Marine infrastructure assessment and fitness monitoring. Previous long-term measurements of the Uldolmok tidal current power station showed that the structure's natural frequencies oscillate with a constant cycle—. In this study, laboratory-scale experiments on a simplified offshore structure as a laboratory-scale test structure were conducted in a circulating water channel to thoroughly investigate the causes of fluctuation of the natural frequencies and to validate the displacement estimation method using multimetric data fusion.
INTRODUCTION
Literature Survey
Many studies have also been conducted to estimate infrastructure damage using atypical methods such as modal strain energy (Huynh et al., 2013). Park et al., (2013) proposed the data fusion method using strain and acceleration responses and validated it by comparing the multimetric displacement of the proposed method with the real displacement of a LASER displacement sensor through the field experiments on Sorok Bridge. It is necessary to conduct experiments to validate this data fusion method in cantilever beams, such as those used in monopole-type tidal turbines such as MCTs (Marine Current Turbines) (Fraenkel, 2007).
Research Objectives and Scope
The use of an LVDT in offshore structures is impossible; a fixed point is needed to measure the relative displacement. However, a strain gauge and accelerometer are easy to install on the structure and obtain strain and acceleration responses. This thesis was written based on the author's article entitled 'Issues in structural health monitoring for fixed-type offshore structures under harsh tidal environments, published in Smart Structures and System (Jung et al., 2015).
THEORETICAL BACKGROUND
Displacement Estimation Using Data Fusion
- Example of Displacement Estimation Algorithm
The estimated displacement can be calculated as Eq. 4), which is the analytical solution of Eq. 2.5), known as Tikhon's arrangement scheme. Theoretically, the acceleration can be obtained from the derivative of the second-order displacement; however, the displacement calculated in this way would be divergent without the boundary condition. Park et al., (2013) developed this method by adding a quasi-static displacement component from the strain response—the so-called data fusion in Eqs.
Estimating the static displacement from drag forces is important because internal forces such as the bending moment and the amount of external force can be determined. Unlike the tidal stream power plant including Uldolmok TCPP, this model represented a wind turbine with a relatively simple structure and under simple external force, because this structure is monopile type and the governing force is the thrust on the top of the wind tower. 2.4, especially in the red box, the dynamic displacement of free vibration after about 150 seconds is almost the same as the displacement under acceleration alone.
Acceleration of the top of the tower and load of the bottom of the tower were measured. By using this algorithm, a graph with the same trend of static displacement can be calculated from dynamic load response as Fig. Alpha is the ratio of amplitudes between the first natural frequency of displacement from load only and that from acceleration only.
Damage index(β)
However, estimating quasi-static displacement is difficult; since, the neutral axis of a structure may depend on the load intensity when the structure does not have perfect symmetrical structural property. Equation (2.17) shows damage index, which has the same base as the normalized damage factor (NDF) proposed by Park et al, where αIntactCase,αDamageCase are the scaling coefficients in Intact Case and Damage Case, respectively.
EXPERIMENTAL STUDY
Information on the Lab-scale Test Structure
Note that the structural behavior of the Uldolmok TCPP was influenced more by the tidal current than by waves. The laboratory scale test structure of the Uldolmok TCPP is designed in the form of a monopile support structure. The natural frequencies of the test structure are higher than those of the Uldolmok TCPP.
However, these similarity rules were not used in this laboratory-scale experiment; as it is beyond the scope of this study. The aim of this study is limited to analyze the effect of added mass and the stiffness of the foundation on the structural dynamic properties. 3.3, the laboratory-scale test structure is made of stainless steel, and its height, diameter, and thickness are 1.1 m, 38 mm, and 2 mm, respectively.
The pump of the circular water channel caused vibration in the experimental field, much more vibration was in the channel due to the flow. Therefore, we could not install a laser displacement meter on the channel to measure the structural behavior in the current direction and instead used a video camera to measure from the outside of the water channel. To reduce error and improve accuracy, the video camera was placed as close as possible to the laboratory testing structure.
The hydrometer was installed at least 5 m from the front of the laboratory-scale test structure to avoid flow interference.
Outline of the Lab-scale Experiment
An image processing technique was used to extract the displacement from the registration data (Yi et al., 2013c). However, the FFT resolution is not high enough to observe the variation of natural frequencies. The second experiment was carried out with the variation of the current speed intact and two cases of damage to the boundary connections.
In Intact Case, the foundation of the laboratory-scale test structure and the bottom of the water channel are connected by fastening four bolts. Because the laboratory-scale test structure was subjected to the current load, which causes ambient vibration, extra excitation was not required.
Analysis of Experimental Results
- Added Mass Effect
- Boundary Damage Effect from Current Velocity Change
- Damage Index (β)
- Displacement Estimation
In the case of the first bending mode, the natural frequency should theoretically decrease as the water level increases. Power spectral density (PSD) data of the acceleration responses at each flow rate in damage case 2 are plotted in Figure. In the intact and damage case 1 cases, there were small changes in the natural frequencies; however, the first natural frequency in damage case 2 decreased from 7.3 to 6.7 Hz - to 8.25% with increasing water speed.
The reason why the natural frequencies decrease as the degree of damage to the boundary joints becomes severe is due to the reduction of stiffness in the boundary joints. Also, no effects of water velocity in Intact Case and Damage Case 1 indicate that the entire system of the laboratory-scale test structure is a linear system. However, if the damage to the boundary connection is severe, as in damage case 2, lowering the natural frequency as the water velocity increases means that the laboratory-scale test structure may behave as a nonlinear system.
Since the boundary connection condition changes only in this experiment, the boundary connection is nonlinear. Therefore, foundation damage is much more influential than the effect of added mass associated with tidal height, as the first TCPP Uldolmok frequency varies by up to 16% per day, and severe foundation damage can cause non-linearity. in the structural system. a) PSD data from acceleration in case of damage 2. b) Specific change of the first natural frequency. Regarding the stability of the overall structure, however, additional research is needed to assess quantitative damage.
The damage index β in the case of damage 1 has a higher value than in the case of damage 2; i.e. border damage becomes severe. It is possible that these vibrations are transmitted to the tripod video camera and introduced by experimental errors due to the maximum displacement of the laboratory scale. We introduced two error factors as shown in Eqs. 3.1) and (3.2): Err1 is the deviation of the maximum actual (based on vision) and estimated displacements (ureal,max and uestimated,max), and Err2 is the root mean square error (RMSE) of the deviation between the actual and estimated displacements (ureal and estimated ). The Err1 and Err2 values of all cases are listed in Table 5.
VERIFICATION OF THE EXPERIMENT
Added Mass Effect
However, in the case of the second bending mode in the fore-aft direction, the experimental and numerical values are very close to each other. Therefore, it can be confirmed that the previous equation for added mass—Eq. 2.2)—is correct for the second bending mode. However, in the case of the first bending mode, the special constant is needed to fit the experimental and numerical values.
We could find the correct added mass coefficient in the first bending mode by updating the model.
CONCLUSION
Future Study
Also, it was difficult to fix the test structure because it is impossible to make holes to fix it;. After checking whether the water is inside the submerged parts of Uldolmok TCPP and the laboratory-scale test structure, a numerical and experimental study should be carried out to compare the real phenomenon, which is the frequency change, of Uldolmok TCPP with the laboratory-scale test structure . According to the word, the laboratory-scale test structure of the connecting column and its foundation with concrete grout can be used.
Before curing, the current load will be subjected to test structure at regular intervals and responses such as natural frequency and displacement can be observed as time passes. For wind turbines, the main load on the tower is the pressure force on top of the tower. It absolutely caused this structure to vibrate as the first state form; however, controlling load under other offshore structures such as tidal current power plant and wave power.
This means that other higher modes become involved in the vibration because the excitation location is not on top of the structure. To improve the accuracy of displacement estimation, higher modes should be used in this algorithm. Fraenkel, P.L., 2007, Marine current turbines: pioneering the development of marine kinetic energy converters, Journal of Power and Energy, 221(2), pp.159-169. amp; Wootton, L.R., 1977, Dynamics of marine structures: methods of calculating the dynamic response of fixed structures subject to wave and current action.
Juang, J.N., & Pappa, R.S., 1985, A eigensystem realization algorithm for modal parameter identification and model reduction, Journal of Guidance, Control, and Dynamics, 8(5), pp.620-627.
