Section: New Results
Handling the temperature effect
Participants : Michael Döhler, Maurice Goursat, Laurent Mevel.
See modules 3.4 , 3.5 and 4.3 .
The PhD thesis of Houssein Nasser completed in 2006 has addressed the problem of rejecting the temperature effect when performing damage detection tests on civil structures. Because of the temperature effects, the test may not react to some damages, and conversely may be too sensitive to some ambient temperature changes.
Some approach is based on the collection of varying reference temperature datasets, and on a reference kernel averaging all the temperature scenarios [25] . This is the topic of the following case studies.
Temperature rejection : case study on a bridge
After the work on a beam under temperature and damages changes in I4S and LCPC [53] , this year, the actual robustness to temperature changes of the proposed methods has been further investigated within LCPC by Dominique Siegert under the supervision of I4S [23] . The extreme reaction of the test to summer/winter and its robustness to small temperature changes have been shown. This work presents the processing methods and the analysis results of ambient vibration data recorded during a six-month period on a highway bridge, the Roberval bridge in France. Data analysis was focused on the variations of the modal parameter related to the accuracy of the estimates and to the temperature effects. The first flexural and torsional modes were estimated with their variance from short acceleration time series records using stochastic covariance driven subspace identification techniques. The frequency variation estimates were compared to the variations induced by structural modifications simulated with a finite element model for assessing the detection threshold level. Subsequently the temperature induced variations in the measured frequencies were analysed. The effectiveness of the temperature robust version of Houssein Nasser Phd thesis was investigated with the available data records. Results of the damage detection test show the response to the temperature effects. Significant increases in the variance of the parameter estimates were also detected by the test. Confidence intervals based on analytical derivation of the subspace equation were compared with the evolution of the damage detection test with success, in collaboration with Katholieke Universiteit Leuven.
Temperature rejection : case study on a bridge mock up
Small localized damages are hardly detected by global monitoring methods. The effectiveness of vibration based detection depends on the accuracy of the modal parameter estimates and is limited by the low sensitivity of the modal parameters to a local stiffness reduction. A local reduction of stiffness related to frequency changes less than 1% was successfully detected on a 10 meter span composite UHPFRC - FRP reinforced timber beam bridge loaded in laboratory conditions up to the serviceability limit state (SLS). Such a small decrease in the stiffness was not detected by the monitoring of the static load-deflection measurements but was confirmed by non-linear local strain measurements. Statistical subspace-based damage detection successfully detected the change of the modal parameters of the investigated structure. Further analysis with a finite element model was conducted for assessing the consistency of the expected location and extent of the damaged elements. Work will be presented in [22] .
