Numerical audit framework of failure modes in combination of multi-rate and GNSS accelerometer for three-component estimation of seismic displacement of dam crest
Numerical audit framework of failure modes in combination of multi-rate and GNSS accelerometer for three-component estimation of seismic displacement of dam crest
Zahra Rafatnejad1 Shamsedin Hashemi2
1) Bachelor s degree in Surveying, Faculty of Civil, Water and Environmental Engineering, Shahid Beheshti University,
2) Associate Professor, Faculty of Civil, Water and Environmental Engineering, Shahid Beheshti University,
IOI Code
:
ABGG-ABKKAZ
Publication :
11th.International Conference on Researches in Science & Engineering & 8th.International Congress on Civil, Architecture and Urbanism in Asia(11icrsie.ir)
Abstract :
The aim of this study is to develop and numerically evaluate a dam-based multi-rater integration of GNSS and accelerometers in the three-component estimation of the seismic displacement of the crest of an ideal gravity concrete dam. The reference model is a reduced modal criterion and should not be interpreted as a three-component finite element model of a real dam. Twelve designed parametric cases with varying excitation intensity, modal frequencies, and damping were investigated; the accelerometer was simulated at a rate of 50 Hz and the GNSS at a rate of 10 Hz, and 120 pairwise runs were performed for each main scenario and 360 runs for the saturation sequence analysis. In the nominal case, the P-KF-9 achieved a three-dimensional RMSE of 4.28 mm, NEES=3.39, and empirical coverage of 92.4%; This combination indicates a slight underdispersion of uncertainty, not a complete calibration. The ratio NEES/3=1.13 was reported only as a posteriori diagnostic indicator, was not applied in the main results, and is not interpreted as a correction to the validated calibration. In the GNSS outlier scenario, IR-HKF-9 and MH-KF-9 reduced the mean RMSE relative to P-KF-9 by 1.15 and 1.23 mm, corresponding to about 20.4% and 21.9%, respectively, and both pairwise comparisons were significant (p<0.001). In the mixed contamination, the error distribution became tail-heavy, and no significant superiority was confirmed for IR-HKF-9 over P-KF-9 (p=0.111), indicating that the robustness of the observation stage to contamination from the forecast path alone is not sufficient. In the 360-run clipping test, simple quality control reduced Q95 and CVaR95 in some configurations, but did not uniformly improve the mean error and large error rate. Therefore, the main contribution of the study is the disaggregated audit of performance boundaries and failure modes within the benchmark. This study is a controlled numerical benchmark and is not equivalent to field validation, shake table validation, or operational qualification of a real dam monitoring system.
Keywords : gnss accelerometer robust kalman multi rate fusion failure mode auditing