铁路沿线室外气象参数的神经网络构成

分析影响室外气象计算参数各主要因素,利用神经网络的非线性和良好的学习能力,以夏季空调室外计算(干球)温度或夏季空调室外相对湿度为例,在全国范围内建立并找到了其最优BP模型结构。在采用国家规范中标准样本进行训练、校核后,认为该模型除在效率和局部精度上仍需提高外,就整体而言输出结果误差很小,并能较好地模拟铁路沿线室外气象参数的变化规律。

Study on Outdoor Weather Parameters Structured by Neural Network Arithmetic along Railways

According to the non-linear character and strong study ability of neural network, an optimal BP model structure is set up in the whole country based on the analysis of important factors affecting the calculation of outdoor weather parameters and taking as an example dry-temperature and relative humidity calculated in the summer outdoor condition. After experimenting and verifying the standard sample of the national specifications, it is found that the global error of the output is very small while the efficiency and part of the accuracies need be improved. The regular pattern of the change of outdoor weather data along railways can be simulated well by the model.