Analyzing Observed Seepage of Wushantou Dam Using BP Method
碩士 === 國立成功大學 === 水利及海洋工程學系專班 === 95 === Earth fill dam needs to monitor seepage to let supervisors be able to assess dam safety. Since seepage is expected to be accordance with reservoir’s water level and an abnormal value of it indicates a potential danger of the dam, a model that can predict norm...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
2007
|
Online Access: | http://ndltd.ncl.edu.tw/handle/28121559354840449741 |
id |
ndltd-TW-095NCKU5083026 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-095NCKU50830262015-10-13T14:16:11Z http://ndltd.ncl.edu.tw/handle/28121559354840449741 Analyzing Observed Seepage of Wushantou Dam Using BP Method 倒傳遞類神經網路法於烏山頭水庫壩體滲流量觀測值分析之研究 YEN-HSING CHEN 陳艷星 碩士 國立成功大學 水利及海洋工程學系專班 95 Earth fill dam needs to monitor seepage to let supervisors be able to assess dam safety. Since seepage is expected to be accordance with reservoir’s water level and an abnormal value of it indicates a potential danger of the dam, a model that can predict normal seepage based on water level and other parameters is important. Wushantou dam’s has four automatic seepage monitoring points, namely A, B, C and D, located at the toe of the dam. Based on the 746 daily records of reservoir water level, rainfall depth and the seepages of the four points, training models are constructed by using the back-propagation neural network. It is expected to use the model to predict reference seepage for the reservoir management unit. The result of analysis shows that D point’s values after training and verification are less than other three points and MSE values are bigger thus the mode expecting values are not applicable. Verification with A, B, and C points shows that that the expecting monitoring seepage values are close to the practical monitoring seepage values during heavy rains and MSE values are smaller. It shows that using BP method to analyze dam’s expecting seepage values during sudden encountered heavy rains is basically applicable. However, during dry or less rain days, the expecting monitoring seepage values are with a larger deviation. In order to achieve more accurate expecting values, more potential dam affected seepage factors and practical observation seepage values should be adapted in further investigation. Jan-Mou Leu chan-Ji Lai 呂珍謀 賴泉基 2007 學位論文 ; thesis 63 zh-TW |
collection |
NDLTD |
language |
zh-TW |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立成功大學 === 水利及海洋工程學系專班 === 95 === Earth fill dam needs to monitor seepage to let supervisors be able to assess dam safety. Since seepage is expected to be accordance with reservoir’s water level and an abnormal value of it indicates a potential danger of the dam, a model that can predict normal seepage based on water level and other parameters is important.
Wushantou dam’s has four automatic seepage monitoring points, namely A, B, C and D, located at the toe of the dam. Based on the 746 daily records of reservoir water level, rainfall depth and the seepages of the four points, training models are constructed by using the back-propagation neural network. It is expected to use the model to predict reference seepage for the reservoir management unit.
The result of analysis shows that D point’s values after training and verification are less than other three points and MSE values are bigger thus the mode expecting values are not applicable. Verification with A, B, and C points shows that that the expecting monitoring seepage values are close to the practical monitoring seepage values during heavy rains and MSE values are smaller. It shows that using BP method to analyze dam’s expecting seepage values during sudden encountered heavy rains is basically applicable. However, during dry or less rain days, the expecting monitoring seepage values are with a larger deviation. In order to achieve more accurate expecting values, more potential dam affected seepage factors and practical observation seepage values should be adapted in further investigation.
|
author2 |
Jan-Mou Leu |
author_facet |
Jan-Mou Leu YEN-HSING CHEN 陳艷星 |
author |
YEN-HSING CHEN 陳艷星 |
spellingShingle |
YEN-HSING CHEN 陳艷星 Analyzing Observed Seepage of Wushantou Dam Using BP Method |
author_sort |
YEN-HSING CHEN |
title |
Analyzing Observed Seepage of Wushantou Dam Using BP Method |
title_short |
Analyzing Observed Seepage of Wushantou Dam Using BP Method |
title_full |
Analyzing Observed Seepage of Wushantou Dam Using BP Method |
title_fullStr |
Analyzing Observed Seepage of Wushantou Dam Using BP Method |
title_full_unstemmed |
Analyzing Observed Seepage of Wushantou Dam Using BP Method |
title_sort |
analyzing observed seepage of wushantou dam using bp method |
publishDate |
2007 |
url |
http://ndltd.ncl.edu.tw/handle/28121559354840449741 |
work_keys_str_mv |
AT yenhsingchen analyzingobservedseepageofwushantoudamusingbpmethod AT chényànxīng analyzingobservedseepageofwushantoudamusingbpmethod AT yenhsingchen dàochuándìlèishénjīngwǎnglùfǎyúwūshāntóushuǐkùbàtǐshènliúliàngguāncèzhífēnxīzhīyánjiū AT chényànxīng dàochuándìlèishénjīngwǎnglùfǎyúwūshāntóushuǐkùbàtǐshènliúliàngguāncèzhífēnxīzhīyánjiū |
_version_ |
1717750967997300736 |