中文摘要：

利用BP神经网络法，建立了再生混凝土的坍落度预测模型，该模型采用了三层网络结构模式，输入层采用再生混凝土的配合比数据，输出层为再生混凝土坍落度数据，模型的转移函数均采用单极性Sigmoid函数，由于Sigmoid函数值为[0，1]间，因此，对再生混凝土输入数据的归一化处理。通过再生混凝土坍落度的大量试验，将试验结果与网络模型预测结果进行比对，证实了该模型对再生混凝土的坍落度预测值与实际试验结果相符合，可以为试验及生产提供参考。

英文摘要：

Through the method of BP neutral network,the article constituted forecast model about recycled concrete slump.The model was three-layer network structure pattern.The inputting-layer was the mix proportions of recycled concrete.The outputting-layer was the recycled concrete slump.The model＇s transfer function was Sigmoid function and the result of Sigmoid was in[0,1].So the article turned the inputting data into the data in[0,1].The experimental work for the recycled concrete slump was designed and performed.The article contrast the model prediction data with the experimental data,and then confirmed that forecast results was match with experimental results and it could provide the referenced for experimentation and manufacture.