Nakkeeran G, Krishnaraj L, Shakor Pshtiwan, Alaneme George Uwadiegwu, Otu Obeten Nicholas
Department of Civil Engineering, Madanapalle Institute of Technology and Science, Madanapalle, 517325, Andhra Pradesh, India.
Department of Civil Engineering, SRM Institute of Science and Technology, Kattankulathur, Kancheepuram, 603 203, Tamil Nadu, India.
Sci Rep. 2024 Oct 14;14(1):24075. doi: 10.1038/s41598-024-74970-9.
In recent years, building materials made from agricultural waste have become popular due to their lower cost and environmental impact. The Bio-Brick is mixed with Cement-Fly Ash and Hydrated Lime and a fine aggregate of groundnut shell in percentages (20%, 30%, 40%, 50%, and 60%). The optimum mix proportions of Bio-Brick and hydrated lime mortar were found from the compressive strength and were further continued to study the dry density, water absorption, and efflorescence. Machine Learning techniques are used to optimize and predict the properties of Bio-Bricks and mortars. Response Surface Methodology (RSM) and Artificial Neural Networks (ANN) are employed to forecast properties such as compressive strength, dry density, and water absorption with exceptional accuracy. The results from RSM models exhibit high degrees of accuracy, with R-squared values exceeding 0.88 for compressive strength, dry density, and water absorption. ANN models further enhance this predictive power, with R-squared values exceeding 0.99 in predicting these critical properties.
近年来，由农业废弃物制成的建筑材料因其成本较低和对环境的影响较小而受到欢迎。生物砖与水泥粉煤灰、熟石灰以及花生壳细集料按（20%、30%、40%、50%和60%）的比例混合。通过抗压强度确定了生物砖和熟石灰砂浆的最佳配合比，并进一步研究了干密度、吸水率和泛霜情况。机器学习技术用于优化和预测生物砖及砂浆的性能。采用响应面法（RSM）和人工神经网络（ANN）以极高的精度预测抗压强度、干密度和吸水率等性能。RSM模型的结果显示出高度的准确性，抗压强度、干密度和吸水率的决定系数（R平方）值超过0.88。ANN模型进一步增强了这种预测能力，在预测这些关键性能时，R平方值超过0.99。
