Strength prediction of paste filling material based on convolutional neural network. (9th July 2020)

• Record Type: Journal Article
• Authors: Cheng, Haigen; Hu, Junjian; Hu, Chen; Deng, Fangming
• Other Names: Srivastava Gautam guestEditor.; Hsu Ching‐Hsien (Robert) guestEditor.; Kumar Priyan Malarvizhi guestEditor.
• Is Part Of: Computational intelligence. Volume 37:Number 3(2021)
• Page Start: 1355
• Page End: 1366
• Publication Date: 2020-07-09
• Subjects: compressive strength -- convolution neural network -- smart cities -- paste filling material -- prediction model; Artificial intelligence -- Periodicals; Computational linguistics -- Periodicals; 006.3
• DOI: 10.1111/coin.12373
• Languages: English
• ISSNs: 0824-7935

Abstract

Abstract: The common backfill mining technology in the green mining industry can be used for the secondary utilization of construction waste in smart cities. This measure has the advantages of low cost, fast results, and less environmental pollution. Over the past few decades, with the continuous advancement of global urbanization, the effective and environmentally friendly construction waste disposal and emission are very important for the development of urban green construction. Construction waste can be prepared as paste filling material, as one of the raw materials for backfill mining. This paper proposes a new method that can quickly and accurately predict the strength of paste filling materials with different compositions. A deep connected convolutional neural network (CNN) that can extract input parameters is used to build a prediction model. The coarse aggregate, fine aggregate, and cementing material are employed as the input variables of the CNN model, and five indicators which are generally used to evaluate the strength of filling material are selected as the output results. The experimental results show that the proposed prediction approach can obtain robust prediction results and high prediction accuracy and speed.