基于神经网络的沥青混凝土路面使用性能预测

Asphalt Concrete Pavement Performance Forecast Based on Neural Network

阐述了应用BP神经网络进行沥青混凝土路面使用性能预测的可行性,以环境因素,即交通量、降雨量、温度等3个直接使路面性能发生变化的主要因素作为BP神经网络的输入层,以沥青混凝土路面的破损率、平整度、弯沉值、摩擦系数等4项性能指标作为输出层,采用MATLAB语言中的BP神经网络模块进行自学习,得出环境因素与路面使用性能之间的关系。研究了基于神经网络的沥青混凝土路面使用性能预测模型的分析和计算方法,通过实例验证了该方法的实用性.

The feasibility of neural network application to asphalt concrete pavement performance forecast is set forth.Three main environmental factors——traffic,rainfall,temperature,which can lead to the direct change of pavement performance are used as the input layer of neural network.The performance indexes of asphaltum concrete pavement,i.e.dilapidation rate,smoothness,deflection,friction coefficient are used as output layer.The relation between environment factor and pavement performance is acquired by introducing the module of BP Neural Network in MATLAB.The forecast method of asphaltum concrete pavement performance is studied.The efficiency of the calculation method is proved through illustration.