Scholarlink

<p>In the present study two different waste plastic materials such as, High Density Polyethylene (HD-PE) and Polypropylene (PP) were collected from dumping yards and was partially replaced with fine aggregate (FA) by 2.5% increment in each mix till 10% replacement. The waste plastic was partially replaced to prepare M30 and M40 grade concrete to find the bending/flexural property. The bending stress of HDPE material initially (7 &amp; 14&#xa0;days) showed better results to that of PP material for M30 grade concrete. The maximum bending stress of 6.1&#xa0;Mpa and 5.6&#xa0;Mpa for 7.5% and 5% of HDPE and PP was observed respectively for M30 grade concrete. The same pattern to that of M30 grade concrete was even observed in M40 grade concrete. Here also the maximum bending stress of 6.99&#xa0;Mpa and 6.0&#xa0;Mpa for 7.5% and 5% of HDPE and PP was observed respectively for M40 grade concrete. The optimum percentage of 5% and 7.5 for PP and HDPE material respectively can be replaced in concrete for bending/flexural strength for both M30 and M40 grade. The same experimental values of M30 &amp; M40 grade concrete were used to statically predict the bending stress value using response surface method (RSM) in minitab and compared with artificial neural network (ANN) in matlab. Regression models are evaluated using ANOVA, Pareto charts, and second-order polynomial equations. By applying the Levenberg–Marquardt algorithm in MATLAB and using Mean Squared Error (MSE) in Minitab, both Response Surface Methodology (RSM) and Artificial Neural Networks (ANN) are effective for predicting concrete's flexural strength. Moreover both RSM and ANN gave nearly the same values to that of experimental values but RSM model is much accurate than ANN. It is observed from the analysis that the actual values and predicted values differ by 5% error for 28&#xa0;days and 60&#xa0;days of curing time. </p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Statistical and experimental prediction of bending strength with partial replacement of waste plastic using response surface method (Minitab) and artificial neural network (Matlab) in concrete

  • 期刊论文

  • 发表日期：

    2025年5月10日

查看全文
  • C. Arvind Kumar,
  • Y. Suvidha,
  • Sumesh Jain

摘要

In the present study two different waste plastic materials such as, High Density Polyethylene (HD-PE) and Polypropylene (PP) were collected from dumping yards and was partially replaced with fine aggregate (FA) by 2.5% increment in each mix till 10% replacement. The waste plastic was partially replaced to prepare M30 and M40 grade concrete to find the bending/flexural property. The bending stress of HDPE material initially (7 & 14 days) showed better results to that of PP material for M30 grade concrete. The maximum bending stress of 6.1 Mpa and 5.6 Mpa for 7.5% and 5% of HDPE and PP was observed respectively for M30 grade concrete. The same pattern to that of M30 grade concrete was even observed in M40 grade concrete. Here also the maximum bending stress of 6.99 Mpa and 6.0 Mpa for 7.5% and 5% of HDPE and PP was observed respectively for M40 grade concrete. The optimum percentage of 5% and 7.5 for PP and HDPE material respectively can be replaced in concrete for bending/flexural strength for both M30 and M40 grade. The same experimental values of M30 & M40 grade concrete were used to statically predict the bending stress value using response surface method (RSM) in minitab and compared with artificial neural network (ANN) in matlab. Regression models are evaluated using ANOVA, Pareto charts, and second-order polynomial equations. By applying the Levenberg–Marquardt algorithm in MATLAB and using Mean Squared Error (MSE) in Minitab, both Response Surface Methodology (RSM) and Artificial Neural Networks (ANN) are effective for predicting concrete's flexural strength. Moreover both RSM and ANN gave nearly the same values to that of experimental values but RSM model is much accurate than ANN. It is observed from the analysis that the actual values and predicted values differ by 5% error for 28 days and 60 days of curing time.