Reinforced concrete member anti-collision performance prediction method based on machine learning

Abstract

The invention discloses a reinforced concrete member anti-collision performance prediction method based on machine learning, and the method comprises the following steps: obtaining impact test parameters and member characteristic parameters of a reinforced concrete member, constructing a data set, and randomly dividing the data set into a training set and a prediction set; determining an artificial neural network topological structure, and obtaining a trained residual performance prediction model; adopting a prediction set to check the prediction performance of the trained model until the appropriate prediction precision is achieved; and for a specific reinforced concrete member, performing predicting by utilizing the trained residual performance prediction model to obtain a vulnerable curved surface of the reinforced concrete member. According to the invention, based on a machine learning algorithm, the residual performance of the RC component after impact is used as an index, and a rapid and reliable method for predicting and evaluating the impact resistance of the RC component is established. By fully using the existing historical test data, the impact resistance of the RC component can be evaluated more quickly and accurately.

Description

technical field [0001] The invention relates to the field of evaluation of the anti-collision performance of civil engineering, in particular to a method for predicting the anti-collision performance of reinforced concrete members based on machine learning. Background technique [0002] The static load-bearing performance evaluation method of reinforced concrete members (RC members) is very mature, but the impact resistance performance evaluation of RC members has been lacking in a reasonable and simple way. This is mainly affected by two reasons. First, there is a lack of reasonable indicators for evaluating the impact resistance of ordinary components. The use of maximum support reaction force, maximum impact displacement, etc. to evaluate the impact resistance of components cannot be directly related to the working performance of the structure after impact. It is easy to use in engineering practice; the second is that the damage mechanism of RC components after impact is ...

AI Technical Summary

Problems solved by technology

[0002] The static load-bearing performance evaluation method of reinforced concrete members (RC members) is very mature, but the impact resistance performance evaluation of RC members has always lacked a reasonable and simple method

This is mainly affected by two reasons. First, there is a lack of reasonable indicators for evaluating the impact resistance of ordinary components. The use of maximum support reaction force, maximum impact displacement, etc. to evaluate the impact resistance of components cannot be directly related to the working performance of the structure after impact. It is easy to use in engineering practice; the second is that the damage mechanism of RC components after impact is relatively complex, and the mechanism of bearing capacity decline is not clear. At present, the performance degradation range under limited parameters can only be obtained by testing and high-precision numerical simulation, while the model test consumes There are many manpower and material resources, and the technical threshold of high-precision numerical simulation is high. The above two points make it difficult to evaluate the impact resistance of RC components.

Images