Yazar "Tanyildizi, Harun" seçeneğine göre listele

Listeleniyor 1 - 8 / 8
  • Küçük Resim Yok
    Öğe
    Comparison of extreme learning machine and deep learning model in the estimation of the fresh properties of hybrid fiber-reinforced SCC
    (Springer London Ltd, 2021) Kina, Ceren; Turk, Kazim; Atalay, Esma; Donmez, Izzeddin; Tanyildizi, Harun
    This paper studied the estimation of fresh properties of hybrid fiber-reinforced self-compacting concrete (HR-SCC) mixtures with different types and combinations of fibers by using two different prediction method named as the methodologies of extreme learning machine and long short-term memory (LSTM). For this purpose, 48 mixtures, which were designed as single, binary, ternary and quaternary fiber-reinforced SCC with macro-steel fiber, two micro-steel fibers having different aspect ratio, polypropylene (PP) and polyvinylalcohol (PVA), were used. Slump flow, t(50) and J-ring tests for designed mixtures were conducted to measure the fresh properties of fiber-reinforced SCC mixtures as per EFNARC. The experimental results were analyzed by Anova method. In the devised prediction model, the amounts of cement, fly ash, silica fume, blast furnace slag, limestone powder, aggregate, water, high-range water-reducer admixture (HRWA) and the fiber ratios were selected as inputs, while the slump flow, t(50) and the J-ring were selected as outputs. Based on the Anova analysis' results, the macro-steel fiber was the most important parameter for the results of slump-flow diameter and t(50), while the most important parameter for the results of J-ring was fly ash. Furthermore, it was found that the use of more than 0.20% by volume of 6/0.16 micro-steel fiber positively influenced the fresh properties of SCC mixtures with hybrid fiber. On the other hand, the inclusion of steel fiber instead of synthetic fiber into SCC mixture as micro-fiber was more advantageous in terms of workability of mixtures as result of hydrophobic nature of steel fibers. This study found that extreme learning machine model estimated the slump flow, t(50) and J-ring with 99.71%, 81% and 94.21% accuracy, respectively, while deep learning model found the same experimental results with 99.18%, 77.4% and 84.8% accuracy, respectively. It can be emphasized from this study that the extreme learning machine model had a better prediction ability than the deep learning model.
  • Küçük Resim Yok
    Öğe
    Deep learning and machine learning-based prediction of capillary water absorption of hybrid fiber reinforced self-compacting concrete
    (Ernst & Sohn, 2022) Kina, Ceren; Turk, Kazim; Tanyildizi, Harun
    Deep auto-encoders and long short-term memory methodology (LSTM) based on deep learning as well as support vector regression (SVR) and k-nearest neighbors (kNN) based on machine learning models for the capillary water absorption prediction of self-compacting concrete (SCC) with single and binary, ternary, and quaternary fiber hybridization were developed. A macro and two types of micro steel fibers having different aspect ratios, and PVA fiber were used. One hundred and sixty-eight specimens produced from 24 mixtures were used in the prediction models. The input was the content of cement, fly ash, silica fume, fine and coarse aggregate, water, superplasticizer (SP), macro and micro steel fibers, PVA, time that the specimen was immersed in water, and splitting tensile strength. Water absorption was used as output. As per the ANOVA analysis of the experiment results, the most effective parameters were macro steel fiber and time for tensile strength and water absorption, respectively. Finally, binary hybridization of 1% macro steel fiber and PVA improved the splitting tensile strength while the use of PVA as binary, ternary, and quaternary fiber hybridization increased the water absorption of SCC specimens. The auto-encoder, LSTM, SVR, and kNN models predicted the water absorption of fiber reinforced SCC with 99.99%, 99.80%, 94.57%, and 95.50% accuracy, respectively. The performance of deep autoencoder in the estimation of water absorption of fiber-reinforced SCC was superior to the other prediction models.
  • Küçük Resim Yok
    Öğe
    Estimation of strengths of hybrid FR-SCC by using deep-learning and support vector regression models
    (Ernst & Sohn, 2022) Kina, Ceren; Turk, Kazim; Tanyildizi, Harun
    In this work, to estimate the compressive, splitting tensile, and flexural strength of self-compacting concrete (SCC) having single fiber and binary, ternary, and quaternary fiber hybridization, the deep-learning (DL) and support vector regression (SVR) models were devised. The fiber content and coarse aggregate/total aggregate ratio (CA/TA) were the variables for 24 designed mixtures. Four different fibers, which were a macro steel fiber, two types of micro steel fibers with different aspect ratio, and polyvinyl alcohol (PVA) fiber, were used in SCC mixtures. The specimens of each mixture were tested to measure the engineering properties for 7, 28, and 90 days. The amount of cement, fly ash, fine aggregate, CA, high-range water-reducing admixture, water, macro steel fiber, PVA fiber, two types of micro steel fibers, and curing time were selected as input layers while the output layers were strength results. The experimental results were compared with the estimation results. The engineering properties were estimated using the SVR model with 95.25%, 87.81%, and 93.89% accuracy, respectively. Furthermore, the DL model estimated compressive strength, tensile strength, and flexural strength with 99.27%, 98.59%, and 99.15% accuracy, respectively. It was found that the DL estimated the engineering properties of hybrid fiber-reinforced SCC with higher accuracy than SVR.
  • Küçük Resim Yok
    Öğe
    Extreme Learning Machine for Estimation of the Engineering Properties of Self-Compacting Mortar with High-Volume Mineral Admixtures
    (Springer Int Publ Ag, 2024) Turk, Kazim; Kina, Ceren; Tanyildizi, Harun
    The utilization of supplementary cementitious materials obtained from industrial by-products or wastes is one of the most effective ways to minimize the costs as well as environmental impact associated with cement production. This work investigated the effects of the replacement of Portland cement (PC) with (25, 30, 35 and 40%) fly ash (FA) and (5, 10, 15, and 20%) silica fume (SF) by weight as binary and ternary blends on the compressive strength (f(c)) and flexural strength (f(ft)) of self-compacting mortars (SCMs) at 28 and 91 curing days. Extreme learning machine (ELM), support vector regression (SVR), artificial neural network (ANN), and decision tree (DT) models were devised to predict these strengths of SCMs containing high-volume mineral admixture (HVMA). The selected input variables were the number of curing days, water-cementitious material (W/CM), PC, FA, SF, and sand contents, while the f(c) and f(ft) were the output variables. ANOVA results show that the curing time was the most effective parameter for determining both strengths. The results also indicated that ELM achieved superior performance for the prediction of f(c) and f(ft) of SCMs with HVMA compared to SVR, ANN, and DT due to having the highest coefficient of determination values of 0.9802 for both strengths.
  • Küçük Resim Yok
    Öğe
    Forecasting the compressive strength of GGBFS-based geopolymer concrete via ensemble predictive models
    (Elsevier Sci Ltd, 2023) Kina, Ceren; Tanyildizi, Harun; Turk, Kazim
    The compressive strength (fc) of the concrete is an important parameter in the structural design. However, the assessment of fc via an experimental program is time-consuming, costly, and needs a labor force. Therefore, the forecasting of fc through different algorithms can accelerate and facilitate this process and also provide guidance for scheduling the progress of the construction. While some studies have explored the use of models for the prediction of fc of concrete, the ensemble models that can predict the fc of GPC with industrial by-products is still lacking. Within this scope, decision tree (DT), Bootstrap aggregating (Bagging), and Least-squares boosting (LSBoost) models were devised to predict fc of ground granulated blast furnace slag (GGBFS)-based geopolymer concrete (GPC). The data points collected to devise a GEP model in the previous study were used and the prediction results of the GEP model were compared with the proposed ensemble models in the current study. The age of the specimen, NaOH solution concentration, natural zeolite (NZ) content, silica fume (SF) content, and GGBFS content were used as input parameters, and fc was used as output parameter. According to ANOVA analysis, the age of the specimen was found as the most influential parameter in the determination of the fc of GGBFS-based GPC. Also, Multiple linear regression equation was proposed to estimate the fc of GGBFS-based GPC with the accuracy of 93%. The most accurate model was introduced through performance metrics and the Taylor diagram. The results proved that the highest accuracy and stable predictions were achieved by the LSBoost model with R-squared value of 98.25% followed by GEP model developed in the previous study, DT and Bagging models. However, it is worth mentioning that due to having a high coefficient of correlation values (>%80), DT and Bagging models also have an acceptable ability for predicting fc of GGBS-based GPC.
  • Küçük Resim Yok
    Öğe
    Hybrid deep learning model for concrete incorporating microencapsulated phase change materials
    (Elsevier Sci Ltd, 2022) Tanyildizi, Harun; Marani, Afshin; Turk, Kazim; Nehdi, Moncef L.
    The inclusion of microencapsulated phase change materials (MPCMs) in concrete promotes thermal energy storage, thus enhancing sustainable design. Notwithstanding this advantage, the compressive strength of concrete dramatically decreases upon MPCM addition. While several experimental studies have explored the origin of this compressive strength reduction, a reliable and practical framework for the prediction of the compressive strength of MPCM-integrated concrete is yet to be developed. The current research proposes a deep learning approach to estimate the compressive strength of MPCM-integrated cementitious composites based on its mixture proportions and the thermophysical properties of PCM. Extreme learning machines (ELMs), autoencoders, hybrid ELM-autoencoder, and extreme gradient boosting (XGBoost) models were purposefully developed using the largest pertinent experimental dataset available to date encompassing 244 mixture design examples retrieved from the open literature. The results demonstrate the capability of the hybrid deep learning and XGBoost models in accurately modeling the compressive strength of PCM integrated concrete with favorably low prediction error. Furthermore, a sensitivity analysis identified the most influential parameters on the compressive strength development to assist the mixture design of concrete incorporating MPCM.
  • Küçük Resim Yok
    Öğe
    Machine Learning Prediction of Residual Mechanical Strength of Hybrid-Fiber-Reinforced Self-consolidating Concrete Exposed to Elevated Temperature
    (Springer, 2023) Turk, Kazim; Kina, Ceren; Tanyildizi, Harun; Balalan, Esma; Nehdi, Moncef L. L.
    Establishing the engineering properties of cement-based composites at elevated temperature requires costly, laborious, and time-consuming experimental work. Data-driven models can provide a robust and efficient alternative. In this study, extreme learning machine (ELM), support vector machine (SVM), artificial neural network (ANN), and decision tree (DT) models were trained to predict the residual compressive, splitting tensile, and flexural strengths of hybrid fiber-reinforced self-compacting concrete (HFR-SCC) exposed to high temperatures. Mixtures including macro and micro steel fibers, polyvinyl alcohol (PVA), and polypropylene (PP) were subjected to different temperature levels, leading to an experimental database of 360 specimens. Eleven input parameters including cement, fly ash, water, sand, gravel, fiber type, water reducer, and temperature were deployed. The residual mechanical strengths were targeted as output parameters. ANOVA was used to explore the influence of input parameters. Temperature was found to be the most influential parameter. Dataset consisting of 114 instances was retrieved from pertinent literature and used along with the authors' experimentally generated dataset for residual strength prediction. The experimental results were compared with predictions of ELM, SVM, ANN, and DT. ELM achieved superior performance and can offer a robust tool for predicting the residual mechanical strengths of HFR-SCC upon exposure to high temperature.
  • Küçük Resim Yok
    Öğe
    Predicting sorptivity and freeze-thaw resistance of self-compacting mortar by using deep learning and k-nearest neighbor
    (Techno-Press, 2022) Turk, Kazi; Kina, Ceren; Tanyildizi, Harun
    In this study, deep learning and k-Nearest Neighbor (kNN) models were used to estimate the sorptivity and freeze -thaw resistance of self-compacting mortars (SCMs) having binary and ternary blends of mineral admixtures. Twenty-five environment-friendly SCMs were designed as binary and ternary blends of fly ash (FA) and silica fume (SF) except for control mixture with only Portland cement (PC). The capillary water absorption and freeze-thaw resistance tests were conducted for 91 days. It was found that the use of SF with FA as ternary blends reduced sorptivity coefficient values compared to the use of FA as binary blends while the presence of FA with SF improved freeze-thaw resistance of SCMs with ternary blends. The input variables used the models for the estimation of sorptivity were defined as PC content, SF content, FA content, sand content, HRWRA, water/cementitious materials (W/C) and freeze-thaw cycles. The input variables used the models for the estimation of sorptivity were selected as PC content, SF content, FA content, sand content, HRWRA, W/C and predefined intervals of the sample in water. The deep learning and k-NN models estimated the durability factor of SCM with 94.43% and 92.55% accuracy and the sorptivity of SCM was estimated with 97.87% and 86.14% accuracy, respectively. This study found that deep learning model estimated the sorptivity and durability factor of SCMs having binary and ternary blends of mineral admixtures higher accuracy than k-NN model.