MODELLING TWO-PHASE FLOW-EXCITED DAMPING AND FLUIDELASTIC INSTABILITY IN TUBE ARRAYS

This paper reports the results of an experimental study of the flow-induced vibration of a heat exchanger tube array subjected to two-phase cross-flow of refrigerant 11. The primary concern of the research was to develop a methodology for predicting the critical flow velocities for fluidelastic instability which better characterize the physics of two-phase flows. A new method is proposed for calculating the average fluid density and equivalent flow velocity of the two-phase fluid, using a newly developed void fraction model to account for the difference in velocity between the gas and liquid phases. Additionally, damping measurements in two-phase flow were made and compared with the data of other researchers who used a variety of modelling fluids. The results show that the two-phase damping follows a similar trend with respect to homogeneous void fraction, and when normalized, agree well with the data in the literature. The fluidelastic threshold data of several researchers who used a variety of fluids, is re-examined using the proposed void fraction model, and the results show a remarkable change in trend with flow regime. The data corresponding to the bubbly flow regime shows no significant deviation from the trend established by Connors' theory. However, the data corresponding to the intermittent flow regime show a significant decrease in stability which is nearly independent of the mass-damping parameter. It is believed that the velocity fluctuations that are inherent in the intermittent flow regime are responsible for tripping the instability, causing lower than expected stability of the bundle.     

管阵流体弹性不稳定性分析

在设计换热器、蒸汽发生器等核电站和石油化学工业设备的管阵时都会遇到横向流作用下的流体诱发振动问题.无论是从引起振动的可能性,还是从产生振动后的危害性来看,流体弹性不稳定性是管阵流体诱发振动的最重要机理.本文在比较、分析现行的流体弹性不稳定性数学模型的基础上,提出了一种新的数学模型,给出了不稳定性临界流速的计算方法,开发了相应的计算机程序.与实验结果和其它理论模型的计算结果比较表明:本文模型应用方便,得到的结果与实验结果基本一致.     

AN EXPERIMENTAL STUDY ON THE FLUIDELASTIC FORCES ACTING ON A SQUARE TUBE BUNDLE IN TWO-PHASE CROSS-FLOW

A tube in a square tube bundle of P/D=1·42 was oscillated in the lift direction in air–water two-phase cross-flow, and fluidelastic forces acting on the oscillated tube were measured. First, the tube amplitude was fixed to 3 mm (=0·136 D), and added mass, damping, and stiffness coefficients were obtained as a function of two-phase mixture characteristics such as nondimensional gap velocity and void fraction. When reference mixture density and velocity were estimated, the drift–flux model, in which the relative velocity between the gas and liquid phases was estimated, generated better results than the homogeneous model. The added mass coefficient was obtained from quiescent two-phase flow as a function of void fraction. Using the added mass coefficient, the added stiffness coefficient converged to zero with decreasing nondimensional gap velocity. This overcame the contradiction in the added stiffness estimation without added mass, in which the added stiffness coefficient did not converge to zero with decreasing nondimensional gap velocity. Next, the effects of the vibration amplitude on the fluidelastic force coefficients were considered. When the tube amplitude was 3 mm (=0·136 D) or less, the equivalent added stiffness and damping coefficients were almost constant and nonlinearity was small. This showed the validity of the fluidelastic force coefficients obtained based on the data of amplitude of 3 mm. The linearity did not exist when the tube displacement amplitude was 4·5 mm (=0·205 D) or more; a remarkable nonlinearity appeared in the equivalent added damping coefficient. A method to estimate the limit-cycle amplitude of the fluidelastic vibration was proposed when only one tube in the tube bundle was able to vibrate in the lift direction. The amplitude could be obtained from the amplitude at which the equivalent added damping coefficient changed from negative to positive with increase in the tube amplitude.     

Modelling of fluidelastic instability in a square inline tube array including the boundary layer effect

Flow-induced vibration (FIV) is a design concern in many engineering applications such as tube bundles in heat exchangers. When FIV materializes, it often results in fatigue and/or fretting wear of the tubes, leading to their failure. Three cross-flow excitation mechanisms are responsible for such failures: random turbulence excitation, Strouhal periodicity, and fluidelastic instability. Of these three mechanisms, fluidelastic instability has the greatest potential for destruction. Because of this, a large amount of research has been conducted to understand and predict this mechanism. This paper presents a time domain model to predict the fluidelastic instability forces in a tube array. The proposed model accounts for temporal variations in the flow separation. The unsteady boundary layer is solved numerically and coupled with the structure model and the far field flow model. It is found that including the boundary layer effect results in a lower stability threshold. This is primarily due to a larger fluidelastic force effect on the tube. The increase in the fluidelastic effect is attributed to the phase difference between the boundary layer separation point motion and the tube motion. It is also observed that a non-linear limit cycle is predicted by the proposed model.     

NEW DESIGN RECOMMENDATIONS FOR FLUIDELASTIC INSTABILITY IN HEAT EXCHANGER TUBE BUNDLES

Design equations are presented to determine the critical velocities for the occurrence of fluidelastic instability in uniform single-phase cross-flow. These equations are an essential part of a new guideline in the “VDI-Wärmeatlas” for estimating vibration excitation in real tube bundle heat exchangers. Six existing guidelines for fluidelastic instability were tested and compared with about 300 experimental data from 34 papers (eight of them being not yet considered in a guideline before). New equations for the stability factor K as a function of the pitch ratio for different tube configurations were derived with statistical methods by a variation of the reference definitions for the structural parameters and the exponents of either the dimensionless mass and the damping or the mass-damping parameter. The criterion used here was first to be on the safe side with a minimum number, and minimum deviation, of experimental data below the recommended threshold line, and second a minimum r.m.s. error for all data considered. The pitch ratio has the strongest influence for the normal triangular array and the influence becomes less for the rotated as well as for the in-line square configuration. This significant gradation of the stability constants as a function of the pitch ratio and the tube bundle configuration enables a reasonable interpolation for non-standard configurations. The exponent of the dimensionless mass-damping parameter for gas cross-flow depends on the tube configuration and is 0·5 for the 30°- and 45°-configuration, and 0·4 for the 60°- and 90°-configuration. For liquid flow, an average exponent of 0·15 has been observed.     

Scaling of damping induced by bubbly flow across tubes

The damping of tubes subjected to two-phase air–water bubbly cross-flow is investigated with the use of an experimental database from several authors. A new definition of damping in stagnant flow is proposed using an extrapolation of the measured values at low dimensionless flow velocities. This approach yields values of damping substantially lower than those currently defined in the literature. They are found to vary continuously with void fraction, within the bubbly flow regime. These data are used to compare several models of the equivalent viscosity of a two-phase mixture. The effect of the flow velocity is then analysed up to fluidelastic instability. It is observed that, using scaling factors based on the characteristics of the liquid phase, fluidelastic effects of bubbly flows are closely related to those known in single-phase flows.     

Fluidelastic instability study in a rotated triangular tube array subject to two-phase cross-flow. Part I: Fluid force measurements and time delay extraction

Fluidelastic instability is a key issue in steam generator tube bundles subjected to cross-flow. The extension to two-phase flow of the existing theoretical models, developed and tested mostly for single phase flow, is investigated in this paper. The time delay is one of the key parameter for modeling fluidelastic instability, especially the damping controlled mechanism. The direct measurement of the time delay between the tube motion and the fluid force faces certain difficulties in two-phase flow since the high turbulence due to the interaction of the two components of the flow may increase the randomness of the measured force. To overcome this difficulty, an innovative method for extracting the time delay inherent to the quasi-steady model for fluidelastic instability is proposed in this study.Firstly, experimental measurements of unsteady and quasi-static fluid forces (in the lift direction) acting on a tube subjected to air–water two-phase flow were conducted. The unsteady fluid forces were measured by exciting the tube using a linear motor. These forces were measured for a wide range of void fractions, flow velocities and excitation frequencies. The experimental results showed that the unsteady fluid forces could be represented as single valued function of the reduced flow velocity. It was also found that for a given frequency, the unsteady fluid force phase was weakly dependent on the void fraction for the range of flow velocities considered.The time delay was determined by equating the unsteady fluid forces with the quasi-steady forces. The results given by this innovative method of measuring the time delay in two-phase flow were consistent with theoretical expectations. The time delay could be expressed as a linear function of the convection time and the time delay parameter was determined for void fractions ranging from 60% to 90%.     

Development of models correlating vibration excitation forces to dynamic characteristics of two-phase flow in a tube bundle

Recent experiments revealed significant quasi-periodic forces in both the drag and lift directions in a rotated triangular tube bundle subjected to two-phase cross-flow. The quasi-periodic drag forces were found to be related to the momentum flux fluctuations in the main flow path between the cylinders. The quasi-periodic lift forces, on the other hand, are mostly correlated to the oscillation in the wake of the cylinders. In this paper, we develop semi-analytical models for correlating vibration excitation forces to dynamic characteristics of two-phase flow in a rotated triangular tube bundle for a better understanding of the nature of vibration excitation forces. The relationships between the lift or drag forces and the dynamic characteristics of two-phase flow are established through fluid mechanics momentum equations. A model has been developed to correlate the void fraction fluctuation in the main flow path and the dynamic drag forces. A second model has been developed for correlating the oscillation in the wake of the cylinders and the dynamic lift forces. Although still preliminary, each model can predict the corresponding forces relatively well.     

Modeling of the phase lag causing fluidelastic instability in a parallel triangular tube array

Fluidelastic instability is considered a critical flow induced vibration mechanism in tube and shell heat exchangers. It is believed that a finite time lag between tube vibration and fluid response is essential to predict the phenomenon. However, the physical nature of this time lag is not fully understood. This paper presents a fundamental study of this time delay using a parallel triangular tube array with a pitch ratio of 1.54. A computational fluid dynamics (CFD) model was developed and validated experimentally in an attempt to investigate the interaction between tube vibrations and flow perturbations at lower reduced velocities U_r=1–6 and Reynolds numbers Re=2000–12 000. The numerical predictions of the phase lag are in reasonable agreement with the experimental measurements for the range of reduced velocities U_g/fd=6–7. It was found that there are two propagation mechanisms; the first is associated with the acoustic wave propagation at low reduced velocities, U_r<2, and the second mechanism for higher reduced velocities is associated with the vorticity shedding and convection. An empirical model of the two mechanisms is developed and the phase lag predictions are in reasonable agreement with the experimental and numerical measurements. The developed phase lag model is then coupled with the semi-analytical model of Lever and Weaver to predict the fluidelastic stability threshold. Improved predictions of the stability boundaries for the parallel triangular array were achieved. In addition, the present study has explained why fluidelastic instability does not occur below some threshold reduced velocity.     

换热器管束中的流体弹性不稳定性

文中重点讨论了管壳式换热器管束中的流体弹性不稳定性．从工程实用的观点出发，应用了Ｌｅｖｅｒ与Ｗｅａｖｅｒ的“流管”模型，在建立管子振动的基本方程时则认为管子作周期性的衰减（或发散）运动，以复数解法进行求解．根据所获得的对比流速与质量阻尼参数之间的关系画出的稳定区图，可用来确定管子的流体弹性不稳定性或该时的临界流速．实验数据表明，理论计算结果是令人满意的．     

FLUIDELASTIC ANALYSIS OF TUBE BUNDLE VIBRATION IN CROSS-FLOW

Tube bundles in cross-flow vibrate in response to motion-induced fluid-dynamic forces; hence, the resultant motions are considered to be a fluidelastic vibration. The characteristics of the vibration depend greatly on the features of the fluid-dynamic forces and the structure of the tube bundle. Therefore, in this study, the equations of motion of the tube bundle are derived. From the viewpoint of vibration, each tube is not independent of the surrounding tubes because its vibration is affected by fluid-dynamic coupling with the neighboring tubes. Thus, the equations are a set of coupled equations and the solution is obtained as an eigenvalue problem. The fluid-dynamic forces, which are indispensable in the calculation, have been obtained by experiments using a vibrating tube in the bundle; it was found that the forces depend strongly on the reduced velocity. Using these equations and the fluid forces, critical velocities of the tube bundle vibration are calculated, and it is found that the critical velocity is strongly dependent on the fluid-dynamic force characteristics, as they vary with the reduced velocity. Vibration tests of the tube bundle have also been conducted, and the critical velocities obtained in the tests are compared with the calculated values; agreement with the calculated values is good, demonstrating that the method of calculation is useful. The effects of mass ratio, frequency deviation and damping deviation of tubes in the bundle on the critical velocity are also examined theoretically. It is found that it is better to treat the mass ratio and the logarithmic decrement separately when the mass ratio is less than 10. Differences in natural frequencies make the critical velocity large. Similarly, differences in logarithmic decrement may distribute the vibration energy to other tubes and make the critical velocity large.     

Fluid-Damping Controlled Instability in Tube Bundles Subjected to Air Cross-Flow

There are different excitation mechanisms that cause fatal damages due to undesirable vibrations in heat exchanger tube bundles subjected to cross-flow. One of them is the fluid-damping-controlled instability (galloping) that is characterised by a sudden appearance of large amplitudes of the tubes exclusively in cross-flow direction. This paper reports on investigations using an experimental set-up in a wind tunnel where the galloping mechanism in a tube bundle can be observed as an isolated phenomenon. The apparatus allows to realise several tube bundle configurations and geometry's of real heat exchangers. The position of a flexible test tube with a linear iso-viscoelastic mounting inside the tube array is variable. The test tube is equipped with dynamical pressure sensors which are placed directly under pressure holes inside the tube. For the investigation of the acting fluid forces the non-stationary pressure distribution is measured simultaneously at 30 points on the circumference in mid plane and at 15 points in line along the tube together with the tube motion. The acting fluid forces are determined by integration of the whole pressure field process. The study gives insights into the effect of the fluid-damping-controlled instability that is still not fully understood. Moreover, a flow visualization gives an impression of the mechanism at relevant Reynolds-numbers. The results show that in case of instability due to galloping the correlation length of the forces acting along the tube axis increases suddenly to large values. The fluid forces are correlated well for the whole tube when galloping is dominant. The exciting fluid forces show harmonic character and lead to a classical resonance behaviour. Instead of a simple free vibration test in vacuum or still air, which is done mostly for fluid excited structures, the damping coefficient of the oscillating system is determined under operating conditions on the basis of the measured fluid forces. A comparison of the results with those of a free vibration test in still air is shown. This revised version was published online in July 2006 with corrections to the Cover Date.     

Modal localization in mistuned tube-array structures

An investigation of mode localization in mistuned tube-array structures is studied in this work. The continuous action of hot-cold fluid shock waves in tube-array heat exchangers results in a significant abrasive wear of the tubes, which in turn alters their dynamic behavior and may introduce an undesirable modal localization effect within the tube bundle. This study performs a numerical investigation into the problem of modal localization in mistuned tube-array heat exchangers, with cross-flow. In conducting the investigation, the heat exchanger is modeled as a bundle of periodically-arranged cooling tubes in which the vibrational displacements of the individual tubes are weakly coupled to those of their neighbors via a squeezed water film in the gap between them. In general, the numerical results reveal that damage to even a single tube within the array is sufficient to introduce a severe modal localization effect. Furthermore, due to the weak coupling effect of the fluid, the vibrational energy induced by modal localization is confined to the defected tube and its immediate neighbors, and hence the risk of further wear defected within the tube bundle is increased. The results suggest that the modal localization phenomenon is alleviated at higher values of the cross-flow velocity, but becomes more severe as the tube wall thickness is increased.     

Fluid-elastic instability in tube arrays subjected to air-water and steam-water cross-flow

Flow induced vibrations in heat exchanger tubes have led to numerous accidents and economic losses in the past. Efforts have been made to systematically study the cause of these vibrations and develop remedial design criteria for their avoidance. In this research, experiments were systematically carried out with air-water and steam-water cross-flow over horizontal tubes. A normal square tube array of pitch-to-diameter ratio of 1.4 was used in the experiments. The tubes were suspended from piano wires and strain gauges were used to measure the vibrations. Tubes made of aluminum; stainless steel and brass were systematically tested by maintaining approximately the same stiffness in the tube-wire systems. Instability was clearly seen in single phase and two-phase flow and the critical flow velocity was found to be proportional to tube mass. The present study shows that fully flexible arrays become unstable at a lower flow velocity when compared to a single flexible tube surrounded by rigid tubes. It is also found that tubes are more stable in steam-water flow as compared to air-water flow. Nucleate boiling on the tube surface is also found to have a stabilizing effect on fluid-elastic instability.     

Numerical simulation of streamwise fluidelastic instability of tube bundles subjected to two-phase cross flow

This work aims to develop and validate a numerical model to simulate the flow-structure interaction in tube bundles subjected to two-phase flow. The model utilizes a mixture multiphase module in which a drift flux formulation is used to account for the slip between the phases. Two methods of numerical flow-structure interaction are used to predict the onset of fluidelastic instability (FEI) in the streamwise direction for a two-phase air–water flow mixture in parallel triangular tube bundles. These models are the hybrid analytical-flow field model and the direct numerical flow/structure coupling model. This work investigates the effects of void fractions in the range of 20% to 80% and several pitch-to-diameter ratios (P/D) in the range of 1.3 to 1.7. The results of the fluidelastic forces and the stability threshold are validated against the experimental data available in the literature and show an excellent agreement. The streamwise FEI threshold shows a significant dependency on the pitch-to-diameter ratio while the void fraction exhibits a lesser effect. Generally, the stability threshold increases as the pitch-to-diameter ratio increases. The model that was developed paves the way for devising of more reliable prediction tools for FEI in steam generators.     

CFD modelling of the cross-flow through normal triangular tube arrays with one tube undergoing forced vibrations or fluidelastic instability

A CFD methodology involving structure motion and dynamic re-meshing has been optimized and applied to simulate the unsteady flow through normal triangular cylinder arrays with one single tube undergoing either forced oscillations or self-excited oscillations due to damping-controlled fluidelastic instability. The procedure is based on 2D URANS computations with a commercial CFD code, complemented with user defined functions to incorporate the motion of the vibrating tube. The simulation procedure was applied to several configurations with experimental data available in the literature in order to contrast predictions at different calculation levels. This included static conditions (pressure distribution), forced vibrations (lift delay relative to tube motion) and self-excited vibrations (critical velocity for fluidelastic instability). Besides, the simulation methodology was used to analyze the propagation of perturbations along the cross-flow and, finally, to explore the effect on the critical velocity of the Reynolds number, the pitch-to-diameter ratio and the degrees of freedom of the vibrating cylinder.     

Drift-flux model for upward two-phase cross-flow in horizontal tube bundles

In relation to void fraction prediction of cross-flow in horizontal tube bundle of shell-tube heat exchangers, a drift-flux correlation has been developed to meet the demand on the study of two-phase flow gas and liquid velocities, two-phase pressure drop, heat transfer, flow patterns and flow induced vibrations in the shell side. Two critical parameters such as distribution parameter and drift velocity have been modeled. The distribution parameter is obtained by constant asymptotic values and taking into account the differences in channel geometry. The drift velocity is modelled depending on the density ratio and the non-dimensional viscosity number. The relationship between the channel averaged and gap mass velocity has been discussed in order to obtain the superficial gas and liquid velocities in the drift-flux correlation. The newly developed drift-flux correlation agrees well with cross-flow experimental databases of air-water, R-11 and R-113 in parallel triangular, normal square and normal triangular arrays with the mean absolute error of 1.06% and the standard deviation of 4.47%. In comparison with other existing correlations, the newly developed drift-flux correlation is superior to other studies due to the improved accuracy. In order to extend the applicability of the newly developed drift-flux correlation to void fraction of unity, an interpolation scheme has been developed. The newly developed drift-flux correlation is able to calculate the void fraction of cross-flow over a full range with different sub-channel configurations in shell-tube type heat exchangers.     

PIV measurement of the vertical cross-flow structure over tube bundles

Shell and tube heat exchangers are among the most commonly used types of heat exchangers. Shell-side cross-flow in tube bundles has received considerable attention and has been investigated extensively. However, the microscopic flow structure including velocity distribution, wake, and turbulent structure in the tube bundles needs to be determined for more effective designs. Therefore, in this study, in order to clarify the detailed structure of cross-flow in tube bundles with particle image velocimetry (PIV), experiments were conducted using two types of model; in-line and staggered bundles with a pitch-to-diameter ratio of 1.5, containing 20 rows of five 15 mm O.D. tubes in each row. The velocity data in the whole flow field were measured successfully by adjusting the refractive index of the working fluid to that of the tube material. The flow features were characterized in different tube bundles with regards to the velocity vector field, vortex structure, and turbulent intensity.     

Flow-induced vibration of curved cylinder arrays subject to loose support

In this paper, an improved model is developed aimed at analyzing the fluidelastic vibration of a single flexible curved tube which is surrounded by rigid cylinders and subjected to cross-flow and loose support. Based on the previous model, the axial extension of the curved tube described by von Karman nonlinearity has been accounted for in the current research. Simulations are performed to explore the effect of quasi-steady fluid force model and velocity-limited friction model on the post-instability behavior. Numerical results show that the out-of-plane response is confined to a limit cycle by the von Karman nonlinearity and the in-plane vibration is induced by the out-of-plane vibration through the nonlinear coupling. When the loose support comes into play, the nonlinear impact forces become dominate. The results are presented; comparisons are made to analyze the parameters influencing the fretting-wear damage, such as normal work rate, contact ratio and impact force level.     

THE CROSSING FREQUENCY AS A MEASURE OF HEAT EXCHANGER SUPPORT-PLATE EFFECTIVENESS

The crossing frequency is the number of times per second the vibration amplitude crosses the zero displacement line from negative displacement to positive displacement. In flow-induced vibration in which the motions are often random and/or a number of modes contribute to the vibration amplitudes, the crossing frequencies are modal-weighted average frequencies of the vibration. It is postulated in this paper that the crossing frequency can be used as a measure of heat exchanger support-plate effectiveness. Using a time-domain, nonlinear analysis technique, the crossing frequencies of a tube vibrating in support plates with oversized holes can be computed as a function of time and the tube-to-support-plate clearances. It was found that the fluid–elastic stability margin of a tube bundle, in the context of the original Connors' equation for tube bundle fluid–elastic instability, should be independent of the tube-to-support-plate clearances. A simple method of estimating the critical velocity based on the time-domain equation of fluid–elastic stability is suggested.