Wave Prediction Research Articles

Non-causal Linear Optimal Control With Adaptive Sliding Mode Observer for Multi-Body Wave Energy Converters

As a non-causal optimal control problem, the performance of wave energy converter (WEC) control relies on the accuracy of the future incoming wave prediction. However, the inevitable prediction errors can degrade WEC performance dramatically especially when a long prediction horizon is needed by a WEC non-causal optimal controller. This paper proposes a novel non-causal linear optimal control with adaptive sliding mode observer (NLOC+ASMO) scheme, which can effectively mitigate the control performance degradation caused by wave prediction errors. This advantage is achieved by embedding the following enabling techniques into the scheme: (i) an adaptive sliding mode observer (ASMO) to estimate current excitation force in real-time with explicitly formulated boundary of estimation error, (ii) an auto-regressive (AR) model to predict the incoming excitation force with explicitly formulated boundary of prediction error using a set of latest historical data of ASMO estimations from (i), and (iii) a compensator to compensate for both the estimation error and the prediction error of excitation force. Moreover, the proposed NLOC+ASMO scheme does not cause heavy computational load enabling its real-time implementation on standard computational hardware, which is especially critical for the control of WECs with complicated dynamics. The proposed NLOC+ASMO framework is generic and can be applied to a wide range of WECs, and in this paper we demonstrate the efficacy by using a multi-float and multi-motion WEC called M4 as a case study, whose control problem is more challenging than the widely studied point absorbers. Simulation results show the effectiveness of the proposed control scheme in a wide range of sea states, and it is also found that the controller is not sensitive to change of ASMO parameters.

Individual Responses to Heat Stress: Implications for Hyperthermia and Physical Work Capacity.

BackgroundExtreme heat events are increasing in frequency, severity, and duration. It is well known that heat stress can have a negative impact on occupational health and productivity, particularly during physical work. However, there are no up-to-date reviews on how vulnerability to heat changes as a function of individual characteristics in relation to the risk of hyperthermia and work capacity loss. The objective of this narrative review is to examine the role of individual characteristics on the human heat stress response, specifically in relation to hyperthermia risk and productivity loss in hot workplaces. Finally, we aim to generate practical guidance for industrial hygienists considering our findings. Factors included in the analysis were body mass, body surface area to mass ratio, body fat, aerobic fitness and training, heat adaptation, aging, sex, and chronic health conditions.FindingsWe found the relevance of any factor to be dynamic, based on the work-type (fixed pace or relative to fitness level), work intensity (low, moderate, or heavy work), climate type (humidity, clothing vapor resistance), and variable of interest (risk of hyperthermia or likelihood of productivity loss). Heat adaptation, high aerobic fitness, and having a large body mass are the most protective factors during heat exposure. Primary detrimental factors include low fitness, low body mass, and lack of heat adaptation. Aging beyond 50 years, being female, and diabetes are less impactful negative factors, since their independent effect is quite small in well matched participants. Skin surface area to mass ratio, body composition, hypertension, and cardiovascular disease are not strong independent predictors of the heat stress response.ConclusionUnderstanding how individual factors impact responses to heat stress is necessary for the prediction of heat wave impacts on occupational health and work capacity. The recommendations provided in this report could be utilized to help curtail hyperthermia risk and productivity losses induced by heat.

Atmospheric Gravity Waves Observed in the Nightglow Following the 21 August 2017 Total Solar Eclipse

AbstractNighttime airglow images observed at the low‐latitude site of São João do Cariri (7.4°S, 36.5°W) showed the presence of a medium‐scale atmospheric gravity wave (AGW) associated with the 21 August 2017 total solar eclipse. The AGW had a horizontal wavelength of ∼1,618 km, observed period of ∼152 min, and propagation direction of ∼200° clockwise from the north. The spectral characteristics of this wave are in good agreement with theoretical predictions for waves generated by eclipses. Additionally, the wave was reverse ray‐traced, and the results show its path crossing the Moon's shadow of the total solar eclipse in the tropical North Atlantic ocean at stratospheric altitudes. Investigation about potential driving sources for this wave indicates the total solar eclipse as the most likely candidate. The optical measurements were part of an observational campaign carried out to detect the impact of the August 21 eclipse in the atmosphere at low latitudes.

Efficient prediction of wind and wave induced long-term extreme load effects of floating suspension bridges using artificial neural networks and support vector machines

Long-term extreme load effects are one of the primary concerns in the design of civil and offshore structures. Such load effects can be evaluated using the accurate but computationally demanding full long-term method or the more efficient but approximate first-order and second-order reliability methods. Monte Carlo based methods enhanced with machine learning algorithms offer efficient alternatives to the traditional methods. Therefore, artificial neural networks and support vector machines are used as surrogate models for the limit state function to speed up the prediction of long-term extreme load effects. A three-span suspension bridge with two floating pylons under combined wind and wave actions is used as a case study. The cumulative density functions of the long-term extreme values corresponding to a bending moment value due to vertical deflections at the critical position of the girder are calculated. It is then shown that the artificial neural network and support vector machine-based approaches require less computational effort and yield more accurate results than the first- and second-order reliability methods.

Walter Heinrich Munk. 19 October 1917—8 February 2019

Born in Vienna, Austria, shortly before the Austro-Hungarian Empire lost its access to the sea, Walter Munk became a leading geophysicist and physical oceanographer after being sent to school in New York State and later moving west to California. He and Harald Sverdrup developed wave prediction schemes that were used by the Allies in World War II to evaluate whether amphibious landings would be feasible. His post-war research included further major contributions to the understanding of ocean waves, as well as ocean circulation, tides, internal waves, mixing processes and many other phenomena. He showed how variations in the Earth's rotation contained a wealth of valuable geophysical information and was an instigator of the ‘Mohole’ project that failed for political reasons but paved the way for the Deep Sea Drilling Project. Although primarily a theoretician, throughout his career he was an enthusiastic adopter of new technologies and data analysis techniques. He showed how acoustic transmissions could be used to map ocean eddies and currents as well as to monitor temperature changes of whole ocean basins. He was a key adviser to the US Navy and a dedicated member of the science advisory group, JASON, that focuses on national security. As well as his scientific contributions, his legacy includes the La Jolla Institute of Geophysics and Planetary Physics. He was liberal in his political outlook, gregarious, hospitable and treated everyone, from undergraduate to admiral, with the same interest and respect.

Modeling Uncertainties of Bathymetry Predicted With Satellite Altimetry Data and Application to Tsunami Hazard Assessments

AbstractModels of bathymetry derived from satellite radar altimetry are essential for modeling many marine processes. They are affected by uncertainties which require quantification. We propose an uncertainty model that assumes errors are caused by the lack of high‐wavenumber content within the altimetry data. The model is then applied to a tsunami hazard assessment. We build a bathymetry uncertainty model for northern Chile. Statistical properties of the altimetry‐predicted bathymetry error are obtained using multibeam data. We find that a Von Karman correlation function and a Laplacian marginal distribution can be used to define an uncertainty model based on a random field. We also propose a method for generating synthetic bathymetry samples conditional to shipboard measurements. The method is further extended to account for interpolation uncertainties, when bathymetry data resolution is finer than ∼10 km. We illustrate the usefulness of the method by quantifying the bathymetry‐induced uncertainty of a tsunami hazard estimate. We demonstrate that tsunami leading wave predictions at middle/near field tide gauges and buoys are insensitive to bathymetry uncertainties in Chile. This result implies that tsunami early warning approaches can take full advantage of altimetry‐predicted bathymetry in numerical simulations. Finally, we evaluate the feasibility of modeling uncertainties in regions without multibeam data by assessing the bathymetry error statistics of 15 globally distributed regions. We find that a general Von Karman correlation and a Laplacian marginal distribution can serve as a first‐order approximation. The standard deviation of the uncertainty random field model varies regionally and is estimated from a proposed scaling law.

Prediction of Wave Breaking Using Machine Learning Open Source Platform

지금까지 연안에서 발생하는 쇄파에 대한 연구는 ì§€ì†ì ìœ¼ë¡œ 수행되었으며, 그에 따른 많은 실험자료가 ì¶•ì ë˜ì–´ 왔다. 또한, 다양한 실험자료로부터 ê³µí•™ì ì¸ ì ìš©ì„ 위한 쇄파 ì •ë³´ë¥¼ ì •ëŸ‰ì ìœ¼ë¡œ 예측하기 위하여 회귀분석에 기반한 다양한 경험식이 ì œì•ˆë˜ì—ˆë‹¤. 그러나 쇄파는 ë‚´ìž¬í•˜ê³ ìžˆëŠ” 변동성이 있으므로 ì„ í˜• 회귀분석과 같은 ì„ í˜•ì  í†µê³„ì ‘ê·¼ 방법에는 한계가 있다. 본 연구에서는 ì‡„íŒŒíŒŒê³ ë° 쇄파수심을 예측하기 위하여 기계학습 중 하나인 ì‹ ê²½ë§ì„ 사용하는 ë¹„ì„ í˜• 방법을 ì œì•ˆí•˜ì˜€ë‹¤. ì‹ ê²½ë§ì€ 구글에서 ë°°í¬í•˜ê³ ìžˆëŠ” ë¨¸ì‹ ëŸ¬ë‹ 오픈소스 플랫폼인 í ì„œí”Œë¡œ(Tensorflow)를 이용하여 구축하였다. ì‹ ê²½ë§ 모델은 수집된 실험자료를 무작위로 ì„ íƒí•˜ì—¬ 학습하였으며, 학습에 이용하지 않은 자료를 사용하여 학습된 ì‹ ê²½ë§ì„ 평가하였다. 학습된 ì‹ ê²½ë§ì— 의해 예측된 ì‡„íŒŒíŒŒê³ ì™€ 쇄파수심에 대한 예측결과는 기존의 경험식에 의한 계산결과에 비해 높은 예측성능을 보였으며, 이는 충분히 학습된 ì‹ ê²½ë§ì€ ì‡„íŒŒíŒŒê³ ë° 수심을 예측하기 위한 ìœ ìš©í•œ 도구로 ì‚¬ìš©ë ìˆ˜ 있음을 보여준다. 핵심용어: ì‡„íŒŒíŒŒê³ , 쇄파수심, ë¨¸ì‹ ëŸ¬ë‹, ì‹ ê²½ë§, í ì„œí”Œë¡œ

Decomposition of Geomagnetic Secular Acceleration Into Traveling Waves Using Complex Empirical Orthogonal Functions

AbstractSatellite observations reveal short pulses in the second time derivative of the geomagnetic field. We seek to interpret these signals using complex empirical orthogonal functions (CEOFs). This methodology decomposes the signal into traveling waves, permitting estimates for the period, angular wave number, and phase velocity. We recover CEOFs from the CHAOS‐6 model, focusing on three geographic regions with strong secular acceleration. Two regions are confined to the equator, while the third is located under Alaska. We find evidence for both eastward and westward traveling waves with periods between 7 and 20 years. There is also evidence for weaker standing waves with complex spatial patterns. Two of the three regions have waves that are compatible with predictions for waves in a stratified fluid. Our results yield estimates for the structure of fluid stratification at the top of the core.

Rogue Wave 예측을 위한 이상지수 모델

Rogue Wave 예측을 위한 이상지수 모델

Phase-resolved wave prediction model for long-crest waves based on machine learning

Phase-resolved wave prediction is important to a vessel for both predicting deterministic motion and assisting decision-making. In this study, both linear and nonlinear physical-based wave models are addressed for achieving phase-resolved wave prediction. These physical-based models are developed based on the fast Fourier transform (FFT) technique. Although nonlinear wave models such as high order spectrum models are capable of nonlinear wave modeling, they still experience nonstationary difficulties. The FFT technique requires wave motions to be linear and stationary. Accordingly, inspired by the capability of machine learning in solving nonlinear dynamic problems, a machine learning method is applied in phase-resolved wave prediction. Thus, an artificial neural network-based wave prediction (ANN-WP) model is proposed. Experimental wave data sets are employed in model training optimization, verification, and comparison studies. The prediction accuracy of the ANN-WP model is verified and the input–output strategy is explored for training optimization of the ANN-WP model. A comparison between the ANN-WP and linear wave prediction (LWP) models is presented using the nonlinear experimental wave data. Furthermore, a preliminary discussion of the extreme wave predictive ability of the ANN-WP model is presented. The comparison results indicate that the ANN-WP model performs better than the LWP model. Although selected problems remain in extreme wave prediction, the proposed ANN-WP model provides a feasible and effective approach for achieving accurate nonlinear phase-resolved wave prediction.

Predicting site-specific storm wave run-up

Storm wave run-up causes beach erosion, wave overtopping, and street flooding. Extreme runup estimates may be improved, relative to predictions from general empirical formulae with default parameter values, by using historical storm waves and eroded profiles in numerical runup simulations. A climatology of storm wave run-up at Imperial Beach, California is developed using the numerical model SWASH, and over a decade of hindcast spectral waves and observed depth profiles. For use in a local flood warning system, the relationship between incident wave energy spectra E(f) and SWASH-modeled shoreline water levels is approximated with the numerically simple integrated power law approximation (IPA). Broad and multi-peaked E(f) are accommodated by characterizing wave forcing with frequency-weighted integrals of E(f). This integral approach improves runup estimates compared to the more commonly used bulk parameterization using deep water wave height H_0 and deep water wavelength L_0 Hunt (Trans Am Soc Civ Eng 126(4):542–570, 1961) and Stockdon et al. (Coast Eng 53(7):573–588, 2006. https://doi.org/10.1016/j.coastaleng.2005.12.005). Scaling of energy and frequency contributions in IPA, determined by searching parameter space for the best fit to SWASH, show an H_0L_0 scaling is near optimal. IPA performance is tested with LiDAR observations of storm run-up, which reached 2.5 m above the offshore water level, overtopped backshore riprap, and eroded the foreshore beach slope. Driven with estimates from a regional wave model and observed beta _f, the IPA reproduced observed run-up with <30% error. However, errors in model physics, depth profile, and incoming wave predictions partially cancelled. IPA (or alternative empirical forms) can be calibrated (using SWASH or similar) for sites where historical waves and eroded bathymetry are available.

Probabilistic Prediction of Significant Wave Height Using Dynamic Bayesian Network and Information Flow

Short-term prediction of wave height is paramount in oceanic operation-related activities. Statistical models have advantages in short-term wave prediction as complex physical process is substantially simplified. However, previous statistical models have no consideration in selection of predictive variables and dealing with prediction uncertainty. This paper develops a machine learning model by combining the dynamic Bayesian network (DBN) with the information flow (IF) designated as DBN-IF. IF is focused on selecting the best predictive variables for DBN by causal analysis instead of correlation analysis. DBN for probabilistic prediction is constructed by structure learning and parameter learning with data mining. Based on causal theory, graph theory, and probability theory, the proposed DBN-IF model could deal with the uncertainty and shows great performance in significant wave height prediction compared with the artificial neural network (ANN), random forest (RF) and support vector machine (SVM) for all lead times. The interpretable DBN-IF is proven as a promising tool for nonlinear and uncertain wave height prediction.

A new empirical equation of shear wave velocity to predict the different peak surface accelerations for Jakarta city

Site condition and bedrock depth play important roles in the determination of peak surface acceleration (PSA) values by earthquake motions. The soil parameters of shear wave velocity (Vs) and standard penetration test-number (N) value for Jakarta city are available up to 100 m below the Earth's surface even though the typical depths to bedrock are in excess of 100 m. This study referred to the base motion peak ground acceleration (PGA) values of 0.100 g, 0.218 g and 0.378 g to predict the PSA values using the Nonlinear Earthquake site Response Analysis (NERA) to analyse a simulated dataset for the bedrock depths of 100 m, 200 m, 300 m, 400 m and 500 m with conditioned by clayey and sandy soils. A new empirical equation of Vs = 102.48 N0.297 (m/s) was proposed to calculate the values of Vs used as an input parameter in the NERA programme for the prediction of seismic wave propagation.The results showed that the PSA values are dependent on the amplitude of seismic waves, depths of bedrock and the local site conditions. Changes in the PSA values from 41.0% to 51.5% and from 46.1% to 79.8% for the bedrocks overlain by sand, from 20.0% to 42.1% and from 45.9% to 58.8% for the bedrocks overlain by clay with increasing of bedrock depths from 200 m to 300 m and from 400 m to 500 m, respectively, were predicted for a 2500-year return period earthquake. Decreases in the PSA values by 41.0%, 51.5%, 46.1%, 79.8% for the bedrocks overlain by sand and by 20.0%, 42.1%, 45.9%, 58.8% for the bedrocks overlain by clay were predicted for a 2500-year return period earthquake due to the bedrock depth changes of 200 m, 300 m, 400 m, 500 m.Large-magnitude earthquake of Jakarta city has a significant effect on an increase or a decrease of the PSA value with depth of bedrock and may cause the vibration damage to buildings and other constructions on the ground. The analysis of the PSA value and PSA ratio influenced by the PGA value, bedrock depth and local soil conditions will make a contribution to the design of earthquake-safe building for Jakarta city in the future.

Simulation and Prediction of Storm Surges and Waves Using a Fully Integrated Process Model and a Parametric Cyclonic Wind Model

AbstractThis paper presents a fully integrated coastal process model and a simple parametric cyclonic wind‐pressure model for simulation of wind, storm surges, waves, tidal currents, and river flows. By sharing one computational grid within all those process modules and no need for switching executable codes from one module to another, this full‐coupling feature eliminates possible errors and loss of information due to interpolation and extrapolation of variables between different grids. To generate better cyclonic wind speed and barometric pressure, this parametric wind model includes nonlinear decay effect on wind intensity after hurricane's landfall. By implementing a new wind energy source term, the wave action model is capable of computing wave growth, propagation, and deformation through a regional‐scale domain from deepwater to shallow waters. Model validation and model skill assessment were performed by hindcasting wind, storm surges, waves, and river flows during Hurricane Gustav (2008) by using a high‐resolution grid covering the northern Gulf Coast. With improved wind fields estimated by the new parametric wind model, this fully integrated process model produced high‐quality wavefields in deep and shallow waters and storm tidal levels in the northern Gulf Coasts. Because of computing efficiency provided by seamless integration of process modules and optimized numerical solution schemes, faster‐than‐real‐time predictions of storm surges for the advisories during Hurricane Isaac (2012) were achieved by running the validated model in a desktop computer.

Experimental and numerical assessment of deterministic nonlinear ocean waves prediction algorithms using non-uniformly sampled wave gauges

We assess the capability of fast wave models to deterministically predict nonlinear ocean surface waves from non-uniformly distributed data such as sampled from an optical ocean sensor. Linear and weakly nonlinear prediction algorithms are applied to long-crested irregular waves based on a set of laboratory experiments and corresponding numerical simulations. An array of wave gauges is used for data acquisition, representing the typical spatial sampling an optical sensor (e.g., LIDAR) would make at grazing incidence. Predictions of the weakly nonlinear Improved Choppy Wave Model are compared to those of the Linear Wave Theory with and without a nonlinear dispersion relationship correction. Wave models are first inverted based on gauge data which provides the initial model parameters, then propagated to issue a prediction. We find that the wave prediction accuracy converges with the amount of input data used in the inversion. When waves are propagated in the models, correctly modeling the nonlinear wave phase velocity provides the main improvement in accuracy, while including nonlinear wave shape effects only improves surface elevation representation in the spatio-temporal region where input data are acquired. Surface slope prediction accuracy, however, strongly depends on the appropriate nonlinear wave shape modeling.

Satellite-retrieved sea ice concentration uncertainty and its effect on modelling wave evolution in marginal ice zones

Abstract. Ocean surface waves are known to decay when they interact with sea ice. Wave–ice models implemented in a spectral wave model, e.g. WAVEWATCH III® (WW3), derive the attenuation coefficient based on several different model ice types, i.e. how the model treats sea ice. In the marginal ice zone (MIZ) with sea ice concentration (SIC) &lt; 1, the wave attenuation is moderated by SIC: we show that subgrid-scale processes relating to the SIC and sea ice type heterogeneity in the wave–ice models are missing and the accuracy of SIC plays an important role in the predictability. Satellite-retrieved SIC data (or a sea ice model that assimilates them) are often used to force wave–ice models, but these data are known to have uncertainty. To study the effect of SIC uncertainty ΔSIC on modelling MIZ waves during the 2018 R/V Mirai observational campaign in the refreezing Chukchi Sea, a WW3 hindcast experiment was conducted using six satellite-retrieved SIC products based on four algorithms applied to SSMIS and AMSR2 data. The results show that ΔSIC can cause considerable wave prediction discrepancies in ice cover. There is evidence that bivariate uncertainty data (model significant wave heights and SIC forcing) are correlated, although off-ice wave growth is more complicated due to the cumulative effect of ΔSIC along an MIZ fetch. The analysis revealed that the effect of ΔSIC can overwhelm the uncertainty arising from the choice of model ice types, i.e. wave–ice interaction parameterisations. Despite the missing subgrid-scale physics relating to the SIC and sea ice type heterogeneity in WW3 wave–ice models – which causes significant modelling uncertainty – this study found that the accuracy of satellite-retrieved SIC used as model forcing is the dominant error source of modelling MIZ waves in the refreezing ocean.

A multi-layer perceptron approach for accelerated wave forecasting in Lake Michigan

A machine learning framework based on a multi-layer perceptron (MLP) algorithm was established and applied to wave forecasting in Lake Michigan. The MLP model showed desirable performance in forecasting wave characteristics, including significant wave heights and peak wave periods, considering both wind and ice cover on wave generation. The structure of the MLP regressor was optimized by a cross-validated parameter search technique and consisted of two hidden layers with 300 neurons in each hidden layer. The MLP model was trained and validated using the wave simulations from a physics-based SWAN wave model for the period 2005–2014 and tested for wave prediction by using NOAA buoy data from 2015. Sensitivity tests on hyperparameters and regularization techniques were conducted to demonstrate the robustness of the model. The MLP model was computationally efficient and capable of predicting characteristic wave conditions with accuracy comparable to that of the SWAN model. It was demonstrated that this machine learning approach could forecast wave conditions in 1/20,000th to 1/10,000th of the computational time necessary to run the physics-based model. This magnitude of acceleration could enable efficient wave predictions of extremely large scales in time and space.

Pile Length Estimate of Impulse Response Testing with Guided Wave Approach

The constant velocity one-dimensional wave theory has been commonly used to determine the pile lengths from the impulse response testing results for several years. In this paper, the three-dimensional guided wave theory was introduced to extract more information from the impulse response spectra on five concrete piles. Combined with resonance concept, this guided wave-based approach was developed to simply and efficiently identify the length estimates with 3% or less errors, similar to those of the conventional one-dimensional wave-based approach. Moreover, the guided wave theory provides more individual resonant frequencies insight over that of the one-dimensional wave theory. All the experimental phase velocities at their corresponding resonant frequencies have a good consistency with the guided wave prediction curve at frequencies up to 5000 Hz. The resulting scattering also presents a slightly decrease trend with frequency. The overall analysis indicates that the guided wave theory can give more useful information than that of the one-dimensional wave theory. In future, one can use this inverse computation approach to evaluate unknown pile lengths.

Wave Data Assimilation to Modify Wind Forcing Using an Ensemble Kalman Filter

In order to improve the predictability of winter storm waves in the East Sea, this article explores the use of the ensemble Kalman filter technique for data assimilation in wave modeling. The nested wave model has been established using SWAN along the east coast of Korea to simulate wave transformation and wave dissipation in coastal areas to obtain a better modeling performance with regard to wind waves and swells in the East Sea. The regional atmospheric model is used to provide high-resolution forcing winds. These are adjusted by directly assimilating measurements of offshore wave heights into the wave model state. The model setup, data assimilation parameters, and validation of prediction are described with optimal conditions during the stormy periods in 2015. The ensemble Kalman filter data assimilation has shown itself to be very efficient, leading to large reductions of up to 40% in the root-mean- square error of the signification wave height compared to the results with and without data assimilation at locations other than those of the observations used. It shows that the wave modeling with ensemble Kalman filter data assimilation is very feasible to predict coastal waves, in particular storm events in the East Sea. kw]Keywords -winter storm waves, wave modeling, data assimilation, wind forcing, East Sea

A Technique of Panels Cutting for Modification of Hull Geometry Hydrodynamic Models

A panel cutting technique is developed for automatic modification of an initial mesh of a ship hull used for hydrodynamic computations leading to improved meshes for the prediction of wave induced vertical load effects. The technique can provide a model with divided panels in any defined position regardless of the initial discretization of the body. The applications of the provided technique include panel distinction and division in predetermined positions, generation of finer mesh based on the initial coarser model of meshes and improvement of vertical load prediction in predetermined positions. The method is applied for case studies of a barge, shuttle tanker and frigate to depict various applications. Finally, the hydrostatic and hydrodynamic vertical shear forces are calculated for two models of initial and modified panels of well-known frigate 5415. The results are compared for is shown for predetermined sections.