COMPARISON OF BINARY AND MULTIPLE LINEAR REGRESSION ANALYSIS IN CONCRETE STRENGTH PREDICTION

Nevbahar Ekin

Article Details

COMPARISON OF BINARY AND MULTIPLE LINEAR REGRESSION ANALYSIS IN CONCRETE STRENGTH PREDICTION

Authors:

Nevbahar Ekin

Publisher:

Süleyman Demirel University

Main Topics:

Yer Bilimleri ve Jeoloji Mühendisliği (Diğer)

Journal:

Mühendislik Bilimleri ve Tasarım Dergisi Vol.13 Issue 1, pp.64-77

Document Type:

Article

Publication Date:

3/2025

Document Language:

Turkish

Document ID (Org No.):

2001342863

pISSN:

1308-6693

Keywords:

Abstract:

 This study aims to estimate the compressive strength of low-medium strength reinforced concretes using non-destructive methods with Multiple Linear Regression MLR analysis. Ultrasonic and resistivity methods were utilized as non-destructive methods. Ultrasonic P and S wave velocities and apparent resistivity values of reinforced concrete obtained from these methods were used. In addition, the curing time of reinforced concrete, reinforcement diameter, weight, and density values of concrete samples were used. Separate relationships were obtained between all these parameters and the uniaxial compressive strength UCS of the concrete. In addition, the dependent variable UCS was tried to be estimated with the help of the mathematical model created with the MLR analysis. Significance levels showing which independent variables are important and effective in UCS estimation were also determined. According to the significance levels, it was understood that the reinforcement diameter, weight, and density parameters were not significant parameters because they were greater than the threshold value of 0.05. On the other hand, curing time, ultrasonic P and S wave velocities, and apparent resistivity parameters were found to be significant. The error rates were revealed by comparing the UCS values obtained from the parameters found to be significant in the MLR analysis and the UCS values determined by binary linear regressions BLR.

 This study aims to estimate the compressive strength of low-medium strength reinforced concretes using non-destructive methods with Multiple Linear Regression MLR analysis. Ultrasonic and resistivity methods were utilized as non-destructive methods. Ultrasonic P and S wave velocities and apparent resistivity values of reinforced concrete obtained from these methods were used. In addition, the curing time of reinforced concrete, reinforcement diameter, weight, and density values of concrete samples were used. Separate relationships were obtained between all these parameters and the uniaxial compressive strength UCS of the concrete. In addition, the dependent variable UCS was tried to be estimated with the help of the mathematical model created with the MLR analysis. Significance levels showing which independent variables are important and effective in UCS estimation were also determined. According to the significance levels, it was understood that the reinforcement diameter, weight, and density parameters were not significant parameters because they were greater than the threshold value of 0.05. On the other hand, curing time, ultrasonic P and S wave velocities, and apparent resistivity parameters were found to be significant. The error rates were revealed by comparing the UCS values obtained from the parameters found to be significant in the MLR analysis and the UCS values determined by binary linear regressions BLR.

 This study aims to estimate the compressive strength of low-medium strength reinforced concretes using non-destructive methods with Multiple Linear Regression MLR analysis. Ultrasonic and resistivity methods were utilized as non-destructive methods. Ultrasonic P and S wave velocities and apparent resistivity values of reinforced concrete obtained from these methods were used. In addition, the curing time of reinforced concrete, reinforcement diameter, weight, and density values of concrete samples were used. Separate relationships were obtained between all these parameters and the uniaxial compressive strength UCS of the concrete. In addition, the dependent variable UCS was tried to be estimated with the help of the mathematical model created with the MLR analysis. Significance levels showing which independent variables are important and effective in UCS estimation were also determined. According to the significance levels, it was understood that the reinforcement diameter, weight, and density parameters were not significant parameters because they were greater than the threshold value of 0.05. On the other hand, curing time, ultrasonic P and S wave velocities, and apparent resistivity parameters were found to be significant. The error rates were revealed by comparing the UCS values obtained from the parameters found to be significant in the MLR analysis and the UCS values determined by binary linear regressions BLR.

 Bu çalışma tahribatsız yöntemler kullanarak Çoklu Doğrusal Regresyon ÇDR analizi ile düşük-orta dayanımlı donatılı betonların basınç dayanımını tahmin etmeyi amaçlamıştır. Tahribatsız olarak ultrasonik ve özdirenç yöntemleri kullanılmıştır. Bu yöntemlerden elde edilen donatılı betonun ultrasonik P ve S dalga hızları ve görünür özdirenç değerlerinden yararlanılmıştır. Ayrıca donatılı betonun kür süresi, donatı çapı, beton numunelerin ağırlık ve yoğunluk değerleri de kullanılmıştır. Tüm bu parametreler ile betonun tek eksenli basınç dayanımı TEBD arasında ayrı ayrı ilişkiler elde edilmiştir. Ayrıca ÇDR analizi ile oluşturulan matematiksel model yardımıyla bağımlı değişken TEBD tahmin edilmeye çalışılmıştır. TEBD tahmininde hangi bağımsız değişkenin önemli ve etkili olduğunu gösteren anlamlılık önem düzeyleri de belirlenmiştir. Anlamlılık düzeylerine göre donatı çapı, ağırlık ve yoğunluk parametrelerinin eşik değer olan 0.05 değerinden büyük olması nedeni ile anlamlı parametreler olmadığı anlaşılmıştır. Buna karşın kür süresi, ultrasonik P ve S dalga hızları ve görünür özdirenç parametreleri anlamlı bulunmuştur. ÇDR analizinden anlamlı bulunan parametrelerden elde edilen TEBD değerleri ve İkili Doğrusal Regresyon İDR ile belirlenen TEBD değerleri karşılaştırılarak hata oranları ortaya konulmuştur.

Article Details

COMPARISON OF BINARY AND MULTIPLE LINEAR REGRESSION ANALYSIS IN CONCRETE STRENGTH PREDICTION

Authors:

Nevbahar Ekin

Publisher:

Süleyman Demirel University

Main Topics:

Yer Bilimleri ve Jeoloji Mühendisliği (Diğer)

Journal:

Mühendislik Bilimleri ve Tasarım Dergisi Vol.13 Issue 1, pp.64-77

Document Type:

Article

Publication Date:

3/2025

Document Language:

Turkish

Document ID (Org No.):

2001342863

pISSN:

1308-6693

Concrete Strength

Ultrasonic

Resistivity

Binary Linear Regression

Multiple Linear RegressionBeton Dayanım

Ultrasonik

Özdirenç

İkili Doğrusal Regresyon

Çoklu Doğrusal Regresyon

Issues

Article Details

Authors: Nevbahar Ekin

Publisher: Süleyman Demirel University

Main Topics: Yer Bilimleri ve Jeoloji Mühendisliği (Diğer)

Journal: Mühendislik Bilimleri ve Tasarım Dergisi Vol.13 Issue 1, pp.64-77

Document Type: Article

Publication Date: 3/2025

Document Language: Turkish

Document ID (Org No.): 2001342863

pISSN: 1308-6693

Issues

Authors:

Nevbahar Ekin

Publisher:

Süleyman Demirel University

Main Topics:

Yer Bilimleri ve Jeoloji Mühendisliği (Diğer)

Journal:

Mühendislik Bilimleri ve Tasarım Dergisi Vol.13 Issue 1, pp.64-77

Document Type:

Article

Publication Date:

3/2025

Document Language:

Turkish

Document ID (Org No.):

2001342863

pISSN:

1308-6693