节点文献
基于混合遗传神经网络的高速公路沥青路面使用性能评价方法研究
Study on the Performance Evaluation in Expressway Asphalt Pavement Based on the Mixed GANN Model
【作者】 黄文雄;
【导师】 刘沐宇;
【作者基本信息】 武汉理工大学 , 市政工程, 2003, 硕士
【摘要】 高速公路把国民经济各领域和各个地区紧密地联系起来,对繁荣经济发展、增进文化交流、促进社会进步都起着重要的推动作用。路面使用性能是高速公路最重要的功能之一,要保证高速公路安全、经济、高效的运营,必须对它实行科学的管理与养护。路面使用性能的正确评价是合理制定养护维护计划、进行投资决策的重要依据之一,是路面养护管理系统中重要的一环。因此,对路面使用性能评价方法的研究在国内外受到极大的重视,是当前道路工程领域研究的热点问题之一。 目前国内外常用的四类路面使用性能评价方法中回归法评价结果与实测数据相关性不太理想,系统分析法客观性不强,灰色理论法存在一定的主观经验性,其它一些方法还处于研究阶段;而且这些方法都是基于普通公路建立,高速公路具有交通量大、汽车轴载重、交通渠化和行车速度高等特点,路面使用性能的特点和发展变化规律与普通公路存在较大差异。所以,对高速公路路面使用性能评价新方法的研究,对确保道路运输的畅通和国民经济顺利的发展有着极其重要的理论和现实意义。 针对这些问题,本文在分析现有沥青路面使用性能评价方法的基础上,对高速公路沥青路面使用性能评价问题进行了比较深入的研究,本文的具体研究成果如下: 1、对国内外路面使用性能评价方法的现状进行了评述,详细论述了各种评价方法的基本思路及其优缺点,并深入探讨了路面使用性能评价的发展趋势,为建立新的评价方法提供了基础与参考; 2、分析了高速公路沥青路面的特点,对常用的路面使用性能评价指标进行了研究,并针对高速公路的特点对各指标进行改进,探讨了一些能充分反映高速公路沥青路面使用性能特点的评价指标; 3、针对遗传算法与神经网络优缺点互补的特点,探讨了GA与NN结合的必要性与可行性,将其结合与完善,建立了能充分发挥遗传算法与神经网络优势的混合GANN模型;选择简单、高效、开放的Matlab语言实现了混合GANN模型; 4、首次提出了高速公路沥青路面使用性能混合GANN评价方法;该方 武汉理工大学硕士学位论文法运用混合GANN模型,以路面最大裂缝率CR、最大车辙深度RD、平整度o、强度系数551、抗滑能力SFC实测数据为基础进行模型建立,实现了对高速公路沥青路面使用性能的评价;通过对京珠高速公路湖北段路面的评价,结果表明该方法准确、科学、高效、实用,评价过程方便、快捷,预测结果科学、准确; 5、首次提出了沥青路面路面状况指数PCI的混合GANN预测方法;该方法运用混合 GANN模型,以路面平整度。、强度系数551与抗滑能力SFC实测数据为基础进行模型建立,实现了对路面状况指标PCI的预测;依据该方法可以对路面破损进行有针对性的调查,减少调查范围,节约人力物力,提高工作效率。 本研究项目得到湖北省科技攻关项目(高速公路路面结构优化方法研究98 1 P 1201)资助。
【Abstract】 Varieties of field and area of national economy are correlated by expressway and expressway does great advantages to the economical developments culture communion and social progress. Pavement performance is one of the most important functions of the expressway, the good administration and maintenance must be put into effect to it if we want it working in a safety > economic and efficient way. Aiming at making a good maintenance plan and investment decision, the correct evaluate of pavement performance is one of the most important reference, it is a great important component of Pavement Management System. So, study on the pavement performance evaluation is attached great importance by various nation, and it is becoming the hot problem in the research of road engineering.Now, there are four evaluation methods in this field, every method has its advantages and shortages and most of them are established based on highway. High weight communication, great loads channelized traffic and high speed is the expressway’s feature, there is great difference in the development law between highway and expressway. We will get some unfitness result by using those evaluation methods to evaluate expressway. So, study on the new evaluation method of the pavement performance is very important and it does great advantages to the economical development and traffic progress.Based on these problems, we study the expressway asphalt pavement performance evaluation methods deeply. The main achievement of this dissertation are briefly mentioned as follows:1 Account for the present situation of the pavement performance evaluation, expatiate on the connotations advantages and shortages of every existing methods of the pavement performance evaluation methods, and probe into its developing direction about the pavement performance evaluation, those establish a good foundation of new evaluation methods.2 After analysis the feature of expressway asphalt pavement, we study thecommon pavement performance evaluation index. Based on the feature of expressway asphalt pavement, we approach some new index which can embody the expressway asphalt pavement feature.3 Based on the complementation of advantages of GA and NN, we probe into the necessity and feasibility of their conjunction and associate them together, a new mixing GANN model which possesses the advantages both GA and NN is established. Matlab, one of the programming assistant which is simple directly efficient and open is selected to fulfill this model.4 It is the first time that A new mixed GANN evaluation method of expressway asphalt pavement performance is set up. Based on the measured data of CR RD σ SSI and SFC, the mixing model is used when this method is established. Thought the evaluation of the Jingzhu expressway, the results indicate that the model is scientific, precise, practical and useful.5 It is the first time that A new mixed GANN forecasting method of asphalt pavement PCI is set up. Based on the measured data of σ SSI and SFC, the mixing model is used when this method is established. Subject to this forecasting method, we can forecast PCI by σ SSL SFC and the workload of pavement distress detection is reduced. So, this model is useful.
【Key words】 expressway; asphalt pavement; pavement performance; evaluation models; the mixed genetic algorithm neural network;
- 【网络出版投稿人】 武汉理工大学 【网络出版年期】2004年 03期
- 【分类号】U416.217
- 【被引频次】50
- 【下载频次】950