scholarly journals Bayesian Experimental Design of Cyber-Physical Tests for Hydrodynamic Model Calibration

2021 ◽  
Author(s):  
Giuseppe Abbiati ◽  
Thomas Sauder

An application of cyber-physical testing to the empirical estimation of difference-frequency quadratic transfer functions is presented. As an alternative to today's procedure based on hydrodynamic tests with broad-banded or realistic (e.g., JONSWAP) wave spectra, tests in bichromatic waves are considered. The laboratory setup is the one developed by Sauder \& Tahchiev (2020) that enables magnifying the sensitivity of the floater response to the low-frequency wave loading by adjusting the stiffness and damping parameters of a virtual soft mooring system. Bayesian experimental design is proposed to optimize the selection of the control variables (frequencies in the bichromatic wave and properties of the virtual mooring system) for a batch of cyber-physical tests. The experimental design algorithm is based on the recent work of Huan \& Marzouk (2013). In a virtual yet realistic case study using an uncertain parametric quadratic transfer function, we demonstrate how the uncertainty of its describing parameters and other calibration parameters (low-frequency added mass and hydrodynamic damping) can be reduced. Results indicate that the proposed procedure has the potential for reducing experimental cost for calibration of hydrodynamic models.

2021 ◽  
Author(s):  
Giuseppe Abbiati ◽  
Thomas Sauder

Abstract An application of cyber-physical testing to the empirical estimation of difference-frequency quadratic transfer functions is presented. As an alternative to today’s procedure based on hydrodynamic tests with broad-banded or realistic (e.g., JONSWAP) wave spectra, tests in bichromatic waves are considered. The laboratory setup is the one developed by Sauder & Tahchiev (2020) that enables magnifying the sensitivity of the floater response to the low-frequency wave loading by adjusting the stiffness and damping parameters of a virtual soft mooring system. Bayesian experimental design is proposed to optimize the selection of the control variables (frequencies in the bichromatic wave and properties of the virtual mooring system) for a batch of cyber-physical tests. The experimental design algorithm is based on the recent work of Huan & Marzouk (2013). In a virtual yet realistic case study using an uncertain parametric quadratic transfer function, we demonstrate how the uncertainty of its describing parameters and other calibration parameters (low-frequency added mass and hydrodynamic damping) can be reduced. Results indicate that the proposed procedure has the potential for reducing experimental cost for calibration of hydrodynamic models.


Author(s):  
Olaf J. Waals

Operability of offshore moored ships can be affected by low frequency wave loads. The low frequency motions of a moored ship may limit the uptime of an offshore structure such as an LNG offloading terminal. The wave loads that cause the main excitation of these low frequency motions are usually computed using second order wave drift theory for long crested waves, which assumes that the low frequency components are only related to waves coming from the same direction. In this method short crested seas are dealt with as a summation of long crested seas, but no interaction between the wave components traveling in different directions is usually taken into account. This paper describes the results of a study to the effect of 2nd order low frequency wave loads in directional seas. For this study the drift forces related to the interaction between waves coming from different directions is also included. This is done by computing the quadratic transfer functions (QTF) for all possible combinations of wave components (frequencies and directions). Time traces of drift forces are generated and compared to the results without wave directional interaction after which the motions of an LNG carrier are simulated. A sensitivity study is carried out towards the number of direction steps and the water depth. Finally the motions of an LNG carrier in shallow water (15m water depth) are simulated and mooring forces are compared for various amounts of wave spreading.


1987 ◽  
Vol 109 (2) ◽  
pp. 133-141
Author(s):  
R. Eatock Taylor ◽  
P. Sincock

This paper investigates methods of simulating the combined wave frequency and low-frequency wave drift forces and motion responses of floating systems. This is motivated by the requirement for estimates of response statistics for systems on nonlinear moorings. Results are given for a linear system for which experimental data are available (an articulated column model); and for a moored barge on mildly and strongly nonlinear moorings. Estimates are obtained for low-frequency spectra, linear and quadratic transfer functions, response probability distributions, and peak distributions.


Author(s):  
Flávia C. Rezende ◽  
Cédric Brun

In the analysis of the mooring system, one of the most important steps concerns the computation of the low-frequency wave loads that will excite the system at its resonance periods. Therefore, over the last decades much work has been devoted to this topic, especially for unidirectional waves although some publications can be found which deal with the computation of second-order loads in multi-directional waves. The common practice in the design of mooring systems is to ignore the interaction between waves coming from different directions even if we know that the unidirectional sea is an idealized situation that does not exist in practice. In Rezende & Chen (OMAE 2010), new method for computation of the low-frequency loading has been presented which is more efficient than the method based on the computations of the complete QTF in multi-directional waves, even though this method is still much more time consuming than the usual approximations in unidirectional seas. In this paper, the effects of the directionality of the waves are evaluated for typical mooring systems in deep water and in shallow water. The contribution of the directional interaction terms in the loads will be assessed in order to conclude on its importance to the mooring design.


1975 ◽  
Vol 15 (06) ◽  
pp. 487-494 ◽  
Author(s):  
J.A. Pinkster

Abstract The influence of the low-frequency-wave-drifting force on the motions of moored vessels and the loads in the mooring system is demonstrated from results of model tests in irregular waves. The origin of the wave drifting force is discussed and methods for calculating the mean drifting force are reviewed. To facilitate calculation of the low-frequency-wave drifting force on an object in irregular waves, an existing method using the mean drifting force in regular waves is generalized. The results of calculations using the method introduced in this paper are compared with previously published test results. Finally, some remarks are added concerning effects that have not been accounted for in existing calculation methods. Introduction A vessel moored at sea in stationary conditions with regard to waves, wind, and current is subjected to forces that tend to shift it from the desired position. For a given vessel and position in the position. For a given vessel and position in the horizontal plane, the motions depend on both the mooring system and the external forces acting on the vessel. In steady conditions, the forces caused by a constant wind and current are constant quantities for a given heading angle of the vessel. The forces caused by a stationary irregular sea are of an irregular nature and may be split into two parts: first-order oscillatory forces with wave parts: first-order oscillatory forces with wave frequency, and second-order, slowly varying forces with frequencies much lower than the wave frequency.The first-order oscillatory wave forces on a vessel cause the well known ship motions whose frequencies equal the frequencies present in the spectrum of the irregular waves. These are the linear motions of surge, sway, and heave and the three angular motions of roll, pitch, and yaw. In general, the first-order wave forces are proportional to the wave height, as are the ensuing motions. The magnitude of the linear oscillatory motions is in the order of the height of the waves.The second-order wave forces, perhaps better known as the wave drifting forces, have been shown to be proportional to the square of the wave height. These forces, though small in magnitude, are the cause of the low-frequency, large-amplitude, horizontal motions sometimes observed in large vessels moored in irregular waves. Tests run in irregular waves in wave tanks of the Netherlands Ship Model Basin revealed a number of properties and effects of the low-frequency-wave properties and effects of the low-frequency-wave drifting force that are discussed here using the results of two test programs.The first of these programs concerns tests run with the model of a 125,000-cu m LNG carrier moored in head seas with an ideal linear mooring system. The second program deals with a 300,000-DWT VLCC moored with a realistic nonlinear bow hawser to a single-buoy mooring in waves, wind, and current coming from different directions.The results of the tests with the LNG carrier are shown in Figs. 1 through 3, while the results of the tests with the 300,000-DWT VLCC are shown in Fig. 4. All results are given in full-scale values. Fig. 1 shows the wave trace and the surge motion of the LNG carrier to a base of time. SPEJ P. 487


2020 ◽  
Vol 143 (1) ◽  
Author(s):  
Nuno Fonseca ◽  
Carl Trygve Stansberg ◽  
Kjell Larsen ◽  
Rune Bjørkli ◽  
Tjerand Vigesdal ◽  
...  

Abstract Model tests have been performed with four mobile offshore drilling units (MODUs) with the aim of identifying wave drift forces and low-frequency damping. The MODUs configuration is different, namely on the number and diameter of columns; therefore, the sample is representative of many of the existing concepts. The model scale is the same as well as the wave and current conditions. The experimental program includes irregular waves with systematic variations of the significant wave height, wave peak period, current velocity, and vessel heading. A nonlinear data analysis technique (cross bi-spectral analysis) is applied to identify the surge and sway quadratic transfer functions (QTFs) of the slowly varying excitation, together with the linearized low-frequency damping. The paper also presents a semi-empirical formula developed in the scope of the EXWAVE JIP to correct potential flow mean wave drift force coefficients of Semis in high seastates with current. The empirical QTFs are then compared with numerical predictions. Comparisons with potential flow coefficients lead to conclusions on the role of viscous drift. The semi-empirical formula is assessed based on comparisons with test results and concluded that it provides a significant improvement compared to potential flow predictions.


Author(s):  
Min Zhang ◽  
Junrong Wang ◽  
Junfeng Du ◽  
Nuno Fonseca ◽  
Galin Tahchiev ◽  
...  

Abstract The paper presents calibration and validation of a time domain numerical model for mooring analysis of a spread moored FPSO in moderate seastates with and without current. The equations of motion are solved in the time domain with a fully coupled method, accounting for linear wave frequency (WF) radiation and diffraction, second order wave drift forces and nonlinear low frequency (LF) damping. The mooring system dynamics is solved by a FEM. Uncalibrated numerical models are based on input from the mooring system, vessel mass, radiation/diffraction analysis, decay tests and current coefficients. WF responses are very well predicted by standard radiation/diffraction linear analysis, therefore the focus is on the LF responses. LF motions are underpredicted by the uncalibrated numerical model. Calibration is performed by comparing simulations with model test data and adjusting hydrodynamic coefficients known to be affected by uncertainty. These include wave drift force coefficients and LF damping. Correction of the drift coefficients is based on empirical quadratic transfer functions (QTFs) identified from the test data by a nonlinear data analysis technique known as “cross-bi-spectral analysis”. The LF damping coefficients are then adjusted by matching low frequency surge and sway spectra from the model tests and from the simulations.


2009 ◽  
Vol 131 (4) ◽  
Author(s):  
S. Ma ◽  
M. H. Kim ◽  
S. Shi

This paper studies the influence of three different calculation methods of the second-order low-frequency (LF) wave-force quadratic transfer functions (QTFs) for a single point mooring (SPM) tanker system in relatively shallow water region. The multivessel-mooring hawser coupled dynamic analysis is used to simulate the floater relative motions and mooring and hawser tensions. Because the SPM tanker is deployed in shallow water region and the slowly varying drift motions are to be dominant in typical operational conditions, the accurate calculation of LF wave-force QTFs become important especially for mooring and hawser-tension prediction. The practically popular Newman’s approximation and another approximation excluding complicated free-surface integrals are used to calculate the LF QTFs on the offloading tanker and they are compared with the complete QTF results. Further comparison is carried out by calculating the resulting LF wave-force spectra and motion time histories and analyzing their impacts on hawser and mooring line tensions. Through the example studies, the limitation of the Newman’s approximation in the case of shallow water and longer period irregular waves is underscored.


Author(s):  
S. Ma ◽  
S. Shi ◽  
M. H. Kim

This paper studies the influence of three different calculation methods of the second-order low-frequency (LF) wave forces on the tanker responses and hawser/mooring tensions in relatively shallow water region. The vessel-mooring-riser coupled dynamic analysis computer program HARP is used to simulate the coupled dynamic responses of offloading tanker moored to a SPM (Single Point Mooring). Because the SPM is supposed to be deployed in shallow water and the slowly varying drift motions of the tanker are to dominate the motion responses in typical operational conditions, the accurate calculation of LF wave-force quadratic transfer functions (QTFs) becomes important especially for mooring and hawser tensions. Like common practice, the so-called Newman’s approximation and another approximation method without including complicated free-surface integrals are first used to calculate the LF QTFs on the offloading tanker and they are compared with the complete QTF results. Further comparison is performed by calculating the resulting LF wave-force spectra and response time series by using the three different methods. The impact of the three different approaches on vessel surge motions and hawser/mooring line tensions is also addressed.


Author(s):  
Nuno Fonseca ◽  
Carl Trygve Stansberg

The paper presents calibration of a time domain numerical model for the motions of the Exwave Semi in high seastates with current. The time domain equations of motion combine linear radiation, linear diffraction and second order wave drift forces, based on MULDIF diffraction code, with nonlinear forces from quadratic damping and from the mooring system. Calibration is performed by comparing simulations with model test data and adjusting hydrodynamic coefficients known to be affected by uncertainty. These include wave drift force coefficients, damping and added mass coefficients. Correction of the drift coefficients is based on empirical quadratic transfer functions (QTFs) identified from the test data by a nonlinear data analysis technique known as “cross-bi-spectral analysis”. Initial “uncalibrated” numerical models are based on input from the mooring, vessel mass, MULDIF hydrodynamic analysis, decay tests and current coefficients. They need adjustments for surge and sway. Empirical drift coefficients, natural periods and damping coefficients are then adjusted by matching low frequency surge and sway spectra. Wave-frequency coefficients need no adjustment. Low frequency wave drift forces, damping and added mass need increase in high sea states, in particular with current. Final motion simulations show 30%–40% underestimation in initial simulations, while final calibrated simulations are close to the measured records.


Sign in / Sign up

Export Citation Format

Share Document