pith. sign in

arxiv: 2606.05026 · v1 · pith:K4HCRZDXnew · submitted 2026-06-03 · 📊 stat.ME · stat.AP

Removal of Multivariate Environmental Influences in Structural Health Monitoring through Conditional Covariances and Supervised Learning

Lizzie Neumann , Philipp Wittenberg , Jan Gertheiss This is my paper

Pith reviewed 2026-06-28 04:54 UTC · model grok-4.3

classification 📊 stat.ME stat.AP
keywords structural health monitoringconditional covarianceenvironmental confoundingsupervised learningresponse surface modelingbridge monitoringeigenfrequencies
0
0 comments X

The pith

Environmental conditions alter not only means but also covariances among SHM sensor outputs, which standard response surface models cannot remove.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that environmental factors such as temperature and load change the variances and correlations between multiple sensor readings in structural health monitoring, beyond any shifts in average values. Standard supervised regression techniques used to subtract environmental effects only handle the means and therefore leave residual confounding in the second-order structure. The authors introduce four supervised learning approaches to estimate the full conditional covariance matrix of the outputs given the environmental covariates. These estimated matrices are then applied within an SHM workflow to produce corrected data. Evaluation on both simulated examples and measurements from two instrumented bridges demonstrates that the corrections can be obtained and inserted into existing pipelines.

Core claim

Environmental influences on SHM data extend to the covariance structure of the multivariate outputs, and this structure can be estimated as a function of the environmental variables by kernel regression, random forests, additive models, or deep networks; the resulting conditional covariance matrices, once subtracted or normalized, remove the higher-order confounding that mean-only response surface models leave behind.

What carries the argument

Supervised estimators (kernel, random forest, additive, deep learning) that map environmental features to the full conditional covariance matrix of the vector of sensor outputs.

If this is right

  • The corrected data fed into damage-detection algorithms will contain fewer spurious changes driven by weather or traffic.
  • Eigenfrequency vectors and multi-location strain readings both become usable after the same covariance adjustment.
  • Method performance can be ranked on artificial data before deployment on the Vahrendorfer Stadtweg and KW51 bridges.
  • Existing SHM pipelines require only an added conditional-covariance step rather than new hardware.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same conditional-covariance step could be inserted into any multivariate monitoring task where correlations carry diagnostic information.
  • Extending the approach to conditional skewness or kurtosis might capture still higher-order environmental effects.
  • Real-time updating of the covariance estimator could support online SHM under continuously changing conditions.

Load-bearing premise

The chosen supervised learning methods recover the true dependence of output covariances on environmental conditions from the observed training samples.

What would settle it

Apply the conditional-covariance correction to a dataset from an undamaged bridge under documented environmental variation and check whether the corrected covariance matrices remain statistically indistinguishable from those recorded under a fixed reference environment.

Figures

Figures reproduced from arXiv: 2606.05026 by Jan Gertheiss, Lizzie Neumann, Philipp Wittenberg.

Figure 1
Figure 1. Figure 1: Schematic drawing of a neural network with multiple inputs (green), two hidden layers (blue), and a multi-dimensional output layer (red). To estimate the conditional covariance of x at z using neural networks, the random vector y from Eq. (6) is used, and the conditional mean of y at z is estimated. The neural network is constructed as follows yˆ = fm(Wmfm−1(. . . f1(W1f0(W0z))), (8) with activation functi… view at source ↗
Figure 2
Figure 2. Figure 2: Simulation study setup. Right: conditional mean µ1(z) (top) and µ2(z) (bottom). Middle: conditional variance σ 2 1(z) (top) and σ 2 2(z) (bottom). Right: conditional covariance σ1,2(z) (top) and resulting correlation (bottom). and Wiener, 2002) is used for the conditional mean, and the CovRegRF (Alakus et al., 2023) is used with 100 trees to estimate the conditional covariance. For the additive models, the… view at source ↗
Figure 3
Figure 3. Figure 3: Conditional correlation as a function of the two covariates z1 and z2, and the corresponding estimates on a single dataset. Top: original/true correlation function (left), Nadaraya-Watson (middle), random forest (right). Bottom: additive model (left), additive model with interactions (middle), neural network (right). In [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Root mean square error (RMSE) of the conditional covariance estimates for the five considered methods over 50 simulation runs. additive models performed best, followed by the neural network and the Nadaraya-Watson kernel estimator. 12 [PITH_FULL_IMAGE:figures/full_fig_p012_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Vahrendorfer Stadtweg bridge from the west side (top left) and with three additional masses (top right). An overview of the sensor placement is found at the bottom. Load tests on the bridge were conducted from February 22nd to March 23rd, 2024. These tests used three large sand-filled bags weighing 680 kg, 740 kg, and 740 kg, positioned at the center of the bridge. Each additional weight was applied for te… view at source ↗
Figure 6
Figure 6. Figure 6: Conditional correlation (left columns) and mean (right columns) of the Vahrendorfer Stadtweg bridge given the material and asphalt temperature (x-axis) and the humidity (y-axis) using (top to bottom) the Nadaraya￾Watson kernel estimator, a random forest, an additive model, an additive model with interactions, and a neural network. 15 [PITH_FULL_IMAGE:figures/full_fig_p015_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Squared conditional Mahalanobis distances for the additive model (top) and the additive model with interactions (bottom), shown on a logarithmic y-axis. The horizontal purple line denotes the empirical control limit, defined as the 99.9 % quantile of the Phase-I reference values. Threshold crossings are highlighted as Phase-I reference exceedances, Phase-IIa false-positive alarm time points, and Phase-IIb … view at source ↗
Figure 8
Figure 8. Figure 8: Squared conditional Mahalanobis distances obtained with the five methods for the KW51 bridge eigenfre￾quency data, shown on a logarithmic y-axis. The horizontal purple line denotes the empirical control limit, defined as the 99.9 % (left) and 99.5 % (right) quantile of the Phase-I reference values. The dashed vertical line marks the start of Phase II, and the solid vertical line marks the start of the retr… view at source ↗
read the original abstract

In structural health monitoring (SHM) systems, data is collected from a multitude of sensors measuring, for example, vibration or strain in the structure, along with additional features that capture environmental or operational information. It is well known that changes in the measured sensor outputs do not necessarily originate from structural damage but are often induced by environmental changes. One popular approach to account for these effects is regressing the system outputs on the confounding factors, also known as "response surface modeling". Afterward, the predicted values are subtracted from the observed ones to obtain corrected data with the environmental effects (supposedly) removed. However, the evaluation of real-world SHM data shows that environmental conditions may affect not only the expected output values but also higher-order statistical moments, particularly the variances of and the covariances and correlations between the output quantities, such as eigenfrequencies of different modes or strain sensors at different locations. By construction, the (supervised) machine learning techniques commonly used for response surface modeling cannot account for those higher-order effects. To address these issues, we present and discuss several approaches for identifying and quantifying multivariate confounding effects on output covariances and correlations: a nonparametric, kernel-based estimator, a random forest, a semiparametric additive model, and a deep learning approach. Furthermore, we show how the resulting conditional covariance matrices can be used in an SHM pipeline. We compare the competing methods on both artificial data and real-world load test data from the Vahrendorfer Stadtweg bridge in Hamburg, Germany, as well as eigenfrequency data from the railway bridge KW51 near Leuven, Belgium.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 0 minor

Summary. The paper claims that environmental conditions in SHM affect not only means but also variances and covariances of sensor outputs, which standard response-surface modeling cannot capture. It proposes four supervised learning estimators (kernel, random forest, additive model, deep learning) for conditional covariance matrices and shows how to integrate them into an SHM pipeline, with comparisons performed on artificial data and real load-test data from the Vahrendorfer bridge and KW51 railway bridge.

Significance. If the conditional-covariance estimates demonstrably improve downstream damage detection, the approach would extend response-surface methods to higher-order moments and could reduce false alarms in operational SHM. The explicit comparison of four distinct estimators on both synthetic and field data is a constructive contribution.

major comments (1)
  1. [Abstract] Abstract: the manuscript states that the four estimators are compared on the Vahrendorfer and KW51 datasets and that the resulting conditional covariance matrices are used in an SHM pipeline, yet reports no quantitative metrics (change in false-alarm rate, ROC-AUC shift, Mahalanobis separation, or similar) that would confirm a detectable gain over mean-only correction. This validation step is load-bearing for the central claim that the methods 'meaningfully improve' the pipeline.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive review and positive assessment of the paper's potential contribution. We address the single major comment below and will incorporate the suggested evaluation in a revised version.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the manuscript states that the four estimators are compared on the Vahrendorfer and KW51 datasets and that the resulting conditional covariance matrices are used in an SHM pipeline, yet reports no quantitative metrics (change in false-alarm rate, ROC-AUC shift, Mahalanobis separation, or similar) that would confirm a detectable gain over mean-only correction. This validation step is load-bearing for the central claim that the methods 'meaningfully improve' the pipeline.

    Authors: We agree that quantitative metrics demonstrating improvement in downstream damage detection (e.g., false-alarm rates or separation measures when using conditional covariances versus mean-only response surface correction) would strengthen the practical relevance of the work. The current manuscript focuses on the estimation accuracy of the conditional covariance matrices themselves (via comparisons on synthetic and field data) and illustrates their use in an SHM pipeline through the Mahalanobis distance, without performing a full damage-injection study. We will revise the manuscript to add such an evaluation on the Vahrendorfer and KW51 datasets, including simulated damage scenarios to report changes in detection performance. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper applies four standard supervised learning estimators (kernel, random forest, additive model, deep learning) to predict conditional covariance matrices from environmental covariates, then inserts those matrices into an SHM pipeline. None of the estimators is defined in terms of the downstream SHM metric, no prediction is obtained by fitting a parameter to a subset and renaming the fit, and no uniqueness theorem or ansatz is imported via self-citation. The artificial-data and real-bridge comparisons constitute independent checks rather than tautological reductions. The derivation chain therefore remains self-contained.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The paper relies on standard statistical assumptions underlying kernel estimation, random forests, additive models, and neural networks for conditional covariance estimation; no new free parameters, axioms, or invented entities are introduced beyond typical machine-learning hyperparameters and the domain assumption that environmental features are observed.

free parameters (1)
  • model hyperparameters
    Random forest, deep learning, and additive models require hyperparameter choices that are typically tuned on data.
axioms (1)
  • domain assumption Environmental and operational variables are observed and serve as valid predictors for conditional second-moment statistics.
    The methods presuppose access to confounding features that can be used to condition the covariance estimation.

pith-pipeline@v0.9.1-grok · 5827 in / 1273 out tokens · 34264 ms · 2026-06-28T04:54:02.593013+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

59 extracted references · 51 canonical work pages · 2 internal anchors

  1. [1]

    Alakus, D

    C. Alakus, D. Larocque, and A. Labbe. Covariance regression with random forests. BMC Bioinformatics, 24 0 (258), 2023. doi:10.1186/s12859-023-05377-y

    work page doi:10.1186/s12859-023-05377-y 2023

  2. [2]

    Bezanson, A

    J. Bezanson, A. Edelman, S. Karpinski, and V. B. Shah. Julia: A fresh approach to numerical computing. SIAM R eview , 59 0 (1): 0 65--98, 2017. doi:10.1137/141000671

    work page doi:10.1137/141000671 2017

  3. [3]

    C. M. Bishop. Pattern Recognition and Machine Learning. Springer New York, NY, 2006. ISSN 978-1-4939-3843-8

    2006

  4. [4]

    Conn and G

    D. Conn and G. Li. An oracle property of the Nadaraya–Watson kernel estimator for high-dimensional nonparametric regression . Scandinavian Journal of Statistics, 46 0 (3): 0 735--764, 2019. doi:10.1111/sjos.12370

    work page doi:10.1111/sjos.12370 2019

  5. [5]

    Cross, K

    E. Cross, K. Koo, J. Brownjohn, and K. Worden. Long-term monitoring and data analysis of the Tamar Bridge . Mechanical Systems and Signal Processing, 35 0 (1-2): 0 16--34, 2013. doi:10.1016/j.ymssp.2012.08.026

    work page doi:10.1016/j.ymssp.2012.08.026 2013

  6. [6]

    E. J. Cross, K. Worden, and Q. Chen. Cointegration: a novel approach for the removal of environmental trends in structural health monitoring data. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 467 0 (2133): 0 2712--2732, 2011. doi:10.1098/rspa.2011.0023

    work page doi:10.1098/rspa.2011.0023 2011

  7. [7]

    E. J. Cross, G. Manson, K. Worden, and S. G. Pierce. Features for damage detection with insensitivity to environmental and operational variations. In Proceedings of the Royal Society A, volume 486, pages 4098--4122, 2012. doi:10.1098/rspa.2012.0031

    work page doi:10.1098/rspa.2012.0031 2012

  8. [8]

    A. Cury, C. Cremona, and J. Dumoulin. Long-term monitoring of a PSC box girder bridge: Operational modal analysis, data normalization and structural modification assessment . Mechanical Systems and Signal Processing, 33: 0 13--37, 2012. doi:10.1016/j.ymssp.2012.07.005

    work page doi:10.1016/j.ymssp.2012.07.005 2012

  9. [9]

    de Ville

    B. de Ville. Decision trees. WIREs Computational Statistics, 5 0 (6): 0 448--455, 2013. doi:10.1002/wics.1278

    work page doi:10.1002/wics.1278 2013

  10. [10]

    F. Deng, S. Wei, X. Jin, Z. Chen, and H. Li. Damage identification of long-span bridges based on the correlation of probability distribution of monitored quasi-static responses. Mechanical Systems and Signal Processing, 186: 0 109908, 2023. doi:10.1016/j.ymssp.2022.109908

    work page doi:10.1016/j.ymssp.2022.109908 2023

  11. [11]

    Dervilis and K

    N. Dervilis and K. Worden and E.J. Cross. On robust regression analysis as a means of exploring environmental and operational conditions for SHM data. Journal of Sound and Vibration, 347: 0 279-296, 2015. doi:10.1016/j.jsv.2015.02.039

    work page doi:10.1016/j.jsv.2015.02.039 2015

  12. [12]

    Farrar and K

    C. Farrar and K. Worden. Structural Health Monitoring: A Machine Learning Perspective. John Wiley & Sons, Ltd, 2013. doi:10.1002/9781118443118

    work page doi:10.1002/9781118443118 2013

  13. [13]

    C. R. Farrar, P. J. Cornwell, S. W. Doebling, and M. B. Prime. Structural health monitoring studies of the alamosa canyon and i-40 bridges. Technical report, Los Alamos National Lab.(LANL), Los Alamos, NM (United States), 2000

    2000

  14. [14]

    Flexa, W

    C. Flexa, W. Gomes, and C. Sales. Data normalization in structural health monitoring by means of nonlinear filtering. In 2019 8th Brazilian Conference on Intelligent Systems (BRACIS), pages 204--209, 2019. doi:10.1109/BRACIS.2019.00044

    work page doi:10.1109/bracis.2019.00044 2019

  15. [15]

    J. H. Friedman. Regularized discriminant analysis. Journal of the American Statistical Association, 84 0 (405): 0 165--175, 1989. doi:10.1080/01621459.1989.10478752

    work page doi:10.1080/01621459.1989.10478752 1989

  16. [16]

    Gasser and H.-G

    T. Gasser and H.-G. M \"u ller. Kernel estimation of regression functions. In T. Gasser and M. Rosenblatt, editors, Smoothing Techniques for Curve Estimation, pages 23--68, Berlin, Heidelberg, 1979. Springer Berlin Heidelberg. ISBN 978-3-540-38475-5

    1979

  17. [17]

    Q. Han, Q. Ma, J. Xu, and M. Liu. Structural health monitoring research under varying temperature condition: a review. Journal of Civil Structural Health Monitoring, 11: 0 149--173, 2021. doi:10.1007/s13349-020-00444-x

    work page doi:10.1007/s13349-020-00444-x 2021

  18. [18]

    Hotelling

    H. Hotelling. Multivariate quality control illustrated by the air testing of sample bombsights. In C. Eisenhart, M. Hastay, and W. Wallis, editors, Techniques of Statistical Analysis, pages 111--184, New York, 1947. McGraw-Hill

    1947

  19. [19]

    Huang, S.-J

    J.-z. Huang, S.-J. Yuan, D. sheng Li, and H. nan Li. A kernel canonical correlation analysis approach for removing environmental and operational variations for structural damage identification. Journal of Sound and Vibration, 548: 0 117516, 2023. doi:10.1016/j.jsv.2022.117516

    work page doi:10.1016/j.jsv.2022.117516 2023

  20. [20]

    M. Innes. Flux: Elegant machine learning with Julia . Journal of Open Source Software, 3 0 (25): 0 602, 2018. doi:10.21105/joss.00602

    work page doi:10.21105/joss.00602 2018

  21. [21]

    Fashionable Modelling with Flux

    M. Innes, E. Saba, K. Fischer, D. Gandhi, M. C. Rudilosso, N. M. Joy, T. Karmali, A. Pal, and V. Shah. Fashionable modelling with flux. CoRR, abs/1811.01457, 2018. doi:10.48550/arXiv.1811.01457

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1811.01457 2018

  22. [22]

    Jesus, P

    A. Jesus, P. Brommer, R. Westgate, K. Koo, J. Brownjohn, and I. Laory. Bayesian structural identification of a long suspension bridge considering temperature and traffic load effects. Structural Health Monitoring, 18 0 (4): 0 1310--1323, 2019. doi:10.1177/1475921718794299

    work page doi:10.1177/1475921718794299 2019

  23. [23]

    Keshmiry, S

    A. Keshmiry, S. Hassani, M. Mousavi, and U. Dackermann. Effects of environmental and operational conditions on structural health monitoring and non-destructive testing: A systematic review. Buildings, 13 0 (4), 2023. doi:10.3390/buildings13040918

    work page doi:10.3390/buildings13040918 2023

  24. [24]

    S. Kim, H. Kim, and Y. Hwang. Data-driven dynamic response forecasting and anomaly detection in long-span bridges. Journal of Civil Structural Health Monitoring, 15: 0 3045--3062, 2025. doi:10.1007/s13349-025-00952-8

    work page doi:10.1007/s13349-025-00952-8 2025

  25. [25]

    Koehncke, Y

    M. Koehncke, Y. Jaelani, A. Mendler, L. Neumann, P. Wittenberg, A. Rode-Klemm, J. Grashorn and S. Kessler. Data Set of Load Tests and Structural Health Monitoring of a concrete boxgirder bridge . Zenodo, mar 2026 b . doi:10.5281/zenodo.20522191

    work page doi:10.5281/zenodo.20522191 2026

  26. [26]

    Koehncke, Y

    M. Koehncke, Y. Jaelani, A. Mendler, L. Neumann, P. Wittenberg, A. Rode-Klemm, and S. Kessler. Data set of load tests and structural health monitoring of a concrete boxgirder bridge. arXiv preprint arXiv:2412.15713, 2026 a . doi:10.48550/arXiv.2412.15713

    work page doi:10.48550/arxiv.2412.15713 2026

  27. [27]

    Kumar, P

    K. Kumar, P. Kumar Biswas, and N. Dhang. Time series-based SHM using PCA with application to ASCE benchmark structure. Journal of Civil Structural Health Monitoring, 10: 0 899--911, 2020. doi:10.1007/s13349-020-00423-2

    work page doi:10.1007/s13349-020-00423-2 2020

  28. [28]

    Liaw and M

    A. Liaw and M. Wiener. Classification and Regression by randomForest . R News, 2 0 (3): 0 18--22, 2002. URL https://CRAN.R-project.org/doc/Rnews/

    2002

  29. [29]

    C. A. Lowry, W. H. Woodall, C. W. Champ, and S. E. Rigdon. A Multivariate Exponentially Weighted Moving Average Control Chart . Technometrics, 34 0 (1): 0 46--53, 1992. doi:10.2307/1269551

    work page doi:10.2307/1269551 1992

  30. [30]

    López Cabrera and F

    B. López Cabrera and F. Schulz. Volatility linkages between energy and agricultural commodity prices. Energy Economics, 54: 0 190--203, 2016. doi:10.1016/j.eneco.2015.11.018

    work page doi:10.1016/j.eneco.2015.11.018 2016

  31. [31]

    J. M. Lucas and M. S. Saccucci. Exponentially weighted moving average control schemes: Properties and enhancements. Technometrics, 32 0 (1): 0 1--12, 1990. doi:10.1080/00401706.1990.10484583

    work page doi:10.1080/00401706.1990.10484583 1990

  32. [32]

    Maes and G

    K. Maes and G. Lombaert. Monitoring data for railway bridge KW51 in Leuven, Belgium, before, during, and after retrofitting. V1.0 . 2020. doi:10.5281/zenodo.3745914

    work page doi:10.5281/zenodo.3745914 2020

  33. [33]

    Maes and G

    K. Maes and G. Lombaert. Monitoring railway bridge KW51 before, during, and after retrofitting. Journal of Bridge Engineering, 26 0 (3): 0 04721001, 2021. doi:10.1061/(ASCE)BE.1943-5592.0001668

    work page doi:10.1061/(asce)be.1943-5592.0001668 2021

  34. [34]

    K. Maes, L. Van Meerbeck , E. Reynders and G. Lombaert. Validation of vibration-based structural health monitoring on retrofitted railway bridge KW51. Mechanical Systems and Signal Processing, 165 0 (3): 0 108380, 2021. doi:10.1016/j.ymssp.2021.108380

    work page doi:10.1016/j.ymssp.2021.108380 2021

  35. [35]

    Magalhães, A

    F. Magalhães, A. Cunha, and E. Caetano. Vibration based structural health monitoring of an arch bridge: From automated OMA to damage detection . Mechanical Systems and Signal Processing, 28: 0 212--228, 2012. doi:10.1016/j.ymssp.2011.06.011

    work page doi:10.1016/j.ymssp.2011.06.011 2012

  36. [36]

    Mu, Z.-J

    H.-Q. Mu, Z.-J. Zheng, X.-H. Wu, and C. Su. Bayesian network-based modal frequency–multiple environmental factors pattern recognition for the Xinguang Bridge using long-term monitoring data . Journal of Low Frequency Noise, Vibration and Active Control, 39 0 (3): 0 545--559, 2020. doi:10.1177/1461348418786520

    work page doi:10.1177/1461348418786520 2020

  37. [37]

    Neumann, P

    L. Neumann, P. Wittenberg, A. Mendler, and J. Gertheiss. Confounder-adjusted covariances of system outputs and applications to structural health monitoring. Mechanical Systems and Signal Processing, 224, 2025. doi:10.1016/j.ymssp.2024.111983

    work page doi:10.1016/j.ymssp.2024.111983 2025

  38. [38]

    Higher-Order Multivariate Environmental Influences in Structural Health Monitoring

    L. Neumann, P. Wittenberg, and J. Gertheiss. Higher-order multivariate environmental influences in structural health monitoring. arXiv preprint arXiv:2603.22950, 2026 a . doi:10.48550/arXiv.2603.229503

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2603.229503 2026

  39. [39]

    Neumann, P

    L. Neumann, P. Wittenberg, A. Mendler, and J. Gertheiss. Signal and feature reconstruction in structural health monitoring under varying environmental conditions. Engineering Structures, 361: 0 122865, 2026 b . doi:10.1016/j.engstruct.2026.122865

    work page doi:10.1016/j.engstruct.2026.122865 2026

  40. [40]

    Y. Ni, X. Hua, K. Fan, and J. Ko. Correlating modal properties with temperature using long-term monitoring data and support vector machine technique. Engineering Structures, 27 0 (12): 0 1762--1773, 2005. doi:10.1016/j.engstruct.2005.02.020

    work page doi:10.1016/j.engstruct.2005.02.020 2005

  41. [41]

    E. Page. Continuous inspection schemes. Biometrika, 41 0 (1-2): 0 100--115, 1954. doi:10.1093/biomet/41.1-2.100

    work page doi:10.1093/biomet/41.1-2.100 1954

  42. [42]

    Petersen, S

    A. Petersen, S. Deoni, and H.-G. M \"u ller. Fréchet estimation of time-varying covariance matrices from sparse data, with application to the regional co-evolution of myelination in the developing brain . The Annals of Applied Statistics, 13 0 (1): 0 393--419, 2019. doi:10.1214/18-AOAS1195

    work page doi:10.1214/18-aoas1195 2019

  43. [43]

    R: A Language and Environment for Statistical Computing

    R Core Team . R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, 2025. URL https://www.R-project.org/

    2025

  44. [44]

    Reynders, G

    E. Reynders, G. Wursten, and G. D. Roeck. Output-only structural health monitoring in changing environmental conditions by means of nonlinear system identification. Structural Health Monitoring, 13 0 (1): 0 82--93, 2014. doi:10.1177/1475921713502836

    work page doi:10.1177/1475921713502836 2014

  45. [45]

    Rezazadeh, A

    N. Rezazadeh, A. De Luca, D. Perfetto, M. R. Salami, and G. Lamanna. Systematic critical review of structural health monitoring under environmental and operational variability: approaches for baseline compensation, adaptation, and reference-free techniques. Smart Materials and Structures, 34 0 (7): 0 073001, 2025. doi:10.1088/1361-665X/ade7db

    work page doi:10.1088/1361-665x/ade7db 2025

  46. [46]

    E., Hinton, G

    D. Rumelhart, G. Hinton, and R. Williams. Learning representations by back-propagating errors. Nature, 323: 0 533–536, 1986. doi:10.1038/323533a0

    work page doi:10.1038/323533a0 1986

  47. [47]

    H. Sohn, M. Dzwonczyk, E. G. Straser, A. S. Kiremidjian, K. H. Law, and T. Meng. An experimental study of temperature effect on modal parameters of the Alamosa Canyon Bridge . Earthquake Engineering & Structural Dynamics, 28 0 (8): 0 879--897, 1999. 10.1002/(SICI)1096-9845(199908)28:8\

    work page doi:10.1002/(sici)1096-9845(199908)28:8 1999

  48. [48]

    D. A. Tibaduiza, L. E. Mujica, J. Rodellar, and A. Güemes. Structural damage detection using principal component analysis and damage indices. Journal of Intelligent Material Systems and Structures, 27 0 (2): 0 233--248, 2016. doi:10.1177/1045389X14566520

    work page doi:10.1177/1045389x14566520 2016

  49. [49]

    Viefhues, M

    E. Viefhues, M. Döhler, F. Hille, and L. Mevel. Fault detection for linear parameter varying systems under changes in the process noise covariance. IFAC-PapersOnLine, 53 0 (2): 0 13668--13673, 2020. doi:10.1016/j.ifacol.2020.12.868. 21st IFAC World Congress

    work page doi:10.1016/j.ifacol.2020.12.868 2020

  50. [50]

    Wang, D.-H

    Z. Wang, D.-H. Yang, T.-H. Yi, G.-H. Zhang, and J.-G. Han. Eliminating environmental and operational effects on structural modal frequency: A comprehensive review. Structural Control and Health Monitoring, 29 0 (11): 0 e3073, 2022. doi:10.1002/stc.3073

    work page doi:10.1002/stc.3073 2022

  51. [51]

    Y. Wei, Z. Chen, Z.-S. Ye, and E. Pan. High-dimensional process monitoring under time-varying operating conditions via covariate-regulated principal component analysis. Reliability Engineering & System Safety, 252: 0 110440, 2024. doi:10.1016/j.ress.2024.110440

    work page doi:10.1016/j.ress.2024.110440 2024

  52. [52]

    T. R. Willemain and G. C. Runger. Designing control charts using an empirical reference distribution. Journal of Quality Technology, 28 0 (1): 0 31--38, 1996. doi:10.1080/00224065.1996.11979634

    work page doi:10.1080/00224065.1996.11979634 1996

  53. [53]

    Wittenberg, L

    P. Wittenberg, L. Neumann, K. Maes, and J. Gertheiss. Covariate-Dependent Functional Principal Component Analysis for SHM . arXiv preprint arXiv:2603.18279, 2026. doi:10.48550/arXiv.2603.18279

    work page doi:10.48550/arxiv.2603.18279 2026

  54. [54]

    S. Wood. Generalized Additive Models: An Introduction with R. Chapman and Hall/CRC, 2 edition, 2017

    2017

  55. [55]

    S. N. Wood. Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 73 0 (1): 0 3--36, 2011. doi:10.1111/j.1467-9868.2010.00749.x

    work page doi:10.1111/j.1467-9868.2010.00749.x 2011

  56. [56]

    Worden, G

    K. Worden, G. Manson, and C. Surace. Aspects of novelty detection. Key Engineering Materials, 347: 0 3--16, 2007. doi:10.4028/www.scientific.net/KEM.347.3

    work page doi:10.4028/www.scientific.net/kem.347.3 2007

  57. [57]

    J. Yin, Z. Geng, R. Li, and H. Wang. Nonparametric covariance model. Statistica Sinica 20, pages 469--479, 2010

    2010

  58. [58]

    Zhou and F

    Y.-H. Zhou and F. A. Wright. Hypothesis testing at the extremes: fast and robust association for high-throughput data. Biostatistics, 16: 0 611--625, 2015. doi:10.1093/biostatistics/kxv007

    work page doi:10.1093/biostatistics/kxv007 2015

  59. [59]

    Zhu, Y.-Q

    Y. Zhu, Y.-Q. Ni, D. Inaudi, and I. Laory. A temperature-driven MPCA method for structural anomaly detection. Engineering Structures, 190: 0 447--458, 2019. doi:10.1016/j.engstruct.2019.04.004

    work page doi:10.1016/j.engstruct.2019.04.004 2019