Title: Study on properties of polyurethane concrete based on BP neural network

Authors: Xiaolian Gai; Dongpo He

Addresses: School of Civil Engineering, Northeast Forestry University, Harbin, 150040, China; School of Civil Engineering, Harbin Institute of Petroleum, Harbin, 150028, China ' School of Civil Engineering, Northeast Forestry University, Harbin, 150040, China

Abstract: In order to improve the research accuracy of fatigue properties of polyurethane concrete and shorten the research time, a new research method of fatigue properties of polyurethane concrete based on BP neural network is proposed in this paper. Firstly, polyurethane materials and aggregates are selected to prepare polyurethane concrete. Secondly, the characteristic parameters of polyurethane concrete are calculated, including age, damage degree, tensile recovery, compressive elastic recovery, compressive strength and stiffness modulus. Finally, the calculated characteristic parameters are input into the BP neural network, and the final fatigue properties of polyurethane concrete are calculated iteratively after the correction of weight and threshold. The experimental results show that the proposed method can accurately analyse the fatigue properties of polyurethane concrete, and the average analysis accuracy of fatigue life reaches 94.281%. The research time of the proposed method is short, and the longest time is not more than ten minutes.

Keywords: BP neural network; polyurethane concrete; fatigue properties of concrete; compressive strength; stiffness modulus.

DOI: 10.1504/IJMPT.2023.132194

International Journal of Materials and Product Technology, 2023 Vol.67 No.1, pp.52 - 65

Received: 26 Oct 2022
Accepted: 08 Feb 2023
Published online: 12 Jul 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article