Prediction of compressive strength of high performance concrete based on GA-BP neural network
Compressive strength is a key indicator for evaluating the performance of high performance concrete(HPC).In order to accurately predict its compressive strength,a new GA-BP-GRA model was developed by combining BP neural network with genetic algorithm(GA),using 181 sets of data for model training,and con-ducting sensitivity assessment through grey relational analysis(GRA).The study shows that the neural network can predict the compressive strength of HPC quickly and accurately;compared with a single neural network model,the re-optimized new model has higher accuracy in predicting compressive strength and smaller disper-sion in the prediction results;GRA can effectively analyze the sensitivity of the main components of HPC.The proposed GA-BP-GRA model can rapidly and accurately predict the compressive strength of concrete,thereby saving test time and reducing design costs.
high performance concretecompressive strengthgenetic algorithmBack Propagation nearal networkgrey relational analysissensitivity
万方数据
维普
