Construction of performance degradation model and development of performance estimation system for stock management of irrigation tank

灌溉水箱库存管理性能退化模型构建及性能评估系统开发

基本信息

批准号：
21380146
负责人：
KOBAYASHI Noriyuki
金额：
$ 9.07万
依托单位：
Ehime University
依托单位国家：
日本
项目类别：
Grant-in-Aid for Scientific Research (B)
财政年份：
2009
资助国家：
日本
起止时间：
2009 至 2011
项目状态：
已结题

项目摘要

This research aimed to establish the method of functional diagnosis and to develop the performance estimation system for stock management of irrigation tanks. The field investigations by two types of the geophysical exploration methods, that is, the electrical resistivity method and the surface wave method, are carried out in three overage irrigation tanks. The spatial distributions of the electric resistivity and the S-wave velocity are individually obtained by the two methods, and the self-organizing map(SOM) is employed to combine the two results organically. SOM is a type of artificial neural network that is trained using unsupervised learning to produce a two dimensional discretized representation called a map. This makes SOM useful for visualizing low-dimensional views of high dimensional data. The soil types are classified automatically and objectively with the S-wave velocity, the electric resistivity and the change ratio of electric resistivity measured in full water level and low water level adopted as the input vectors. This procedure also enables us to estimate the grouting area. A computational method, incorporating the finite element model into identification using the Monte Carlo filter, is suggested in order to estimate deteriorated area within embankment of irrigation tank based on measured data on the settlement.

这项研究旨在建立功能诊断的方法，并开发用于灌溉罐的库存管理的绩效估计系统。通过两种类型的地球物理勘探方法进行的现场研究，即电阻率法和表面波方法，是在三个过度灌溉罐中进行的。电阻率和S波速度的空间分布是通过两种方法单独获得的，并且使用自组织图（SOM）有机地组合两个结果。 SOM是一种人工神经网络，使用无监督的学习训练，以产生称为地图的二维离散表示形式。这使得SOM可用于可视化高维数据的低维视图。土壤类型以S波速度自动和客观地分类，电阻率和电阻率的变化比在全水位和低水位中测得的电阻率和低水位作为输入向量。此过程还使我们能够估算灌浆区域。建议使用蒙特卡洛过滤器将有限元模型纳入识别识别，以估算灌溉罐的路堤内的恶化区域，该区域基于测量的沉降数据。