Thermal fatigue evaluation of piping system "tee"- connection

The task of the load determination is to detect and quantify the thermal turbulence occurring in the fluid by experiments and by numerical fluid calculations. The determination of the heat transfer coefficient (fluid to wall) is considered to be one crucial point. After a collation of the existing load determination experience thermo-hydraulic tests have been performed to measure the relevant load parameter. In addition, thermo-hydraulic analyses with CFD (Computational Fluid Dynamics)-codes have been carried out to simulate the turbulent fluid behaviour. Generalised guidelines for load determination have been derived from the experimental and numerical results. As an additional sophisticated aspect, the development of virtual sensors based on neural networks and fuzzy logic tools has been included to determine the temperature loads for a selected NPP-PWR-component (the pressurizer surge line). In experimental tests temperature differences and load cycle frequencies due to the turbulent fluid flow occurring at various mixing Tees have been simulated, illustrated and quantified by experimental approach in small scale Tee-tests models in various positions and configurations. These tests have been performed on plexi-glass models for flow visualization and metal mock-ups for direct measurements of temperature distributions in the fluid and through the wall pipe. In addition to the experimental tests, numerical thermo-hydraulic calculations (CFD-analyses) have been performed for benchmark cases and for other Tee-configurations. The Benchmark calculations have been performed to determine the turbulent fluid distribution in a 50:50 mm Tee-connection with different CFD-Code approaches in comparison to experimental testing results. The CFD-Code calculations are based on different simulation approaches (Large Eddy-Simulation, Ke-approach or other code methods), which have been applied in THERFAT. Agreement between analytical predictions and experimental results is reasonable. Guidelines for an analytical load assessment that integrate velocity and temperature fields and the calibration of the heat transfer modelling between fluid and wall can be derived from the obtained results. But the CFD-analyses turned out to be very time consuming; even short time history calculations of a few seconds require a tremendous amount of computer time (weeks or even months). Powerful computers are required for the execution of the calculations. An additional highly sophisticated aspect in the load determination field covers the development of virtual sensors based on neural network and fuzzy logic tools to simulate the dependency of thermal fluctuations on from transient mass flow and temperature distributions of the surge lines in the Vandellos NPP, Spain and the Mochovce NPP, Slovakia. The THERFAT-work comprises the development and training of the virtual sensors, the pilot installation, the online data acquisition of the surge line temperature transients and the evaluation of the results.

As a first step in the THERFAT project, a survey of fatigue susceptible Tee-configuration with hot and cold temperature mixing in various nuclear power plants has been carried out by the utilities E.ON, EDF and Fortum. From these a variety of Tee-geometries were selected to be investigated (DN 50/DN 50, DN 100/DN 25, DN 80/ DN 80) with different flow velocities and temperature differences. A reduced number of configuration is selected as the starting point for the upcoming activities in the project: performance of representative experiments, visualization and quantification of relevant parameters, verification of experimental results by numerical calculation, development of a numerical evaluation approach applicable for Tee-configurations, determination of existing safety margins using common codes procedures, determination of load bearing capabilities, comparison of results, combination of the results and development of an overall integrity concept framework.

By means of verification tests, damage tests on small-scale specimen (investigation of fatigue behaviour and fracture mechanics parameter) and on large-scale components using realistic mock-up configurations have been carried out. The selection of the large-scale test configuration was based on a comprehensive compilation of existing thermal fatigue laboratory test data and of information about the most fatigue relevant parameter (e.g. surface finishes, mean stress effects). In most fatigue tests performed so far mechanical loads were applied, only a few thermal load tests are known. Therefore, the THERFAT large-scale damage tests were focused on (severe) thermal shock loads. Tests on small specimen with mechanical loads were performed to compare the effects of mechanical to thermal loads. The experimental damage test pattern established in THERFAT was supposed to investigate realistic component or specimen resistance capabilities concerning specific degradation mechanisms (e.g. thermal shocks) to provide a better understanding of failure modes and the quantification of safety margins in terms of the load/stress-fatigue-crack initiation interaction and to single out the most relevant parameter that can serve as criteria for the verification and calibration of integrity procedures. The achieved results in the experimental damage tests concerning the load bearing capability of components and regarding selected aspects of fatigue assessments confirm the predicted tendencies of higher load bearing capabilities than expected by application of common code analysis procedures. In terms of fatigue usage, all the experimental test results showed strain rates above or in accordance with the ASME-design fatigue curve, proving the currently applied fatigue curves (ASME design curve) as appropriate. The tests confirmed known tendencies, that the number of load cycles up to crack initiation is comparable between mechanical and thermal loads. But in the further development from crack initiation to failure, the current experience shows, that cracks caused by mechanical loads tend to propagate, while initiated cracks due to thermal loads appear to come to a crack arrest. The results of the damage tests with the related analytic investigation are generally valid for all respective components subjected to cyclic thermal loads and are not only limited to Tee-connections.

The THERFAT project has been initiated to provide improved thermal fatigue evaluation approaches for managing thermal fatigue risks in Tee-connections susceptible to high cyclic temperature fluctuations. It was intended to close a gap regarding one specific fatigue issue. Although the THERFAT-project scope covered only this small-scale fatigue issue, a conceptual approach was required to frame and to condense the activity results of several technical disciplines to provide useful conclusions for practical engineering application. Furthermore, the THERFAT-consortium consists of 16 partners with different technical and scientific background and objectives. Thus, a common understanding had to be developed, which parameter were fatigue significant and important compared to those of less importance. To reach this goal, a conceptual approach was required for the THERFAT project processing focused on the "small scale" Tee-connection investigation. In NPP, the same configuration of different technical disciplines and persons with different technical background exists. Simple guide lines are helpful how to mitigate or even avoid fatigue significant loads before they occur e.g. by optimising the system operation procedure or by ISI-measures to stop a valve leakage. Plant operation people who are not very much familiar with fatigue analyses should get an idea which temperature differences may be stress and fatigue relevant. The applied THERFAT-processing approach can be regarded as a small example for the application of a general integrity concept. Enlarged for general application it can be used for all kinds of integrity evaluations (e. g. "Leak before Break"-Assessments, or Ageing Management-activities). The "Road-Map" explains the different elements of this overall integrity concept and how they are linked together. The approach may help to improve a harmonized understanding for the people involved and can be regarded as a first step on the road to a future European fatigue evaluation methodology. The objective is to demonstrate that an unnecessary accumulation of conservatisms in single integrity concept elements can be avoided and that a balanced interaction between the different elements can be achieved by increasing the importance level of one element compensating the reduced significance level of another element, if appropriate.

The "integrity evaluation" is based on the thermal spectrum loads determined by the experimental and numerical results deduced in THERFAT. Different integrity evaluation procedures concerning the assessment of stresses, fatigue usage, crack initiation and crack propagation issues for components or specimen subjected to cyclic thermal loads has been executed. Code and standard procedures and more refined analysis models have been applied for more accurate predictions about the influence of the degradation mechanisms. Assigned partners used the load determination input from experiments and numerical analyses to evaluate stresses/fatigue usage, crack initiation and crack propagation ("forward integrity evaluation approach"). Other partners already performed stress/fatigue usage, crack initiation and crack propagation analyses bases on so-called damage tests to predict and to verify the experimental results. The results of the integrity evaluation demonstrate the feasibility of the introduced integrity evaluation procedures. Code evaluation methods and other common approaches are appropriate to deal with the fatigue related issues. This applies for the stress calculation with provided thermal spectrum loads (from experimental tests and in a smaller scale from CFD-analyses or from damage tests) and for the related fatigue evaluation. For the fracture mechanics assessments (crack initiation and crack propagation) common approaches based on elastic material behaviour have been confirmed as suitable. More sophisticated approaches, e.g. using elastic/plastic material laws are available for safeguarding reasons, if appropriate. As conclusion can be stated that effective simplified and more sophisticated engineering tools are available to deal with high cyclic turbulent loads.

Livrables

Partager cette page

Télécharger