Study On Prediction Of Concrete Strength Based On Artificial Neural Network

Recently, the safety of existing civil engineering structures attracts more and moreattention. The strength of concrete plays a key role during the assessment of safety and durabilityfor civil engineering structures. The strength of concrete will gradually decrease during theservice of civil engineering structures. It is significant to accurately predict the strengthdeterioration of concrete for correctly evaluating the safety of structures.In this paper, based on concrete test data, using atificial neural network model and neuralnetwork toolbox in MATLAB, the prediction of concrete compressive strength has been studied.At the very beginning of this paper, the theory of neural network, structure of BP network,the theory of BP argorithm and process of the implementation of BP network program wasexpounded. Three-layer BP neural network was established，the neural network was trained andtested by the samples collected from testing data at home and abroad. After analysis，it is foundthat accurately predicting the strength of concrete is feasible.Secondly, this paper summarized the effect factors for concrete strength. Based on theanalysis method of sensitivity, the paper anslysised the sensitivity of several factors.At last, A pilot studies of the mechanical properties of concrete have been finished througha series of experiments. And based on the results of experiments, the model of strengthprediction of concrete has been established. From the test results, the trained neural network canaccurately predict the long-term strength of concrete with the error less then7%.