment of the Interaction of the Combined Piled Raft Foundation Elements Based on Long-Term Measurements
Understanding the complex phenomena of interactions between the elements of a combined piled raft foundation (CPRF) is essential for the proper design of such foundations. To evaluate the effects of mutual influence among the CPRF’s elements, a series of long-term measurements of selected physical quantities related to the performance of the foundation were conducted on a building with a frame structure, stiffening walls, and monolithic technology, consisting of seven aboveground stories and one underground story. The analysis distinguishes the real deformations resulting from temperature changes and from stress strains resulting from load changes. The two types of deformations were subjected to further interpretation of only changes in the stress and strain over time. Changes in stress values in the subsoil, as well as strain measurements in the vertical direction of concrete columns, were recorded to assess the load distribution between the CPRF’s components. The numerical analysis results obtained for a fragment of the monitored foundation were compared with actual measurement results to verify the numerical model of interaction between the structure and the soil. Field monitoring and FEA methods were used to compare the long-term deformation analysis, and they helped to minimize the monitoring time. This comparison also served to supplement and simultaneously expand the dataset of test results on a real-world scale.
- Open Access
- 31 March 2025
- Published by MDPI AG in Preprints.org
Assessment of the Interaction of the Combined Piled Raft Foundation Elements Based on Long-Term Measurements
Understanding the complex phenomena of interactions between the elements of a Combined Piled Raft Foundation (CPRF) is essential for the proper design of such foundations. To evaluate the effects of mutual influence among CPRF elements, a series of long-term measurements of selected physical quantities related to the performance of the foundation were conducted on a building with a frame structure, stiffening walls, and monolithic technology, consisting of seven above-ground stories and one underground story. The analysis distinguishes the real deformations resulting from temperature changes and from stress strains resulting from load changes. The two types of deformations were subject to further interpretation of only changes in stress strain over time. Changes in stress values in the subsoil, as well as strain measurements in the vertical direction of concrete columns, were recorded to assess the load distribution between the CPRF components. The numerical analysis results obtained for a fragment of the monitored foundation were compared with actual measurement results to verify the numerical model of interaction between the structure and the soil. This comparison also served to supplement and sim-ultaneously expand the dataset of test results on a real-world scale.
- 31 March 2025
- Published by MDPI AG in Preprints.org
Open AccessAssessment of the Interaction of the Combined Piled Raft Foundation Elements Based on Long-Term Measurements
Understanding the complex phenomena of interactions between the elements of a Combined Piled Raft Foundation (CPRF) is essential for the proper design of such foundations. To evaluate the effects of mutual influence among CPRF elements, a series of long-term measurements of selected physical quantities related to the performance of the foundation were conducted on a building with a frame structure, stiffening walls, and monolithic technology, consisting of seven above-ground stories and one underground story. The analysis distinguishes the real deformations resulting from temperature changes and from stress strains resulting from load changes. The two types of deformations were subject to further interpretation of only changes in stress strain over time. Changes in stress values in the subsoil, as well as strain measurements in the vertical direction of concrete columns, were recorded to assess the load distribution between the CPRF components. The numerical analysis results obtained for a fragment of the monitored foundation were compared with actual measurement results to verify the numerical model of interaction between the structure and the soil. This comparison also served to supplement and sim-ultaneously expand the dataset of test results on a real-world scale.
- Open Access
- 20 November 2023
- Published by MDPI AG in Construction Materials
- Vol. 3 (4) , 414-433
Soil Consolidation Analysis in the Context of Intermediate Foundation as a New Material Perspective in the Calibration of Numerical–Material Models
This paper presents the authors’ research results from an analysis of intermediate foundations as well as slab and pile foundations in the context of soil consolidation. Looking at soil as a building material that changes its properties over time is very important from the point of view of the safety of construction, implementation, and operation of building structures. In addition, soil can be parameterized in such a way as to accurately describe its possible behavior under service loading. Of great interest is the phenomenon of consolidation, which is based on the reduction of soil volume over time under constant loading. This study explores existing piles and replicates soil conditions to understand individual and grouped pile behavior in combined pile–raft foundations (CPRF). To assess pile settlement from primary and secondary consolidation phases, 13 field measurements on concrete columns in gyttja clay were conducted. Analyzing data from these tests allowed engineers to accurately calibrate a numerical model. This calibrated model was instrumental in designing high-rise buildings, ensuring stability and safety. This study emphasizes the importance of understanding soil behavior, particularly consolidation phenomena, in optimizing foundation design and construction practices.
- 20 November 2023
- Published by MDPI AG in Construction Materials
- Vol. 3 (4) , 414-433
Open AccessSoil Consolidation Analysis in the Context of Intermediate Foundation as a New Material Perspective in the Calibration of Numerical–Material Models
This paper presents the authors’ research results from an analysis of intermediate foundations as well as slab and pile foundations in the context of soil consolidation. Looking at soil as a building material that changes its properties over time is very important from the point of view of the safety of construction, implementation, and operation of building structures. In addition, soil can be parameterized in such a way as to accurately describe its possible behavior under service loading. Of great interest is the phenomenon of consolidation, which is based on the reduction of soil volume over time under constant loading. This study explores existing piles and replicates soil conditions to understand individual and grouped pile behavior in combined pile–raft foundations (CPRF). To assess pile settlement from primary and secondary consolidation phases, 13 field measurements on concrete columns in gyttja clay were conducted. Analyzing data from these tests allowed engineers to accurately calibrate a numerical model. This calibrated model was instrumental in designing high-rise buildings, ensuring stability and safety. This study emphasizes the importance of understanding soil behavior, particularly consolidation phenomena, in optimizing foundation design and construction practices.