Pith. sign in

REVIEW 3 major objections 4 minor 60 references

Data-Driven Modeling Approaches for Optimal Control and Control Co-Design of Floating Offshore Wind Turbines

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A wind-speed-scheduled linear surrogate of floating offshore wind turbine dynamics, trained on simulated state-derivative data, reproduces the controller design-space shape at nearly 48x lower simulation cost and beats subspace…

desk verdict A useful incremental surrogate method for FOWT controller optimization, with a credible 48x speedup but a design-space trend claim that needs quantitative support. read the letter →

arxiv 2505.14515 v2 pith:2M3SBJA6 submitted 2025-05-20 eess.SY cs.SY

Athul K. Sundarrajan , Daniel R. Herber This is my paper
classification eess.SYcs.SY
keywords derivativefunctionsurrogatemodellinearparameter-varyingsystemsfloatingoffshorewindturbinesdata-drivenlow-fidelitymodelingsystemidentificationlongshort-termmemorynetworksdamageequivalentloadcontrolleroptimization
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper proposes building a derivative function surrogate model (DFSM) for a floating offshore wind turbine by fitting a wind-speed-scheduled linear state-space model to state-derivative data extracted from high-fidelity simulations. The authors argue that this linear parameter-varying DFSM balances simulation time and prediction accuracy better than a subspace state-space identification baseline (n4sid) or an LSTM network, reducing a roughly 20-minute closed-loop high-fidelity simulation to about 25 seconds. They show that the DFSM reproduces the shape and trends of the tower-base damage-equivalent-load design space over pitch-controller gains, even though it underpredicts the absolute damage value. If correct, the approach makes multi-hundred-evaluation controller optimization studies for floating offshore wind turbines feasible on a personal computer rather than a computing cluster.

What carries the argument

The load-bearing machinery is the derivative function surrogate model built through a spline-derived identification pipeline. Simulated state trajectories are approximated by cubic splines, and the exact polynomial derivatives give approximate state-derivative time series. For each mean wind speed, an optimization problem fits the state matrices $(A(w), B(w))$ to minimize the derivative error subject to the constraint that all eigenvalues of $A(w)$ have negative real part, and a second least-squares problem fits $(C(w), D(w))$ to the measured outputs. Because wind-speed trajectories overlap, the converged solution for one mean wind speed seeds the next, so the identification cost drops from 460 seconds for the first wind speed to about 2.6 seconds for each neighbor. Interpolating these matrices across wind speed yields the continuous LPV plant that is then coupled with the reference open-source controller for closed-loop simulation.

What would settle it

Train the same DFSM on a floating turbine with a different platform or controller architecture and compare the damage-equivalent-load contour over controller gains with the high-fidelity simulator; if the contour shape changes or the predicted optimal gain moves outside the contour spacing, the LPV structure is insufficient. A sharper test is to simulate the DFSM-chosen optimal controller in the high-fidelity model and check whether its damage-equivalent load is close to the true optimum across several wind seeds.

Watch

Extended reading notes

Core claim

The central claim is that a continuous-time linear parameter-varying model, with six physically meaningful states (platform pitch, tower-top displacement, generator speed, and their first derivatives) and wind speed as the sole scheduling variable, can serve as an accurate low-fidelity plant for control studies. The system matrices $A(w)$, $B(w)$, $C(w)$, $D(w)$ are identified by minimizing the error between spline-extracted state derivatives and model predictions, with a stability constraint on the eigenvalues of $A(w)$, followed by a least-squares fit of the output map. In closed-loop tests the surrogate tracks blade pitch and generator power closely, captures the mean of the tower-base moment but underpredicts its range, and produces a damage-equivalent-load contour over controller gains whose shape matches the high-fidelity simulator even when the DFSM was trained at a different controller gain. The paper's quantitative result is a nearly 48-fold speedup, about 25 seconds per simulation instead of roughly 20 minutes, with lower response variance than n4sid and lower simulation time than the LSTM baseline.

Load-bearing premise

The turbine's dynamics are assumed to be fully represented by six measured states evolving under a linear model whose matrices depend only on the instantaneous wind speed, with no unmeasured internal states.

Editorial extensions

If this is right

  • A 25-point design-of-experiments scan of pitch-controller gains drops from roughly 250 CPU-hours with the high-fidelity simulator to about 5 CPU-hours with the DFSM, turning a cluster job into a workstation task.
  • The DFSM supplies closed-loop time series of blade pitch, generator power, and tower-base moment, so performance metrics such as damage equivalent load and annual energy production can be post-processed from its outputs.
  • The surrogate predicts responses for controller gains it was never trained on and preserves the shape of the DEL design space, making it usable for early-stage controller optimization before high-fidelity verification.
  • The DFSM is a continuous-time model with physically meaningful states, unlike the n4sid baseline whose states are abstract and whose required model order changes with wind speed, which blocks a direct LPV extension.
  • Training the DFSM takes about 509 seconds on top of roughly 11.6 CPU-hours of simulation data, comparable to LSTM training time but with much faster evaluation.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper, the same spline-derivative-plus-LPV pipeline should transfer to other expensive dynamic simulators, such as wave energy converters or marine hydrokinetic turbines, whenever the scheduling variable is measurable and the key states appear among the outputs.
  • The known shortfall—underprediction of the tower-base moment range—suggests a testable extension: keep the linear state equation but add a small static nonlinear correction to the output map; if the DEL contour improves without losing the speedup, the output map, not the derivative identification, was the limiting factor.
  • A stronger validation the paper leaves implicit is to run the DFSM-selected optimal controller in the high-fidelity simulator and compare its DEL to the true optimum; if the gap stays small across seeds, the surrogate could replace full-fidelity DOE in early design stages.
  • The authors note the DFSM was trained at one controller gain point; a plausible generalization is that design-space fidelity degrades gradually as the training gain moves away from the optimum, which could be checked by training at several gains and comparing contour shapes.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper develops a linear-parameter-varying (LPV) derivative function surrogate model (DFSM) for the IEA 15-MW floating offshore wind turbine. State derivatives are extracted from OpenFAST simulation time series using cubic splines, and wind-speed-scheduled state-space matrices are fit by solving an optimization problem with a stability constraint. The DFSM is compared in closed-loop simulation with a subspace-identification model (n4sid) and an LSTM model, then evaluated for two use cases: predicting key time series and performance metrics (DELt, AEP) over multiple wind speeds, and reproducing the DELt design space over blade-pitch controller gains (omegaPC, zetaPC) in a 25-point design-of-experiments study. The authors report that the DFSM offers the best trade-off among the three low-fidelity models and provides roughly a 48x speedup over OpenFAST while preserving the qualitative shape of the DELt design space.

Significance. If the quantitative claims survive scrutiny, the LPV-DFSM is a useful engineering contribution: it is a continuous-time, stable-by-construction surrogate with physically meaningful states, and it can be coupled to a controller for closed-loop simulation. The paper deserves credit for evaluating the DFSM on held-out seeds and, in the DOE study, on controller gains not used in training; that generalization test is genuinely external. The main significance limitation is that the central design-space-preservation claim rests on visual comparison of contour plots and on an uncontrolled comparison of the three modeling approaches, so the claimed utility for controller optimization is not yet established at the level of a journal publication.

major comments (3)
  1. [Sec. 5.3] The headline comparison of the three modeling approaches is not controlled: the DFSM is trained on simulations at all seven wind speeds (5 seeds each), the n4sid model is trained only at 14 m/s, and the LSTM model is trained at 12, 14, and 16 m/s. Fig. 9 then compares their closed-loop blade-pitch MSE on ten 14 m/s test cases and concludes that the DFSM achieves the best accuracy/simulation-time trade-off. Because training data coverage strongly affects both accuracy and variance, this comparison does not isolate modeling approach; please retrain the n4sid and LSTM baselines on the same 7-wind-speed training set (or at least on the same subset as the DFSM) and repeat the MSE/time comparison, or explicitly reframe the claim as 'best under the given per-model training-data budget.'
  2. [Sec. 6.2, Figs. 14-15] The central optimization-utility claim is that the DFSM 'preserves the shape' of the DELt design space, but this is supported only by visual inspection. Fig. 14 shows DFSM DELt levels of roughly 50-85 MNm against OpenFAST's 111-135 MNm, a 30-40% underprediction, and Fig. 15 removes the bias by normalizing each panel to a different scale, making the shapes look more similar than an unscaled overlay would. The two DFSMs trained from different xc values also produce visibly different normalized contours. Since DELt in Eq. (14) is computed from Mt,y, the quantity the DFSM predicts least accurately (Fig. 13c), the paper needs a quantitative evaluation of design-space fidelity: report a rank correlation (e.g., Spearman) between DFSM and OpenFAST DELt values over the 25 DOE points, the location and value of the predicted optimum xc, and an unscaled error map; otherwise the conclusion that the DFSM can be used in controller optimization is not established.
  3. [Sec. 5.3] In the closed-loop validation, the only plotted comparison between low-fidelity models and OpenFAST is the blade-pitch signal beta. However, the ROSCO controller uses omega_g and x_ddot_t as feedback variables (Sec. 2.2), so errors in those variables are the direct cause of any beta error; the sentence in Sec. 6.1 that 'the efficacy of the DFSM in predicting omega_g and x_ddot_t can be inferred from Fig. 11a' is not a substitute for a direct comparison. Please add time-series or error metrics for omega_g and x_ddot_t in the closed-loop validation (and in Sec. 6.1), since these are the quantities the plant model must provide for correct control action.
minor comments (4)
  1. [Title/Abstract] The arXiv title promises 'optimal control and control co-design', but the manuscript does not perform a co-design study and Sec. 7 explicitly lists nested control co-design as future work; please align the title and abstract with the actual scope.
  2. [Sec. 5.2] The LSTM model is stated to predict y_k = f_LSTM(u_k) with a 0.5 s timestep while the other models use dt = 0.01 s; please specify whether the LSTM was trained on downsampled data and how outputs are compared at common time instants.
  3. [Sec. 5.3, Fig. 9] The scatter plot of MSE versus simulation time would be easier to read on a logarithmic time axis, since the DFSM and LSTM points at 25 s and 70 s are compressed against the n4sid points at 2.5 s.
  4. [Sec. 6.1, Eq. (14)] In the definition of DELt, the dependence of DEL(wbar) on the seed and on the controller gains xc is not made explicit; please clarify the notation so that the weighted integral is unambiguous.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the DFSM is fitted to derivative data and every headline evaluation (held-out seeds, all wind speeds, and a controller-gain DOE) is a genuine extrapolation against OpenFAST simulations not used in the fit.

full rationale

The derivation chain is self-contained. The DFSM parameters are obtained by solving Eq. (10), which minimizes the error between spline-extracted state derivatives from OpenFAST and the LPV prediction, subject only to a stability constraint. None of the headline results is a restatement of the fitted objective. In Sec. 5.3, the closed-loop validation uses ten wbar = 14 m/s test cases that were not part of the five training seeds; in Sec. 6.1, the model is evaluated on all 35 load cases spanning wbar = 6-18 m/s; and in Sec. 6.2, the decisive DOE study varies controller gains xc = [omega_PC, zeta_PC] over a 25-point grid that the training data never sampled. The paper explicitly states that the DFSM 'was not trained on these' gains, and it reports the resulting DELt underprediction instead of hiding it (Figs. 11c, 13c, 14). The only self-citations are contextual: Ref. [9] (same research group) is cited for the spline-derivative extraction validation and for LPV-based control co-design, and Refs. [21, 37] are related prior work; none supplies a load-bearing uniqueness theorem or an assumption whose validity is imported from the self-citation. The LPV structure A(w) xi + B(w) u is an ansatz, but it is argued in-text from the wind-speed dependence of turbine dynamics and is then fitted to the paper's own data rather than assumed to hold. The concern that design-space 'shape preservation' is assessed only by visual contour inspection is a matter of evidence strength and correctness risk, not circularity: the DOE predictions are not implied by any fitted parameter or by construction.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central claim rests on the fitted LPV matrices and on strong model-structure assumptions: a six-state LPV model with wind speed as the only scheduling variable, no unmeasured internal states, and accurate spline-based derivative extraction. These are domain assumptions that are not independently verified for this FOWT system. The DOE generalization test partially checks one of them but only at two training points and with qualitative comparisons.

free parameters (3)
  • LPV state-space matrices A_i, B_i, C_i, D_i = not reported
    Fitted per wind speed using a hybrid GA plus gradient-based solver (Eqs. 10 and 11). These are the model parameters that produce the surrogate; values are not listed in the paper.
  • State vector composition = theta_p, delta_tt, omega_g and their first derivatives (6 states)
    Chosen by hand in Sec. 4.3. The assumption that these outputs form the state vector determines the model structure and is not directly derived from data.
  • n4sid model order n_x = 6
    Selected via a sensitivity study in Sec. 5.1. Used only for the comparison baseline, but affects the fairness of the comparison.
assumptions (5)
  • domain assumption The FOWT dynamics can be represented as a linear parameter-varying system with wind speed as the only scheduling parameter.
    Invoked in Sec. 4.4, Eq. (9). If the dynamics depend strongly on other conditions or on controller gains, the LPV model will miss them.
  • domain assumption The selected output variables form the full state vector, with no unmeasured internal states (xi is a subset of y).
    Stated in Sec. 4.3. OpenFAST has many additional degrees of freedom, so this is a strong simplification.
  • domain assumption Cubic spline interpolation of sampled state trajectories gives accurate state derivatives for training.
    Section 4.3 uses spline and fnder; derivative error is not quantified for this system.
  • domain assumption A model trained at one controller gain x_c, train generalizes to other gains without retraining.
    Assumed in the Sec. 6.2 DOE study; the paper tests it at only two training points.
  • domain assumption Closed-loop stability of the identified linear models is sufficient for stable closed-loop surrogate simulation.
    The stability constraint in Eq. (10c) only constrains the open-loop A matrix; it does not guarantee closed-loop behavior with the ROSCO controller.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Data-Driven Modeling Approaches for Optimal Control and Control Co-Design of Floating Offshore Wind Turbines." pith.science (2026). https://pith.science/paper/2M3SBJA6

@misc{pith2026250514515,
  author       = {Pith},
  title        = {Pith review of: Data-Driven Modeling Approaches for Optimal Control and Control Co-Design of Floating Offshore Wind Turbines},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2M3SBJA6}},
  note         = {Machine review of arXiv:2505.14515}
}
read the original abstract

Models that balance accuracy against computational costs are advantageous when designing wind turbines with optimization studies, as several hundred predictive function evaluations might be necessary to identify the optimal solution. We explore different approaches to construct low-fidelity models that can be used to approximate dynamic quantities and be used as surrogates for design optimization studies and other use cases. In particular, low-fidelity modeling approaches using classical systems identification and deep learning approaches are considered against derivative function surrogate models ({DFSMs}), or approximate models of the state derivative function. This work proposes a novel method that utilizes a linear parameter varying (LPV) modeling scheme to construct the DFSM. We compare the trade-offs between these different models and explore the efficacy of the proposed DFSM approach in approximating wind turbine performance and different design optimization studies. Results show that the proposed DFSM approach balances computational time and model accuracy better than the system identification and deep learning-based models. Additionally, the DFSM provides nearly a fifty times speed-up compared to the high-fidelity model, while balancing accuracy.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

60 extracted references · 44 canonical work pages

  1. [1]

    Design of dynamic systems using surrogate models of derivative functions

    Deshmukh, A. P., and Allison, J. T., 2017. “Design of dynamic systems using surrogate models of derivative functions”. J. Mech. Design, 139(10), Aug. doi: 10.1115/1.4037407

    doi:10.1115/1.4037407 2017

  2. [2]

    A trajectory-based sampling strategy for sequentially refined metamodel management of metamodel-based dynamic optimization in mechatronics

    Lefebvre, T., et al., 2018. “A trajectory-based sampling strategy for sequentially refined metamodel management of metamodel-based dynamic optimization in mechatronics”.Optim. Control Appl. Meth- ods, 39(5), May, pp. 1786–1801. doi: 10.1002/oca.2442

    work page doi:10.1002/oca.2442 2018

  3. [3]

    Control co-design of 13 MW down- wind two-bladed rotors to achieve 25% reduction in levelized cost of wind energy

    Pao, L. Y ., et al., 2021. “Control co-design of 13 MW down- wind two-bladed rotors to achieve 25% reduction in levelized cost of wind energy”. Annu. Rev. Control, 51, pp. 331–343. doi: 10.1016/j.arcontrol.2021.02.001

    work page doi:10.1016/j.arcontrol.2021.02.001 2021

  4. [4]

    [Online]

    OpenFAST. [Online]. url: https: //github.com/OpenFAST/openfast

    work page

  5. [5]

    WEC-Sim /WEC-Sim: WEC-Sim v5.0.1

    Ruehl, K., et al., 2022. WEC-Sim /WEC-Sim: WEC-Sim v5.0.1. doi: 10.5281/zenodo.7121186

    doi:10.5281/zenodo.7121186 2022

  6. [6]

    Wind energy generation systems - part 3-2: Design requirements for floating o ffshore wind turbines

    IEC, 2025. Wind energy generation systems - part 3-2: Design requirements for floating o ffshore wind turbines. IEC 61400-3- 2:2025, International Electrotechnical Commission, Jan

    work page 2025

  7. [7]

    Floating wind turbine control op- timization

    Zalkind, D., et al., 2022. “Floating wind turbine control op- timization”. J. Phys.: Conf. Ser., 2265(4), May, p. 042021. doi: 10.1088/1742-6596/2265/4/042021

    work page doi:10.1088/1742-6596/2265/4/042021 2022

  8. [8]

    Concurrent probabilistic con- trol co-design and layout optimization of wave energy converter farms using surrogate modeling

    Azad, S., and Herber, D. R., 2023. “Concurrent probabilistic con- trol co-design and layout optimization of wave energy converter farms using surrogate modeling”. In Design Automation Confer- ence, ASME. doi: 10.1115/detc2023-116896

    work page doi:10.1115/detc2023-116896 2023

Show all 60 references
  1. [9]

    Open-loop control co-design of semisubmersible floating o ffshore wind turbines using linear parameter-varying models

    Sundarrajan, A. K., et al., 2023. “Open-loop control co-design of semisubmersible floating o ffshore wind turbines using linear parameter-varying models”. J. Mech. Design, 146(4), Nov. doi: 10.1115/1.4063969

    2023 doi

  2. [10]

    FAST modularization framework for wind turbine simulation: full-system linearization

    Jonkman, J. M., and Jonkman, B. J., 2016. “FAST modularization framework for wind turbine simulation: full-system linearization”. J. Phys.: Conf. Ser., 753, Sept., p. 082010. doi: 10.1088 /1742- 6596/753/8/082010

    2016

  3. [11]

    A reference open-source controller for fixed and floating offshore wind turbines

    Abbas, N. J., et al., 2021. “A reference open-source controller for fixed and floating offshore wind turbines”. Wind Energy Sci., 7(1), Apr., pp. 53–73. doi: 10.5194/wes-2021-19

    2021 doi

  4. [12]

    Added-mass effects on a horizontal-axis tidal turbine using FAST v8

    Murray, R. E., et al., 2018. “Added-mass effects on a horizontal-axis tidal turbine using FAST v8”. Renew. Energy, 126, Oct., pp. 987–

    2018

  5. [13]

    Design and modeling of an open- source baseline floating marine turbine

    Wiley, W., et al., 2023. “Design and modeling of an open- source baseline floating marine turbine”. In University Marine Energy Research Community Conference. https: //www.osti.gov/ biblio/2202791

    2023

  6. [14]

    A tutorial on the dynamics and control of wind turbines and wind farms

    Pao, L. Y ., and Johnson, K. E., 2009. “A tutorial on the dynamics and control of wind turbines and wind farms”. In American Control Conference. doi: 10.1109/acc.2009.5160195

    2009

  7. [15]

    Functional requirements for the WEIS toolset to enable controls co-design of floating o ffshore wind tur- bines

    Jonkman, J., et al., 2021. “Functional requirements for the WEIS toolset to enable controls co-design of floating o ffshore wind tur- bines”. In International O ffshore Wind Technical Conference, ASME. doi: 10.1115/iowtc2021-3533

    2021 doi

  8. [16]

    Development of a control co-design modeling tool for marine hydrokinetic turbines

    Ross, H., et al., 2022. “Development of a control co-design modeling tool for marine hydrokinetic turbines”. In Interna- tional Mechanical Engineering Congress & Exposition, ASME. doi: 10.1115/imece2022-94483

    2022 doi

  9. [17]

    Control of floating wind turbines

    van der Veen, G. J., et al., 2012. “Control of floating wind turbines”. In American Control Conference. doi: 10.1109/acc.2012.6315120

    2012

  10. [18]

    Data-driven surrogate model for wind turbine damage equivalent load

    Haghi, R., and Crawford, C., 2024. “Data-driven surrogate model for wind turbine damage equivalent load”. Wind Energy Sci., 9(11), Nov., pp. 2039–2062. doi: 10.5194/wes-9-2039-2024

    2024 doi

  11. [19]

    Survey of multifidelity methods in un- certainty propagation, inference, and optimization

    Peherstorfer, B., et al., 2018. “Survey of multifidelity methods in un- certainty propagation, inference, and optimization”. SIAM Review, 60(3), Jan., pp. 550–591. doi: 10.1137/16m1082469

    2018 doi

  12. [20]

    Provably convergent multi- fidelity optimization algorithm not requiring high-fidelity deriva- tives

    March, A., and Willcox, K., 2012. “Provably convergent multi- fidelity optimization algorithm not requiring high-fidelity deriva- tives”. AIAA J., 50(5), May, pp. 1079–1089. doi: 10.2514/1.j051125

    2012 doi

  13. [21]

    Development of an opti- mal variable-pitch controller for floating axial-flow marine hy- drokinetic turbines

    Sundarrajan, A. K., et al., 2024. “Development of an opti- mal variable-pitch controller for floating axial-flow marine hy- drokinetic turbines”. In Design Automation Conference, ASME. doi: 10.1115/detc2024-141271

    2024 doi

  14. [22]

    Influence of wind turbine design pa- rameters on linearized physics-based models in OpenFAST

    Jonkman, J. M., et al., 2022. “Influence of wind turbine design pa- rameters on linearized physics-based models in OpenFAST”. Wind Energy Sci., 7(2), Mar., pp. 559–571. doi: 10.5194/wes-7-559-2022

    2022 doi

  15. [23]

    Beamdyn: a high-fidelity wind turbine blade solver in the fast modular framework

    Wang, Q., et al., 2017. “Beamdyn: a high-fidelity wind turbine blade solver in the fast modular framework”. Wind Energy, 20(8), Mar., pp. 1439–1462. doi: 10.1002/we.2101

    2017 doi

  16. [24]

    Control co-design optimization of nat- ural gas power plants with carbon capture and thermal storage

    Vercellino, R., et al., 2022. “Control co-design optimization of nat- ural gas power plants with carbon capture and thermal storage”. In 15 Copyright © 2025 by ASME Design Automation Conference, ASME. doi: 10.1115 /detc2022- 90021

    2022

  17. [25]

    Nested control co-design of a spar buoy horizontal-axis floating offshore wind turbine

    Bayat, S., et al., 2025. “Nested control co-design of a spar buoy horizontal-axis floating offshore wind turbine”. Ocean Eng., 328, June, p. 121037. doi: 10.1016/j.oceaneng.2025.121037

    2025

  18. [26]

    Modeling and control co-design of a float- ing offshore vertical-axis wind turbine system

    Lee, Y . H., et al., 2024. “Modeling and control co-design of a float- ing offshore vertical-axis wind turbine system”. In Design Automa- tion Conference, ASME. doi: 10.1115/detc2024-143495

    2024 doi

  19. [27]

    Surrogate modeling of nonlinear dynamic systems: A comparative study

    Zhao, Y ., et al., 2022. “Surrogate modeling of nonlinear dynamic systems: A comparative study”. J. Comput. Inf. Sci. Eng., 23(1), May. doi: 10.1115/1.4054039

    2022 doi

  20. [28]

    The performance of LSTM and BiLSTM in forecasting time series

    Siami-Namini, S., et al., 2019. “The performance of LSTM and BiLSTM in forecasting time series”. In International Con- ference on Big Data, IEEE, pp. 3285–3292. doi: 10.1109 /big- data47090.2019.9005997

    2019

  21. [29]

    Frequency domain system identifi- cation toolbox for MATLAB: Improvements and new possibili- ties

    Kollár, I., et al., 1997. “Frequency domain system identifi- cation toolbox for MATLAB: Improvements and new possibili- ties”. IFAC Proceedings Volumes, 30(11), July, pp. 943–946. doi: 10.1016/s1474-6670(17)42968-5

    1997 doi

  22. [30]

    V ., and Moor, B

    Overschee, P. V ., and Moor, B. D., 1996. Subspace Identification for Linear Systems. Springer. doi: 10.1007 /978-1-4613-0465-4

    1996

  23. [31]

    Past, present and future of nonlinear sys- tem identification in structural dynamics

    Kerschen, G., et al., 2006. “Past, present and future of nonlinear sys- tem identification in structural dynamics”. Mech. Syst. Signal Pro- cess., 20(3), Apr., pp. 505–592. doi: 10.1016/j.ymssp.2005.04.008

    2006 doi

  24. [32]

    On evolutionary system identification with applications to nonlinear benchmarks

    Worden, K., et al., 2018. “On evolutionary system identification with applications to nonlinear benchmarks”.Mech. Syst. Signal Pro- cess., 112, Nov., pp. 194–232. doi: 10.1016/j.ymssp.2018.04.001

    2018 doi

  25. [33]

    System Identification

    Ljung, L., 1998. System Identification. Birkhäuser Boston, pp. 163–

    1998

  26. [34]

    A new subspace identification method for open and closed loop data

    Jansson, M., 2005. “A new subspace identification method for open and closed loop data”. IFAC Proceedings Volumes, 38(1), pp. 500–

    2005

  27. [35]

    A novel surrogate model-based solving framework for the black-box dynamic co-design and optimization problem in the dynamic system

    Zhang, Q., et al., 2022. “A novel surrogate model-based solving framework for the black-box dynamic co-design and optimization problem in the dynamic system”. Mathematics, 10(18), Sept., p. 3239. doi: 10.3390/math10183239

    2022 doi

  28. [36]

    A new sequential sampling method of surrogate models for design and optimization of dynamic systems

    Qiao, P., et al., 2021. “A new sequential sampling method of surrogate models for design and optimization of dynamic systems”. Mech. Mach. Theory, 158, Apr., p. 104248. doi: 10.1016/j.mechmachtheory.2021.104248

    2021

  29. [37]

    On the uses of linear- quadratic methods in solving nonlinear dynamic optimization prob- lems with direct transcription

    Herber, D. R., and Sundarrajan, A. K., 2020. “On the uses of linear- quadratic methods in solving nonlinear dynamic optimization prob- lems with direct transcription”. In International Mechanical Engi- neering Congress & Exposition, ASME. doi: 10.1115/IMECE2020- 23885

    2020 doi

  30. [38]

    Co-design of an active suspension us- ing simultaneous dynamic optimization

    Allison, J. T., et al., 2014. “Co-design of an active suspension us- ing simultaneous dynamic optimization”. J. Mech. Design, 136(8), June. doi: 10.1115/1.4027335

    2014 doi

  31. [39]

    LPV-based torque control for an extreme-scale morphing wind turbine rotor

    Martin, D. P., et al., 2017. “LPV-based torque control for an extreme-scale morphing wind turbine rotor”. In American Control Conference, IEEE. doi: 10.23919/ACC.2017.7963146

    2017

  32. [40]

    Prediction error method for identification of LPV models

    Zhao, Y ., et al., 2012. “Prediction error method for identification of LPV models”. J. Process Control, 22(1), Jan., pp. 180–193. doi: 10.1016/j.jprocont.2011.09.004

    2012 doi

  33. [41]

    LPV /LFT modelling and iden- tification: overview, synergies and a case study

    Casella, F., and Lovera, M., 2008. “LPV /LFT modelling and iden- tification: overview, synergies and a case study”. In International Conference on Computer-Aided Control Systems, IEEE, pp. 852–

    2008

  34. [42]

    Modeling and Identification of Linear Parameter- Varying Systems

    Tóth, R., 2010. Modeling and Identification of Linear Parameter- Varying Systems. Springer Berlin Heidelberg. doi: 10.1007 /978-3- 642-13812-6

    2010

  35. [43]

    IEA wind TCP task 37: definition of the IEA 15-megawatt o ffshore reference wind turbine

    Gaertner, E., et al., 2020. IEA wind TCP task 37: definition of the IEA 15-megawatt o ffshore reference wind turbine. Tech. Rep. NREL/TP-5000-75698, National Renewable Energy Labora- tory, Mar. doi: 10.2172/1603478

    2020 doi

  36. [44]

    Definition of the UMaine V olturnUS-S ref- erence platform developed for the IEA wind 15-megawatt o ffshore reference wind turbine

    Allen, C., et al., 2020. Definition of the UMaine V olturnUS-S ref- erence platform developed for the IEA wind 15-megawatt o ffshore reference wind turbine. Tech. Rep. NREL/TP-5000-76773, National Renewable Energy Laboratory, July. doi: 10.2172/1660012

    2020 doi

  37. [45]

    [Online, version gd3ea7969]

    WEIS. [Online, version gd3ea7969]. url: https: //github.com/ WISDEM/WEIS/tree/develop

  38. [46]

    MathWorks. spline. [Online]. url: https: //www.mathworks.com/ help/matlab/ref/spline.html

  39. [47]

    CubicSpline

    Scipy. CubicSpline. [Online]. url: https: //docs.scipy.org/doc/scipy/ reference/generated/scipy.interpolate.CubicSpline.html#scipy. interpolate.CubicSpline

  40. [48]

    Closed-loop issues in system iden- tification

    Van den Hof, P., 1998. “Closed-loop issues in system iden- tification”. Annu. Rev. Control, 22, Jan., pp. 173–186. doi: 10.1016/s1367-5788(98)00016-9

    1998 doi

  41. [49]

    TensorFlow: Large-scale machine learning on heterogeneous systems

    Abadi, M., et al., 2015. TensorFlow: Large-scale machine learning on heterogeneous systems. Software available from tensorflow.org

    2015

  42. [50]

    Determining equivalent dam- age loading for full-scale wind turbine blade fatigue tests

    Freebury, G., and Musial, W., 2000. “Determining equivalent dam- age loading for full-scale wind turbine blade fatigue tests”. In ASME Wind Energy Symposium, AIAA. doi: 10.2514/6.2000-50

    2000 doi

  43. [51]

    A machine learning method for model- ing wind farm fatigue load

    Miao, Y ., et al., 2022. “A machine learning method for model- ing wind farm fatigue load”. Appl. Sci., 12(15), July, p. 7392. doi: 10.3390/app12157392

    2022 doi

  44. [52]

    Deep state space models for nonlinear system identification

    Gedon, D., et al., 2021. “Deep state space models for nonlinear system identification”. IFAC-PapersOnLine, 54(7), pp. 481–486. doi: 10.1016/j.ifacol.2021.08.406

    2021 doi

  45. [53]

    Deep networks for system iden- tification: A survey

    Pillonetto, G., et al., 2025. “Deep networks for system iden- tification: A survey”. Automatica, 171, Jan., p. 111907. doi: 10.1016/j.automatica.2024.111907

    2025

  46. [54]

    Accelerating simulation of stiff nonlinear systems using continuous-time echo state networks

    Anantharaman, R., et al., 2020. Accelerating simulation of stiff nonlinear systems using continuous-time echo state networks. arXiv:2010.04004

    2020 arXiv

  47. [55]

    Continuous-time echo state networks for predicting power system dynamics

    Roberts, C., et al., 2022. “Continuous-time echo state networks for predicting power system dynamics”. Electr. Power Syst. Res., 212, Nov., p. 108562. doi: 10.1016/j.epsr.2022.108562

    2022

  48. [56]

    Chen, R. T. Q., et al., 2018. Neural ordinary di fferential equations, June. arXiv:1806.07366 16 Copyright © 2025 by ASME

    2018 arXiv

  49. [173]

    doi: 10.1007/978-1-4612-1768-8_11

    doi

  50. [505]

    doi: 10.3182/20050703-6-cz-1902.00084

    1902 arXiv

  51. [857]

    doi: 10.1109/cacsd.2008.4627358

    2008

  52. [1002]

    doi: 10.1016/j.renene.2018.04.023

    2018 doi

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.