摘要
鉴于新拌混凝土流变参数与硫酸盐侵蚀劣化指标之间存在的非线性映射关系,基于新拌混凝土流变参数,利用BP神经网络建立了硫酸盐干湿循环侵蚀后混凝土抗压强度损失率、质量损失率及相对动弹性模量的预测模型,预测了混凝土受硫酸盐侵蚀程度。结果表明,BP神经网络模型预测的硫酸盐侵蚀结果与实测结果吻合较好,且预测误差较小,说明BP神经网络模型对混凝土受硫酸盐侵蚀后的劣化程度具有良好的预测效果。
There is complex nonlinear mapping relationship between fresh concrete rheological parameters and its re sistance to sulfate erosion. Based on the fresh concrete rheological parameters, BP neural network (BP-ANN) is used to establish the prediction model of strength loss ratio, mass loss rate and relative dynamic modulus loss. And then the de gree of sulfate erosion to concrete is predicted. The results show that the BP-ANN prediction model is of good forecast precision and the prediction error is small, which indicated that the BP-ANN model is fit for the degradation prediction of concrete suffered sulfate erosion.
出处
《水电能源科学》
北大核心
2013年第11期102-104,112,共4页
Water Resources and Power
关键词
混凝土
流变参数
硫酸盐侵蚀
BP神经网络
concrete
theological parameter
sulfate erosion
BP neural network