基于BP神经网络的机制砂混凝土抗压强度预测

Prediction of Compressive Strength of Mechanical Sand Concrete Based on BP Neural Network

机制砂混凝土抗压强度受多种因素影响,为了提高混凝土品质,需要对其强度特性进行分析。针对传统的机制砂混凝土抗压强度检测方法,利用具有非线性特性、学习能力和自适应能力的BP神经网络进行分析。将石粉含量、水泥、粉煤灰、水、机制砂、碎石和养护龄期作为输入参数,抗压强度作为输出参数,构建了一个包含6个隐含层节点的BP神经网络模型。通过仿真结果表明,平均相对误差为3.47%,线性相关系数大于0.99,该模型具有良好的预测性。

The compressive strength of machine-made sand concrete is influenced by various factors,and it is necessary to analyze its strength characteristics in order to improve the quality of concrete.In this paper,a BP neural network with nonlinear characteristics,learning ability,and adaptive ability was used to analyze the compressive strength of machine-made sand concrete.Stone powder content,cement,fly ash,water,machine sand,gravel,and curing age are taken as input parameters,while compressive strength was taken as the output parameter.A BP neural network model with 6 hidden layer nodes is constructed.The simulation results showed an average relative error of 3.47%and a linear correlation coefficient greater than 0.99,indicating that the model has good predictive performance.