The low-carbon significance of COOA-RBFNN prediction ability in green buildings: rapid detection of concrete microcracks

Abstract

Abstract In the context of a low-carbon economy, emphasizing energy efficiency and sustainability in buildings has become a crucial goal. The performance of concrete, one of the most commonly used building materials in the construction industry, is directly related to the energy efficiency of buildings. By optimizing the fracture properties of concrete, the durability and safety of structures can be significantly improved, which in turn reduces the overall energy use and environmental ipact. There are a large number of uneven micro-cracks in concrete materials. The existence of these mcrocracks makes the concrete material produce nonlinear behavior from the beginning of loading deformation. The traditional prediction method of fracture performance of concrete is mainly linear, represented by the water-cement ratio rule, and adopts linear regression formula, which holds that the strength is completely controlled by water-cement ratio and has nothing to do with other factors. Using neural network technology to predict the fracture performance of concrete has the advantages of strong adaptability, accuracy and effectiveness. COOA-RBFNN (Computation Offloading Optimization Algorithm-RBF neural network-RBF neural network) has the characteristics of simple structure, fast convergence speed, and can approximate any nonlinear function. Therefore, this paper attempts to use COOA-RBFNN to predict the fracture performance of concrete. According to the input-output relationship of practical problems, determine the number of nodes in the input layer and the number of nodes in the output layer. The results show that the maximum relative error of COOA-RBFNN in predicting compressive strength is 0.8161%, the minimum relative error is 0.3999%, the maximum relative error of COOA-RBFNN in predicting flexural strength is 3.6664%, and the minimum relative error is 1.7268%, so COOA-RBFNN has high accuracy in strength prediction. COOA-RBFNN has the advantages of simple structure, strong adaptability, fast convergence, less workload and high accuracy.