[en] Tube failures due to excessive flow-induced vibration must be avoided to assure the reliable performance of heat exchangers. Such components must be thoroughly analysed for vibration at the design stage. Several flow-induced vibration mechanisms are possible. In this paper, particular attention is given to fluidelastic instability of tube bundles subjected to liquid and two-phase cross-flow. The performance of ten operating heat exchanger components is reviewed to validate our recommended vibration analysis guidelines. Both tube failure histories and satisfactory performances are considered. The results show that a fluidelastic instability constant K=3.3 is a reasonable design criterion. (auth)

[en] This paper presents a flow-induced vibration analysis technique and a procedure to predict fretting wear damage from experimental data. Examples are used to illustrate these techniques. (auth.)

[en] Tube and shell heat exchange components such as steam generators, heat exchangers and condensers are essential parts of CANDU nuclear power stations. Excessive flow-induced vibration may cause tube failures by fatigue or more likely by fretting-wear. Such failures may lead to station shutdowns that are very undesirable in terms of lost production. Hence good performance and reliability dictate a thorough flow-induced vibration analysis at the design stage. This paper presents our approach and techniques in this respect. (author)

[en] Steam generator thermalhydraulic codes are frequently used to calculate both global and local parameters inside a stern generator. The global parameters include heat transfer output, recirculation ratio, outlet temperatures, and pressure drops for operating and abnormal conditions. The local parameters are used in further analyses of flow-induced vibration, fretting wear, sludge deposition, and flow-accelerated corrosion. For these purposes, detailed, 3-dimensional 2-phase flow and heat transfer parameters are needed. To make the predictions more accurate and reliable, the codes need to be validated in geometries representative of real conditions. One such study is an international co-operative experimental program called CLOTAIRE, which is based in France. The CANDU Owners Group (COG) participated in the first two phases of the program. The results of the validation of Phase 1 were presented at the 1994 Steam Generator and Heat Exchanger Conference, and the results of the validation of Phase II are the subject of this report. THIRST is a thermalhydraulic, finite-volume code used to predict flow and heat transfer in steam generators. The local results of CLOTAIRE Phase II were used to validate the code. The results consist of the measurements of void fraction and axial gas-phase velocity in the U-bend region. The measurements were done using bi-optical probes. A comparison of global results indicates that the THIRST predictions, with the Chisholm void fraction model, are within 2% to 3% of the experimental results. Using THIRST with the homogeneous void fraction model, the global results were less accurate but still gave very good predictions; the greatest error was 10% for the separator pressure drop. Comparisons of the local predictions for void fraction and axial gas-phase velocity show good agreement. The Chisholm void fraction model generally gives better agreement with the experimental data, whereas the homogeneous model tends to overpredict the void fraction and underpredict the gas velocity. (author)

[en] A thorough flow-induced vibration analysis of nuclear components such as heat exchangers and steam generators is essential at the design stage to ensure good performance and reliability. This paper presents the authors' approach and techniques in this respect. In a steam generator, for example, the flow may be liquid or two-phase. In general, parallel and cross-flow exist in the tube bundles of heat exchange components. In cross-flow three basic vibration excitation mechanisms are considered, namely fluidelastic instability, periodic wake shedding resonance, and forced response to random flow turbulence. The latter may need to be considered in parallel flow. These vibration excitation mechanisms and the dynamics of multispan tubes are formulated in a computer model which is used to predict the vibration response of the tubes. The computer model and the parameters required to formulate the vibration excitation mechanisms are discussed. Examples of vibration analysis of steam generators and heat exchangers are outlined. It is concluded that most flow-induced vibration problems may be avoided by proper analysis at the design stage. (Auth.)

[en] Steam generator thermalhydraulic codes are used frequently to calculate both global and local parameters inside the steam generator. The former include heat transfer output, recirculation ratio, outlet temperatures, and pressure drops for operating and abnormal conditions. The latter are used in further analyses of flow-induced vibration, fretting wear, sludge deposition, and flow accelerated corrosion. For these purposes, detailed, three-dimensional two-phase flow and heat transfer parameters are needed. To make the predictions more accurate and reliable, the codes need to be validated in geometries representative of real conditions. One such study is an international cooperative experimental program called CLOTAIRE based in France. COG participated in the first two phases of the program; the results of the validation of Phase 1 were presented at the 1994 Steam Generator and Heat Exchanger Conference, and the results of the validation of Phase II are the subject of this paper. THIRST is a thermalhydraulic, finite volume code to predict the flow and heat transfer in steam generators. The local results of CLOTAIRE Phase II have been used to validate the code. These consist of the measurements of void fraction and axial gas-phase velocity in the U-bend region. The measurements were done using bi-optical probes. A comparison of global results indicates that the THIRST predictions, with the Chisholm void fraction model, are within 2 to 3% of the experimental results. Using THIRST with the homogeneous void fraction model, the global results were less accurate but still well predicted with the greatest error of 10% for the separator pressure drop. Comparisons of the local predictions for void fraction and axial gas-phase show good agreement. The Chisholm void fraction model generally gives better agreement with the experimental data while the homogeneous model tends to overpredict the void fraction and underpredict the gas velocity. (author)

[en] Tube and shell heat exchange components such as steam generators, heat exchangers and condensers are essential parts of CANDU (CANadian Deuterium Uranium) nuclear power stations. Excessive flow-induced vibration may cause tube failures by fatigue or more lkiely by fretting-wear. Such failures may lead to station shutdowns that are very undersirable in terms of lost production. Hence good performance and reliability dictate a thorough flow-induced vibration analysis at the design stage. This paper presents our approach and techniques in this respect. The vibration excitation mechanisms and the dynamics of multispan tubes are formulated in a computer model. The model predicts tube vibration response and critical velocities for fluidelastic instability. A description of the model is given. The vibration analysis of a steam generator is outlined as an example. (Auth.)

[en] The deposition of solid particles from two-phase flow on surfaces is a complicated problem. It is further complicated in steam generator analysis by the need to simultaneously model the heat transfer and its response to deposition. This paper describes a three-dimensional transient simulation of magnetite particulate fouling, and the accumulation of particles on the heat transfer surfaces in a nuclear steam generators

[en] This paper provides an example of a combined thermal-hydraulic and vibration analysis of a nuclear U-tube steam generator using the THIRST code to calculate the multidimensional two-phase flow field and the PIPO code to calculate the free and forced response of a specific U-tube. Several hypothetical U-bend support configurations are considered to demonstrate the effects active and inactive tube supports. (author)

[en] Tube and shell heat exchange components such as steam generators, heat exchangers and condensers are essential parts of CANDU (CANadian Deuterium Uranium) nuclear power stations. A flow-induced vibration analysis is presented. In cross-flow three basic flow-induced vibration excitation mechanisms are considered, namely: fluidelastic instability, forced vibration response due to random flow turbulence and periodic wake shedding (the first two mechanisms in both liquid and two-phase cross-flow). Periodic wake shedding has not been detected in two-phase flow but is possible in liquid flow. It is only significant for upstream tube rows. Random flow turbulence is he dominant excitation in both liquid and two-phase axial flow. These vibration excitation mechanisms and the dynamics of multispan tubes are formulated in a computer model. The model predicts tube vibration response and critical velocities for fluidelastic instability. A description of the model is given. The vibration analysis of a steam generator is outlined as an example. The parameters required to formulate the vibration excitation mechanisms are discussed. Periodic wake shedding excitation is formulated in terms of a Strouhal No. and a lift coefficient which is generally less than unity. Fluidelastic instability thresholds are related to dimensionless flow velocity and dimensionless damping for both liquid and two-phase cross-flow. Some statistical parameters to describe random flow turbulence excitation are deduced from experimental data. The power spectral density of the latter is related to a power of the flow velocity. The velocity exponent is roughly two for liquid flow and near unity for two-phase flow. For a given mass flux, the random excitation reaches a maximum at a steam quality of roughly 15%. Damping in two-phase flow is found to be at least four times greater than in liquid flow