Title: Strength prediction of concrete with large amount of fly ash based on improved random forest
Authors: Fangfang Zhang; Huajian Fang; Pengyun Li; Yuandong Qiao
Addresses: School of Architecture and Surveying Engineering, Shanxi Datong University, Datong, Shanxi, 037003, China ' ZheJiang ZheFeng Engineering Consulting Co. Ltd., Hangzhou, Zhejiang, 310021, China ' Shandong Civil Air Defense Architectural Design Institute Co. Ltd., Jinan, Shandong, 250000, China ' School of Architecture and Surveying Engineering, Shanxi Datong University, Datong, Shanxi, 037003, China
Abstract: In order to solve the problems of the large gap between the predicted results and the actual strength value, the high MSE value and the long time in the traditional method, a strength prediction method for concrete with large amount of fly ash based on an improved random forest is proposed. Firstly, genetic algorithm and BP neural network are used to screen suitable prediction indicators. Then, the bee mating intelligent optimisation algorithm is used to improve the traditional random forest algorithm. Finally, taking the strength grade and cement consumption as the input variables of the model and the concrete strength value as the output variables, the improved random forest algorithm is used to establish the concrete strength prediction model. Experimental results show that the difference between the predicted results of the proposed method and the actual strength value is small, the MSE value is low, and the time is short.
Keywords: improved random forest; concrete strength; intelligent optimisation algorithm for bee mating; BP neural network.
DOI: 10.1504/IJMMP.2023.134772
International Journal of Microstructure and Materials Properties, 2023 Vol.16 No.6, pp.467 - 481
Received: 14 Feb 2023
Accepted: 14 Jul 2023
Published online: 09 Nov 2023 *