Real-time Strength Prediction of Different Types of Concrete Based on BP Neural Network

In terms of information processing, artificial neural network is similar to the synaptic connection structure in the human brain. As a mathematical model for information processing, it is widely used in various fields like biology, medicine and construction engineering. This paper adopts the back propagation neural network and predicts the strength of concrete considering a variety of impact factors. Firstly, it introduces the basic principle of the neural network algorithm based on brain neural mechanism, and summarizes the safety problems in concrete projects and concrete strength prediction methods. With the real-time strength prediction BP model for different types of concrete proposed in this paper, a nonlinear mapping relationship between concrete strength and impact factors can be established, which effectively reduces the number of concrete trials and improve the quality of concrete. The measured results show that the accuracy of concrete strength prediction is higher than 96%. Therefore, this model can provide theoretical guidance for better application of concrete in construction projects.