论文部分内容阅读
道路混凝土弯拉强度受多种因素共同影响,具有非线性特征。考虑集料最大粒径、水泥用量、用水量、砂和粗集料含量对混凝土弯拉强度的影响,基于径向基神经网络理论,通过优化网络扩散系数和神经元数,建立了用于预测不同配合比下道路混凝土抗弯拉强度模型,给出的预测值与真实值相比在可接受的误差范围内,而且训练网络的数据覆盖范围广,因此该网络预测模型普适性强,对于优化道路混凝土的配合比设计具有重要的意义。
The flexural strength of road concrete is affected by many factors and has nonlinear characteristics. Considering the influence of aggregate maximum size, cement content, water consumption, sand and coarse aggregate content on the flexural strength of concrete, based on RBF neural network theory, by optimizing the network diffusion coefficient and neuron number, The models of flexural tensile strength of road concrete under different mix ratios show that the predicted value of the road network is within acceptable error range compared with the real value, and the data coverage of the training network is wide. Therefore, the network prediction model is universally applicable. Optimizing the mix design of road concrete is of great significance.