A TRIAL STUDY TOWARDS THE UTILIZING OF AI TECHNOLOGY FOR ANOMALY DETECTION OF DAM SAFETY MANAGEMENT DATA

In safety management of dams in their maintenance works, visual inspection and monitoring of various measurement data such as water leakage and deformation are basic and important means for anomaly detection. However, the number of dams in long-term service increases, while the number of experienced...

Full description

Saved in:
Bibliographic Details
Published inInfrastructure Maintenance Practices Vol. 1; no. 1; pp. 363 - 371
Main Authors KOBORI, Toshihide, NIKAIDO, Ryohei, MATSUSHITA, Tomoaki, KONDO, Masafumi
Format Journal Article
LanguageJapanese
Published Japan Society of Civil Engineers 2022
Subjects
AI
anomaly detection
dam
LSTM
safety management
Online AccessGet full text

Cover

More Information
Summary:In safety management of dams in their maintenance works, visual inspection and monitoring of various measurement data such as water leakage and deformation are basic and important means for anomaly detection. However, the number of dams in long-term service increases, while the number of experienced and skilled staff is decreasing. In order to deal with such a situation, it is necessary to make more effective use of various measurement data so that the manager can more accurately judge whether or not there is a safety problem in the dam. In this study, the possibility of using AI technology for supporting dam managers’ judgement was discussed on a trial basis focusing on time-series deformation data obtainded in concrete dams. As a result, it was shown that LSTM (Long Short Term Memory), which is a kind of recurrent neural network, can be expected to be used for predicting the behavior of dams for anomaly detection.
ISSN:2436-777X
DOI:10.11532/jsceim.1.1_363