基于SVM与灰色理论的大坝基岩安全监测动态评判研究

发布时间：2018-04-21 10:35

本文选题：大坝坝基 + 灰色理论　；参考：《合肥工业大学》2014年硕士论文

【摘要】：大坝监测数据分析理论和方法的研究与应用已经取得了相当的进展，为保证大坝安全运行发挥了巨大的作用。但是，在数据分析方面依然存在许多问题和不足。大坝坝基是大坝的主要组成部分，目前对坝基做出综合评价的研究仍然较少，大多数研究主要集中在坝基单项指标基础上的单项指标评价。针对现有分析方法和分析模型中存在的问题和不足，建立坝基综合评价体系中各指标的权值模型，结合灰色理论综合评判框架，对坝基运行状态做出安全评价，并在此基础上将支持向量机模型应用到大坝坝基监测数据的分析及预测中，形成支持向量机与灰色系统理论相结合的动态评价体系，更加有效合理地实现对坝基运行现状及发展趋势做出评价，，满足实际工程应用的需要。影响坝基安全的因素众多，目前技术水平还难以获取坝基运行状态全部信息。由于人们的认识有限或受到伪信息的干扰，这使得坝基监测信息不完全，即具有灰性，同时为了避免坝基单项指标评价的不足，本文采用灰色聚类综合评判模型实现对坝基运行状态做出安全综合评价。针对综合评价中权值对评判结果的重要程度，在现有权值计算模型的基础上，为充分挖掘大坝监测数据所携带的信息，并考虑影响大坝基岩安全的因素特征，从投影寻踪(PP)方法的优化函数和约束条件两个方面改进传统算法，从而得到约束型最大熵投影寻踪耦合权值模型，并采用此权值模型对坝基运行安全现状做出了综合评判。由于大坝运行管理者不仅仅希望了解大坝坝基运行现状，还希望能够了解坝基运行的发展趋势，本文将支持向量机模型引入到坝基位移预测中，建立了坝基位移与环境量之间的支持向量机模型，并对大坝坝基位移做出预测。本文在监测数据为小样本的基础上，将灰色系统理论与支持向量机两种适合于小样本的数学模型相结合，利用动态支持向量机对坝基位移做出预测，并利用约束型最大熵-投影寻踪模型计算坝基指标的动态权重，再在两者的基础上利用灰色系统综合评价理论，构建了完整的动态安全评价框架。
[Abstract]:Considerable progress has been made in the research and application of dam monitoring data analysis theory and method, which plays an important role in ensuring dam safe operation. However, there are still many problems and shortcomings in data analysis. The dam foundation is the main part of the dam. At present, the research on the comprehensive evaluation of the dam foundation is still few, and most of the studies are mainly focused on the single index evaluation based on the single index of the dam foundation. In view of the problems and shortcomings in the existing analysis methods and models, the weight model of each index in the comprehensive evaluation system of dam foundation is established, and the safety evaluation of the operation state of the dam foundation is made by combining the comprehensive evaluation frame of grey theory. On this basis, the support vector machine model is applied to the analysis and prediction of dam foundation monitoring data, and a dynamic evaluation system combining support vector machine and grey system theory is formed. It is more effective and reasonable to evaluate the present situation and development trend of dam foundation operation and to meet the needs of practical engineering application. There are many factors influencing the safety of the dam foundation, so it is difficult to obtain all the information of the operation state of the dam foundation at present. Because people's understanding is limited or disturbed by false information, the monitoring information of dam foundation is not complete, that is, it is grey, and in order to avoid the deficiency of single index evaluation of dam foundation, In this paper, the grey cluster comprehensive evaluation model is used to evaluate the operation state of dam foundation. In view of the importance of weight value to the result of comprehensive evaluation, based on the existing calculation model of weight value, the information carried by dam monitoring data is fully excavated and the characteristics of factors affecting the safety of dam bedrock are considered. The traditional algorithm is improved from two aspects of optimization function and constraint condition of projection pursuit PP) method, and the coupling weight model of constrained maximum entropy projection pursuit is obtained, and the safety situation of dam foundation is evaluated synthetically by using this weight model. Because the managers of dam operation not only want to know the current situation of dam foundation operation, but also want to know the development trend of dam foundation operation, this paper introduces support vector machine model into dam foundation displacement prediction. The support vector machine model between dam foundation displacement and environmental quantity is established, and the dam foundation displacement is predicted. Based on the small sample of monitoring data, this paper combines the grey system theory with the support vector machine (SVM) to predict the displacement of dam foundation by using dynamic support vector machine (DSVM). The dynamic weight of the dam foundation index is calculated by using the constrained maximum entropy projection pursuit model, and the comprehensive evaluation theory of grey system is used to construct a complete dynamic safety evaluation framework.
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TV698.1

【参考文献】