Multivariate Adaptive Regression Splines for CompressiveStrength of Self-Compacting Concrete Modeling and Prediction

Publish Year: 1396
نوع سند: مقاله کنفرانسی
زبان: English
View: 477

This Paper With 9 Page And PDF Format Ready To Download

  • Certificate
  • من نویسنده این مقاله هستم

استخراج به نرم افزارهای پژوهشی:

لینک ثابت به این Paper:

شناسه ملی سند علمی:

IRCIVIL02_035

تاریخ نمایه سازی: 22 دی 1396

Abstract:

Concrete is one of the most important materials in construction. In recent years, researchers have conducted various investigations on different types of concretes [1]. Introduction of self-compacting concrete has brought huge technological advances. Use of SCC facilitated the concrete placing between the rebars, without need of external vibration, and just through the weight of concrete itself. Utilizing self-compacting concrete results in reducing construction time and cost in addition to reducing the noise in construction sites [2]. Concrete workability is an important factor for proper execution, which after widespread application of reinforcing bars in concrete in the beginning of the 20th century and necessity of utilizing high workability concrete, it was maintained for a long time by addition of water to the cement. But in latter research it was found that use of high amounts of the water and cement would bring about negative results [3]. In selfcompacting concrete, super plasticizers and binder materials are important to achieve high workability and proper viscosity while eliminating the separation, and some solutions for optimal mix design of concrete like reducing the aggregate to cement materials ratio, increase in the amount of cement- paste with a certain water to cement ratio, and control of the largest coarse aggregate size have been proposed [4]. The volume of binder materials used in self-compacting concrete, in comparison to conventional concretes, is higher and this indicates the importance of utilizing proper type of the material and weight combination of these materials to provide higher durability and strength of concrete and also its corresponding effects like reduced generation of pollutant gazes during cement production and participation in the sustainable development [5]. With respect to this issue that consumption of high amounts of cement and super plasticizers requires huge expenses, utilization some alternative supplementary cementitious materials (SCMs) like metakaolin as a replacement for Portland cement has been in consideration. The environmental concerns over extraction of raw materials and emission of CO2 during cement production; urge us to reduce the amount of consumed cement by application of additives. Utilizing metakaolin increases the concrete strength and durability against chemical attacks, alkali silica reaction and freeze-thaw cycles. Metakaolin also is effective in some mechanical properties of concrete including compressive strength, early age and flexural strength [6-12]. The wide range of materials andsubstances used in this type of concrete and complexity of its corresponding mix design which is affected by various parameters, also difficulty in finding existing relationships between these parameters have made it necessary to present a model for mix design of the self-compacting concrete incorporated metakaolin. Today, utilizing the artificial intelligence methods for modeling and predicting problems in civil engineering hasbecome widespread due to their advantages. The experience and studies of the researchers have revealed that in addition to various experimental research works, use of various artificial intelligence methods in investigating and predicting the fresh and hardened properties of the concrete has become a necessity [13-18]. There is limited research literature concerning modeling of the metakaolin contained self-compacting concrete.Safarzadegan Gilan et al. utilized the SVR-PSO hybrid method and also ANFIS method for predicting the compressive strength and RCPT test of 25 samples of metakaolin contained self-compacting concrete at ages of 7, 28, 90 and 180 days. The results exhibited the desired and highly accurate performances of both models [19]. Sarıdemir modeled 179 mortar samples containing metakaolin utilizing the two artificial intelligencemethods. The results of the statistical error indices in 60 test samples and 119 trained data, reported significant precision and validity of both methods in predicting the compressive strength [20]. In this study, the effect of largest grain size in the mix design (Dmax) also was investigated for the first time as an input in the modeling of concrete and predicting behavior of the self-compacting concrete. The main goal in this paper is utilizing the novel methods (MARS) in the field of modeling and predicting the concrete properties. Hence, with respect to the limited artificial intelligence research conducted on the metakaolin contained SCC, this type of concrete was selected for development of the proposed models.

Authors

Ali Ashrafian

Department of Civil Engineering, TabariUniversity, Babol, Iran

Javad berenjian

Department of Civil Engineering, Babol Noshirvani University of Technology, Babol, Iran

Mohammad Javad Taheri Amiri

Department of Civil Engineering, Babol Noshirvani University of Technology, Babol, Iran