Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Performance of data-based models for early detection of damage inconcrete dams
KTH, School of Architecture and the Built Environment (ABE), Civil and Architectural Engineering, Concrete Structures.ORCID iD: 0000-0003-2594-4107
KTH, School of Architecture and the Built Environment (ABE), Civil and Architectural Engineering, Concrete Structures.ORCID iD: 0000-0003-3586-8988
KTH, School of Architecture and the Built Environment (ABE), Civil and Architectural Engineering, Concrete Structures.
(English)Manuscript (preprint) (Other academic)
Abstract [en]

A failure of a massive concrete dam could cause catastrophic consequences. Thepurpose of monitoring is to detect anomalies and damage at an early stage to preventfailure. Data-based models for anomaly detection are based on the hypothesis thatthe behaviour of an undamaged dam will follow an expected pattern, and deviationfrom this pattern is an indication of damage. In this study, simulations were usedto create time series for an undamaged dam and three different damage scenariosat three different locations in the dam body. Three common data-based modelswere used to predict a dams crest displacements, both on the generated artificialdata and corresponding measurements from the dam. Prediction bands for futuremeasurements were created, and the ten time-series were used to test the ability todetect damage. All models could detect instantaneous damage but struggle to detectprogressive damage; the Neural network outperforms the two regression models. Thechoice of the mathematically optimal threshold limit leads to a large number of falsealerts. Requiring three consecutive values outside the threshold before an alert isissued, increases the possibility to receive an early alert compared to the standardapproach where observations are classified individually.

Keywords [en]
dam safety; dams; data analysis; monitoring; statistical models
National Category
Civil Engineering
Research subject
Civil and Architectural Engineering, Concrete Structures
Identifiers
URN: urn:nbn:se:kth:diva-259718OAI: oai:DiVA.org:kth-259718DiVA, id: diva2:1353194
Note

QC 20190925

Available from: 2019-09-20 Created: 2019-09-20 Last updated: 2019-09-25Bibliographically approved
In thesis
1. Condition assessment of concrete dams in cold climate
Open this publication in new window or tab >>Condition assessment of concrete dams in cold climate
2019 (English)Licentiate thesis, comprehensive summary (Other academic)
Alternative title[sv]
Tillståndsbedömning av betongdammar i kallt klimat
Abstract [en]

Dams in many countries are approaching their expected service life. Proper assessment of the aging dams structural health increase the knowledge of the current safety, and allow for better planning of renovation and rebuilding investments. The behavior of concrete dams is, to a great extent, governed by the ambient variation in temperature and water level. In cold regions, the ice sheet formed in the reservoir may subject a pressure load on the dams. Theoretically, this load has a significant impact on the structural behavior of dams. Despite this, the maximum magnitude, as well as the seasonal variation of the ice load, constitute the most considerable uncertainty in the safety assessment of dams.

This thesis presents research that examines how to model the expected behavior of dams in cold climate. The underlying problem is to predict the response of dams due to variation in the external conditions. Since the ice load is such a vital part of the external conditions in cold climate, the understanding and modeling of ice loads have been given extra attention. Models suitable to predict the long-term behavior of dams can be divided between theoretical, data-based, and hybrid. Prediction accuracy is essential to set alert thresholds, and in that regard, the data-based models are generally superior.

The major contribution of this thesis is the design and installation of a prototype ice load panel with direct measurement of the ice pressure acting on a dam. The panel is attached on the upstream face of the dam and is large enough so that the whole thickness of the ice sheet is in contact with the panel. The predicted ice load from the best available model that includes loads from both thermal events and water level changes did not correspond to the measured ice loads. As there are no validated models or measurement methods for ice load on the dam, continued research is necessary, both through further measurements to increase knowledge and development of models.

Abstract [sv]

I många länder närmar sig vattenkraftsdammarna deras förväntade tekniska livslängd. Korrekt utvärdering av dammens strukturella status ökar kunskapen om det nuvarande säkerhetsläget och möjliggör för bättre planering av renoveringar och ombyggnadsinvesteringar. Betongdammarnas beteende styrs till stor del av variationen i omgivande temperaturer och vattennivå. I kalla regioner kan is som bildas i magasinet utsätta dammen för en tryckande last. Teoretiskt har denna belastning en betydande inverkan på dammarsstrukturella beteende. Trots detta är den maximala storleken såväl som säsongsvariationen för islasten en av de mest betydande osäkerheterna vidsäkerhetsbedömningen av dammar.

Denna uppsats presenterar forskning som undersöker hur förväntat beteendehos dammar i kallt klimat kan modelleras. Den underliggande frågeställningen är att förutsäga en damms respons orsakad av variation i de yttre förhållandena. Extra fokus har lagts på förståelsen och modelleringen av islasten då den är en viktig del av de yttre förhållandena för dammar i kalla klimat. Modeller som är lämpliga för att förutsäga dammars beteende kan delas upp i teoretiska, databaserad, och hybridmodeller. Förutsägbarhetsnoggrannhet är avgörande för att ställa in varningsnivåer, och i det avseendet är de databaserade modellerna i allmänhet överlägsna.

Det huvudsakliga bidraget från detta projekt är utvecklandet och installationen av en prototyp av en islast panel, fäst på en damms uppströmssida. Panelen mäter istrycket direkt mot dammen och är tillräckligt stor så att helaisens tjocklek förblir i kontakt med panelen. Panelen bidrar till ökad kunskap om säsongsvariationen i istryck och mekanismen för islaster orsakad av variation i vattennivån. Den predikterade islaten från den bästa tillgängliga islastmodellen som inkluderar termiska islaster och islaster från vattennivåförändringar överensstämmer inte med de uppmätta islastpanelen som uppmätts med panelen. Då det saknas validerade modeller eller mätmetoder för islast är det viktigt med fortsatt forskning, dels genom ytterligare mätningar för att öka kunskapen men även fortsatt modellutveckling.

Place, publisher, year, edition, pages
Stockholm: KTH Royal Institute of Technology, 2019. p. 104
Series
TRITA-ABE-DLT ; 1934
Keywords
dam safety, dams, data analysis, monitoring, statistical models, hydraulic structures, static ice loads, concrete dams, cold regions
National Category
Civil Engineering Civil Engineering
Research subject
Civil and Architectural Engineering, Concrete Structures
Identifiers
urn:nbn:se:kth:diva-259719 (URN)978-91-7873-301-9 (ISBN)
Presentation
2019-10-23, B25, Brinellvägen 23, Stockholm, 10:00 (English)
Opponent
Supervisors
Note

QC 20190926

Available from: 2019-09-26 Created: 2019-09-25 Last updated: 2019-09-26Bibliographically approved

Open Access in DiVA

No full text in DiVA

Authority records BETA

Hellgren, RikardMalm, RichardAnsell, Anders

Search in DiVA

By author/editor
Hellgren, RikardMalm, RichardAnsell, Anders
By organisation
Concrete Structures
Civil Engineering

Search outside of DiVA

GoogleGoogle Scholar

urn-nbn

Altmetric score

urn-nbn
Total: 36 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf