Please use this identifier to cite or link to this item:
http://172.22.28.37:8080/xmlui/handle/1/226| Title: | Modelling Compressive Strength of Recycled Aggregate Concrete Using Neural Networks and Regression |
| Authors: | Deshpande Neela, Kulkarni S.S. Londhe Shreenivas |
| Keywords: | Recycled concrete aggregates Artificial Neural Networks Non-linear regression 28 days Compressive strength input parameters and non-dimensional parameters |
| Issue Date: | Jun-2013 |
| Publisher: | ISSR Journals (Concrete Research Letters) |
| Abstract: | Recycled aggregates are used in concrete and the concrete is called as Recycled Aggregate Concrete (RAC). This paper aims at predicting the 28 day compressive strength of RAC using two techniques namely Artificial Neural Network (ANN) and Non-linear regression (NLR). Five ANN and NLR models were developed with input parameters as per cubic proportions of cement, sand, natural coarse aggregate, recycled coarse aggregate, water , admixtures used in the mix designs and non-dimensional parameters sand aggregate ratio, water to total materials ratio and the replacement ratio of recycled coarse aggregates to natural aggregates(by volume) in concrete. The effects of each parameter on networks in both techniques were studied. Comparing the techniques shows that ANN performs better than NLR equations. With limited amount of input parameters also, ANN1 predicted the strength of RAC better as compared to NLR1. The performance of ANN models and NLR models improved with the use of non-dimensional parameters. |
| Description: | Construction and demolition waste (C&D) is increasing day by day and is a cause of concern owing to its harmful effect on the surroundings. A possible solution to these problems is reuse of these C&D waste in concrete. Recycled aggregates (RA) which can be used in concrete as aggregates is a material derived from waste concrete is generally produced by two stage crushing of demolished concrete, screening and removal of contaminants such as reinforcement, wood,plastic etc [1]. Many studies were carried out which prove that concrete made with RA’s can have mechanical properties similar to those of conventional concretes and even high strength concrete is nowadays possible [2, 3]. Recycled concrete aggregates (RCA) has mortar and old cement paste attached to it and accounts for 20 to 30% of total volume of aggregates, which increase water absorption capacity and reduces density and becomes governing criteria for the compressive strength of concrete with recycled aggregates [2,4-6]. In a study done using different recycled aggregates (RA) replacement ratios, w/c ratios and RAs with different strengths and different moisture conditions, concluded that the strength of (RAC) was about 10–25% lower than that of Natural aggregate concrete (NAC) and thus 100% replacement of RCA tends to lower the strength of concrete [2,8]. The study also concluded that the compressive or tensile strength loss of RAC prepared with low strength RA was more significant than that of concrete prepared with high strength RA, and the extent of the reduction was related to many parameters, such as the type of concrete used for making the RA (high, medium or low strength), replacement ratios, water-cement ratios and the moisture conditions of the RA [2]. According to another study made [7] the strength characteristics of recycled aggregate concrete are not affected when w/c ratio is higher. Due to this diverse behaviour of RAC the task of predicting 28th day Compressive strength of RAC becomes tedious and requires both extensive testing and time. A number of efforts were done on using multivariable regression models to improve the accuracy of predictions. In a study, relationships among demolished concrete characteristics, properties of their RA and strength of their RAC were done using regression analysis [9]. An attempt was also made to predict compressive strength of concrete using multiple regressions [10]. The conventional methods for predicting the compressive strength of concrete are based on statistical analysis and involve estimating and the choice of an appropriate regression equation. [11, 17]. ANN has also been a popular technique in predicting the compressive strength of concrete. ANN models were developed to predict the strength and slump of ready mixed concrete and high strength concrete, in which chemical admixtures and or mineral additives were used [12]. ANN is also used to predict the slump flow of concrete [13]. Particularly in the field of RCA, ANN was used to predict strength of recycled aggregate concrete [14]. |
| URI: | http://localhost:8080/xmlui/handle/1/226 |
| Appears in Collections: | Faculty Publication |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| Modelling Compressive Strength of Recycled Aggregate Concrete Using Neural Networks and Regression.pdf | 402.32 kB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.