论文部分内容阅读
国外轨道不平顺的研究思路和方法大多是建立在对轨道动态检测数据的不同分析手段和方法上,然而由于各国检查数据的不同使得国外模型无法直接应用于国内轨道状态预测。基于昆明铁路局的实际数据,提出了灰色预测模型和灰色-马尔可夫预测模型。通过将2种模型的预测结果与一元线性回归预测结果对比分析,证明灰色预测模型不仅适用于较长区段的TQI(轨道质量指数)预测,对于更长区段的TQI预测也有比较好的效果,同时证明灰色-马尔可夫预测模型不仅能够预测随机性较强的TQI数据,也能应用于较长时间范围内的预测。
Most of the research ideas and methods of track irregularity abroad are based on different analysis methods and methods of orbit dynamic detection data. However, due to the different inspection data in different countries, foreign models can not be directly applied to the prediction of domestic orbit status. Based on the actual data of Kunming Railway Administration, the gray forecasting model and gray-Markov forecasting model are proposed. By comparing the prediction results of the two models with the results of the univariate linear regression, it is proved that the gray prediction model is not only suitable for the prediction of TQI (Longitudinal Section Index) but also for the longer interval TQI prediction , At the same time it is proved that the gray-Markov forecasting model not only can predict TQI data with strong randomness, but also can be applied to the prediction in a longer time range.