摘要
当前,国内已经构建起完善的高速公路网,而各地针对高速公路的养护模式与技术手段各不相同,且缺乏一个有效的技术手段。在针对高速公路养护技术现状下,以路面使用性能预测评价为基础,构建了基于神经网络和马尔科夫组合预测的路面使用性能预测模型框架。针对路面状况指数PCI、行驶安全指数SRI、行驶质量质量RQI,通过建立组合预测模型,进行路面使用性能的预测,研究结果表明:采用加权算数平均组合、加权平方和平均组合、加权比例平均组合相较于单一的马尔科夫预测以及神经网络预测,在预测精度和预测误差范围上都要得到了很大提高,尤其是加权平方和平均组合所构成的路面使用性能预测模型获得了很好的使用性能预测结果。
At present, China has build up a system of highway construetion and network,The mod- els of the highway maintenance and technical means are not identical, In addition, there still lack of an ef- fective technicai standards. Based on Present situation of China, s highway maintenance management, the pape evaluation for the technical requirements for pavement performance prediction, constructed of pave- ment'performance prediction model framework Based on the neural network and markov prediction mod- el. In this paper,pavement condition index PCI, driving safety index SRI RQI, driving quality quality, through establisht a combination forecast model to predict the pavement performance, the results show that compared with sum of squares of a single markov prediction and neural network prediction,the combina-tion of weighted arithmetic average,weighted average, weighted average combination proportion at the pre- diction accuracy and prediction error range have greatly improved, especially the average weighted sum of squares of combination of pavement performance prediction model is obtained obtained the very good use performance prediction results.
出处
《公路工程》
北大核心
2015年第6期264-270,共7页
Highway Engineering
关键词
路面使用性能预测
PCI
组合预测模型
pavement performance prediction
PCI
combination forecast model